electro-mechanical world - Students Science Conference

1 downloads 153 Views 17MB Size Report
graphical software configuring tool, which gene ... ate software. ...... cedes, which offer vehicles equipped with autop
Key words: LEM BULLET, wiring harness, Controller area network, STM F303, ARM

Tomasz DZIUBAK*

THE PROCESS OF REALIZATION OF THE ELECTRIC BUNDLE BASED ON CAN INTERFACE IN THE LEM BULLET MOTORCYCLE

Introduction: LEM Bullet is a innovative electrical motorcycle which is designed for a Police use in the urban area. Components that are used in the motorcycle made electric bundle modular and easily upgradable and reduces the overall cost of production. Aim of the study is to simplify electric bundle and to minimize amount of excessive cables in the motorcycle and presentation of each stage of production carried during the realization of the project. Materials and methods: In the process of realization of the project the STM F303 microcontrollers based on the ARM architecture with build in CAN controller and external transceiver are used. In the system self created and crafted PCB boards are applied. Results of the usage of the system are immensely decreased amount of the cables mounted on the motorcycle frame and in addition curtail the montage time. Conclusions: System should be developed in the stability of reaction by insensitivity inputs of the regular motorcycle switches and furthermore by the minimalisation of the main ECU.

1. INTRUDUCTION 1.1. THE LEM BULLET

The LEM Bullet is a light electric motorcycle built by the PiRM Student Scientific Club operating at Wroclaw University of Science and Technology. Theaim of building the motorcycle was to create a light vehicle for Wroclaw police that would have been able to deal with heavy traffic and operating in restricted area for vehicles with fuel

*Wroclaw University of Science and Technology, 50-370 Wrocław, ul. Wybrzeże Stanisława Wyspiańskiego 27

1

Tomasz DZIUBAK

engines. The 6kW BLDC hub engine with a 2.14kWh battery allows to travel up to 100km with maximum speed of 110km/h..

1.2. CONTROLLER AREA NETWORK

The Controller area network (CAN) was developed in the 80s at Robert Bosch GmbH company to allow microcontrollers and other devices to communicate with each other without one central host [1].The CAN interface is one of the most common used communication in automotive and industry [2], notwithstanding that in Polish publications inadequately described.

2. AIM OF STUDY 2.1. USED STANDARDS

The aim of study was to simplify wiring harness to achieve increased possibility of using the standard motorcycle switches and to minimize amount of excessive cables in the motorcycle. The CAN interface implemented in the LEM Bullet was based on 2.0A standard1 and was built initially with 3 nodes. That standard used 11-bit identifier what was more than enough for motorcycle interface. 2.2. INNITAL ASSUMPIOTNS

The maximum length of the wiring on motorcycle was calculated for amount of 5m. It gave opportunity to determine the speed value to 100kbit/s. The speed was dictated by safety reasons as also by requirements of quantity of data transfer. The amount of data transfer allow to set DLC (Data length code) from 1 to 4 bytes of data. Due to non-built-in CAN controllers in peripherals used in LEM Bullet resulted in self-created array of data transfer and its identifiers.

3. BASE PLATFORM AND ADDITIONAL BOARDS In the project were used the STM F303 microcontrollers, built in the nucleo development board. The F303 based on the Cortex M4 ARM architecture, have a built-in CAN controller and core speed up to 84Mhz. Due to above mentioned issues, the

2

HE PROCESS OF REALIZATION OF THE ELECTRIC BUNDLE BASED ON CAN INTERFACE IN THE LEM BULLET MOTORCYCLE

shield compatible with morphopinout on the nucleo board to cope with standard I/O in the motorbikehas been designed. The design of the board allowed making one type of shield for all purpose.To provide best stability of the system, a transil diode operating at 12,7V have been mounted from the main power source to the ground of the circuit. For the same purpose each input on the board have been secured by the Zener diodes. Due to the problem of bouncing switches on the inputs, the capacitors connected to the ground ensures reliability of reading the state. The Power part of the board consists of ULN2003 darlington’s IC a two mosfet controlled by the bipolar transistors.The first of three microcontrollers used in LEM BULLET and also the main board is placed in front of the motorcycle, just before the head of the frame. The second, which commucommunicates with the motor controller, is positioned centrally in the vehicle and the third is located on the back in the trunk [4].

Fig. 1 Design of the addition printed circuit boards

The centre box includes one microcontroller and also two DC/DC converters. Each converter is able to deliver up to ten Amps of power constantly. The main bundle consists of shielded twisted pair of wires for the CAN communication and two, four millimetres square, wires for power transfer. The LED technology light beam consumed only 4 Amps of current in peak when main and police lights were on. Average usage of the power in the system is less than 40W. The design of the software started with

3

Tomasz DZIUBAK

the selection of the language and the library delivered by the producer of the micromicrocontrollers. The most common language in microcontrollers is the ANSI C. The ST corporation delivers two type of libraries, STDlib (Standard Library for the ST) and and newer one HAL Library (Hardware Abstraction Layer). In this project the HAL Library was selected to create software. The HAL library is much more similar to asassembler and also uses object-oriented programming. programming. The ST corporation ensures the graphical software configuring tool, which generates C initialization code, called CubeMX. The main IDE was a SW4STM32 (System Workbench for stm32) based on the Eclipse.

Fig. 2. Location of nodes in LEM BULLET

In order to debug the code the STM Studio was used. To handle with the higher amount of inputs, outputs and to ensure the real-time operating, the system was created on the interrupts event handlers and two timers. One of them clocked every 1ms updates the outputs, when second one checks the synchronization every 1s.The CAN communication based on the MSP2551 transceivers with the 120 ohm terminal resistors. Each transceivers have no plugged in slew rate resistors to ensure maximum speed of communication. Due to safety reasons, the lowest lowest ID have the emergency messages and first ten ID were used for broadcasting.

4

HE PROCESS OF REALIZATION OF THE ELECTRIC BUNDLE BASED ON CAN INTERFACE IN THE LEM BULLET MOTORCYCLE

4. RESULTS The final result of the whole project was a created electric bundle, that was able to cope with difficult weather conditions, resistant to jamming and furthermore with universal purpose. The bundle can be easily changed from the police wiring to a normal daily use motorbike bundle or even to other specialist wiring for motorbike. In comparison to old LEM vehicles’bundles this one shorten the mounting time by 40% and used 25% less wires.

REFERENCES

[1] CAN in Automation, History of CAN technology : HTTP://WWW.CAN-CIA.ORG/CANKNOWLEDGE/CAN/CAN-HISTORY/ [Access in 19 of July 2016] [2] GABRIEL LEEN , DONAL HEFFEMAN , and ALAN DUNNE; Digital Networks in the Automotive Vehicle [3] ISO 11898-1:2015 Road vehicles -- Controller area network (CAN) -- Part 1: Data link layer and physical signaling [4] PETRUS A. JANSE VAN RENSBURG and HENDRIK C. FERREIRA Automotive Power-Line Com-munications : Favourable Topology for Future Automotive Electronic Trends, 7th International Symposium on Power-Line Communications and Its Applications Kyoto, Japan, March 26-28, 2003 Session A3: Cable Modeling

5

Key words: wind turbine, PMSG, stand-alone supply system, simulation studies

Piotr GAJEWSKI*

SENSORLESS CONTROL OF A STAND-ALONE WIND TURBINE SYSTEM WITH PMSG

The paper presents the stand-alone power supply system with wind energy conversion system. The Permanent Magnet Synchronous Generator (PMSG) is connected to the three-phase load through a switch mode rectifier and load side converter. The switch mode rectifier consists of three-phase diode bridge and DC/DC boost converter. The control strategy for the switch mode rectifier has been used in order to extract the maximum power form wind turbine. In the control of DC/DC boost converter the sensorless control strategy has been applied. The sensorless control method is based on the rotor speed estimation, which is determined by measuring of the DC voltage and DC current of the diode bridge. The excess power of wind energy conversion system is dissipated in the damp resistor controlled by the chopper. The considered control strategy of load side converter allows regulate the load voltage in terms of magnitude and frequency under wind speed variations.

1. INTRODUCTION The stand-alone wind energy conversion systems (WECS) can be considered as an effective way to provide of the supply energy for the small customers [1, 3, 5, 6]. The main objectives of stand-alone wind energy conversion systems are the controls of the value of the frequency and the value of the amplitude of supply voltage at the variable load. In stand-alone system the output AC voltage has been controlled in terms of value of amplitude and frequency by using the Load Side Converter (LSC). The aim of this paper is the analysis of the small scale stand-alone supply power installation based on the wind energy conversion system with permanent magnet synchronous generator (PMSG).

__________ * Wroclaw University of Science and Technology, Department of Electrical Machines, Drives and Measurements, ul. Smoluchowskiego 19, 50-372 Wroclaw, Poland, [email protected]

6

Piotr GAJEWSKI

2. CONTROL OF STAND-ALONE SYSTEM The considered stand-alone wind energy conversion system with PMSG and control circuits is shown in Figure 1. The system presented in this Figure consists of the following elements: PMSG, which is directly driven by the wind turbine, Switch Mode Rectifier (SMR), which consists of three phase diode bridge and DC/DC boost converter, Chopper, which consists of IGBT switch and damp resistance Rd, Load Side Converter (LSC), LC filter, which is connected at the output of the LSC converter, Three phase load. Switch Mode Rectifier (SMR)

Ld

Chopper Load Side Converter (LSC)

T1

id Cd 1

PMSG

LC Filter

D1 T3

AC bus

T5

Lg Load

K

Load Load

Cd

Rd

T4

T6

T2

Cf i gabc

vd vd id

Rotor speed estimation

PWM

ωm vd id

v gabc

v dc

D

PWM Chopper cotroller

v*ga v*gb v *gc

SMR control with MPPT

v dc max vdc

LSC controller

i gabc v gabc

vdc

Fig. 1. Control diagram of the stand-alone wind energy conversion system with direct-driven PMSG

The application of SMR with DC/DC boost converter allows to control of the electromagnetic torque of PMSG. In the control of DC/DC boost converter the Maximum Power Point Tracking (MPPT) algorithm should be used in order to maximize the power that the turbine extracts from the wind [2, 3, 6, 7]. The output mechanical power captured by the wind turbine can be expressed through the following relation [4, 5]:

Pt = 0.5 ρ A C p (λ , β ) vw3

(1)

where: ρ - air density; A = πR2 - area swept by the rotor blades; R -radius of the turbine blade; Cp - power coefficient of the wind turbine; β - blade pitch angle; vw wind speed, λ – tip speed ratio, which can be calculated as:

λopt =

ωm R vw

(2)

where: ωm – angular rotor speed. The power of wind turbine can be expressed as the function of tip speed ratio:

7

Analysis of the operation of direct-driven wind energy conversion system during voltage sags 3

ω R  Pt = 0.5 ρ A C p max  m  = K optωm3 λ   opt 

(3)

where: Kopt – coefficient of wind turbine. In the control of DC/DC boost converter the sensorless method with MPPT algorithm has been applied. The angular rotor speed ωmest can be estimated by measuring the output DC voltage vdc and DC current idc of the diode bridge. The angular rotor speed ωmest can be estimated based on the parameters of PMSG and the base of measurement of the variables of diode bridge, which can be calculate as [4]:

ωmest =

2π (vd + 2 Rs id )

(4)

3 3  60 K M − pb Ls id   π 

where: vd - the output DC voltage of diode bridge, id - the DC current of diode bridge, Rs - stator resistance, KM – peak of line-to-neutral back emf constant, pb - number of pole pairs, LS - stator inductance. The next control step determines the reference power Pt* of the wind turbine in accordance with equation (3). This reference power is then used to calculate the reference DC current id* of rectifier as:

id* =

Pt* vd

(5)

The reference DC current id* is then compared with measured DC current of DC/DC boost converter. The error signal is sent to PI controller. The control signal of PI controller determines the required signal to the transistor switch K. The block scheme of DC/DC boost converter with MPPT algorithm has been presented in Figure 2. vd vd id

ωmest

i d*

* g

P

id

Fig. 2. Control diagram of the DC/DC boost converter [4]

8

Piotr GAJEWSKI

3. CONTROL OF LOAD SIDE CONVERTER The control objective of Load Side Converter (LSC) is to regulate the value of the amplitude of load voltage and the value of the voltage frequency at the load [1, 3]. The block scheme of vector control for LSC system has been presented in Figure 3. The control scheme of LSC consists of four control loops with PI controllers. The outer control loop regulates the DC link voltage. The output value from PI controller determines the reference value vd* of the amplitude of the load voltage. The inner loops are designated to obtain the operation at unit power factor. The instantaneous reactive power is set to zero.

Rd

Cd

i gabc

Sa Sb Sc

v gabc vd

* ga

v

* gb

* gc

v v

θg

i d*

vd*

* vdc

v*gd

vdc max

vdc

v*gq

ω g Lg iq

id

vd

ωg Lg id

iq

vq

i q*

vdc v q* = 0

Fig. 3. Control diagram of the Load Side Converter (LCS)

The output values from both PI controllers determine the reference values of voltage vgd* and vgq* for LSC. These reference voltages are then transformed to the abc system and are sent to the PWM block. In order to obtain better performance of the control method, the decoupling terms are added as shown in Figure 3. In order to dissipate the excess power, the chopper with damp resistor Rd has been applied. The application of chopper allows to protect and control the voltage in DC link. If the DC link voltage exceeds the maximum value of this voltage, the DC link will be short-circuited through damp resistance Rd.

3. SIMULATION RESULTS The proposed control strategy of stand-alone wind energy conversion system with PMSG has been simulated using MATLAB/Simulink. Digital situation studies were

9

Analysis of the operation of direct-driven wind energy conversion system during voltage sags

a)

12

vw [m/s]

performed for the system with wind turbine data and parameters: rated power Pt 5kW; rotor radius R -2,8m; air density ρ -1,225 kg/m3 and for 3-phase PMSG data and parameters: rated power Pg -5kW; stator resistance Rs -1,5 Ω; stator dq-axis inductance Ld, Lq - 14,04mH; stator rated phase current Isn -15A, total moment of inertia J=0.4kgm2; number of pole pairs pb=8. Figure 4a. shows the waveforms of the response of the system for a step change of wind speed, during 4s simulation. The estimated ωmest and measured angular rotor speed ωm of PMSG are presented in Figure 4b. Form this Figure, it can be noticed that the estimated angular rotor speed track the reference speed accurately. Figure 4c shows the waveforms of tip speed ratio λ and power coefficient Cp. The value of power coefficient and value of tip speed ratio have the maximum value, which is equal λ=6.8 and Cp=0.47. It is clear that the MPPT control technique works very well and with the high accuracy.

