Thermal Parameter Identification And PID Parameter Optimization Based On The Golden Section Method

Thermal parameter identification and PID parameter optimization can be transformed into an optimization problem at some angle. With the rapid development of swarm intelligence theory in various fields, there have been many scholars use swarm intelligence algorithm for thermal parameter identification and PID tuning. However, swarm intelligence algorithm have strong randomness, slow convergence, and time-consuming when it been used in thermal parameter identification and PID parameter optimization. Aiming at these defects, This paper aims at applying certainty and speed method to thermal parameter identification and PID parameter optimization.First of all, This paper introduce how to transform thermal parameter identification and PID parameter optimization into an optimization problem. Secondly, simulation verifies the two problems' target function in a characteristic of unimodality to each parameter. And then, respectively combine the coordinate rotation method or the Powell algorithm with the golden section method, of direct formation of two new deterministic algorithm, formed into two new deterministic algorithm, it's binding process is first based on the coordinate alternation method or the Powell algorithm to determine the direction of a one-dimensional search, then using golden section method to one-dimensional optimizing. Finally, used the two new algorithm into thermal parameter identification and PID parameter optimization, and compared with PSO to highlight the efficiency of the two new algorithm. In addition, this paper combines the internal model control PID, for the prevalent one order-delay plus dead time, second order-delay plus dead time thermal process, this paper offers a method to ITAE index as an objective function, use the golden section method optimized the IMC-PID filter parameters.Based on the application of field data indicate, the method on the quickness, accuracy is superior to the particle swarm algorithm, take short time, quick early convergence and occupy less memory, more suitable for thermal online identification of the object and on-line optimization of PID control system, then combine two application provide, provides a new method for thermal control system's adjusts the parameters online fast auto-tuning.