Evolution of Industrial Motor Control

Transcription

1 Evolution of Industrial Motor Control Loic LeCoz MITSUBISHI EUROPE B.V. Travellers Lane Hatfield, Hertfordshire AL10 8XB UK Embedded System Show TECHNOLOGY FOR LIFE MITSUBISHI Page 1 of 12

2 Abstract Over more than a century, many different types of motors have been developed, with most usually dedicated to a particular application. The reason for this was that there was no easy way of regulating the supply voltage or frequency to control the speed, and designers had therefore to find ways of providing this control within the motor itself. All this changed in the 1960s, first with the thyristor, providing a relatively cheap and easily controlled variable-voltage supply for DC motors and later in the 1970s with the development of variablefrequency inverters suitable for induction motors. These major developments resulted in the discontinuation of many of the special motors, leaving the majority of applications to only few types for example; DC and Brushless DC motors; Induction motors and recently Switched Reluctance motors. Complexity has now shifted from the motor itself to the external drives and control circuit. Choosing a complete drive system requires not only knowledge about motors, but also the associated power electronics and the control options, this paper will examine first five fundamental machines used in industrial drives: DC, Brushless DC and AC, Induction and Switched Reluctance motors. The topics such as how each motor and drive system works and for which application they are suitable will then be considered. In the second section, the focus will be on two main solutions offering a high level of performance. The control of induction machines will be explained starting with the basics. The different types of PWM will be described, mentioning their practical implementation using microprocessors, and explanation of the famous Space Vector Modulation and Vector control will be provided. Finally, as switched reluctance drives are rapidly expanding into new applications, the development of a closed loop system solution will be presented. TECHNOLOGY FOR LIFE MITSUBISHI Page 2 of 12

3 Introduction Electric motors now form a big part in our daily life especially with domestic applications like hairdryers, fans, mixers and drills that we do not really pay any attention to them. We just expect all these items to do their job without even giving a thought to the motor itself and its rather complex electronics. Because of this, and considering the fact that electric motors use 60% of the world wide energy generated, this paper will provide the reader first with a basic understanding of how the different technologies operate and what kind of applications each one can be associated with. This paper is then targeted for all engineers wishing to get a general understanding of the main technologies on the market. Assumptions To generate the torque required to produce rotation, most motors use the force created on a current-carrying conductor placed in a magnetic field. This paper assumes that the reader already knows that when a current-carrying carrying conductor is placed in a magnetic field, it experiences a force depending on the current in the wire, the intensity of the magnetic field as shown below. Force Current Magnetic Field Fig 1: Electromagnetic force created on a currentcarrying conductor in a magnetic field DC Brushless DC Fig2: Classification of electrical motors 1-1 DC Motors MOTORS Synchronous Brushless AC AC Reluctance Asynchronous Induction Although DC motors (also called universal motors) now tends to be less popular than inverter-fed induction motors, there are still a number of applications such as cranes, fork lift trucks where DC motors are still used. When it is important to hold a load, customers still appreciate the accurate and high torque at zero speed of the classical solution: DC Motor + Drive system. The basic principle is described in Fig 3. Basically, the torque is produced by the interaction between the axial current-carrying conductors on the rotor and the radial magnetic flux generated by the stator. A universal motor can be divided into 3 main parts: Stator: Permanent magnets (in Fig3) or fieldwindings are used to generate the radial magnetic flux. Rotor: Coils wound in the rotor (armature) supplied in current via the carbon brushes Carbon Brushes: Mechanical commutators linking DC power supply and rotor windings 1- Motor technologies The following diagram gives an overview of the different families of electric motors. TECHNOLOGY FOR LIFE MITSUBISHI Page 3 of 12