10

620

e) vdc [V]

610

8

0.5 30

1

1.5

2

2.5

3

3.5

570

4

0.5

1

1.5

2

2.5

3

3.5

4

3.5

4

250

f)

ωm

25

ωmest

20 15 10 0 10

c)

0.5

1

1.5

2.5

3

3.5

200

150

4

g)

Load change

0.5

1

1.5

2

2.5

3

60

λopt= 6.8

8 6

Cp= 0.47

4 2 0

0.5 0.4 0.3 0.2 0.1

0.5

1

1.5

2

2.5

3

3.5

h)

Te 0.5

1

1.5

2

t [s]

2.5

3

3.5

4

0

-5

20 0.6

500

vabcload [V]

0 -200

40

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

1.1

0

4

Tt

200

-400

iabcload [A]

TSR λ,

Cp

5

d) 400 Te ,Tt [Nm]

2

RMS=230V

vload [V]

[rad/s]

ωm , ωmest

590 580

6

b)

600

0.5

1

1.5

2

2.5

3

3.5

4

0

-500 0.6

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

1.1

t [s]

Fig. 4. Waveforms of: a) wind speed vw; b) estimated angular rotor speed ωmest and measured angular rotor speed ωm of PMSG; c) tip speed ratio λopt and power coefficient Cp; d) electromagnetic torque Te of PMSG and mechanical torque Tt of wind turbine; e) DC-link voltage vdc; f) RMS voltage at load; g) load current iabcload; h) load voltage vabcload

The responses of electromagnetic torque Te of PMSG and mechanical torque Tt of wind turbine have been presented in Figure 4d. The DC link voltage vdc is maintained at its reference value as shown in Figure 4e. The average value of DC link voltage is quite constant even if the wind is changed. Figure 4f shows the waveform of instantaneous value of rms load voltage. From Figure 1f, it can be noticed that the average

10

Piotr GAJEWSKI

value of rms load voltage is maintained at reference value, which is equal to 230V. Figure 4g and 4h show the responses of load current and load voltage. From these waveforms, it can be noticed that the application of LSC allows to maintain the load voltage and frequency at reference value in spite of variations in wind speed and load. Hence, in the case that, load is suddenly changed, the rise in load voltage is very small. The high quality of voltage supplied to the load has been achieved by application of LC filter.

4. CONCLUSIONS The control strategy for stand-alone wind energy conversion system with PMSG has been presented. In the control of DC/DC boost converter the sensorless control method with MPPT algorithm has been applied. In the control of Load Side Converter the vector control technique to maintain the constant value of the amplitude and constant frequency of the output voltage has been applied. The simulation results demonstrate that the applied controller works with high accuracy and fast responses of considered system. The application of MPPT algorithm shows that the maximum power from the wind is obtained. It can be noticed from simulation results the tip speed ratio and power coefficient are kept at reference and maximum value. On the base of simulation results, it can be stated that through the application of considered control method the constant load voltage is obtained, even at the changes of the wind speed and the load values. REFERENCES [1] BHENDE C.N., Stan-alone wind energy supply system, International Conference on Power Systems, ICPS ‘09, 2009, 1-6. [2] GAJEWSKI P., PIEŃKOWSKI K., Analysis of a wind energy converter system with PMSG generator, Czasopismo Techniczne, 2015, Elektrotechnika Zeszyt 1-E (8), 2015, 219-228. [3] HAQUE M.E., NEGNEVITSKY M., MUTTAQI K.M., A novel control strategy for a variable speed wind turbine with a permanent magnet synchronous generator, IEEE Transactions on Industry Applications, Vol. 46, No.1, 2010, 331-339. [4] HUSSEIN M.M., SENJYU T., ORABI M., WAHAB A.A., HAMADA M.M., Simple maximum power extraction control for permanent magnet synchronous generator based wind energy conversion system, Conference on Electronics Communications and Computers (JEC-ECC), 2012, 194-199. [5] HUSSEIN M.M., SENJYU T., ORABI M., WAHAB A.A., HAMADA M.M., Load power management control for a stand alone wind energy system based on the state of charge of the battery, IEEE International Conference on Power and Energy (PECon), 2012, 93-98. [6] SERBAN I., Small wind turbine control with frequency support for integration in microgrids, Bulletin of the Transilvania University of Brasov, Engineering Sciences, Vol.6 (55), No.2, 2013, 89-96. [7] WU B., YONGQIANG L., NAVID Z. SAMIR K., Power Conversion and Control of Wind Energy, John Wiley & Sons, INP., Publication 2011.

11

Key words: wind turbine, PMSG, low voltage ride through, chopper, simulation studies

Piotr GAJEWSKI*

CONTROL OF THE WIND TURBINE SYSTEM WITH PMSG UNDER GRID VOLTAGE SAGS

The paper presents the control method of the Wind Energy Conversion System (WECS) with variable speed wind turbine during the grid voltage sags. The considered system of wind turbine consists of Permanent Magnet Synchronous Generator (PMSG) and back-to-back converter. The back-toback converter system consists of Machine Side Converter (MSC) and Grid Side Converter (GSC). The Low Voltage Ride-Through (LVRT) requirements are necessary to ensure the stability of the power systems under voltage sags. During the voltage sags, the surplus output power of wind energy conversion system has been dissipated in a DC-link by damp resistor controlled through the chopper. In this paper, the control scheme for the back-to-back converter with damp resistor is presented. The simulation studies of considered converter system has been performed and discussed. The results of simulation studies demonstrate the high efficiency and high accuracy of the investigated control technique during the three-phase voltage sags.

1. INTRODUCTION The increased trends of application of the wind energy conversion systems (WECS) have an influence of the operation of the AC grid. Nowadays, the requirements of grid connection of WECS are a very important and trend topic [1, 4, 7, 10]. These requirements describe the behavior of the wind turbine systems and the power flow during the grid fault and disturbance, especially the stability of systems at the connection point. During the low voltage ride-through (LVRT), the power injected to AC grid by the wind turbine system is limited by the voltage reduction. However, due to restrictions, the wind turbine system should remain connected to the AC grid for specific time duration, because the voltage dip will become deeper [1, 9, 10]. __________ * Wroclaw University of Science and Technology, Department of Electrical Machines, Drives and Measurements, ul. Smoluchowskiego 19, 50-372 Wroclaw, Poland, [email protected]

12

Piotr GAJEWSKI

During the voltage sag, the voltage in DC-link will increase to the excessively high value. For this reason, the surplus power in DC-link should be dissipated. The surplus power can be dissipated by application of chopper with damp resistor. Hence, the chopper circuit will absorb the unbalanced power during the grid fault [7, 10]. The aim of this paper is the analysis of the control method for enhancing the LVRT capability of fixed speed WECS during the three-phase symmetrical voltage sag.

2. CONTROL OF BACK-TO-BACK CONVERTER SYSTEM The considered back-to-back converter system with permanent magnet synchronous generator (PMSG) and control circuits has been presented in Figure 1. In this configuration the direct-driven PMSG is connected to the power system through the fully controlled back-to-back converter system. The back-to-back converter system consists of Machine Side Converter (MSC), Grid Side Converter (GSC), grid filter and chopper with damp resistor. The detailed description of mathematical model of PMSG and wind turbine model are presented in the following paper [2, 3, 5, 6, 8].

i ga i gb i gc Lg , Rg

isa i sb i sc

v ga

vgb

Cd

vgc

Rd

ωm

θe

encoder

θe

ωm

Te*

i ga

i gb i gc

i sa* i sb* i sc*

vdc

* i ga

Vabc=

vgabc

(

)

2 2 v +v2 +v2 3 ga gb gc

Va V b V c

* i gb

isd* =0

* I grid

* i gc

2 3n pψ PM

ω m*

vga vgb vgc

Sa Sb Sc

Sa Sb Sc i sc i sb isa

i sq*

Pdc

idc

vdc

vdc * v dc

Fig. 1. Control diagram of the wind energy conversion system with direct-driven PMSG

The main function of MSC is to control the generator speed and the electromagnetic torque of PMSG in order to extract the maximum power from the wind. To obtain the maximum power from the wind, the special algorithm of Maximum Power Point Tracking (MPPT) should be applied. The control scheme of MSC consists of two

13

Analysis of the operation of direct-driven wind energy conversion system during voltage sags

control loops. The outer control loop regulates the generator speed ωm to follow the optimum speed of wind turbine ωm* at which the operation of maximum power point is fulfilled. The reference speed ωm* can be determined by using power speed curve of generator [8]. The instantaneous power in DC-link is obtained by measuring the DC-link vdc voltage and DC current idc. The reference speed is calculated as:

ωm* = 3 Pdc / K opt

(1)

where: Pdc –DC-link power, Kopt – coefficient of wind turbine, which can be calculated as [3, 9]:

K opt = 0.5

C p max

λ3opt

ρπR 5

(2)

where: ρ – air density, Cpmax – maximum power coefficient of the wind turbine, λopt – optimal tip speed ratio, R - radius of the turbine blade. The output value from PI speed controller determines the reference value of electromagnetic torque Te* of PMSG. The reference component of stator current vector isq* can be calculated as [3, 8]:

2 Te* i = 3(n pψ PM ) * sq

(3)

where: np – number of pole pairs of PMSG, ψPM - flux linkage established by the permanent magnets. In the control strategy, it is assumed that, the component isd of the stator current vector is set to zero in order to obtain the maximum torque at minimum stator current [9]. The results control signals determine the reference stator currents isabc* of PMSG. These reference currents are compared with the three phase measured stator currents isabc of PMSG. The error signals are then sent to three hysteresis current controllers. The control signals are the required switching signals for MSC. The control objective for Grid Side Converter (GSC) is to stabilize the DC-link voltage between MSC and GSC and to send the maximum power generated from PMSG to the AC grid. The control scheme of GSC consists of two control loops. The outer control loop with PI controller regulates the DC-link voltage vdc. The output value from this PI controller determines the reference value of magnitude Igrid* of grid phase currents. The reference magnitude of grid currents are then multiplied by the signal units Va, Vb, Vc obtained from supply in-phase voltages of AC grid. These signals are the reference grid phase currents igabc* for inner control loop. This condition allows to achieve the unit power factor operation. The difference of three phase reference currents igabc* and measured grid currents igabc are sent to hysteresis current con-

14

Piotr GAJEWSKI

trollers. The required switching signals for Grid Side Converter are generated through hysteresis current controllers. The operation of the GSC is directly affected by the grid voltage sag, where the power delivered to AC grid by the wind turbine system is restricted. During the voltage sag, the wind turbine and PMSG are operated at its normal condition. Therefore, the voltage in DC-link may increase to the excessively high value [2, 10]. For this reason, the surplus power during the voltage sag should be dissipated by application of damp resistor controlled through chopper. The chopper is active only when DClink voltage is above the rated value. When the DC-link voltage will increase to the value of 1.05vdcref the control signal will active the IGBT transistor and the chopper is turned on. Then the surplus energy is short circuited through the damp resistance Rd.

3. SIMULATION RESULTS The proposed control strategy of wind energy conversion system with PMSG during voltage sag has been simulated using MATLAB/Simulink. Digital situation studies were made for the system with wind turbine data and parameters: rated power Pt 20kW; rotor radius R -4,4m; air density ρ -1,225 kg/m3 and 3-phase PMSG data and parameters: rated power Pg -20kW; stator resistance Rs -0,1764 Ω; stator dq-axis inductance Ld, Lq - 4,48mH; rated speed nn -211rpm; stator rated phase current Isn -35 A. The conducted simulation results of considered wind energy conversion system are presented in Figure 2-5. Figure 2a shows the variation of wind speed during 1.5s simulation. Figure 2b shows the measured angular rotor speed ωm of PMSG and optimal speed obtained ωopt from MPPT algorithm. The proper operation of MPPT algorithm is shown in Figure 3a. The tip speed ratio λopt and power coefficient Cp are maintained at reference values. The responses of the stator current vector components isd, isq have been presented in Figure 3b. The daxis stator current vector component isd is set to zero. The three phase grid voltage during the voltage sag has been presented in Figure 4a. The following data are considered for the aim of simulation: the drop of three phase grid voltage is equal to 50% of the rated value voltage and the duration time is equal to 0.15s. The voltage sag starts at 0.7s. In 0.85s the grid voltage starts to be recovered. The waveform of DC-link voltage has been presented in Figure 4b. The instantaneous values of voltage are quite constant at variations of wind speed. The DC-link voltage has been increased, when the voltage sag occurs. However, the proper operation of damp chopper can be observed in 0.7-0.85s during the voltage sag. The application of damp resistor allows to consume the extra energy in DC-link and allows to avoid overvoltages and overcurrents.

