Testing motion controllers robustness by emulating electrical and mechanical parameter variations of motor drives

https://doi.org/10.1016/j.simpat.2011.04.011Get rights and content

Abstract

The manuscript extends the method of testing the motion controllers’ robustness by emulating uncertainties and disturbances, caused by real loads and mechanical and electrical parameter variations to multiloop motor drives. Such a contribution is important since multiloop drives are widely used in the industrial environment. The proposed technique requires neither connecting physical load nor actually changing any parameter of the electric drive under test. The microprocessor of the motor drive itself creates the effect of the load torque and parameter variations by applying equivalent feed-forward commands to the corresponding loop controller using digital signal processing software. As a result, the electric drive is virtually perturbed, and the controllers’ behavior resembles the performance of a really perturbed drive controllers. The present method, which can be related to a class of Hardware-in-the-Loop Simulations, can be used to test the implementation of robust control algorithms or at didactic purposes using motion control kits found on the market. Simulations and experimental results validate the proposed algorithm, applied to a Brushless DC motor drive, demonstrating an excellent agreement.

Introduction

Servo control, which is also referred to as “motion control”, is used in industrial processes to move a specific load in a controlled fashion. Servo systems can use pneumatic, hydraulic, or electromechanical actuation technology. Electromechanical systems are typically used in high precision, low to medium power, and high-speed applications. A typical motor drive (Fig. 1) consists of a transmission (T) – load (L) driving electric motor (M), supplied by a power electronics converter (P). The converter is managed by a Digital Signal Processor (DSP), typically executing a two level control algorithm [1]: a high level operator (O), or an application controller, determining the type of operation and supplying reference commands to the low-level controller (C), which ensures the fulfillment of high-level requirements. Basic reasons for using servo systems have not changed for years and include the need to improve transient response times, reduce the steady state errors and reduce the sensitivity to load and parameter variations.

Motor control can in general be broken into two fundamental classes of problems. The first class deals with the command tracking. It addresses the question of how well the actual motion follows the commanded trajectory. The second class determines the disturbance rejection characteristics of the system. Disturbances can be anything from torque disturbances on the motor shaft to incorrect motor parameters estimation. Modern control theory focuses on a variety of subjects, related to motion control. Repetitive control theory deals with tracking and rejecting periodic non sinusoidal inputs [2], while active disturbance rejection theory focuses on rejection load signals of any type [3]. Input loop shaping theory tries to improve the performance of torque-limited drives [4]. Robust Uncertainty and Disturbance Estimation (UDE) [5] and quantitative feedback [6] control theories concentrate their efforts on reducing the influence of parameter variations and noises on drive performance, while stochastic optimal control theory assures optimal operation in presence of input-, output- and state-dependent noises [7].

Recently, there has been an increasing demand in many applications (laboratory equipment and industrial instrumentation, for example) for precision motion control algorithms. Equipment such as automated sample dispensers [8], X/Y tables [9] and laser systems [10] require positioning to be carried out in two or more axes with high accuracy and smoothness. In addition, optical disk drives [11] also require high level of precision and disturbance rejection capability. Hence a lot of research effort was turned into the high precision motion control field.

In order to satisfy the tough requirements of accuracy, the control algorithm and the motor drive as a whole should be tested under conditions of a variety of noises, disturbances and parameter drifts at the development stage. However, a real mechanical load, for example, is quite difficult to implement even it is as simple as a constant torque one. The real mechanical loading becomes even more cumbersome when the load torque should change in time, e.g., a sinusoidal change.

The use of torque-controlled load dynamometers, which is common in engine test beds or in the testing of electrical machines [12], [13], while suitable for a commercial company laboratory, is often impossible in an educational laboratory due to the high price and complexity. In addition, there are servo drives applications where the inertia has a large variation, e.g., in robotics the inertia range could be as wide as 1:10 [14]; moreover, phase resistance may change as a result of high temperature. However, these changes are difficult to be physically implemented at the development stage in order to verify the control algorithm and test the motor performance. Mechanical parameters variations are also almost impossible to apply in a controlled way in real time, since inertia and friction are inherent physical characteristics of a motor. Load simulation techniques are usually employed in order to emulate different loads and parameter drifts by adding more cumbersome hardware to the system [15], [16], [17]. A more recent approach to evaluate system’s performance prior to including all the hardware components is Hardware-in-the Loop (HIL) Simulation [18], [19]. HIL simulation is a technique that is used increasingly in the development and test of complex real-time embedded systems. HIL simulation provides an effective platform for developing and testing real-time embedded systems by adding the complexity of the plant (or a part of the plant) under control to the test platform by adding a mathematical representation of all related dynamic systems. An HIL simulation may also include electrical emulation of sensors and actuators. These electrical emulations act as the interface between the plant simulation and the embedded system under test. The value of each electrically emulated sensor is controlled by the plant simulation and is read by the embedded system under test. Likewise, the embedded system under test implements its control algorithms by outputting actuator control signals. Changes in the control signals result in changes to variable values in the plant simulation. HIL has been widely used in different fields such as motor drives [20], engine control [21], manufacturing systems [22], power quality assessment [23], power electronics and energy conversion [24], [25].