4 Stator S Permanent Magnet Brushes around 100KW up to few Mega Watts for industrial pumps or rollers. They are used as well in electric vehicle and machine tools in a range of power up to 50KW. 1-2 Brushless DC Motors N Windings Rotor Brushless DC motors are part of the synchronous motors family. They are also called permanent magnet motors due to the structure of the rotor. Although they do not operate from a DC voltage source, their name comes from the fact they operate in a similar way as universal DC motors but turned inside out. Phase 1: ON Fig 3: DC motor principle Stator Rotor Regarding the power topologies used for DC motors, the two main solutions (chopper or thyristor bridge) are shown below in Fig 4. DC N S Windings Permanent Magnet DC Fig5: Brushless DC principle Fig4: Power Topologies The benefits offered by this technology are: Separate control of Speed and Torque Torque at 0 speed Widely used in the industry However the disadvantages are: Expensive Maintenance required(brushes) Commutations Sparks Variable Duty-cycle As previously mentioned, DC motors are used in high power applications such as lifts, cranes up Regarding its architecture presented in Fig5, a Brushless DC motor has: Rotor: Permanent magnets Stator: Windings arranged to produce an airgap flux density wave having a trapezoidal shape. The windings being separate poles, they are energised in a pattern rotating around the stator in order to produce rotation. The rotor magnets are then lead by the excited windings to the alignment position where the next commutation will occur. As shown in Fig6, the motor is fed from an inverter producing rectangular currents waveforms. The switching pattern is easily obtained using generally three Hall effect sensors associated with logic gates. TECHNOLOGY FOR LIFE MITSUBISHI Page 4 of 12

5 T1 T2 T1 T2 T3 T3 T4 BDC T5 T4 T5 T6 T6 Logic Gates 3 Hall Effect sensors Fig 6: Brushless DC Drive principle Though Brushless DC motors retain some of universal motors benefits, they offer distinctive advantages as well like: Brushless technology (maintenance free and no spark) High torque to speed ratio High speed (up to 70000rpm) The potential disadvantages of this technology come from the structure itself: Expensive( cost of the rotor permanent magnets) Commutated Torque ripples Brushless DC motors are really popular in consumer applications (vacuum cleaners ), white goods (washing machines ) in a range of power below 20KW. 1-3 Brushless AC Motors Brushless AC motors are also known as Synchronous motors. When looking at their structure, the difference with Brushless DC motors might not be obvious, especially the permanent magnet version. Fig7 presents the basic drive principle for synchronous motors. They are driven with sine wave voltages, and whether permanent magnets or windings are used on the rotor, the motor rotates synchronously with the stator s rotating magnetic field. In variable speed applications, the power topology will often be the one shown above. Though the inverter structure remains the same, the driving strategy will change to use now PWM signals. The different ways to produce PWM signals will be detailed in part two. This type of motor has mainly the same advantages as Brushless DC motors. Another characteristic makes them even attractive: Speed is independent of the load, i.e. it only relates to current. The down side is if the maximum load is reached, the motor will suddenly stall. These motors are used for example in applications like trains, ships, or pumps in a range of power going up to few Mega Watts. As previously mentioned, a very similar structure is then used for both rotor and stator. The main difference is a different organisation of the stator windings. Instead of having different poles as seen before, they are arranged in a sinusoidal distribution. TECHNOLOGY FOR LIFE MITSUBISHI Page 5 of 12

6 T1 T2 T1 T2 T3 T3 T4 B.AC T5 T4 T5 T6 T6 Control Unit Fig 7: Brushless AC Drive principle 1-4 Induction Motors Induction motors play a key role in industrial society as they convert to mechanical energy roughly 30% [1] of all the electricity generated. Using windings on both stator and rotor(instead of mechanical commutators), they are one of the most cost-effective solutions on the market. Fig8 presents the structure of an induction machine. Stator Rotor Rotor: Windings as well Like a DC motor, the torque is created by the interaction of a radial magnetic field produced by the stator and axial currents induced on the rotor. The rotor is dragged around by the rotating field of the stator but it can never run as fast as the field, it just slips as the field rotates. The drive principle is not presented here but it is really similar to brushless AC one. The power topology generally consists in three half-bridges driven by PWM signals. Induction machines main benefits are: Robust mechanical structure Cheaper than most other technologies The weak points are the following: Torque & speed are dependant Tends to slow down when overloaded Windings Fig8: Induction motor principle Windings Induction machines are applicable in a wide range of power, from a few hundred Watts up to a few Mega Watts. Although previously used in constant speed applications (pumps, fans), thanks to new control strategies such as vector control, induction machines can now be used in variable speed applications (Trains, machine tools). Stator: Windings arranged in a sinusoidal distribution. When connected to a 3 phase AC voltage power supply, a rotating magnetic field is then generated TECHNOLOGY FOR LIFE MITSUBISHI Page 6 of 12

