Pipe Friction Parameters Identification Of Heat-supply Networks Based On Matrix Theory

Centralized heating systems have become important modern municipal infrastructures and will be applied widely when we focus on energy saving and emission reduction for those energy saving, environmental protection and good heating quality. Heat-supply network is an important part of centralized heating system, whose further study has drawn great attention because of its large scale, complicated structure, enormous investment and significant social impact. Nowadays the study of heat-supply networks focuses on hydraulic loading condition computation, reliability analysis, operation and regulation, but there is less study on pipe friction parameters identification of heat-supply networks. However, if actual pipe friction parameters can not be identified, the study mentioned above would just be theoretic simulation or qualitative analysis. In real-world engineering, theoretic simulation results can not always accord with practical operation conditions, so the application value could be reduced seriously. In the study of pipe friction parameters identification, there are a few identification models and identification methods of heat-supply networks pipe friction parameters. Identification accuracy is not high, and consuming time of identification computation is great. These deep reserarches of pipe friction parameters identification have theoretical significance and practical value.This thesis focuses on pipe friction parameters identification of heat-supply networks, and proposes some effective identification methods. Pipe friction parameters identification is expressed as matrix equations. Combined with system observability theory, it is proposed that a necessary condition for observability of pipe friction parameters which is all the nodes pressure heads values are available. Then, a sufficient condition for observability of pipe friction parameters is proposed under multiple hydraulic loading conditions, which provides a new approach to theoretic analysis and field measurement of pipe friction parameters.Considering the observability conditions of pipe friction parameters identification can not be satisfied under common measured conditions, based on solution analysis and pseudo-inverse matrix theories of linear equations, an identification method based on pseudo-inverse matrix theory is proposed. This identification method can be applied under the common measured conditions. This thesis presents a progressive linearization iteration methodology for solving the pseudo-inverse solution of the identification equations, and considers the pipe friction parameters values in the pseudo-inverse solution as the identification results which are approximate to real pipe friction parameters values. Then, this identification method is applied in a case study.Furthermore, if more known operational parameters of heat-supply networks can be obtained, a method by using the initial values of some pipe friction parameters (a group of link pipe friction parameters generally) can be used in pipe friction parameters identification. Considering spacial heat-supply network models can represent heat-supply networks more accurately, we presents pipe friction parameters identification problem of spacial heat-supply networks by the form of partitioned matrixes. In the identification process, combined with characteristics of spacial heat-supply networks, an option method of link pipes which is suitable for spacial heat-supply networks is proposed.Combined with clustering theory, it is proposed that an optimal arrangement method of pipe flow measured points and node pressure head measured points based on reflecting the change of pipe friction parameters. By using matrix differential calculation method, the linear results of pipe flows and node pressure heads reflected by the change of pipe friction parameters can be studied. Then, measured points optimal arrangement is studied by pipe flows and node pressure heads relative sensitivity which can be expressed by design values of heat-supply networks.Finally, the pipe friction parameters identification methods are applied on a multi-function experiment platform and a centralized heating system. By using field measured values, the identification processes are introduced under different field measured conditions, which can show the identification method has higher efficiency and better prospect of real-world engineering.It is proposed that pipe friction parameters identification of heat-supply networks may satisfy observability conditions under multiple hydraulic loading conditions based on system observability theory. When pipe friction parameters observability conditions can be satisfied, a new approach for solving pipe friction parameters identification directly is proposed. Considering the real-world nodes pressure heads and pipes flows measured conditions, an identification method based on pseudo-inverse matrix theory is proposed, which can be used in real-world pipe friction parameters identification in heat-supply networks. We consider the pipe friction parameters values in the pseudo-inverse solution as the identification results which are approximate to real pipe friction parameters values. When more known operational parameters of heat-supply networks can be obtained, we could calculate pipe friction parameters which are more closed to their real values by using a group of link pipe friction parameters initial values. On this basis, some measured values with high relative sensitivity can be used. The research results of the thesis which can be used in future study are theoretical foundation of heat-supply network pipe friction parameters identification and real-world application.