基于改进引力搜索算法的水轮机调节系统仿真

Simulation of hydraulic turbine regulation system based on improved gravitational search algorithm

针对现阶段水电机组存在多种复杂工况、工程计算受限于算法本身的复杂性等问题,提出一种改进的引力搜索算法(改进PSOGSA),以此提高水轮机控制参数的优化性能,弥补传统控制策略难以满足动态需求的不足.首先,结合PSO算法,在GSA的速度更新公式中引入学习因子进行改进.其次,应用一种权重系数优化其位置更新公式,提高算法的自适应性.最后,结合相关仿真建模试验,使用所提改进PSOGSA对水轮机调节系统PID参数进行优化调节.仿真结果表明,在5%空载频率扰动下,改进PSOGSA的PID控制器明显优于上述传统算法,所调节的模型系统能在更短时间内趋于稳定,此时的超调量远低于传统算法,表明此改进PSOGSA在后续迭代中具备更高的迭代效率,并且改善了常规算法中易陷入局部最优的问题,从而证明了改进PSOGSA的合理有效性,水轮机调节系统的控制效果在一定程度上得到优化.

In response to the current challenges faced by hydropower units,such as multiple complex operating conditions and limitations in engineering calculations due to algorithm complexity,an improved gravitational search algorithm(improved PSOGSA)was proposed to enhance the optimization performance of hydraulic turbine control parameters and address the shortcomings of traditional control strategies in meeting dynamic demands.Firstly,combined with the PSO algorithm,a learning factor was introduced into the velocity update formula of GSA for improvement.Secondly,a weighting coefficient was applied to optimize the position update formula to enhance the algorithm′s adaptability.Finally,combined with relevant simulation modeling experiments,the proposed improved PSOGSA was utilized to optimize the PID parameters of the hydraulic turbine regulation system.The simulation results demonstrate that,when subjected to a 5%no-load frequency disturbance,the improved PSOGSA PID controller is significantly superior to the conventional algorithm mentioned above.The regulated model system achieves stability within a shorter time period,and its overshoot at this time is much lower than that of conventional algorithms.These findings indicate that the improved PSOGSA exhibits higher ite-ration efficiency in subsequent iterations,and improves the problem of easily converging to local optimum in conventional algorithms.Thus,the reasonable and effectiveness of the improved PSOGSA are validated,and control effect of the hydraulic turbine regulating system is optimized to a certain extent.