15

Analysis of the operation of direct-driven wind energy conversion system during voltage sags

The waveforms of instantaneous active pg and reactive power qg delivered to AC grid have been presented in Figure 5a. The GSC is generally operated at unit power factor. This means that the reactive power has been maintained at zero values qg=0. From this Figure it can be stated that, the grid voltage sag caused the drop of instantaneous active power pg, while the instantaneous reactive power is kept at zero value. The waveform of three-phase grid currents igabc has been shown in Figure 5b. From this Figure it can be noticed, that the three-phase grid currents rise when voltage sag occurs. However, the currents do not exceed the rated value of the GSC current. After clearance of the voltage sag, the wind energy conversion system has been returned to normal operation. 12 25

10 20

8

15

6 0

0.5

1

1.5

10 0

0.5

1

1.5

Fig. 2. Waveforms of: a) wind speed vw; b) reference speed ωopt and measured speed ωm of PMSG 15 20 10

0 -20

5

0 0

0.5

0.6 0.4 0.2 0 1.5

1

-40 0

0.5

1

1.5

Fig. 3. Waveforms of: a) tip speed ratio λopt and power coefficient Cp; b) stator current vector components isd, isd 720 600 710 400 700

200 0

690

-200

680

-400 0

0.5

1

1.5

670 0

0.5

1

1.5

Fig. 4. Waveforms of: a) three phase grid voltage vgabc; b) DC link voltage vdc 4

x 10

50

0

-1 0 -2

-3 0

0.5

1

1.5

-50 0

0.5

1

1.5

Fig. 5. Waveforms of: a) instantaneous active pg and reactive power qg; b) three phase grid current igabc

16

Piotr GAJEWSKI

4. CONCLUSIONS In this paper the control scheme of wind energy conversion system with PMSG has been proposed in order to fulfill the LVRT requirements. The three phase symmetrical grid voltage sag has been discussed in this paper. The surplus power during the voltage sag has been dissipated by damp resistor controlled through the chopper. For this reason, the wind turbine system with PMSG can remain connected to the AC grid during the voltage sag. Simulation results show that the proposed control technique of chopper with damp resistor can enhance the stability and reduce the increase of DC-link voltage during the system failure. The simulation results also show no overcurrent of grid phase currents during the symmetrical voltage sags. REFERENCES [1] DONG S., LI H., WANG Y., Low voltage ride through capability enhancement of PMSG-based wind turbine, International Conference on Sustainable Power Generation and Supply (SUPERGEN), 2012, 1-5. [2] ERRAMI Y., MAAROUFI M., OUASSAID M., Modeling and control strategy of PMSG based variable speed wind energy conversion system, International Conference on Multimedia Computing and Systems (ICMCS), 2011, 1-6. [3] GAJEWSKI P., PIEŃKOWSKI K., Analysis of a wind energy converter system with PMSG generator, Czasopismo Techniczne, 2015, Elektrotechnika Zeszyt 1-E (8), 2015, 219-228. [4] IBRAHIM R.A., HAMAD M.S., DESSOUKY Y.G., WILLIAMS B.W., A review on recent low voltage ride-through solutions for PMSG wind turbine, International Symposium on Power Electronics, Electrical Drives, Automation and Motion (SPEEDAM), 2012, 265-270. [5] KUN H., GUO-ZHU Ch., A novel control strategy of wind turbine MPPT implementation for directdrive PMSG wind generation imitation platform, IEEE 6th International Conference on Power Electronics and Motion Control, IPEMC, 2009, 2255-2259. [6] MOLINA M.G., SANCHEZ A.G., LEDE A.M.R., Dynamic modeling of wind farms with variablespeed direct-driven PMSG wind turbines, Transmission and Distribution Conference and Exposition: Latin America (T&D-LA), IEEE/PES, 2010, 816-823. [7] SHUJU H., JIANLIN L., HONGHUA X., Modeling on converter of direct-driven WECS and its characteristic during voltage sags, IEEE International Conference on Industrial Technology, ICIT, 2008, 1-5. [8] SINGH M., CHANDRA A., Power maximization and voltage sag/swell ride-through capability of PMSG based variable speed wind energy conversion system, 34th Annual Conference of IEEE Industrial Electronics, IECON 2008, 2206-2211. [9] WU B., YONGQIANG L., NAVID Z. SAMIR K., Power conversion and control of wind energy, John Wiley & Sons, INP., Publication 2011. [10] YAN L., YONGNING Chi., ZHEN W., LINJUM W., CHAO L., Study on LVRT capability of D-PMSG based wind turbine, IEEE Power Engineering and Automation Conference (PEAM), 2011, 154-157.

17

Key words:battery storage systems, copper demand, renewable enegy systems

Justyna HERLENDER *

ANALYSIS OF COPPER DEMAND FOR ASSEMBLING BATTERY STORAGE SYSTEMS

The paper begins with the explanation why energy storage systems are now an important issue in European Union. It presents battery storage systems as a possible technology for implementation in electric grid applications. Area of interest is especially directed to the analysis of one basic material – copper which is used to build batteries and additional devices dedicated for storage systems. Moreover, this paper describes forecasted increase of new renewable energy installations based on wind and sun radiation in upcoming years and also the potential of energy storage systems in EU. Possible scenarios of battery storage systems development in the next years in EU are presented as well. The paper ends with the analysis of possible changes in copper demand in years 2020-2050, which aimed to fulfill European Union RES Energy Target.

1. INTRODUCTION The main aim of installing energy storage systems to the power grid is to gather and keep overproduced amount of energy (during off-peak times) and be able to use it during times when it is actually needed. Such systems can basically balance the discrepancy between energy supply and its demand. Secondly, electricity produced from renewable energy sources such as wind or solar radiation plays an increasing influence on the overall energy balance of the world. However, now the green energy producers have to face some challenges. The main is related to ensuring a stable, continuous production without its characteristic moments of deficiency and overproduction. It means that the influence of natural conditions (windless periods and lack of sunlight) has to be overcome. The solution can encompass integration of battery storage systems. __________ * Department of Electrical Power Engineering, Wroclaw University of Science and Technology, Poland

18

Justyna Herlender

2. COPPER DEMAND IN BATTERY STORAGE SYSTEMS The grid scale storage system is comprised of various elements. Obviously in case of Battery Energy Storage Systems (BESS), battery is a crucial part of storage system, but for proper operation of the system some additional components are needed, such as: transformer, corresponding power converter, switches, breakers, control system, wiring and also container to locate batteries. The main material which is used in this system is copper (Cu), because every component of system contains copper in its construction. The main large presently used applications of storage systems with their average copper intensity are [1]: • Integration of renewables (0.3-3 tons Cu/MW) • Community Energy Services (4 tons Cu/MW) • Other distributed (0.04 tons Cu/MW) It is possible to indicate the data concerning the intensity of copper in storage unit. The amount of copper differs according to the battery technology. In Tab.1 the results of Cu demand per 1MW storage unit with comments in which part it is used are presented. Tab.1. Copper in chosen storage unit [1]. Technology Cu[tons/MW] where

Lithium-ion Sodium 0.08 0.01 limited or no use in battery pack

Flow 0.27

Lead Acid 0.25 battery pack

Nickel 0.22

The demand of copper in storage installation varies according to the considered application. The list of these parameters is shown in Tab.2. Tab.2. Copper demand [tons/MW] in battery storage installations [1]. Ancillary Serv. Li-ion 2.7

Li-ion 4.4

CES PbA 4.2

Ni 4.3

Renewables Li-ion PbA 2.6 1.4

Transmissions Ni Na Flow 1 3 2.5

Other Li-ion PbA 3.5 0.4

3. MATERIALS DEMAND FOR REACHING EU RES TARGET 3.1. ELECTRICITY GENERATION AND INSTALLED POWER CAPACITY IN EU 2020 – 2050

The EU’s electricity demand has been rising and the situation will be the same during next years. Due to the increasing of electricity demand in upcoming years it is obligatory to install additional, new power capacity in whole EU. The total increase in electricity production by 2050 exceeds 40% in comparison with 2005. It is assumed

19

Analysis of needed copper for assembling battery storage systems

that almost 50% of this energy will be produced using renewable energy systems (RES) [2]. The rise of wind energy generation during this time is more than 20% and the contribution of this generation will achieve even 26% by 2050. Taking into account generation from PV, the aim is to obtain finally in 2050 - 9%.

Fig.1. New net installed solar and wind capacity 2020-2050.

Fig.1 concerns the new installed capacity based on the wind and solar energy. The total wind capacities rise from 85 GW in 2010 to 413 GW in 2050 [3]. More than 100 GW is planned to be added till 2020. The plans include that more than 25% of wind installations there will be off-shore wind farms. The PV capacity is much smaller than wind capacity. In 2010 there was only 30 GW of installed photovoltaic units. Assuming that by 2050 there will be 231 GW of installed PV capacities [3], it means that the growth is significant. 3.1. STORAGE TECHNLOGIES CURRENTLY INSTALLED

Presently most of installed storage units is based on the Pumped Hydro (PHES). More than 95% of currently installed storage systems in EU are based on the technology mentioned above [4]. The number of energy storage systems already installed in EU does not exceed 5% of total installed capacity and it is almost based on the pumped hydro storage systems located mainly in mountains areas. If only battery storage systems are taken under consideration, it can be shown which of technologies are used the most frequently. The outcomes are presented in Fig.2.

Fig.2. Battery technologies used for electrical energy storage systems.

20

Justyna Herlender 3.2. NEW STORAGE SYSTEMS IN EU IN THE YEARS 2020-2050

Energy storage systems enable to supply the network more flexible, balance the grid and what is very important can be used as a back-up in intermittent renewable energy. According to the growth of renewable system in EU, the more energy storage system will be installed in upcoming years [5]. The three different trends of developing energy storage systems in EU by 2050 were considered. The first and second are based on predictions provided in the ETP 2014 scenarios published by IEA in Technology Roadmap - Energy Storage, 2014 [6]. In the third scenario it is assumed that in the 28 EU’s countries by 2050 the power capacity of installed storage systems will equal approximately 30% of total power capacity from RES.

Installed capacity [GW]

200 150 153

100 50 50

25

0 2010

2050

2050

2050

Fig.3. Installed storage capacity for electricity storage in EU by 2050.

In 2010 there was 40 GW of storage systems installed in EU. Fig. 3 indicates new planned storage installations due to the three chosen scenarios. According to this scenarios 4-11% of total produced electrical energy will be stored. Relating to scenario 1 and 2, share of storage system in 2050 will be at the same level as today. Obviously, due to increasing of energy demand, new storage system will be build. 3.3. NEW BATTERY STORAGE SYSTEMS IN EU 2020-2050

Results presented above point out how many of new storage systems will be required in upcoming years. The next issue relates to the technologies which can be used to build future storage systems. As it was mentioned previously in EU 40 GW of storage systems in 2010 (99% of all storage systems) was installed. It is also claimed that 75% of the total potential for hydropower in EU has already been developed [7]. That is the chance for BESS. In accordance to this information to further consideration it is assumed that all of remaining hydropower potential will be used by 2050. It means that in EU there will be about 54 GW of hydropower storage systems. The rest of new storage installation will be based on other technologies. The next assumption concerns that share of flywheel storage systems will be on the fixed level equal 3% of total installed storage instead of pumped hydro. Due to this assumptions, three different forecasts for developing of BESS were implemented.

21

Analysis of needed copper for assembling battery storage systems

The first was based on current share of particular storage technologies. It was assumed that till 2050 there would not be any changes in application of the most developed technologies. In the second, share of individual storage technologies instead of pumped hydro would remain the same. The last forecast was different because of the prediction that in a future sodium - sulfur battery (NaS) would be used in 45-50% of storage systems, excluding pumped hydro technology. According to numbers, in a future NaS battery storage technology may be more often used as any other type of energy storage except for PHES [8]. In Tab.4 maximum new power capacities of BESS obtained from all scenarios which would be installed during the next three decades in European Union are presented. Tab.4. Installed power of new storage systems by 2050 BESS Sodium-sulfur Lithium-ion

Installed power [GW] 69.500 13.900

BESS Lead-Acid Nickel Cadmium

Installed power [GW] 9.730 4.170

Based on the obtained results it can be concluded that if in the next years the current trend in constructing new storage installation would be the same, the growth in new BESS will be insignificant. Although, if this limitation of using hydropower for new storage systems is valid, especially the applications of NaS battery will increase. 3.4. Copper demand in battery storage systems in EU in the years 2020-2050

Based on the proposed scenarios and data presented in the section 2, it can be estimated how much copper should be used to fulfill individual scenarios. The most significant results are shown in the Tab. 5. Tab.5. Copper needed to build new battery storage systems in EU in years 2010 - 2050. Battery Cu[tons]

NaS 208500

Li-ion 36140

Lead-Acid 13622

Nickel-Cadmium 4170

Flow Batter. 3475

Total max. 232200

If the trend would be kept the amount of copper used for battery storage systems will be insignificant in comparison with total world copper demand. The highest copper demand would be noted for sodium - sulfur battery storage systems. Considerations were focused on copper demand for storage systems dedicated for RES. However, in order to use it, firstly wind and solar power plants have to be built. For 1MW offshore wind power unit 2.8 tons of copper is needed, for onshore – 1.8, new PV installation requires even 5 tons of Cu. According to this numbers and data from Fig.1 it can be assumed that 658,000 and 1.040,000 tons of copper for wind and solar installation will be needed till 2050, respectively.

22

Justyna Herlender

From the presented results it is seen that the amount of copper which could be used to build battery storage systems is significantly lesser than the copper demand for construction of new wind and solar power units. 4. CONCLUSIONS The storage market for integration of renewables is supposed to grow, due to the fact that the number of new intermittent renewable grid installation increases. Policy targets of European Union and other incentives are promoting green generation investments. That will affect on the further look of energy storage market dedicated for RES integration. Many factors indicate that in upcoming years the battery storage technologies will constitute majority of storage installations. To obtain intended growth of battery storage systems, firstly, many challenges have to be defeated. From the technological point of view: capacities, efficiencies and life time of batteries have to increase. Another problem concerns creation of appropriate market regulations which will be necessary for storage application on a grid level. Above all, the key thing is to reduce the cost of storage systems. This is the factor which could block fast development of battery storage technologies in next years. Many different materials are necessary for building BESS. Some of raw materials are dedicated only for individual storage system, depending on the applied battery. Materials as copper, steel or plastic are used independently on the chosen battery. As it was presented, that is enough copper to fulfill the most optimistic scenario of battery storage systems development. LITERATURE

[1] [2] [3] [4] [5] [6] [7] [8]

KEMA, Inc. Fairfax, Market Evaluation for Energy Storage in the United States, Virginia, 2012 EUROPEAN COMMISSION, Energy Roadmap 2050: Impact assessment and scenario analysis, Brussels, 15.12.2011, SEC(2011) 1565 final EUROPEAN COMMISSION, EU Energy, transport and GHG emissions, trends to 2050,Reference Scenario December 2013 UGARTE S, LARKIN J., ECONOMIC AND SCIENTIFIC POLICY, Energy Storage: Which Market Designs and Regulatory Incentives Are Needed, Study for ITRE Committee, October 2015 EUROPEAN COMMISSION, DIRECTORATE-GENERAL FOR ENERGY, DG Ener Working Paper, The future role and challenges of Energy, 2013 IEA, Energy Storage Technology Roadmap, March 2014 EUROPEAN COMMISSION, Technology Information Sheet – Electricity storage in the power sector, 2011 SUN S, Grid - scale energy storage: state of the market, Bloomberg Energy Finance Summit, 2011

23

Key words: differential, fatigue fracture, carburizing, notch

Mikołaj KATKOWSKI, Aleksandra KRÓLICKA, Mateusz JAKOWSKI*

HYPOTHESIS CONCERNING DAMAGE OF RACING CAR DIFFERENTIAL

The paper deals with causes that lead to damage of gear in sport car differential (power above 300 HP). Loads and environment which occurred during exploitation caused fatigue fracture. There are few reasons of that, invalid material, miscalculation in strength, geometry or technology (during production). Research was conducted using light and electron microscopy. Hardness was measured using Vickers method in many places on the cut surface of broken gear. This research allowed to draw hypothesis of damage.

