Implementing a Hardware-Based Big-Bang Big-Crunch optimized controller with Fractional-Order for a Heating, Ventilation, and air conditioning system

Industrial and commercial buildings are significant energy consumers, worldwide, emphasizing the need for effective control techniques to reduce power consumption. This paper presents an FPGA implementation of a fractional order proportional-integral-derivative (FOPID) controller for a Heating, Ventilation and Air Conditioning (HVAC) system. The FOPID model enhanced improved dynamic performance and system robustness compared to classical PID controllers. However, tuning the FOPID controller is challenging due to the need to adjust several parameters, i.e., the proportional, integral, and derivative gains of the PID model, as well as the fractional differential and integral orders. To address this, the Big Bang-Big Crunch (BB-BC) evolutionary algorithm is adopted to tune the controller parameters optimally. The BB-BC algorithm is executed to minimize six different performance indices. The HVAC system is represented by a second-order plus time delay (SOPTD) model, yielding six different cases are obtained by combining three delay values and two gain values. A highperformance pipelined implementation of the BB-BC algorithm andthe FOPID controller is designed and realized on a Xilinx FPGA device. The implementation has successfully reduced the oscillations of the HVAC output signal by employing auto-tuning of control parameters based on BB-BC, while improved results have been also achieved in terms of rapid convergence. Additionally, a substantial reduction of 80 % in the output control signal and 73 % in error has been observed when using a window size (L) of L = 256, in comparison to L = 32. The simulation results demonstrate the highly efficient performance of the FOPID-BB-BC control unit after optimal tuning.

Automated summary

This paper presents an FPGA implementation of a FOPID controller for HVAC systems. The controller parameters are optimized using the BB-BC evolutionary algorithm, resulting in reduced oscillations and improved convergence. The method has potential importance in reducing energy consumption.