7 1-5 Switched Reluctance Motors Switched Reluctance Motors (SRMs) have recently attracted much attention due to their potentially wide-ranging applications as a result of advances in microelectronics. These advances, together with the motor s intrinsic mechanically robust and thermally stable structure, have fuelled the momentum of using SRMs in favour of other motor types in many cost-sensitive and competitive industrial and consumer markets. SRMs have the simplest mechanical structure compared with other types of electrical machines, making them one of the most interesting from an economic point of view. Stator Windings Fig9: 3 Phase 6/4 symmetrical SRM Fig9 presents the structure of a SRM: Rotor Flux Stator: Windings only located on the stator teeth. Quantity determined by the number of phases. Rotor: Steel laminations stacked onto the shaft. The fact that the rotor is only made with steel laminations is the main mechanical difference with conventional motors such as DC or induction motors having either rotor windings or permanent magnets. Indeed, if we consider these machines on the principle of how the torque is produced, they can be classified into two different classes: In the first category, which includes Induction and DC motors, the torque is generated by the interaction of two magnetic fields, one on the rotor and one on the stator. Then, these machines could be differentiated by the geometry used, and on the different method of generating the two fields, with permanent magnets, energised windings or with induced currents. In the second one, with SRMs, the rotation of the rotor is created by the tendency of the motor teeth to align with the excited stator teeth. This is because when a stator winding is energised, a reluctance torque is produced as the rotor moves to its minimum reluctance position. As the first excited phase makes the rotor teeth move to the aligned position, then, the next phase to be excited is chosen to be the most aligned stator teeth, with respect to the required position. As far as the working principle is concerned we can draw comparison between the phenomenon involved in the production of torque for AC & DC motors with the one that makes like poles of bare magnets repel. We can also compare the reluctance torque with the force that attracts iron to permanent magnets. Mechanical simplicity is probably the main SRM s advantage but we could sum up its most interesting characteristics as being: Cheaper than the other motors Motor virtually maintenance free As there is neither permanent magnet nor winding on the rotor, very high speeds can be reached without risk of damage, relatively to comparable motors Tolerant to high temperatures Size. However, SRM drives suffer from different problems. First the rotating field theory is not applicable here, and because of its structure, this motor has highly non-linear characteristics. The disadvantages for that motor could be summarised in the following points[2]: Difficult to control Often noisy Torque ripples Not available yet as standard of the shelf motors. Most of these problems can be compensated by a better understanding of SRM mechanical design and the development of specific algorithms. SRMs are not very common on the actual market place compared with traditional DC and AC motors. However, industries are getting more and more interested in this technology because of the advantages shown above. TECHNOLOGY FOR LIFE MITSUBISHI Page 7 of 12

8 T1 T2 T3 T4 T5 T6 Control Unit Fig 10: SRM Drive principle 2- Introduction to Induction and Switched Reluctance Motor Control For these last few years, two motor technologies have attracted a lot of interest: Induction and Switched Reluctance motors. The reason is that these two solutions seem to be the most cost effective on the market. The aim of this part will be to explain the basics of the key strategies in motor control. 2-1 Induction motor control Although PWM is a very flexible technique, it tends to generate many harmonics. The minimisation of these harmonics will have to be considered in any industrial development. Here are the three main PWM techniques: i ) Triangular wave modulation This strategy is illustrated in Fig 11. V/2 When talking about induction motor control, a lot of people have already heard the expressions PWM, SVM or Vector Control but are still confused about their real meaning. The target for this part is to clarify each of these expressions to provide the reader a good overall understanding. 2-1-a Pulse Width Modulation (PWM) V/2 Carrier Wave Reference Voltage For any analog or digital industrial drive, the desired voltage across the motor phases relates to the output signals of the controller. Pulse Width Modulation is a technique to recreate these desired waveforms from a DC voltage using a direct converter. The control technique consists in first, changing the turn on and turn off time of this converter s power switches but also in controlling the strategy modifying these times. V/2 -V/2 PWM output Fig 11: Triangular wave modulation [3] TECHNOLOGY FOR LIFE MITSUBISHI Page 8 of 12