1. INTRODUCTION There are a lot of reasonscausing damage of machines and its parts which origins are different and mainly come from incorrect exploitation.However, there are some failure really hard to predict and they eventually drive to disaster. Identification of all of the aspects of machine’s working condition and any potential risk which hides behind them helps avoiding damage of particular and single part and the whole device as well. The pinion of driveshaft was taken under evaluation which transmits torque from the main car’s gearbox to the back axis differential (fig. 1). The car logged around 15 000 kilometers – which is very short period of time referring to terms of the warranty – and then the decohesion of three cogs of driveshaft happened and consequently caused damage of hypoid gear assembly and removal of car from exploitation. The examined part comes from WRC racing car propelled by high power combustion engine (around 380 HP/ 283 kW power).The vehicle was manufactured by one of the well-known Japanese automotive concerns.

Wroclaw University of Technology and Science, Faculty of Mechanical Engineering, I. Łukasiewicza 5, 50-371 Wroclaw

24

Mikołaj KATKOWSKI, Aleksandra KRÓLICKA, Mateusz JAKOWSKI

Figure 1. The differential gearbox assembly [1]

The differential works in good lubrication condition oiled by gearbox oil which creates slim oil film on contact surfaces of tooth what decreases friction and in result reduces wear and minimizes power loss. This kind of environment protects from rustirustiness. The main load performing on gear are bending of teeth, shearing at its footing and contact stress on the active surfaces. These all derive from torque transmitted between gears. The part should have high durability due to cyclic and nonsymmetrical character of loading.

2. FRACTURE SURFACE On the active surfaces no unstable tribological wear was noticed what unambiguously shows correct working condition and adequate cogs’ hardness. There were no untribological aspects of wear e.g.: corrosion, cavitation or erosion. Three neighboring cogs broke and their fracture represents typical fatigue character (fig. 2 and 3). What is more, there were seen big area, relating to the whole fracture, with distinct lines of crack propagation (fig. 3). In case of other teeth which did not break, it was stated that there are cracksat their base which ascend around them. This phenomenashows that all of the cogs were facing fatigue crack propagation, hence, the life of the other teeth would not have been longer than the damaged ones.

Figure 2. Whole hypoid gear

Figure 3.Selected fractured tooth

25

HYPOTHESIS CONCERNING DAMAGE OF RACING CAR DIFFERENTIAL

3. ASPECTS OF FATIGUE The gear of driveshaft of the differential worked in the range of low-cycle fatigue (fig. 4) and was damaged due to sum of all cyclic stress amplitudes according to Palmgren-Miner’s rule (eq. 1). According to this rule, sum of damages is linear, proproportional to number of cycles what presents the equation below:



,

(1)

where Ni is number of cycles to damage for particular stress amplitude and constant mean stress, niload cycle sequence for particular amplitude.If DPM rate is higher or equal to 1 the element is going to be damaged.

Figure 4. Ranges of fatigue: fatigue I – quasi-static, II – low-cycle, III – high-cycle

As the damage has character of low-cycle fatigue, there was probably and locally exceeded yield strength due to curb at the bas basee of tooth. Material stops being elastic, so that Hook’s law does not work and it starts nonlinearly hardening. Therefore, Therefore, there should be implemented Neuber’s rule if the case were analytically investigated. What is more, the final life of the gear should be assumed according to Coffin-Manson curve which concerns low-cycle phenomenon unlike Woehler’s curve which concerns highhighcycle fatigue. On the other hand, the load has asymmetrical character, hence, here should be utiutilized mean stress correction according Smith-Watson-Topper, Ince-Glinka orMorrow.

26

Mikołaj KATKOWSKI, Aleksandra KRÓLICKA, Mateusz JAKOWSKI

4. RESEARCH METHODOLOGY The polished sections were prepared in order to identify microstructure and possible structural defects. Samples were taken from selected tooth which is claimed to be the most loaded on the basis of macrostructural observation.Normalizing was conducted to evaluate heat treatment (carburizing) using metallographic method. Observ Observaation of microstructure were done using optical microscope Nikon Eclipse MA200. To capture of figures camera Nikon CCD Ds.-Fi1 was utilized. The microstructure investigation were covered using scanning electron microscope FENOM detector BSE. BSE.The The Hardness was measured at particular locations on damaged zone of cog using Vickers method. In the end, the chemical composition of steel was checked using optical specspectrometer.

5. RESULTS The analysis of chemical composition showed that driveshaft’s gear was made of low-alloy steel. Measurement of hardness in figure 5 and 6.

Figure 5. Vickers’s hardness at points of measurement according to fig. 6

Figure 6. Course of hardness measurement

Figures 7-11present microstructure of particular zones of cog and show evaluation carburized area using metallographic method.

27

HYPOTHESIS CONCERNING DAMAGE OF RACING CAR DIFFERENTIAL

Figure 7. Core’s microstructure, lath martensite. Seen secretion of carbides and grain boundaries. SEM, detector BSE, FENOM, zoom x930

Figure 10. Active side microstructure after normalizing, high carbon perlite on the outside and ferrite in the core. Very distinct carburizing zone smoothly transforming into core,Nikon zoom x100

Figure 8. Microstructure of carbonized zone, martensite of mixed morphology, mostly lath martensite. SEM, detector BSE, FENOM, zoom 5600x

Figure 11.. Crack from outer Surface to core, what may be caused by quenching. Length of the crack around 180 µm. SEM, detector BSE,FENOM, zoom 2059x

28

Mikołaj KATKOWSKI, Aleksandra KRÓLICKA, Mateusz JAKOWSKI

6. CONCLUSION Hypothesis of damage is not unambiguous because there were plenty of damage factors which simultaneously affecting the element caused fracture of three differential’s cogs. One of them are teeth with curbs that remained after grinding process – here should be implemented corrected machining to obtain better surface finish and bigger radius at the basis what will help to avoid local stress accumulation. On the other hand, there is seen, at low magnification, roughness of surface at the base’s round of cog – the surface should be polished. The better quality of the surface, the higher life expectation is. The crack at round was created due to residual stress after heat treatment – carburizing. The crack was fulfilled with oxides. Optimization of heat treatment process is recommended. The investigation exhibited that the gear was manufactured using unusual method. Idle and active surfaces performs asymmetrical because active side always presents wider carburizing zone. It can be claimed that process of carburizing was conducted thanks special method e.g.: using special paste which cover selected region of cog. To sum up, it can be claimed that considering mechanical properties the material for gear has enough hardness for teeth. It represents adequate hardness and Ultimate Tensile Strength. On the other hand, material does not fulfil technological criteria e.g.: heat treatment, curbs after grinding, low quality of surface at base of tooth where it is expected the most loaded place of the gear due to complex form of stresses.

REFERENCES [1] LEE Y-L., PAN J., HATHAWAY R. B., BARKEY M. E.,Fatigue testing and Analysis; Elsevier, 2005 [2] DZIAMA A., Przekładnie zębate, PWN, Warszawa 1989 [3] KRAJCZYK A., Podręczny atlas mikrostruktur metali i stopów, Oficyna Wyd. PWr, Wrocław 2005

29

Key words: TEM, electron diffraction, light field, dark field, isothermal hardening

Aleksandra KRÓLICKA*, Natalia BIAŁA*, Paulina BIAŁA*, Andrzej ŻAK*, Łukasz KONAT*

USAGE OF TEM METHODS FOR PHASE FOR RESEARCH OF ISOTHERMAL TREATED BORON STEEL

The article contains the basic description of electron transmission microscopy methods. The analysis of composition and the operating rules of electron transmission microscope was focused on more beneficial it’s application in comparison to light microscope or scanning microscope. The work presents complex preparation of thin foil, manufactured after conducting detailed heat treatments: the method of indicating of electron diffraction patterns and the possibilities created by the application of TEM methods in the identification of phases and structures occurring in the research material.

1. INTRODUCTION Nowadays, the electron microscopy has wider and wider practical application in all science disciplines i.a.: technical, natural and in the medical diagnosis. Due to its research possibilities, it is broadly applied in the observation of various materials such as biological, metals, alloys and other formulas. The electron microscopy became the crucial technique allowing for observing the structures and materials. Due to the use of transmission electron microscope (TEM), it is possible to show the various elements of the structure, such as: dislocations, borders of grains, failures of positioning, stress fields and coherent release with warp. In the electron microscopy the possibility of connecting the microstructure observations with the crystallographic information which is contained by the diffractive pictures of the particular structure fragment. The samples used for the observation in the electron microscopes must be appropriately prepared. It results from the fact of high vacuum present in the microscope and effect of the electron beam on the sample which are used to create the image. The transmission electron microscopy gives the vast possibilities in researching of materials as *Wroclaw University of Technology, Faculty of Mechanical Engineering, Łukasiewicza 5,50-371 Wrocław

30

Aleksandra Królicka, Natalia Biała, Paulina Biała, Andrzej Żak, Łukasz Konat

well as in their structure and broadening the knowledge concerning the changes occurring in them. The basic advantage of electron microscopy over the traditional light microscopy is the possibility to obtain much greater magnifications. The table 1. presents the overview of light and electron microscopies in reference to the particular criteria. Table 1. The comparison of light and electron microscopy

1.1. TRANSMISSION ELECTRON MICROSCOPE

Electron Transmission Microscope is the device in which the electron beam is created and formed in order to x-ray the formulation. The main task of transmission electron microscope is creating the zoomed image of the x-rayed formulation, the image created by the lenses and the beam going through or by the secondary electron detector, backscattered electron detector (BSE) or scanning transmission electron microscopy (STEM). In the transmission microscope, the image is created by the reacting of sample with the electron beam which results in creating the contrast. The possibility of creating the images is dependent on the thickness of sample. Among the methods, TEM shall distinguished the possibility of registering the image in both light and dark fields. The image in the bright field is created by the beam of electrons being perpendicular to the formulation and being partially dispersed. The image visible in the screen obtains contrast because the areas where the electrons were dispersed are visible as darker and other as lighter. The image in the dark field is generated by one of beams which was scattered. Its characteristic feature is high contrast. In the image in the dark field, the white areas can be observed in the dark surrounding. These areas come from

31

USAGE OF TEM METHODS FOR PHASE RESEARCH OF ISOTHERMAL TREATED BORON STEEL

the scattered beam i.e. from the chosen beam which went through diffraction on the phase [2,4].

2. ELECTRON DIFFRACTOGRAMS The diffraction is curving the electron beam being absorbed by the material of the crystal structure. The part of electrons which go through the formulation experience the diffraction and create a certain number of curved beams. Other electrons create the non-curved beam. The diffraction allows to obtain the diffraction patterns i.e. the diffractive images which present not only the most curved electron beam but also the number of spots (signs of curved beams). The location of reflexes in reference to the sign of curved beam (central point) indicates the orientation of crystallographic surfaces. Conducting of the diffractive research allows to create the possibility of phases identification existing in the particular formulation [1,2]. 2.1. INDICATING OF ELECTRON DIFFRACTION PATTERNS

Indicating is the procedure which allows to determine which atomic surfaces the particular reflexes or rings occurring on the diffraction patterns come from. The procedure is based on selecting the elementary cell of diffraction pattern, measuring the distance of reflexes creating the cell from the center of diffraction pattern (point 0,0,0) and the angles between the vectors r1, r2, r3 (fig. 3.) Furthermore, they shall be compared with the values of distances between the atomic planes, which are obtained from the appropriate data bases. The solution of such diffraction patterns allows to identify the phases occurring, determine the crystallographic orientation of each crystallite in reference to the incident beam as well as the crystallographic orientation between the crystallites [2].

Figure 3. The description elements of monocrystalline diffraction pattern and The description elements of monocrystalline diffraction pattern [2]

32

Aleksandra Królicka, Natalia Biała, Paulina Biała, Andrzej Żak, Łukasz Konat

3. RESEARCH METHODOLOGY

The samples of low-alloy steel obtained from the steel sheet of 10 mm thick were used in the research. The samples were cut by the method which does not change the initial structure of material and had dimensions of 10x10x20 mm. The chemical composition of steel is presented in Table. 2. The heat treatment were carried out in the Department of Materials, Endurance and Welding of Wroclaw University of Technology in the muffle furnaces by Czylok, FCF 12 SM and FCF 12 SHM type. The heat treatment did not cover the usage of additional protective atmospheres. Table 2. The chemical comThe isothermal hardening with tempering in tin was position of low-alloy steel conducted in the temperature of 300°C during 3 hours. with boron (%wt) [4] After conducting the thermal processing, the specialist analysis of thin foil was carried out so as to observe the microstructure in the next stage. The research was performed with the use of the transmission electron microscope, Hitachi H-800. Materials for the study must have a well-defined dimensions, so the preparation was essential. The preparation of samples consisted of embodiment (directly from the test material) of thin foil having a thickness of approx. 80-100μm, which was then cut in the form of discs with a diameter of 3 mm. The ready discs were polished by the electrolyte method with the use of the TENUPOL device of Struers. The images were registered both in the dark and light field. The phase identification was made on the basis of electron diffraction patterns recorded by means of a computer program DYFR, developed in the Department of Materials.

4. RESULTS

The research with the application of electron microscopy allowed to determine clearly the presence of other phases. In the pictures below, the chosen microstructures of low-alloy steel with boron and the electron diffraction patterns are presented including the solutions, used to identify other phases. The images were registered in bright and dark field. Fig. 4-9 present the results of observation confirming the presence of low-carbon martensite of 0,1% and cementite.

