An Application of Feedback Linearizing Controller with Disturbance Observer for DC Motor Real-Time Speed Control

The main objective of this paper is the hardware implementation of robust speed and torque control of DC motors. The proposed speed controller is a feedback linearizing controller (FLC) with a disturbance observer (DO). The designed control circuit can withstand non-linearities of the system and other unpredictable disturbances. In addition, the exact mathematical modelling of the controller is used and thus the system can be easily reproduced to any higher rating. The converter used to interface the DC motor is a bi-directional H bridge converter that can supply and accept power at any voltage conditions. The developed system is experimentally tested under different case studies of motoring and regenerating conditions. From the results, it is verified that the objective is met, and the designed prototype system can work accurately to follow the speed and load torque. The designed system can be used for variety of industrial applications including electric traction, robotics, and various fixed as well as variable speed electric drives.


Introduction
Due to its simple construction, ease of control, low cost, and high power to weight ratio, the direct current (DC) motor is the most popular motor in the modern industrial era.In order to achieve the anticipated drive needs in some application areas, control of the DC motor's speed and load torque is very crucial and thus the controller design is therefore still one of the difficult problems [1].In the literature, various control strategies are put forth.The most frequently used control techniques in previous works include proportional-integral (PI) controller [2] , anti-windup PI controller [3] , fuzzy PI controller [4], linear state feedback controller (LSFC) [5], LSFC with disturbance observer (DO) [6] and sliding mode controller [7].Designing a closed loop controller for a DC machine that deals with nonlinearities and perturbations remains a serious challenge, especially when strong dynamic performance is required.The majority of DC motor control schemes assumed that the system is linear, hence non-linear behaviour and other system uncertainties were not thoroughly examined in earlier publications.Damped dynamics and nonlinear behaviour are integrated into the system as a whole through the switching action of DC-DC converters.Additionally, a variety of elements, including model errors, process noise, and sensor noise, may impact the system's dynamic transient behaviour.In this paper, DS1202 microlabbox control desk is adopted for speed controller implementation of the DC traction motor.The suggested control system is created and simulated in the MATLAB/Simulink environment, and linked to the converter using the DS1202 control desk.
The overall system is described in section 2, The converter controller design in section 3 and the implementation of controller in the Simulink for hardware interface using DS1202 in section 4. The results and discussion are demonstrated in section 5 and finally the conclusion in section 6.

System description
The schematic of the implemented DC motor control hardware circuit is shown in Figure 1 with different components marked.Bidirectional DC source (Chroma 62060D-600 programmable DC power supply) is used as the input power supply which can deliver and recover power.The SEMITEACH IGBT, a multi-function IGBT converter module with an integrated isolated IGBT driver, is used as the converter.Capacitors are connected in parallel across the input and output of the converter terminals.The speed sensor E40H12-2000-3-V-24 is attached to the shaft of the DC motor to sense the motor's output speed.This study uses an MBL/MBZ brake system with a controller to control the motor input load torque.The field winding of the DC motor is coupled to a Chroma 62000H programmable DC power supply.An isolated voltage and current sensor module (USM-31V) is used to measure the current and voltage in various parts of the overall system.The voltage-current sensor output (analog) and speed sensor output (digital) are sent as input to the real-time simulator DS1202.The gate pulses that control the switches in the power electronic converters are produced based on the input signals and the model built using the Simulink interface.A level shifter is used since the pulse output of the DS1202 is lower than the input voltage of the IGBT driver circuit.The developed system can supply and recover power and therefore, a bidirectional H bridge converter is used in the system as depicted in Figure 1.The converter can operate in forward boost mode, forward buck mode, reverse boost mode and reverse buck mode depending on the demanded voltage at the DC link capacitor (C) and input voltage.

Converter controller design
The schematic diagram of the controller used in Simulink to provide the necessary switching pulses for the H bridge converter is shown in Figure 2. Because the circuit uses a H bridge converter and compares the reference voltage generated by the FLC (VDC_ref) with the input voltage to determine the H bridge converter's switching action, the usage of the three control loops is crucial.Here, the FLC with DO controller compares the reference speed requirement to the actual motor speed and generates the DC reference voltage necessary at the output DC link capacitor to maintain the needed speed and load torque.This generated reference voltage will decide the mode of operation of bidirectional H bridge converter in the circuit by correlating to input voltage.To determine the necessary duty ratio for the converter's switches, generated reference current from the voltage PI controller is compared to the converter inductor current.