However, most of the cases of using HIL belong to the vehicle electronic control units [26], [27], [28] and electric drives industry [29], [30], where it has proven to be one of the best ways to develop embedded systems.

The goal of this manuscript is to build up a method which allows testing motion controllers without disconnecting the electric motor and the power converter. The suggested technique utilizes an inverse to the well-known disturbance estimation and cancellation approach [31], [32], where the uncertainties and disturbances, reflected to the plant input, are estimated and added with an opposite sign to the control signal to be cancelled out. Instead of estimating and cancelling the uncertainties and disturbances, the presented approach creates uncertainties and disturbances and adds them to the nominal plant, thus making the controllers “see” a plant with parameter variations and disturbances. The presented method, initially proposed in [33] and elaborated in [34], [35], can be associated with a class of HIL Simulations, however unlike in the general HIL case; the nominal physical plant is present in the system. Only plant variations and external mechanical loads are emulated by the DSP. The main contribution of this paper is presenting a novel multiloop approach, allowing decoupling of emulating mechanical and electrical variations and hence testing each control loop independently using a simple first-order description.

The rest of the manuscript is arranged as follows. Section 2 reviews the basics of a typical Brushless DC motor drive; the proposed algorithm is derived in Section 3 and its application to a Brushless DC motor drive is presented in Section 4. The implementation aspects and the results are discussed in Section 5; the manuscript is concluded in Section 6.

Section snippets

Typical Brushless DC motor drive

A typical cascaded multiloop closed-loop position controlled drive is shown in Fig. 2 [14]. This is a feedback system, which is used to control position and/or velocity, and/or current/torque (sometimes the acceleration is important, it can also be controlled by inserting an additional loop between the speed and current loops). The DSP executes the algorithms to close the desired loops and also handle machine interfacing with inputs/outputs, terminals, etc. Alternatively, a separate amplifier

The proposed algorithm

Consider a linear, time invariant, single input single output first order nominal system defined byxn=anxn+bnuc,where xn is the nominal state vector, uc is the control input, an and bn are known constants. If the nominal system is perturbed by parameter variations and disturbance inputs, the resultant system can be represented asxd=(an+Δa)xd+(bn+Δb1)uc+Δb2ud,where xd is the state vector, Δa, Δb1, and Δb2 are the system uncertainties and ud is the unknown disturbance. Eq. (4) can be rewritten

Application to a BLDC drive

In order to apply the presented algorithm to a multiloop controlled BLDC drive, the signals iDIST and uDIST are created in the software, reflecting the electrical and mechanical disturbances, respectively, as shown in Fig. 6. The signals are added to the outputs of the speed and current controllers, which were tuned according to the nominal motor parameters to create the effect of real parameter variations and make the controllers “see” a disturbed plant and respond accordingly. Note that the

Implementation and discussion

Consider a BLDC motor with nominal parameters given in Table 1. The motor is a part of MCK240 motion control kit. The MCK240 is a complete stand-alone system that allows to experiment with and use the low cost TMS320F240 (‘F240) 16-bit fixed point DSP controller for digital motion control (DMC) applications. The ‘F240 DSP controller is designed to implement advanced DMC applications, by integrating high performance of a DSP core and the on-chip peripherals of a microcontroller into a single

Conclusion

A method of testing the controllers employed by emulating uncertainties and disturbances caused by real loads and parameter variations was extended in the manuscript to multiloop drives. It is implemented by creating signals, reflecting the electrical and mechanical disturbances, in the DSP software and does not require any hardware changes and add-ons to the motor drive. The signals are added to the outputs of the speed and current controllers, which are tuned according to the nominal motor

References (35)

  • O. Yaniv

    Quantitative Feedback Design of Linear and Nonlinear Control Systems

    (1999)
  • D. Bertsekas et al.

    Stochastic Optimal Control: The Discrete-time Case

    (1996)
  • K-H. Kim et al.

    A nonlinear speed control for a PM synchronous motor using a simple disturbance estimation technique

    IEEE Trans. Ind. Elec.

    (2002)
  • J. Solsona et al.

    Disturbance and nonlinear Luenberger observers for estimating mechanical variables in permanent magnet synchronous motors under mechanical parameters uncertainties

    IEEE Trans. Ind. Elec.

    (2003)
  • B. Kim et al.

    Advanced disturbance observer design for mechanical positioning systems

    IEEE Trans. Ind. Elec.

    (2003)
  • K. Oshishi et al.

    Robust tracking servo system considering force disturbance for the optical disk recording system

    IEEE Trans. Ind. Elec.

    (2006)
  • E.R. Collins et al.

    A programmable dynamometer for testing rotating machinery using a three-phase induction machine

    IEEE Trans. Energy Convers.

    (1994)
  • Cited by (3)

    View full text