33

USAGE OF TEM METHODS FOR PHASE RESEARCH OF ISOTHERMAL TREATED BORON STEEL

Figure 4. The image of low-alloy steel microstructure after isothermal hardening with 3hkeeping. Light field, TEM, zoom app. 12 000x

Figure 6. The image of low-alloy steel microstructure after isothermal hardening with 3h-keeping. Dark field, reflex (2 6 -2) marked in fig 5. TEM, zoom app. 12 000x

Figure 8. Electron diffractogram including the solution from the field of lowalloy steel microstructure presented in the Fig. 7. Martensite 0,1% C, TEM, 150kV

Figure 5. Electron diffractogram including the solution from the field of low-alloy steel microstructure presented in the Fig. 4. Bainitic ferrite, Fe3Mo3C carbide coherent with the metallic warp, TEM, 150kV

Figure 7. The image of low-alloy steel microstructure after isothermal hardening with 3hkeeping. Light field, TEM, zoom app. 20 000x

Figure 9. The image of low-alloy steel microstructure after isothermal hardening with 3h-keeping. Dark field, reflex (0 1 1) marked in fig 8. TEM, zoom app. 20 000x

34

Aleksandra Królicka, Natalia Biała, Paulina Biała, Andrzej Żak, Łukasz Konat

5. CONCLUSION

The project presents selected results of research with the use of the transmission electron microscopy. The application of TEM is definitely undisputed because it clearly determines the phases and structures co-existing in the research material after conducting the variations of thermal processing: isothermal hardening with tempering within 3h. Together with the images in the light and dark fields, the solutions of diffractive images were compared. The computer program DYFR was applied to solve the diffractograms. The research revealed the presence of low-carbon martensite of 0,1%C and down low temperature bainite after isothermal hardening within 3h. Receiving of the results was possible by execution of complicated and time-consuming preparation. It should be noted that the preparation of the research must be carried out according to strict terms and conditions, and the quality of preparation is closely linked with the quality of observation. Currently, the transmission electron microscopy allows to obtain the images of truly high quality, in the magnification reaching up to a few million times which places it on one of the first places when concerning the applied methods of observation and recognizing the microstructures.

REFERENCES [1] WILLIAMS D., CARTER C., Transmission Electron Microscopy. A Textbook for Materials Sci ence, Springer Science+Business Media, New York, 2009. [2] BARABACKI A., Mikroskopia elektronowa, Wydawnictwo Politechniki Poznańskiej, Poznań, 2007. [3] BIAŁA N., Analiza procesu produkcji niskostopowej stali z borem w kontekście wybranych własności strukturalnych, Bachelor/engthesis, Wrocław, 2016. [4] KRÓLICKA A., Wpływ procesu wytwarzania niskostopowej stali z borem na mikrostrukturę i wybrane właściwości mechniczne, Bachelor/engthesis, Wrocław, 2016. [5] DUDZIŃSKI W., Materiały konstrukcyjne w budowie maszyn, Oficyna Wydawcza Politechniki Wrocławskiej, Wrocław 1994. [6] Hardox. Data sheet, SSAB-Oxelösund, 2013.

35

six-phase induction motor, Direct Field-Oriented Control, MRAS estimator, Fuzzy-Logic Control, simulation studies

Jacek LISTWAN*

SENSORLESS DIRECT FIELD-ORIENTED CONTROL OF SIX PHASE INDUCTION MOTOR WITH ADAPTIVE FUZZYLOGIC SPEED CONTROLLER

The paper presents the Direct Field-Oriented Control (DFOC) of six-phase squirrel-cage induction motor with adaptive Fuzzy-Logic speed controller and MRAS speed estimator. The mathematical models of the six-phase squirrel-cage induction motor and the MRAS estimator have been presented. The scheme of adaptive Fuzzy-Logic speed controller has been designed and the tasks of the control blocks of applied adaptive Fuzzy-Logic controller have been described. The analyzed DFOC control system has been shown and explained. The simulation model of the control system has been developed and the simulation studies of the sensorless DFOC with adaptive Fuzzy-Logic speed controller were performed. The waveforms of the motor speed, motor electromagnetic torque, load torque and other electromagnetic variables have been investigated. The results of the simulation studies have been presented and discussed.

1. INTRODUCTION The interest in the multi-phase induction motors is caused by the improved reliability, the possibility of reducing the ripple of the electromagnetic torque and reduction of the power per phase in comparison with the three-phase induction motors. Multiphase machines are usually used in high power applications [2, 3, 5]. The sensorless drive systems have been very popular, because the elimination of the mechanical sensors increases the reliability of the drive systems [4]. In the control system considered in this article, the rotor flux and the motor angular speed are estimated by the Model Reference Adaptive System (MRAS) presented in detail in [4]. The adaptive Fuzzy-Logic speed controller has been applied in the analyzed drive __________ * Wroclaw University of Technology, Department of Electrical Machines, Drives and Measurements, Smoluchowskiego 19, 50-372 Wroclaw, [email protected]

36

Jacek LISTWAN

system with six-phase induction motor. This type of controller can give much better control results in comparison with conventional linear controllers [1].

2. MATHEMATICAL MODEL OF SIX-PHASE INDUCTION MOTOR The mathematical model of the six-phase squirrel-cage induction motor has been formulated on the basis of commonly used simplifying assumptions [2, 3, 5]. The motor model formulated in the stator and rotor phase coordinate system is described by differential equations with changing coefficients. The phase coordinate system can be converted to new system, which is described by differential equations with constant coefficients. The general equations of six-phase squirrel-cage induction motor after transformations take the following form [2, 3, 5]: -the voltage equations of the stator and rotor in the rectangular x-y coordinate system, which rotates relative to the stator at arbitrary angular speed ωk:

u sx = Rs isx − ω k ψ sy + pψ sx

0 = Rr irx − (ω k − ωe )ψ ry + pψ rx

u sy = Rs isy + ω k ψ sx + pψ sy

0 = Rr iry + (ω k − ωe )ψ rx + pψ ry

(1) (2)

-the stator voltage equations in the additional coordinate system z1-z2:

u sz1 = Rs isz1 + pψ sz1

u sz 2 = Rs isz 2 + pψ sz 2

(3)

-equation of the electromagnetic torque:

Te =

6 pb (ψ ry isx −ψ rx isy ) 2

(4)

where: usx, usy - components of the stator voltage vectors in the x-y coordinate system; usz1, usz2 - components of the stator voltage vectors in the additional z1-z2 coordinate system; isx, isy, irx, iry - components of the stator and rotor current vectors in the xy coordinate system; isz1, isz2 - components of the stator current vector in the additional z1-z2 coordinate system; ψsx, ψsy, ψrx, ψry - components of the stator and rotor flux linkage vectors in the x-y coordinate system; ψsz1, ψsz2 - components of the stator flux linkage vectors in the additional z1-z2 coordinate system; ωk - arbitrary angular speed of the coordinate system relative to the stator; ωe - the electrical angular speed of the motor; Te - the motor electromagnetic torque; Rs, Rr - stator and rotor phase resistance; pb - the number of motor pole pairs; p=d/dt - derivative operator.

37

Sensorless Direct Field-Oriented Control of six-phase induction motor with adaptive Fuzzy-Logic speed controller

3. ADAPTIVE FUZZY-LOGIC CONTROLLER The scheme of the considered Adaptive Fuzzy-Logic Controller (AFLC) has been presented in Figure 1 [1].

wi ai

ai

Fig. 1. The scheme of the adaptive Fuzzy-Logic controller

The Adaptive Fuzzy-Logic Controller consists of five layers [1]: Layer 1 is responsible for calculation of two control signals. The first control signal e is the difference between the reference and estimated motor speed. The second control signal Δe is the time derivative of the first input signal. In the Layer 2 the output signals from Layer 1 are fuzzified with the use of triangle-shaped membership functions. The multiplication of the appropriate output signals from Layer 2 is carried out in Layer 3 of the FLC. Layer 4 is responsible for multiplication of the activation levels of the rules and weight coefficients. The values of weight coefficients are updated according to the adaptation procedure [1]. Layer 5 is responsible for calculation of the output signal u from FLC. The output signal is calculated according to the method of Center of Gravity. 4. MATHEMATICAL MODEL OF THE MRAS ESTIMATOR The scheme of the Model Reference Adaptive System estimator is presented in Figure 2 [4].

38

Jacek LISTWAN

Fig.2. The scheme of the MRAS estimator

The model of the MRAS estimator contains of two mathematical models: the mathematical model of the simulator based on the stator circuit model (the voltage model) and the mathematical model of the simulator based on the rotor circuit model (the current model) [4]. In the considered MRAS estimator the voltage model is treated as the reference model. The estimated components of the rotor flux vector in the α-β coordinate system are given to the adaptation mechanism with the PI controller. The output signal from PI controller is the estimated value of the motor speed (ωeest). The value of the estimated motor speed retunes the current model. The estimated value of the motor speed is calculated according to the equation [4]: ωeest = K P (Ψrαc Ψrβv − Ψrβc Ψrαv ) + est

est

est

est

1 Ti

∫ (Ψ α t

r c

est

Ψrβv

est

est

− Ψrβc Ψrαv

est

)dt

(5)

0

where: ψrαcest, ψrβcest - the estimated components of the rotor flux linkage vectors obtained from the current model in the α-β coordinate system; ψrαvest, ψrβvest - the estimated components of the rotor flux linkage vectors obtained from the voltage model in the α-β coordinate system; KP, TI - the coefficients of the proportional and integral parts of the PI controller, respectively. 5. SENSORLESS DIRECT FIELD-ORIENTED CONTROL SYSTEM WITH ADAPTIVE FUZZY-LOGIC SPEED CONTROLLER The scheme of the Direct Field-Oriented Control (DFOC) of six-phase squirrelcage induction motor with adaptive Fuzzy-Logic speed controller and MRAS speed estimator is shown in Figure 3. The MRAS estimator determines the estimated value of motor angular speed (ωeest) and the instantaneous magnitude of the rotor flux vector (ψrest). In the considered DFOC control structure four control loops are applied: the control loop for motor angular speed with adaptive Fuzzy-Logic controller, the control loop for magnitude of the rotor flux vector with PI controller and two control loops for x and y component of stator current vector. The reference values of the stator voltage vector components determined in the x-y coordinate system are transformed to the α-β coordinate system and are given to the Space Vector Modulation (SVM) block. The Space Vector Modulator sets the switching states of the six-phase Voltage Source Inverter. The applied SVM sets the reference voltage vector by using

39

Sensorless Direct Field-Oriented Control of six-phase induction motor with adaptive Fuzzy-Logic speed controller

the appropriate switching times of two long voltage vectors situated in the same sector and additionally two zero voltage vectors.

γψ

γψ

Fig.3. Sensorless DFOC control system of six-phase induction motor with AFLC

6. SIMULATION RESULTS Simulation studies of considered control system have been carried out for the sixphase induction motor with the parameters: PN=3 kW, UfN=230 V, fN=50 Hz, ωeN=295 rad/s, pb=2, Rs=1.9 Ω, Rr=2.1 Ω, Lls =Llr =0.013 H, Lm=0.6 H. The waveforms of the reference, calculated and estimated electrical speed of the six-phase induction motor are shown in Figure 4a) and the waveforms of the electromagnetic and load torque are presented in the Figure 4b). It can be stated that the estimated and calculated electrical speed follows the reference speed with great accuracy. The values of the electromagnetic torque depend on the working states of the drive system. 300

10 200

5

100 0

0

-100

-5

-200

-10 -300 0

1

2

3

4

5

6

7

8

0

1

2

3

4

5

6

7

8

Fig. 4. Simulation results: a) the waveforms of reference, calculated and estimated electrical speed of sixphase induction motor; b) the waveforms of the electromagnetic torque and the load torque

The waveform of the stator phase current of the six-phase induction motor has

40

Jacek LISTWAN

been presented in Figure 5a) and the trajectory of the estimated magnitude of the rotor flux vector has been presented in Figure 5b). The magnitude of the rotor flux vector is regulated at the nominal value. The values of the stator current depend on the operating states. 4

1

2

0

0.5 -2

-4

0

0.5

1

1.5

2

2.5

3

3.5

4

0

4

-0.5

2

0

-1 -2

-4

-1 4

4.5

5

5.5

6

6.5

7

7.5

-0.5

0

0.5

1

8

Fig. 5. Simulation results: a) the waveforms of stator phase current of the six-phase induction motor b) the trajectory of the estimated magnitude of the rotor flux vector

7. CONCLUSIONS The mathematical model of the six-phase squirrel-cage induction motor and the MRAS estimator have been presented. The scheme of adaptive Fuzzy-Logic speed controller has been designed and the tasks of the control blocks of applied AFLC have been described. The sensorless DFOC method with six-phase induction motor and adaptive Fuzzy-Logic speed controller has been presented and discussed. The different states of the drive system have been investigated and the results of the simulation studies have been presented and discussed. The obtained results confirmed the good operation and properties of the considered control system. REFERENCES [1] DERUGO P., SZABAT K., Adaptive neuro-fuzzy PID controller for nonlinear drive system, COMPEL: The International Journal for Computation and Mathematics in Electrical and Electronic Engineering, Vol.34, Iss.3, 2015, 792-807. [2] LEVI E., BOJOI R., PROFUMO F., TOLIYAT H.A., WILLIAMSON S., Multiphase induction motor drives – a technology status review, IET Electr. Power Appl., 2007, V.1, Iss.4, 489–516. [3] LISTWAN J., PIEŃKOWSKI K., Analysis of Vector Control of Multi-Phase Induction Motor, Zeszyty Problemowe - Maszyny Elektryczne (Komel), Nr 3, 2014, 235-240 (in Polish), [4] ORLOWSKA-KOWALSKA T., Sensorless Induction Motor Drive, Wroclaw University of Technology Press, 2003 (in Polish). [5] PIEŃKOWSKI K., Analysis and Control of Multi-Phase Squirrel-Cage Induction Motor, Prace Naukowe Instytutu Maszyn, Napędów i Pomiarów Elektrycznych Politechniki Wrocławskiej, Nr 65, 2011, 305-319 (in Polish).