Figure 2. Schematic diagram of the controller for H bridge converter
In the earlier works, the majority of DC motor control situations presupposed a linear system.A linearization method used in feedback linearizing controller differs significantly from the traditional Jacobian linearization in that it approximates linear features over a narrow working range.The feedback linearizing method cancels out the system's non-linearities using algebraic transformation, making the linearization valid over a wide operational region.In addition, the DO can impact the controller operation in two distinct ways.FLC can only add a proportional action to the controller, whereas DO can also add an integral action, which means that fine-tuning the associated performance indicator can aid in lowering the steady state error.Furthermore, DO can compromise the effect of mathematical approximation in the design of the entire system controller and can reduce the number of sensor requirements within the system.
The speed controller FLC with DO can be defined by the expression [8], Where,  = () −1 ;   =  0 ;   = ( 0 +  1 );   =  1 + ;  0 =  0 Here, ξ0 and ω0 are the damping ratio and natural frequency of oscillation of outer speed control loop.Ra is the armature resistance, La is the armature inductance, Kb is the back emf constant, KT is the torque constant of traction motor.Here armature current (Ia) and actual speed output (ω) are the state feedback components of the controller.μ is the constant associated with DO and this value can be tuned to reduce the steady state error of the system.PI controllers are employed as voltage controller and current controller.The design of voltage PI controller parameters is adopted from reference [8].

Implementation of controller in the Simulink
Simulink can be used to build model-based designs, such as control system designs.The DS1202's I/O ports can be accessed through the Simulink library browser.FLC with DO controller is used in this work, with actual speed output, reference speed, and armature current as input to the controller as shown in Figure 3.The reference armature voltage required to maintain the reference speed at any load torque is generated by the speed controller.The real-time simulator DS1202 receives analog input from USM-31V sensor module in the form of analog voltage and current outputs.

Results and discussion
The experimental setup is depicted in Figure 5 and different case studies are performed to test the protype system developed in the laboratory.Case study 1: This study aims to verify the performance of the designed motor control system to withstand sudden increase in speed.In this study the speed of DC motor is varied in steps and increase till 650 rpm. Figure 6 (a) shows the reference speed given (green) and the speed follower characteristics (pink).The speed is plotted as voltage in the oscilloscope with a ratio of 1 rpm/V.From Figure 6(a), it is evident that the developed system can follow the sudden change in speed.
The objective of FLC with DO controller is to produce the required DC link voltage to follow the demand ed speed.Therefore, it is important to check the reference voltage generated by FLC and the generated voltage at the DC link by the controller.Therefore, the controller is able to quickly reach the reference speed even while abrupt acceleration causing the real speed output to overshoot.The FLC with DO controller can track the reference speed and the operation of converter in forward boost and buck mode is also specified in Figure 6(b).This study also aims to show the working of the H-bridge converter in boost and buck mode.The performance of the controller is assessed for Case Study 1, where the system undergoes a sudden shift in speed.The time it takes the system to stabilize within a 1% tolerance band following an input stimulus, commonly known as the settling time, is found to be 0.9 seconds.

Case study 2:
This study aims to verify the ability of the designed system to follow the change in speed and load torque.In this study, the speed is varied continuously and on reaching 500 rpm, the speed is maintained constant.At this constant speed region, the input load torque to the motor is varied in steps, which results in increase in resulting input armature current and resulting torque output of the motor.The developed controller follows the steady increase in reference speed as shown in Figures 7(a).The reference speed is represented in green colour and the actual speed in pink colour and it is evident that the system can follow the speed with variation in speed and also can maintain constant speed despite change in load torque.Figure 7(b) represents the voltage follower characteristics of the designed system.As seen in Figure 7(b), at constant speed region, when the load torque is increased, there is increase in the demanded DC link voltage.The designed system can follow the change in voltage to maintain the speed constant in that region.Figure 8 shows the rise in armature current as a result of the constant speed increase and load torque's abrupt change.The armature current variation at variable speed area and at the region of load torque change is depicted in this figure .The system can withstand sudden increase in armature current demand by the motor.

Conclusion
In this paper the design of the speed and load torque control of DC motor is implemented as hardware prototype using FLC with DO.The features of the designed circuit are: 1) Forward and reverse power flow is possible and therefore, the model can be used for electric vehicle traction motor to recover the braking energy.
2) The technique used here is armature voltage control, that is variable DC link voltage controller.
3) With this technique of controller, multiple sources can be incorporated at the input side using a power split ratio.
The prototype hardware solution is put to the test using various speed and load torque case studies.Both the motoring mode and the regenerative braking mode are verified for functionality.Experimental results show that the designed system can sustain speed despite a sudden increase in load torque.

Figure 1 .
Figure 1.Over all system description for the speed control of DC motor

Figure 4
demonstrates how to utilize the I/O port to measure current and voltage.The analog input is converted to a digital signal in Simulink's ADC block, which may then be used as a reference input by the controller that has been configured for real-time operation.

Figure 3 .
Figure 3. Simulink model of closed loop speed controller

Figure 4 .
Figure 4. Real time Simulink model for measuring voltage

Figure 6 (
b) shows the reference voltage as a red line, and the voltage output from the controller as a blue line.The obtained graph as shown in Figure 6(b), represents the voltage follower characteristics of the designed system.

Table 1 .
Transient performance analysis using fine-tuned PI controller