9 For the controller to generate the correct PWM output that mimics a sine wave for example, it must use a carrier wave. In motor control applications, the carrier wave is the signal used to generate the PWM signal (either triangular or sawtooth). The PWM signal results from the comparison between the carrier wave and the reference signal. This technique has been implemented on Mitsubishi 8 and 16 bit microcontrollers. ii ) Pre-calculated modulation This PWM is pre-calculated relating to the requirements of a specific application such as the minimisation of particular harmonics. The precalculated pattern is then periodically applied on the switches. iii) Space Vector Modulation V α', V β' are the vector coordinates, that represents the three phases system. According to the eight configurations of the inverters power switches, six main vectors with the same modulus define six different sectors (see Fig 13). V V β V 2 ωt V 1 V V ' = V V α α ' β ' V 4 V 1 V 1 K 1 K 2 K V 2 V 3 U 5 K 1 K 2 K 3 V 5 V 6 Fig 12: Power Topology for Induction motors This kind of modulation is based on the fact that the three wanted voltages could be represented in real time by only one vector. During each modulation period, this vector is approximated, using the six power switches commutation sequences. The vector co-ordinates are given by CLARKE transform. This transformation consists in obtaining the CLARKE components: V α', V β', V o' (V o' =0) using the real components: V A', V B' and V C' (the three desired voltages across the motor). These voltages are given by: (1) V V α' β' = /2 3 2 V 1/2 V 3 2 V V A V B V C Induction Motor A' B' C' Fig13: Space Vector diagram [4] The rotation speed of this vector is ω( ω =2πf 0, with f 0 the output voltages frequency). A 2π rotation of this vector represents one rotor revolution. V' is going to be Fig 14 shows how the vector sampled in the first sector. k is the number of modulation periods per sector. (k=3 in this example). V 2 V 1,3 V 1,2 V 1,1 V 1 Fig 14: Sector sampling During a modulation period, each sampled vector can be expressed to a base of four vectors. ω 1 st modulation period TECHNOLOGY FOR LIFE MITSUBISHI Page 9 of 12

10 Therefore, knowing the six switches configurations for each vector, a switching sequence can be expressed by: V '.T c = V 0'.t a + i V '.t b + V ' i+ 1.t c + V 7'.t d The switching times(t a, t b, t c, t d ) for the i th sector are then calculated with : t b = T c.r. t c = T c.r. 3 π sin ( -ωt ) sin ωt 2 t a = t d = 2 1. [Tc -t b -t c ] 2-1-b Vector control (PWM) Vector control is not a PWM technique. It is a modern regulation technique used to control induction machines. This fairly recent control has to be compared to an older one: V/F control. With this technique, for a sensed or even sensorless system, very high performance can be reached with inductance motors. For example, in open loop, the nominal torque can be obtained down to very load speeds (around 30rpm) when having a 1% speed accuracy. In closed loop, the nominal torque is still available at zero speed with an overall 0.01% accuracy on speed. Vector control consists of building a mathematical model of the motor and use the different equations to obtain a separate control of flux and torque. The fairly complex induction machine normally has flux and torque relating to each other, so can then be simply driven like a conventional DC motor where these two parameters are independently controlled. V/F is a simpler control used mainly in applications where constant torque is required. Because this technique is independent from the motor characteristics constantly changing, the same performance, i.e. response time, accuracy can not be reached. 2-2 SRM control 2-2-a Global principle The dynamic control of a motor depends a lot on the torque control. A simple expression of the SRM torque T is: T = 1 i² 2 dl dθ (9) (3) (4) (5) (2) Where - Current ( i ), - Inductance ( L ) - Position ( θ ) The global principle used to control and drive the SRM is then based on a good synchronisation between the time when a phase is excited (i > 0) and the rotor position. In a motoring sequence where a positive torque is required (T >0), each phase has to be excited when the inductance increases as dl/dθ is positive as shown in Fig 15. From equation (1), constant torque can be obtained by maintaining constant current, provided if dl/dθ is constant. Inductance L a L u θ ON Current θ OFF Rotor position Fig 15: Basic Principle for torque control in one phase This part is now going to present the speed control of a three phase SRM implemented with a 16-bit microcontroller, the MITSUBISHI M16C-62. A solution to make the motor run in open loop will be presented. The results obtained in open loop will then be used to draw conclusions about that method. Finally, by implementing the position and speed control, the implementation of the closed loop program will be detailed. 2-2-b Implementation i) Open loop sytem As previously mentioned, the principle chosen to drive the motor is similar to the one used to drive stepper motors. The first objective was to create a simple program to drive the motor TECHNOLOGY FOR LIFE MITSUBISHI Page 10 of 12