41

dual stator induction motor, DTC-ST control, MRAS estimator, simulation studies

Jacek LISTWAN*

DIRECT TORQUE CONTROL OF DUAL STATOR INDUCTION MOTOR WITH MRAS ESTIMATOR

The paper presents the new method of the Direct Torque Control with Switching Table (DTC-ST) of dual stator winding induction motor with MRAS speed estimator. The mathematical models of the special dual stator winding squirrel-cage induction motor and the MRAS estimator have been developed and presented. The control structure of the DTC-ST system with MRAS estimator has been designed. The simulation model of considered control system has been developed and implemented in Matlab/Simulink® Software. Simulation studies of the DTC-ST control system with dual stator winding induction motor have been carried out. The results of simulation studies for the different operating conditions of the drive system have been presented and discussed. The conducted simulation studies confirmed the good performance of the considered control system. The author contribution includes analysis and studies of considered control method with MRAS estimator.

1. INTRODUCTION The interest in induction machines with more than three phases has increased, especially for high-power applications. In the literature the different solutions of the multiphase motors have been discussed [1-3, 5, 6]. One of the interesting and widely presented is the dual stator induction motor (DSIM) having two sets of three-phase stator windings shifted by 30 electrical degrees with two isolated neutral points. The induction motors with dual stator windings could be the interesting alternative to the conventional three-phase induction motors because the redundancy of the stator windings can be used in the case of failure. The multiphase drives have also a lot of advantages in comparison with conventional three-phase systems [1-3, 5, 6]. In this paper the system of Direct Torque Control of dual stator induction motor has been considered and described. The aim of the Direct Torque Control method is __________ * Wroclaw University of Technology, Department of Electrical Machines, Drives and Measurements, Smoluchowskiego 19, 50-372 Wroclaw, [email protected]

42

Jacek LISTWAN

the control of the motor angular speed, the magnitude of the stator flux vector and the value of the motor electromagnetic torque. In the presented control system the Model Reference Adaptive System (MRAS) estimator has been applied. The application of this estimation system allows for the elimination of the mechanical sensors and increasing of the reliability of the drive system with DSIM [4]. 2. MATHEMATICAL MODEL OF DUAL STATOR INDUCTION MOTOR The mathematical model of the dual stator induction motor has been formulated on the basis of commonly used simplifying assumptions [1-3, 5, 6]. The analysis and control of this type of motor in the stator and rotor phase coordinate system can be difficult, because the motor model is then described by the equations with the time dependent coefficients. For this reason in this paper, the appropriate transformation matrices have been used and the improved model of the DSIM with constant coefficients has been formulated [5]. The general equations of the dual stator induction motor after transformations take the following vector form [5]: - the vector voltage equations:

u s1 = Rs1 ⋅ i s1 + pψ s1 + jω k ψ s1

(1)

u s 2 = Rs 2 ⋅ i s 2 + pψ s 2 + jω k ψ s 2

(2)

0 = Rr ⋅ i r + pψ r + j (ω k − ω e )ψ r

(3)

- the vector flux linkages equations:

ψ s1 = Ls1 ⋅ i s1 + Lm ⋅ i s 2 + Lm ⋅ i r

(4)

ψ s 2 = Lm ⋅ i s1 + Ls 2 ⋅ i s 2 + Lm ⋅ i r

(5)

ψ r = Lm ⋅ i s1 + Lm ⋅ i s 2 + Lr ⋅ i r

(6)

- the electromagnetic torque equation in the vector form:

[ (

)

(

Te = pb (Lm / Lr ) ⋅ Im i s1 ⋅ψ r + Im i s 2 ⋅ψ r *

*

)]

(7)

where: subscripts s, r - denote the variables and parameters of the stator and rotor, respectively; subscripts 1, 2 - denote variables and parameters of the stator 1 and 2 respectively; u - the voltage; i - the current; ψ - the flux linkage; Rs1, Rs2, Rr - resistance of stator 1, stator 2 and rotor, respectively; Ls1, Ls2, Lr - inductance of stator 1,

43

Direct Torque Control of dual stator induction motor with MRAS estimator

stator 2 and rotor, respectively; Lm – motor magnetizing inductance; ωk - arbitrary angular speed of the coordinate system relative to the stator; ωe - the electrical angular speed of the motor; Te - the motor electromagnetic torque; pb - the number of motor pole pairs; p=d/dt - derivative operator.

3. MATHEMATICAL MODEL OF THE MRAS ESTIMATOR The scheme of the applied Model Reference Adaptive System (MRAS) estimator is presented in Figure 1 [4].

Fig.1. The scheme of the MRAS estimator

The algorithm of MRAS estimator is based on application of two forms of mathematical models: the voltage model and the current model of induction motor. These models are combined by the adaptation mechanism with the PI controller [4]. In the considered MRAS estimator the voltage model is treated as the reference model. The voltage model is described by the following equations:

Lr (u sα − Rs isα − σLs pisα ) Lm L = r (u sβ − Rs isβ − σLs pisβ ) Lm

pΨrest αv =

(8)

pΨrest βv

(9)

where: α, β - the designation of the variables in the α-β coordinate system; ψrαvest, ψrβv - the estimated components of the rotor flux vector obtained from the voltage model; σ=1-Lm2/LsLr - total motor leakage factor. The current model is described by the following equations: est

pΨrest αc =

(

)

(10)

pΨrest βc

(

)

(11)

Rr est est Lmisα − Ψrest αc − ω e Ψrβc Lr R est est = r Lm i sβ − Ψrest βc + ω e Ψrαc Lr

44

Jacek LISTWAN

where: ψrαcest, ψrβcest - the estimated components of the rotor flux linkage vectors obtained from the current model. The estimated components of the rotor flux vector in the α-β coordinate system obtained from the voltage and current models are given to the adaptation mechanism. The adaptation mechanism determines the estimated value of the motor speed (ωeest). The estimated value of the motor speed is calculated according to the equation [4]: ωeest = K P (Ψrαc Ψrβv − Ψrβc Ψrαv ) + est

est

est

est

1 Ti

∫ (Ψ α t

r c

est

Ψrβv

est

est

− Ψrβc Ψrαv

est

)dt

(12)

0

where: KP, TI - the coefficients of the proportional and integral parts of the PI controller, respectively. The value of the estimated motor speed retunes the simulator based on the rotor circuit model. 4. DIRECT TORQUE CONTROL SYSTEM WITH MRAS ESTIMATOR The scheme of Direct Torque Control system with Switching Table (DTC-ST) of dual stator induction motor is shown in Figure 2 [5].

Fig.2. The scheme of DTC-ST control system of DSIM with MRAS estimator

To implement the DTC-ST control system the MRAS estimator has been used. The estimation block determines the estimated value of motor angular speed (ωeest), the estimated magnitudes of the stator flux vectors (ψs1, ψs2), the estimated value of the electromagnetic torque (Te) and the numbers of sectors N1 and N2,which determine the positions of the stator flux vectors ψs1 and ψs2. In the considered control structure with dual stator induction motor four control loops are applied: the control loop for motor

45

Direct Torque Control of dual stator induction motor with MRAS estimator

angular speed with PI controller, the two control loops for magnitudes of the stator flux vectors with hysteresis controllers and the control loop for electromagnetic torque with hystheresis controller. In order to apply the DTC-ST control system, the two stators of DSIM should be supplied by the two separate Voltage Source Inverters. The voltage vectors of Voltage Source Inverters are determined by the described above hysteresis controllers. The stator voltage vectors are chosen from the appropriate switching tables. 5. SIMULATION RESULTS Simulation studies of DTC-ST control system have been carried out for the dual stator induction motor with the data and parameters given in the Appendix. The transients of the reference, calculated and estimated electrical speed are shown in Figure 3. Load torque was activated after setting the rated speed. It can be stated that the transients of the estimated and calculated speeds coincides exactly with the transients of the reference motor speed. This confirms the great accuracy of the control system and its good performance. 300 200 100 0 -100 -200 -300 0

1

2

3

4

5

6

7

8

9

Fig. 3. The transients of reference speed, calculated speed and estimated electrical speed of dual stator induction motor for DTC-ST with MRAS estimator

The transients of the electromagnetic torque of the dual stator induction motor and the load torque have been presented in the Figure 4a). The trajectory of the stator flux vector is presented in Figure 4b). The DTC-ST control system allows for precise control of the motor electromagnetic torque and magnitude of the stator flux vector. 1

6 4

0.5 2 0

0

-2 -0.5

-4 -6 0

-1 1

2

3

4

5

6

7

8

9

-1

-0.5

0

0.5

1

Fig. 4. Simulation results: a) the transients of the electromagnetic torque and the load torque; b) the trajectory of the estimated magnitude of the stator 1 flux vector

46

Jacek LISTWAN

6. CONCLUSIONS The mathematical model of the dual stator induction motor and the MRAS estimator have been presented. The Direct Torque Control with Switching Table with MRAS estimator and with dual stator induction motor has been presented and discussed. The simulation studies of the considered control system have been carried out in Matlab/Simulink® Software. The results of simulation studies for the different operating conditions of the drive system have been presented and discussed. The conducted simulation studies confirmed the good performance of the considered DTC-ST control system with dual stator induction motor. It can be stated, that the DTC-ST control system allows for precise control of the motor speed, motor electromagnetic torque and stator flux vector.

APPENDIX The dual stator induction motor data and parameters: V=230V, f=50Hz, n=1450rpm, Rs1= Rs2=3.72Ω, Rr=2.12Ω, Lls1= Lls2= Llr =0.022H, Lm=0.3672H, pb=2.

REFERENCES [1] AMIMEUR H., ABDESSEMED R., AOUZELLAG., MERABET E., HAMOUDI F., A Sliding Mode Control associated to the Field-Oriented control of Dual-Stator Induction Motor Drives, Journal of Electrical Engineering, Vol. 10, Nr 3, 2010, 7-12. [2] LEVI E., BOJOI R., PROFUMO F., TOLIYAT H.A., WILLIAMSON S., Multiphase induction motor drives – a technology status review, IET Electr. Power Appl., 2007, V.1, Iss.4, 489–516. [3] LISTWAN J., PIEŃKOWSKI K., Analysis of Vector Control of Multi-Phase Induction Motor, Zeszyty Problemowe - Maszyny Elektryczne (Komel), Nr 3, 2014, 235-240 (in Polish), [4] ORLOWSKA-KOWALSKA T., Sensorless Induction Motor Drive, Wroclaw University of Technology Press, 2003 (in Polish). [5] PIEŃKOWSKI K., Analysis and Control of dual stator winding induction motor, Archives of Electrical Engineering, Vol. 61, 2012, 421-438. [6] ZHAO Y., LIPO T. A., Space vector PWM Control of Dual Three-Phase Induction Machine Using Vector Space Decomposition, IEEE Transactions on Industry Applications, Vol. 31, Nr 5, 1995, 1100-1108.

47

Key words: suspension plasma spraying, substrate preparation, microstructure, topography

Monika MICHALAK*, Paweł SOKOŁOWSKI*, Andrzej AMBROZIAK*

THE INFLUENCE OF SUBSTRATE PREPARATION ON THE MICROSTRUCTURE AND TOPOGRAPHY OF SUSPENSION PLASMA SPRAYED ZIRCONIA COATINGS

Conventional powder plasma spraying is a well-known and widely used thermal spraying technology. Nowadays, novel plasma spraying methods that use liquid feedstock instead of dry powder are tested for the industrial purpose. One of these processes is suspension plasma spraying (SPS). SPS method uses suspensions of fine grained powders as a feedstock and can provide coating with finer microstructures than those obtained by conventional plasma spraying. Nevertheless, the microstructure and properties of coatings are influenced by many process variables. The purpose of this article was to observe the relationship between the substrate preparation and coating microstructure and topography as well. The studies involved the following steel substrates: (i) sand blasted, (ii) grinded and (iii) laser treated. It was determined whether the substrate preparation prior to the spraying would affect the structure of coating. The microstructure, topography and build-up mechanisms of the SPS coatings were analyzed and discussed. The studies were carried out mostly with the use of Scanning Electron Microscopy (SEM). Novel technology, called Shape From Shading, was also applied to visualize coating topography in 3D mode.

1. INTRODUCTION Coatings draw much attention of the surface engineering nowadays. They play significant role in many fields of science – including civil engineering, energy, biomechanics, metallurgy and many others [1-5]. Irrespective of their application, the final performance of coated component highly depends on the quality of manufactured coating. This implies that the microstructural investigations are necessary to predict the usefulness of coating. The coating microstructure is influenced by many variables but mainly by the coating deposition method. Many different methods of coating ap__________ * Wrocław University of Science and Technology, Faculty of Mechanical Engineering, Department of Materials Science, Welding and Strength of Materials

48

Monika Michalak, Paweł Sokołowski, Andrzej Ambroziak

plication can be distinguished. In case of thermal spray technologies, the most important deposition variables are: the operating temperature, velocity of deposition, form and size of feedstock material etc. [3,6-8]. The object of interest in this article will be one of the most universal thermal spraying method, called plasma spraying. Usually plasma spraying process is realized under open atmosphere. Depending on the type of precursors, three main spraying methods are distinguished: PPS (powder plasma spraying), SPS (suspension plasma spraying) and SPPS (solution precursor plasma spraying). Conventional powder plasma spraying method [4] is widely used due to variety of available materials (ceramics, metals, polymers, cermets, others). Unfortunately, the limitation of PPS method is the size of sprayed particles. If they are too small (and too light at the same time), they do not have enough momentum to be introduced into the high speed plasma jet. Only micrometer sized powders can be used and the variety of commercially available feedstock materials is limited also [4,8]. The drawbacks of PPS technique resulted in the development of suspension plasma spraying method. Due to the use of precursor in a form of suspension the injection of fine powder particles was solved. The use of liquid precursor allowed also to avoid the problem of clogging of ultra-small powders at the end of injector nozzle. The suspension is a mixture of solid phase (powder particles) and solvent (water, alcohol or their mixture) [6,9]. Relevant advantage of SPS method is the possibility of obtaining coatings having sub-micro and nano-sized grains [7,10]. The size of lamellas can be reduced by the use of submicrometer or nanometer sized precursor. Another plasma spraying method is solution precursor plasma spraying. SPPS also bases on the liquid feedstock – usually in the form of solution of different salts, nitrides and acetates with the addition of solvent such as water or ethanol [3,6]. In this method, particles of powder are completely replaced by liquid droplets. Therefore SPPS is single-step process [5] – on the contrary to SPS technique, stage of powders preparation is dismissed – it significantly shortens time of spraying. Nevertheless, both SPS and SPPS techniques aim to overcome limitations of APS spraying, by deposition of coatings with refined microstructure [3,6,9]. Beyond the deposition method, an important factor influencing final coating microstructure is proper substrate preparation [11]. This can have a great effect on the substrate/coating interface and then on the coating adhesion. In this study, the influence of substrate topography and preparation on coating growth-up, microstructure and topography is analyzed. The main goal is to determine, whether the preparation prior to the spraying could affect the structure of the coating.

