一种热工过程数据协调与显著误差检测同步处理方法

A Simultaneous Data Reconciliation and Gross error Detection Method for Thermodynamic Systems

提出了一种基于冗余解约束遗传算法的鲁棒数据协调方法。引入鲁棒估计作为数据协调问题中的目标函数,不仅对测量数据随机误差的分布形式不敏感,而且抑制了显著误差对协调结果的影响。将数据协调与显著误差检测看作模型辨识与参数估计问题,采用AIC准则调整参数获得最优估计模型。针对鲁棒数据协调目标函数复杂和热工能量平衡约束可能出现隐函数的情况,结合测量冗余的概念提出冗余解约束的遗传算法求解鲁棒数据协调模型。仿真计算表明该方法能够克服显著误差的影响,给出准确的参数估计值,同时检测出系统中的显著误差。在现场热力实验的应用结果进一步验证了方法的有效性。

A robust data reconciliation method based on redundant solved float genetic algorithm was proposed.In order to restrict the influence on reconciliation result caused by random error distribution and gross error in measured data,the robust estimation was introduced as the objective function in data reconciliation problem.The data reconciliation and gross error detection can be regarded as model identification and parameter estimation problem,the Akaike information criterion was used to adjust the model parameter to obtain the optimal model.According to the complex robust objective function and the implicit function as equilibrium constraint in thermal process,combining measurement redundancy concept the redundant solved float genetic algorithm was proposed which was used to solve the robust data reconciliation model.Simulation showed that this method can overcome the gross error influence and provided accurate parameter estimation while all the gross errors were detected.The method was also applied to the field test data and the result verified this method is effective.