一元和二元函数的数据建模方法及其在火电机组中的应用

研究了一元函数和二元函数的数据建模问题,给出了2种一元函数数据建模方法(基于给定间隔点横坐标的分段直线拟合方法和基于曲线拟合的分段直线拟合方法)及1种二元函数数据建模方法(分段直线拟合插值混合型建模方法),并将其应用于某600 MW超临界火电机组负荷特性的建模中。研究结果表明:这些方法在建模中是有效的; 2种一元函数数据建模方法均具有直线分段数可控、数据适用性广等优点。相比基于给定间隔点横坐标的分段直线拟合方法,基于曲线拟合的分段直线拟合方法具有算法简单、可自动确定分段节点等优点,相比已有的二元函数数据建模方法,分段直线拟合插值混合型建模方法具有灵活性高、通用性强、模型形式简单等优点。将给出的数据建模方法开发成软件,可以通过交互的方式十分方便地完成从历史数据到一元或二元函数模型的建模全过程。

Data Modeling Methods for Unary Function and Binary Function Models and Their Applications in a Thermal Power Generating Unit

The data modeling problem of unary function and binary function is investigated,and two unary function data modeling methods,i.e.,the piecewise linear fitting method based on the given spaced point abscissa and the piecewise linear fitting method based on curve fitting,and one binary function data modeling method,i.e.,the hybrid method combining piecewise linear fitting and interpolation,are proposed and applied to the modeling of load characteristics of a 600 MW supercritical thermal power unit.The results show that these methods are effective in the data modeling of unary and binary functions; both the proposed unary function data modeling methods have the advantages of controllability in number of straight line segments and good applicability to data; compared with the piecewise linear fitting method based on the given spaced point abscissa,the piecewise linear fitting method based on curve fitting has the advantages of algorithm simplicity and automatic determination of piecewise nodes,which has strong practicability; compared with the existing binary function data modeling methods,the hybrid method combining piecewise linear fitting and interpolation has the advantages of high flexibility,strong universality,and simplicity in model form.Once the proposed data modeling methods are developed into software,the whole modeling process from historical data to a unary or binary function model can be accomplished conveniently by means of interaction.