49

The influence of substrate preparation on the microstructure and topography of suspension plasma sprayed zirconia coatings

2. MATERIALS AND METHODS The material used for a preparation of suspension was yttria stabilized zirconia powder (ZrO2 + 14 wt.% Y2O3), manufactured by Tosoh (Tokyo, Japan). The mean volume diameter was equal to dV50 = 398 nm [9]. The suspension contained also water and ethanol (in ratio 1:1) and dispersant agent preventing the suspension beyond agglomeration and sedimentation. The suspensions having two different concentration of solid phase were tested – appropriately with 5 wt.% and 10 wt.% of powder. The substrate material was stainless steel 304L. The substrates were prepared prior to spraying by: (i) sand blasting, (ii) grinding, (iii) laser treatment with linear pattern and (iv) laser treatment with mesh pattern (see Fig. 1).

(a)

(b)

(c)

(d)

Fig. 1. 3D topography of prepared substrates: (a) sand blasted, (b) grinded, (c) laser treated (linear pattern), (d) laser treated (rectangular pattern); results partially published elsewhere [9]

The spraying process was carried out by the use of SG-100 plasma torch of Praxair (Indianapolis, IN, USA). The suspension injection was realized by liquid stream injector mounted internally in plasma torch. The spray process details are described elsewhere [9,10]. The coating morphology was characterized at the coating surface and coating cross-section by two SEM microscopes: Philips XL30 and Phenom G2 Pro. The coating topography was investigated based on novel technology Shape From Shading. Due to the use of four various local 2D images of the specimen and the shadowing effect, 2D and 3D views of substrates and coatings were obtained [11].

50

Monika Michalak, Paweł Sokołowski, Andrzej Ambroziak

3. RESULTS As indicated already in Fig. 1, the topography images revealed different substrates appearances: inhomogeneous (sand blasting), flat (grinding) and with paternal (laser treatment). In order to verify the influence of substrate topography on the coating microstructure, the morphology of sprayed coating and appearance of its cross-section were investigated. It was possible to observe that: • Sand-blasted substrate (Fig. 2) resulted in the most irregular morphology of deposited coating among all prepared samples. Although the substrate/coating interface is bonded, surface prepared for spraying is not smooth (see defected steel surface in Fig. 2b, caused by sandblasting). Notwithstanding the concentration of powder in suspension, a big amounts of micropores can be observed. • The profile of grinded substrate (Fig. 3) is far more smooth than in the previous case. But the coating is still very well bonded to the substrate. Top layer observations revealed very fine grained microstructure of the coating. It seems also that lower powder concentration (Fig. 3b) resulted in more irregular coating than in the case of suspension with 10 wt.% of powder (Fig. 3d). • The laser treatment of substrates allowed creating of columnar-like coatings (Fig. 4, 5). All coatings fits to the prepared surface very well – the coating/substrate interface do not show any cracks or delamination. It should be emphasized that the coating profile is similar to the shape of substrate prior to the spraying (see Fig. 4a). The regular substrate profile resulted in regular shape of coating but only if low concentrated suspensions were used. The greater concentration of suspension resulted in denser microstructure and the columns started to disappear. Furthermore, the Shape From Shading method can be a useful tool to prepare quantitative analysis of columnar-like coatings (to count for example “columns per area” ratio, see Fig. 5a). (a)

(b)

Fig. 2. The influence of sand-blasted substrate on coating morphology; 5 wt.% of solid suspension

51

The influence of substrate preparation on the microstructure and topography of suspension plasma sprayed zirconia coatings (a)

(b)

(c)

(d)

Fig. 3. The influence of grinded substrate on coating morphology; 5 wt.% solid concentration in suspension (a), (b) and 10 wt.% (c), (d)

(a)

(b)

Fig. 4. The influence of linear laser treated substrate on coating morphology; 5 wt.% solid concentration

(a)

(b)

(c)

Fig. 5. The coatings morphologies after rectangular laser treatment: (a) top view of coating with 5 wt.% of powder (by Shape From Shading), (b) cross-section of coating with 5 wt.% of powder, (c) crosssection of coating with 10 wt.% of powder

52

Monika Michalak, Paweł Sokołowski, Andrzej Ambroziak

4. CONCLUSIONS The studies confirmed that Suspension Plasma Spraying is an interesting technology that enables deposition of fine grained coatings. The application of different technologies for substrate preparation prior spraying allows creating different substrate topography. Afterwards, this have an impact on the coating microstructure and surface as well. The build-up mechanism of SPS coatings is affected also by the concentration of suspension. Small amount of solid phase resulted in rough and porous coatings. The reversed phenomenon was observed for higher content of powder – coatings porosity was lower. Moreover, the more irregular substrate, the more likely that also deposited coating will be of irregular shape. It clearly implies that proper selection of deposition technique, substrate preparation and process parameters determines coating microstructure, which finally decides on the functionality of sprayed coating. REFERENCES [1] APEDO K.L., MONTGOMERY P., SERRES N., FOND C., FEUGEAS F.: Geometrical roughness analysis of cement paste surfaces using coherence scanning interferometry and confocal microscopy, Materials Characterization 118, 2016, 212–224. [2] HENNE R.: Solid oxide fuel cells: a challenge for plasma deposition processes, Journal of Thermal Spray Technology 16, 2007, 381-403. [3] CANDIDATO R., SOKOŁOWSKI P., ŁATKA L., KOZERSKI S., PAWŁOWSKI L., DENOIRJEAN A.: Plasma spraying of hydroxyapatite coatings using powder, suspension and solution feedstocks, Przegląd Spawalnictwa 87.10, 2015, 64-71. [4] DOSTA S., ROBOTTI M., GARCIA-SEGURA S., BRILLAS E., GARCIA CANO I., GUILEMANY J.M.: Influence of atmospheric plasma spraying on the solar photoelectro-catalytic properties of TiO2 coatings, Applied Catalysis B: Environmental 189, 2016, 151–159. [5] FAN W., BAI Y.: Review of suspension and solution precursor plasma sprayed thermal barrier coatings, Ceramics International, 2016. [6] JOULIA A., BOLELLI G., GUALTIERI E., LUSVARGHI L., VALERI S., VARDELLE M., ROSSIGNOL S., VARDELLE A.: Comparing the deposition mechanisms in suspension plasma spray (SPS) and solution precursor plasma spray (SPPS) deposition of yttria-stabilised zirconia (YSZ), Journal of the European Ceramic Society 34, 2014, 3925–3940. [7] PAWŁOWSKI L.: Suspension and solution thermal spray coatings, Surface & Coatings Technology 203, 2009, 2807–2829. [8] FAUCHAIS P., ETCHART-SALAS R., RAT V., COUDER J.F., CARON N., WITTMANN-TE´NE`ZE K.: Parameters controlling liquid plasma spraying: solutions, sols or suspensions, Journal of Thermal Spray Technology 17.1, 2008, 31-59. [9] SOKOŁOWSKI P., KOZERSKI S., PAWŁOWSKI L., AMBROZIAK A.: The key process parameters influencing formation of columnar microstructure in suspension plasma sprayed zirconia coatings, Surface & Coatings Technology 260, 2014, 97–106. [10] KOZERSKI S., ŁATKA L., PAWŁOWSKI L., CERNUSCHI F., PETIT F., PIERLOT C., PODLESAK H., LAVAL J.P.: Preliminary study on suspension plasma sprayed ZrO2 + 8 wt.% Y2O3 coatings, Journal of the European Ceramic Society 31, 2011, 2089–2098. [11] SOKOŁOWSKI P., PAWŁOWSKI L., DIETRICH D., LAMPKE T., JECH D.: Advanced Microscopic Study of Suspension Plasma-Sprayed Zirconia Coatings with Different Microstructures, Journal of Thermal Spray Technology 25, 2016, 94-104.

53

Key words:LEM Bullet, motorcycle, production, process, police, SmartMoto Challenge, electric, chassis, frame

Wojciech PAWLAK*, Piotr KONIECZNY*

DESIGN AND PRODUCTION PROCESS OF LIGHT ELECTRIC MOTORCYCLE’S CHASSIS

Main goal of this article is to present design and production process of LEM Bullet’s frame with general issues that occurred during creation. Article presents FEM analysis, weldment equipment and conclusions with development goals. LEM Bullet is a motorcycle built for SmartMoto Challenge Barcelona 2016 competition which main theme was to create police motorcycle.

1. INTRODUCTION 1.1. LEM BULLET

LEM Bullet is another motorcycle designed and produced by Scientific Association of Mobile Robots and Vehicles of Wrocław University of Technology and Science. This motorcycle was designed for police applications due to theme of the SmartMoto Challenge Barcelona 2016 in which LEM Wrocław team have taken 2nd place in business plan category, 2nd place in dynamic category and 3rd place in general classification. LEM Bullet weighs 75 kg, has 6 kW BLDC motor in hub which allows to travel with max. velocity of 120 km/h. Li-Ion 60V battery lets achieve range of 100 km.

____________ *Mechanical Department, Wrocław University of Scienceand Technology, Wybrzeże Wyspiańskiego 27, 50-370 Wrocław

54

Wojciech Pawlak, Piotr Konieczny

2. DESIGN PROCESS 2.1. BOUNDARY CONDITIONS

First step to build a motorcycle is to choose and calculate boundary conditions, type, general characteristics and design. Every electric motorcycle’s core is a frame, to which battery, fairings and suspension are attached. In the Fig. 1. design of LEM Bullet is presented. Frame composition was based on typical supermoto construction with considered changes for electric motorcycle such us impossibility of using engine as a part of a supporting structure.

Fig. 1. LEM Bullet’s design Table 1. Boundary conditions and main loads Type Wheelbase: Front wheel radius: Rear wheel radius: Mass with driver: Vertical comp. of CoG: Horizontal comp. of CoG: Mass percentage on front wheel: Mass percentage on rear wheel: Maximum velocity: Caster angle: Motor torque: Static coefficient of friction:

Value = 1340 = 310 = 300 = 160 = 650 = 425 % = 31,7% % = 68,3% = 140 /ℎ = 24° = 110 ! "# = 1

55

Design and production process of Light Electric Motorcycle’schassis "$ = 0,7

Dynamic coefficient of friction: Main loads: Load on a front wheel Load on a rear wheel Static braking force on front wheel: Static braking force on rear wheel: Total braking force: Centrifugal force:

= 498 ! = 1072 ! & = 498 ! & = 1072 ! ' = 1570 ! ( = 1570 !

During selection of these boundary conditions the geometry of front and rear suspension was taken under consideration and preliminary ergonomic analysis was conducted. Table 2. Material of a frame – aluminum 6063 T6 Mg [%] 0,35 – 0,6

Si [%] 0,3 – 0,6

Fe [%] 0,1 – 0,3

Tensile strength Rm 270 MPa

Chemical composition: Zn [%] Cu [%] max 0,15 max 0,1 Strength properties:

Mn [%] max 0,1

Ti [%] max 0,1

Cr [%] max 0,05

Yield strength Re 170 MPa

Fig. 2. LEM Bullet’s main dimensions

56

Wojciech Pawlak, Piotr Konieczny 2.2. LOAD CASES

FEM analysis was made on a shell model. Mesh created for analysis consisted mostly linear quadrilateral elements S4R (144 197 elements) and linear triangular elements S3 (105 elements). Below are presented all of the load cases that were conducted on a frame construction. a) G1 – the simplest case in which frame is loaded in G acceleration conditions by its main masses: driver and battery. All the other masses are distributed or too low to be considered. b) G2 – case identical to one above with 2G acceleration acting on a frame. c) Acceleration – during this case, motorcycle’s frame is not only loaded with G acceleration but also with maximum possible acceleration of a motorcycle with direction along a motorcycle. d) Emergency Braking – similar case as one above with opposite direction of acceleration along a motorcycle. e) Longitudinal stiffness – case based on a Rinard Test [1] popular bicycle frame test. Load is put on the fictional point on the head tube axis distant from the head tube by 230 mm. Load is calculated of dynamic braking force on a front wheel: '∗ 1570! ∗ 650 (1) + = 498! + = 1259! )& = 1340 ,-

=

' ∗ . 1259! ∗ 230 = / 570

= 508!

(2)

)& – Dynamic braking force on a front wheel, . – Rinard’s test length, / – fork’s length from axle to head tube. f) Torsional stiffness –a test in which to the same point as above but in the transverse to the motorcycle direction is put maximum load acting on a front wheel in turning conditions.

0-

=

∗ . 1570! ∗ 230 = /1 867

(

= 416,5!

(3)

/1 – fork’s length from head tube to the ground on the axis of the head tube g) Shock absorbers load case –this case is actually inversion of a G2 test. Using MSC Adams environment LEM engineers calculated maximal loads acting on springs of the suspension and on the bearing joints.

57

Design and production process of Light Electric Motorcycle’schassis

Fig. 3. MSC Adams calculation model

Table 2. FEM analysis results Type of analysis G1 G2 Acc EB LS TS Shocks’s response

Stress [MPa] 49,3 98,6 98,6 44,0 45,3 79,6 106,9

Displacements [mm] 0,28 0,55 0,66 0,76 0,40 5,30 2,25

58

Wojciech Pawlak, Piotr Konieczny

Fig. 4. Shocks’s response FEM results

3. PRODUCTION Production of LEM Bullet’s frame consisted also designing of welding stations that ensured that the final product has been welded with good enough accuracy. Simultaneously to the process of welding the equipment specific parts of the frame were bended, cut and machined.

