INTEGRATED BATTERY-ELECTRIC VEHICLE MODELING AND PID OPTIMAL CONTROL BASED ON LINEAR QUADRATIC PARTIAL STATE FEEDBACK FOR ENERGY EFFICIENCY OF ELECTRIC VEHICLE

• Main Author: Ristiana, Rina
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/43973
• Institution: Institut Teknologi Bandung

Currently, electric vehicles (EVs) are not widely used as a means of mass transportation because EVs have limited energy stored in batteries. One solution to overcome the limitations of battery energy storage is energy optimization based on an optimal control of energy consumption. An optimal control is designed to minimize the performance index by taking into account the constraint variables. The performance index describes the goal to be achieved by the system, while the variable constraints explain the characteristics of the physical constraints and the system model. Therefore, selecting the optimal control system model and method is an important part of designing an optimal control. This research proposes an energy optimization strategy with an optimal control approach through the determination of the system modeling and the method of designing an optimal control for energy efficency in electric vehicles. The most current model of electric vehicles is called the motor-vehicle model, which includes motor dynamics and longitudinal vehicles. In this dissertation study, the motor-vehicle model is simplified by ignoring the wind resistance. This model is called the ordinary model. As a further development of the model, a vehicle model that takes into account the dynamics of the battery and wind resistance is built into the motor-vehicle model. This model is called the integrated battery-electric vehilce model or simply the integrated model. The integrated model was also built by considering variations in control variables such as torque or speed, gear train, and battery type used. Currently, the design of controllers using motor-vehicle models assumes that the source of battery energy is fixed at all times. In fact, the battery energy changes. Therefore, we need a model that takes into account the interaction between the battery and the motor-electric vehicle system, thus building an integrated battery-electric vehicle model. System model validation is done through open-loop and closed-loop system validation. Open-loop system validation includes an analysis of the model for each subsystem and the overall system. First, real parameters from electric vehicles were applied in the electric vehicle model. Then, a simulation is carried out to observe the open system responses in each subsystem and the complete system. The simulation results were then compared with real system responses of electric vehicles obtained from Molina ITB Model-3 test-drive results. The closed-loop system validation involved comparing several cases of controller designs using the model by means of computer simulations as well as simulations on a hardwarebased minimum vehicle (EV testbed platform) system based on hardware in a loop simulator (HILS). To see the energy saving potential, regular and integrated models were used for the controller design. The controller used in this study was an existing linear controller. The ordinary model was a linear model, whereas the integrated model was a non-linear model. Specifically, for the nonlinear model linearization was carried out, called linearization of the integrated model. The regular model and the linearized integrated model were used in the same controller design to obtain a response time that could be considered the same. All cases of the controller design were applied to nonlinear integrated models or complete models of the electric vehicle system. When the response time is kept the same or similar and the environmental conditions, the speed pattern, and the formulation of the energy calculation are kept the same, different energy consumption values will be obtained for each controller case. This principle was used as the basis for consideration in the analysis of the energy consumption in this study. The simulation, validation and test results showed that the use of an integrated model has the potential to be more energy-efficiency than the use of an ordinary model. The optimal control design proposed in this study is a linear-quadratic optimal PID control design based on partial status feedback and its combination. PID control is a conventional controller with a simple structure that is widely used. The main problem of PID control lies in tuning each of its control constants (Kp, Ki, Kd). For high-order systems, the PID controller tuning design method is used by reducing the system order or reducing the control order, whereas in this study the method proposed is the PID controller tuning design method or optimal PID control without reducing the system order and selecting the variables to be feedback into the system without the need for an observer. This optimal PID control has a linear quadratic performance index. The determination of the performance index is based on several alternative energy formulations available in electric vehicles, such as mechanical energy, electrical energy, control energy, and battery energy. The combination of several alternative energy formulations provides a linear quadratic performance index formulation that contains multiplication between the status variable and the control variable. The performance index formulation provides potential energy savings if used as a basis for the optimal control design method. For systems expressed in canonical form, optimal PID tuning can be done by adopting a linear quadratic method based on partial status feedback by solving linear inequality matrices, which are developed into linear-quadratic optimal PID controls based on partial status feedback and their combinations. For comparison in the energy consumption analysis, various linear-quadratic optimal controls based on full status feedback were also designed. All linear controller designs, both full-status feedback-based controls with observer and optimal PID controls based on partial status feedback and their combinations, were applied to the integrated model. The simulation and testing results showed that the linear-quadratic optimal PID tracking control based on partial status feedback was more energy-efficient than the other controllers. The contributions of this study are the integrated battery-electric vehicle model and the optimal PID control based on linear quadratic partial status feedback method. In general, the integrated model and the optimal PID control method can be applied to electric vehicles, such as cars, that are equipped with automatic driver systems such as an auto-drive system or an assisted drive system including cruise control.