基于Kalman滤波的换热器非线性状态参数校正

炼化工艺系统中换热网络数据的准确提取将直接影响到集成优化方案和优化控制的性能。针对换热器非线性状态参数的数据校正，构建了换热器分段线性集总参数传热过程模型，有效地解决了换热器流股物性非线性变化所引起的非线性状态空间方程求解的问题；提出了分段线性的Kalman滤波状态空间方程建立和换热器状态参数校正方法，并通过蜡油加氢装置反应流出物高压换热器工业实例阐述了所提出方法的实现过程和效果。研究表明：换热器分段线性集总参数模型中分段数对Kalman滤波的计算收敛性具有重要影响，随着分段数的增大，换热器状态参数收敛于固定值，分段数需根据计算精度通过试差确定。本文方法可对换热器非线性状态参数实施有效的数据校正，对流股物性进行分段线性化处理具有较高的计算精度，可用于大温差或物性变化较剧烈情况下换热器非线性状态参数的数据校正。

Data reconciliation for nonlinear state parameters of heat exchangers using Kalman filtering

In a refinery accurately extracting data of heat exchanger network has intimate influence on the scheme of heat integration and the performance of optimal operation.For data reconciliation of nonlinear state parameters of a heat exchanger,a lumped parameter heat transfer model,which featured partitioning the heat exchanger with piecewise linear sub-models,was proposed to overcome the difficulty of nonlinear state space equations induced by nonlinear change of stream properties in the heat exchanger.A set of state space equations on the basis of piecewise linear Kalman filtering was established.A method of data reconciliation for nonlinear state parameters of heat exchangers was proposed.The procedure of the proposed approach and its behavior were exemplified by data reconciliation for a high pressure heat exchanger in a wax oil hydrotreating unit.The results showed that the convergence of the proposed model was strongly dependent upon the number of partitioning segments for a heat exchanger.The state parameters of heat exchanger tended to be convergent to a constant as the number of partitioning segments increased.And the number of partitioning segments should be determined by trial-and-error according to requirements of computational precision in practice.When the inlet temperature of heat exchanger was measured, the outlet temperature could be rectified using the proposed method.And data reconciliation on the inlet and outlet temperatures of the heat exchanger could be performed iteratively.The proposed method could be effectively used for data reconciliation of nonlinear state parameters of heat exchangers.The proposed approach had higher accuracy for data reconciliation of a larger difference between inlet and outlet temperatures or dramatic changes of physical properties of streams in heat exchangers.