Fig. 5. Main welding station

59

Design and production process of Light Electric Motorcycle’schassis

4. CONCLUSIONS 1) The biggest issue during design phase was not knowing all of the internal parts of the motorcycle. It was caused by simultaneous work of people responsible for mechanical, electrical and fundraising part of the project, which was conducted by the date of the publication of a rulebook of SmartMoto Challenge. Due to this situation most of the mountings couldn’t be predicted in the design stage. 2) Another problem occurred during production stage – some of the bendings of the tubes were not possible to be made in Lower Silesia district, which forced engineers to redesign some parts of the frame. 3) Final version of the frame met all of the requirements. Its final weight was equal 4,3 kg, which is about 53% lighter than one of the previous motorcycle’s frame. Such big improvement was one of the reasons why LEM Bullet has such a good power to weight ratio and can achieve velocity of 100 km/h in 5,3 seconds.

REFERENCES

[1] RINARD D., http://www.sheldonbrown.com/rinard/rinard_frametest.html, 1996 [2] RUSIŃSKI E., Metoda elementów skończonych. System COSMOS/M. Wydawnictwo Komunikacji i Łączności, Warszawa 1994 [3] GODLEWSKI T., PAWLAK W. Design of supporting frame for Light Electric Motorcycle, 2015 Wrocław

60

Motorcycle, rear suspension, linkage system, swing arm, finite-element method

Paweł STABLA* ,Piotr KONIECZNY*, Wojciech PAWLAK*

LIGHT ELECTRIC MOTORCYCLE CONSTRUCTIONAND FEM ANALYSISOF MOTORCYCLE’S SHOCKARM

The main goal for this article is to describe reasons and whole creating process of the linkagesystem in LEM Bullet. LEM Bullet is the first motorcycle made by members of Scientific Association of Mobile Robots and Vehicles that has its rear suspension equipped with the linkagesystem. The main attention was paid for the optimization of the geometry of a shock arm. The shock arm is the specific element that connects a damper, a swingarm, and a frame. The optimization was made in the direction to decrease the total mass of the element as well as to improve the manufacturability, especially to facilitate the welding process.

1. INTRODUCTION 1.1 MOTORCYCLE LEM BULLET

LEM Bullet is the latest motorcycle from the series LEM – Light Electric Motorcycles made by students from Wroclaw University of Science and Technology. The motorcycle was designed and manufactured by members of Scientific Association of Mobile Robots and Vehicles and was widely awarded on the international competition in Barcelona. This time the motorcycle’s functionality and attributes were adjusted for police application. In comparison to Bullet’s predecessors the newest version is distinguished by the engine of higher power – 6kW, which enables the motorcycle to ride as fast as 120 km/h with the weight of only 75 kg. Thanks to the use of Li-Ion batteries the range is up to 90 km. Moreover, Bullet is equipped with removable shields which are a part of motorcycle skin and may be used during police interventions. The motorcycle is also equipped with lights and siren characteristic for police. Simultaneously with the designing process a special mobile application was __________ *

Mechanical Department, Wrocław University of Science and Technology

61

Paweł Stabla, Piotr Konieczny, Wojciech Pawlak

written in order to enable the policeman to contact with a dispatcher using a mobile phone attached to the dashboard. 1.2 MAIN ASSUMPTIONS

The primary goal during the whole projecting process of the rear suspension, which is presented in Figure 1, was to determine boundary conditions - restrictions that decrease construction geometry and shape possibilities. The most important restrictions were the shape and the geometry of the motorcycle frame which was designed to usage single central shock absorber. Taking into consideration the wheelbase and the wheel size, the H-shape, half meterswingarm with the linkagesystem was chosen. The swingarm as well as the frame was made of aluminum alloy 6063 T6 and the linkagesystem, because of the bigger forces, was made of steel S355.

Figure 1. LEM Bullet’s linkagesystem

2. THE OPTIMISATION 2.1 RELATION BETWEEN GEOMETRY AND FORCES

The next step of creating the linkagesystem was to set the major geometric points the joints between elements. As the geometry of this connection is strongly related with forming the direction and value forces in the swingarm – linkage – frame system

62

Light Electric Motorcycle construction and FEM analysis of motorcycle’s shock arm

there was created a method, thanks to which this relationship can be studied. Very helpful in this case turned out to be the program MSC Adams in which the whole rear suspension was created on 2D surface. Knowing the value of the force that is working on the swingarm as a reaction between the rear wheel and ground, thanks to the MSC Adams model it can be set every direction and value of forces in joints. It is possible to assume the main conditions like the stiffness of damper’s spring. By changing the location of joints of the linkagesystemwith the rest rigid points (connection with frame and swingarm and its length) it is possible to come up with the most desirable directions and value forces in the linkagesystem. By this process the most optimal solution was obtained. 2.2 FEM ANALYSYS ASSUMPTIONS

The major and last part of designing parts was to bring conceptions and 3D models to program that can conductFEM analysis – Abaqus. As a base for applied force to calculate joints forces that occur in the linkagesystem were the forces from the G2 simulation. It was the maximum restricted value. The results are presented in the table below. Table 1. Forces in main joints in G2 simulation Forces in joints

Value

Swingarm

2144 N

Damper

5813 N

Swingarm axis

5287 N

Swingarm and linkage point connection

7200 N

Linkage and chassis frame point connection

4441 N

In case of a very big value of forces it was decided to use material that have better mechanical properties than the aluminum – the steel grade S355. The safety factor ,,k” was set to level 1.5, which based of this type of steel gave the maximum stress level in the finite elements won’t exceed 236 MPa. The value of the safety factor was determined by the precision of forces’ values and the desire to create a light motorcycle. 2.3 SHOCK ARM

In case of characteristic – symmetric carried force the shock arm was analyzed in only one the most strict case: G2 simulation. It was the situation that can be compared

63

Paweł Stabla, Piotr Konieczny, Wojciech Pawlak

to hit the rear wheel into a hole in a road or drive down the curb. Simulation was created in the dynamic explicit step. The three forces, that were assessed from MSC Adams,were applied onto element without boundary conditions. It gave the shock arm an option to move in the symmetric directions of surface elements as it takes place in the reality. Due to use ofFEM, there was created a discrete model with total number of 2816 elements which consist of2787 linear quadrilateral elements of type S4R and 29 linear triangular elements of type S3. This discrete model and model with forces are presented in Figure 2 below.

Figure 2. Discrete model and forces and its directions in shell model

2.3 OPTIMALIZATION SHOCK ARM

The next simulations of shock arm were conducted in need to compare the results of simulations on the shell and solid model. The next step was to search lighter option of created yet shock arm, which also can be more technologic construction. The shock arm was examined by the possibility of having hole in center that could reduce the element’s mass, and could give a good access to weld the steel axis to a steel sheet. The thickness of sheet was set to 4 mm. Solid models are believed to be more precise than the shell model. For example the area of applied force for a bearing is different. There were 4 models under consideration. They are presented from the

64

Light Electric Motorcycle construction and FEM analysis of motorcycle’s shock arm

Figure 3 to Figure 6. All the simulations were run in the explicit step as the previous main one. Explicit mode was used because of very fast time of changing forces that can run on shock arm. The results of the largest von Mises stresses are presented in the table below. Table 2. Stress level in every model

Model 1 – Full model 2 – The most thin edge 3 – 10 mm edge 4 – model with cuts

Shell [MPa] 190,2 365,6 355,1 232,4

Solid [MPa] 188,6 366,2 345,3 225,3

Figure 3. Model 1. Number of finite elements: shell 9963, solid 23372

Figure 4. Model 2. Number of finite elements: shell 5105, solid 15966

65

Paweł Stabla, Piotr Konieczny, Wojciech Pawlak

Figure 5. Model 3. Number of finite elements: shell 7753, solid 26870

Figure 6. Model 4. Number of finite elements: shell 8489, solid 30274

3. CONCLUSIONS Judging form the results of shell and solid models, it is noticeable that the maximum von Mises stresses of both models are comparable. However, maximum stresses vary considerably depending on the geometries. Models 2 and 3 do not the strength condition, their maximum stresses exceed the permissible stress value, which is 236 MPa. In case of model 4, maximum stresses do not exceed the critical value. It is worth-mentioning, that for such an element with big area of hholes oles should also be simulated to prevent buckling. The model 4 has the advantage that the process of welding of the shock arm elements is simplified due to holes.

66

Light Electric Motorcycle construction and FEM analysis of motorcycle’s shock arm REFERENCES [1] RUSIŃSKI E., Metoda elementów skończonych. System COSMOS/M. Wydawnictwo Komunikacji i Łączności, Warszawa 1994 in Polish. [2] PAWLAK W., KONIECZNY P., Lekkie motocykle elektryczne, Projektowanie i Konstrukcje Inżynierskie, 2016in Polish.

67

Key words: Line follower, transport, engineering, robotics, electronics, industry, electric vehicle, DC motor

Krystian SZEWCZYŃSKI, Bogdan GÓRA, Henryk BĄKOWSKI *

PRESENTATION OF MODEL OF THE DEVICE USED TO PERFORM INDUSTRIAL TASKS

This paper presents transport method used in automated manufacturing plants, warehouses and other industrial purposes, wherever there is a need of transport from point A to point B. All parts used in the device such as electronic platform with microcontroller, reflective sensors, DC motor drivers and DC motors allows the simple robot to move independently, exactly along the line. Possibilities that creates this device are extensive and their continuous development has an impact on the current structure of production and storage.

1. INTRODUCTION Nowadays electric vehicles play an important role in industrial production, storage, rail transport and are increasingly used for road transport. Each factory of motor vehicles, production plants in electrical and electronic industry, have in their fleet units, which speed up workflow and make work easier. Proper programming of such vehicles results in a greater precision and lack of errors in tasks entrusted to them, impossible to obtain by man.

__________ * Silesian University of Technology, Faculty of Transport, Department of Automotive Vehicle Service, Krasińskiego Street 8, 40-019 Katowice

68

Krystian Szewczyński, Bogdan Góra, Henryk Bąkowski

2. CONSTRUCTION AND PRINCIPLE OF OPERATION 2.1. CONSTRUCTION

Our project is a line follower robot, which can overcome distance without human intervention using series of sensors to stay on the path made by scotch tape. The main feature of the design of this model is its simplicity. For its construction were used the following components: - Arduino in version Mega as an electronic platform - One L293 driver used as DC motor driver, attached directly to the Arduino - Two Pololu DC motors with wheels - Seven TCRT5000 reflective sensors In addition to these parts needed were: laminated board as the base and only one 12 V power supply, which powers whole device. This voltage value is appropriate value to run the device (manufacturer recommends input voltage of 7 to 12 volts). Having regard to the drop in voltage while connecting battery cells in series, ten R6 1.5 V batteries were applied. Whole structure is illustrated below.

Fig. 1. Construction scheme

69

Presentation of model of the device used to perform industrial tasks

2.2. PRINCIPLE OF OPERATION

The main component of the device is Arduino which controls the whole system, equipped with the microcontroller Atmel AVR which can be programmed in Arduino Software, integrated development environment. The Arduino IDE supports the programming languages C and C++. Arduino, into which other elements are plugged, controls the entire operation of the line follower. Based on the data obtained from the TCRT5000 reflective sensors send to the driver pulse. These sensors are able to operate on the basis of the distance from the obstacle (up to 12 mm) and contrasting color surfaces (e.g. white and black). Upon receiving the pulse, L293 driver sets in motion motors with a precise rotational speed depending on the situation, for example at left turn right motor has to rotate faster. The algorithm uploaded to the Arduino platform has the task to receive information about which sensor is currently detecting the black line under the device and execution the proper response. For example, if the central sensor detects the path the motors rotate at the same speed. During cornering, when one of the side sensor become active, there is a change in speed of one of the engines, in order to correct the path (similar to the mechanism employed in the tanks). The entire connections between components are illustrated in the scheme below.

Fig. 2. Scheme of connections between components

70

Krystian Szewczyński, Bogdan Góra, Henryk Bąkowski

3. CURRENT CONSUMPTION OF THE DEVICE MODEL The table presents a summary of selected electrical parameters (rated current) measured during tests. Table 1. Statement of electrical parameters of the line follower Function Arduino Mega 2x DC motor 6V 7x TCRT5000 sensors L293 motor driver Other Sum:

Maximum values [mA] 30 480 140 1300 30 1980

4. APPLICATION OF THE DEVICE The technical solutions presented in the article, or very similar to the one shown, find their application in the aforementioned areas of the current industry. After determining the routes, it is possible to efficiently move the transport vehicle, e.g. In warehouses without direct human intervention. Due to the electric drive, the vehicle does not adversely affect health and safety of employees in companies, because the noise created by this solution is small. It is also worth mentioning the modern limousines brands such as BMW or Mercedes, which offer vehicles equipped with autopilot systems, through which the driver while driving on the highway or a crowded city can rest. The operation of these systems is based on a series of sensors responsible for the safety of the vehicle and its surroundings.

71

Presentation of model of the device used to perform industrial tasks

5. CONCLUSIONS The biggest problem for autonomous vehicle is identification of the road. Various external factors (such as temperature, humidity, landform, quality of marked routes) affect the speed and precision of overcoming the road. Appropriate use of the control program and the reliability of the sensors has a decisive influence on safety. From the sensor is required the most accurate detection of the road, and the device must move at a speed that prevents the appearance of disturbances detection path at which it will move. In addition it would be needed the installation of security systems along with a procedure that allows to take a decision in the case of the intersection of the road with a man (or other living organism) or the appearance of unplanned object lying on the road crossing the device in order to avoid a collision. Another situation worth considering is state of battery below the allowable voltage level at which the vehicle would direct to the charging stations.

REFERENCES [1] OXER J., HUGH B.: Practical Arduino, Springer-Verlag, New York 2009 [2] MONK S.: Hacking Electronics: An Illustrated DIY Guide for Makers and Hobbyists, The McGrawHill Companies, New York 2013 [3] SESHAN S., SESHAN A. :: Line Followers: Basic to Proportional, 2015 http://ev3lessons.com/translations/en-us/advanced/LineFollower.pdf [4] NOWAK D.: PRZEPIS NA ROBOTA – PROGRAMOWALNY LINEFOLLOWER, 2009 http://hobbyrobotyka.pl/przepis-robota-programowalny-linefollower/

72