基于matlab-simulink联合并行仿真计算

%% ======================= 三容水箱闭环控制系统 ========================== %% 1、完成三容水箱系统参数辨识。2、完成PID控制器的参数整定 clc; clear; close all; %通过辨识得到确定的参数，获取一组输入输出值 A = 130; %水箱横截面积 cm^2 Sn = 2; %连接阀横截面积 cm^2 a1 = 0.6; %阀门流量系数 假设阀门规格相同 a2 = 0.6; a3 = 0.6; g = 981; %cm/s^2 Up_Lim_1 = 20; %三容水箱输入流量上限 Low_Lim_1 = 0; %水箱目标水位 cm Ts = 50000; H_in = [10*ones(2000,1);linspace(10,40,10)';40*ones(Ts,1);linspace(40,15,10)';20*ones(6000,1)]; sim_time = 1:1:length(H_in); h_in = [sim_time', H_in]; simtime = length(H_in); % 仿真时间 % level_cons = 0; %水箱初始水位 cm modelName = 'ThreeTank_Control2017a'; nfspath = '\\192.168.3.7\f\NFSD\'; %设置输入格式 ；两输入 TC = Simulink.SimulationData.Dataset;%输入结构体名称 element1 = Simulink.SimulationData.Signal; element1.Name = 'InputSig1'; element1.Values = timeseries(H_in); %输入时间序列 TC{1} = element1; element2 = Simulink.SimulationData.Signal; element2.Name = 'InputSig2'; element2.Values = timeseries(H_in); TC{2} = element2; % p, i, d 参数序列 K_p = 30:1:35; K_d = 0:1:5; K_i = 0:1; %生成仿真任务 numTasks = length(K_p) * length(K_d) * length(K_i); outputs = cell(numTasks, 1); %创建一个单元格数组存储仿真结果 tic; %创建并行池 delete(gcp('nocreate')); if isempty(gcp('nocreate')) pool = parpool('j124',8); end %提交并行计算仿真任务 simCount = 0; for p = K_p for d = K_d for i = K_i simCount = simCount +1; simResult(1,simCount) = p; simResult(2,simCount) = d; simResult(3,simCount) = i; %定义仿真函数 % simFcn = @(modelName, TC, p, d, i) simulateModel(modelName, TC, P, D, I); %提交任务 使用并行计算工具箱 f(simCount) = parfeval(pool, @simulateModel, 1, modelName, TC, p, d, i); %参数：p,fcn,numout,in1,in2,... end end end %获取并存储仿真结果 minSSE = Inf; h_level = []; for i = 1:simCount %等待仿真任务完成并获取结果 [completedIdx, outputData] = fetchNext(f); %存储仿真结果 outputs{completedIdx} = outputData; h_level = outputData.simout.Data; %计算拟合优度指标 SSE = sum((h_level(1:end) - H_in).^2); %误差平方和，SSE越小，曲线拟合程度越 if(minSSE >= SSE) minSSE = SSE; minIndex = simCount; end end %关闭并行计算池 delete(gcp('nocreate')); toc; %使用最优的 P I D 值，计算输出响应 Kp = simResult(1,minIndex); %49 Kd = simResult(2,minIndex); %8 Ki = simResult(3,minIndex); %0.8 % assignin('base','Kp_1',Kp); % assignin('base','Kd_1',Kd); % assignin('base','Ki_1',Ki); % assignin('base','Kp_2',Kp); % assignin('base','Kd_2',Kd); % assignin('base','Ki_2',Ki); % % %仿真 % sim(modelName); % figure(1); % plot(outputs{minIndex, 1}.simout.Time, outputs{minIndex,1}.simout.Data); %输出响应 液位高度 % hold on; % plot(sim_time', H_in); %期望液位高度输出 cm % title("期望液位高度与输出响应液位高度对比"); % legend("实际输出液位", "期望输出液位"); % xlabel("仿真时间/s");ylabel("液位高度/cm"); %仿真模型函数 function outputData = simulateModel(modelName , TC, P, D, I) %加载simulink模型 load_system(modelName); stoptime = TC{1}.Values.Time(end); %设置仿真参数 simIn = Simulink.SimulationInput(modelName); %模型名称 simIn = setExternalInput(simIn, TC); simIn = simIn.setModelParameter('StartTime', '0', 'StopTime', num2str(stoptime), 'FixedStep', '1'); simIn = simIn.setBlockParameter([modelName '/PID Controller'], 'P', num2str(P)); %运行变量名称 simIn = simIn.setBlockParameter([modelName '/PID Controller'], 'I', num2str(I)); simIn = simIn.setBlockParameter([modelName '/PID Controller'], 'D', num2str(D)); simIn = simIn.setBlockParameter([modelName '/PID Controller1'], 'P', num2str(P)); simIn = simIn.setBlockParameter([modelName '/PID Controller1'], 'I', num2str(I)); simIn = simIn.setBlockParameter([modelName '/PID Controller1'], 'D', num2str(D)); %执行仿真 outputData = sim(simIn); close_system(modelName); %获取仿真结果 % outputData = simOut.get('outputData'); end