11 phase after phase, being able to adjust and The idea has been to create a LUT where the output pattern is stored, i.e. different combinations of the state of the power switches(i.e. on or off). The advantage of using a LUT is that it becomes easier to compensate problems like overlapping. Then, with a pointer pointing to that table, the different values are copied from memory to the output port. To increase or decrease the speed of rotation, the incrementation speed of the pointer is modified changing a timer value. This value is obtained from the result of the Analogue/Digital conversion of the potentiometer s output voltage. reverse the speed. Overall, the results obtained with the threephase motor in open loop were not as good as expected, especially the torque quality. The main reason for this is the lack of position feedback. This parameter appears to be very critical as contrary to conventional stepper motors, the torque generation depends on the inductance profile ( L(θ) ). Then, if we consider the rotor position, it is possible to define an optimal configuration to supply the motor, this configuration being the objective of the position. Fig 17 illustrates this. Inductance profile in the 3 phases Keeping the same basis the program can be improved using the DMA(Direct Memory Access) instead of a pointer to transfer data from memory to the output port. The benefit of using the DMA is that it takes no CPU time. Although it is not a problem in the open loop system, this time can become important in closed loop where many tasks have to be done within a short time. Fig 16 shows the principle of the DMA transfer. Position Position Fig18: position areas to supply the motor Fig 16: Data transfer using the DMA For practical reasons, i.e. to avoid too much current overlapping, it is interesting to modify the LUT to allow each phase current to decrease before starting to supply the next phase as shown on Fig 17. Phase A Phase B T/3 Fig 17: Phase current From the previous diagram, a condition on the position can be raised for an optimised torque generation: Each phase could be supplied only on 1/3 rd of the electrical period. The dwell angle (θ ON - θ OFF ) will then have to be smaller than 120 electrical degrees to avoid any overlapping. As this condition is not checked in open loop, the worst case can be reached if on one period the average torque is positive(t>0) and on the next one it will be negative(t<0). ii) Closed loop In order to improve the results obtained in open loop, a closed loop control called Bang- Bang control introducing position feedback via Hall effect sensors has been implemented. From the signals generated by these sensors, two main operations are executed by the microcontroller. TECHNOLOGY FOR LIFE MITSUBISHI Page 11 of 12

12 Position control The control principle is based on the fact that each phase should not be excited on more than 1/3 rd of the inductance profile period. Changing the commutation speed modifying the data transfer speed to match a position condition is the first operation. Speed control The technique used is called Bang-Bang control. Basically, it consists in fully supplying the different phases until the motor reaches the reference speed, and then stop, repeating always these sequences of On-Off, On-Off; On-Off... Inductance profile in phase 1 Generally, for low power applications, all types of machines can be considered. In high power, the choice will be reduced to universal motors or brushless technologies (synchronous or asynchronous motors). For very low power, DC motors remain the first. Recently, thanks to high performance controllers the medium power market has evolved a lot. New high performance technologies like Switch reluctance motors or Inductance machines using vector control offer low cost solutions for industrial motor control. Mitsubishi Electric offer a wide range of microcontrollers ideally suited for motor control applications, especially the M16C family. Application notes and demo-systems are available for induction, brushless DC and switched reluctance motors. References Output sensor Driving signal Phase 1 Phase 2 Phase 3 N/3 samples TA1 modified by position info N samples Fig 18: Speed control principle Ω Ω ref [1] Austin Hughes, Electric Motors and Drives, Butterworth Heineman, Second Edition 1990, pp [2] T.J.E. Miller, Switched Reluctance Motors and their Control, Oxford Science Publications, 1993 [3] Guy Grellet, Guy Clerc, Actionneurs Electriques, Principe, Modeles, Commande, Eyrolles Edition,1997, pp 211 [4] Guy Grellet, Guy Clerc, Actionneurs Electriques, Principe, Modeles, Commande, Eyrolles Edition,1997, pp Fig 18 illustrates the global principle of the SRM controller implemented. Conclusion Choosing a motor and the drive associated is a fairly complex process. It does not consist only of checking that parameters such as nominal speed, torque or power match the applications requirements. This is only the first step. The second step will be to study the dynamical performance required by the application (acceleration time, response time, maximum torque and speed allowed). And finally, the environment where the system will be used will help choosing the right technology to match specific needs like no spark, low harmonics, or even maintenance free... TECHNOLOGY FOR LIFE MITSUBISHI Page 12 of 12