Modeling Of Thermal System And Performance Analysis Prediction Of Power Plant Boilers Based On Fluid Network Method

Combined with the performance analysis platform of supercritical power station boiler, this paper uses the fluid network to build a general mathematical model of boiler thermal system basing on the topology of the boiler heating system described by the correlation matrix and the algorithm model of the boiler heat exchanger components. And the general model algorithm is proposed. An analytical method to solve the heat transfer calculation of the tri-sectional air heater with hot and cold layer is provided, heat transfer model is established and computer programs and algorithms are designed to achieve it. For the difficulty to measure some thermal parameters of power plant boilers, soft sensing is used alternatively. Combined with the mechanism model of boiler thermal system, the performance analysis and online prediction system of utility boilers are established, achieving the purpose of performance prediction and control of the boiler unit. Use the browser/server architecture model to develop the Utility Boiler Performance Analysis and Prediction System supporting online monitoring, which is applied to the predictive control of steam temperature and optimization of boiler soot blowing.The main contents are as follows:This paper introduces the "node" and "branch" concept of the fluid network theory, describe the fluid network by incidence matrix, adjacency matrix and loop matrix, respectively. Then the working principle and structure of the boiler are introduced, the boiler thermal system is analyzed. The network topology of smoke flow, air flow and water flow in boiler thermal system are described by correlation matrix and adjacency matrix.The theoretical basis of the calculation of boiler thermal is introduced, the mass balance equations, the energy balance equation, momentum balance equations in the processes of the flue gas flow, air flow, and water flow are provided. On this basis, Kirchhoffs First Law and Kirchhoff’s Second Law in power theory are used to establish the model of the boiler thermal system and the solving method is provided. Then the analytical method to solve the transfer calculation of the tri-sectional air heater is introduced.Many important technical and economic parameters of the thermal power plants are difficult to measure in real-time. Selecting the other parameters that are able to be measured as auxiliary variables and basing on the boiler operating condition monitoring system, operation history database and coal online monitoring system, using the mechanism or historical data, soft sensing models were established in order to achieve the measurement of the boiler thermal parameters not easily to be measured directly. For the establishment of soft sensing, BP neural network model are used to predict the carbon content of fly ash basing on the historical data; as for the ash thermal pollution of the heating surface, mechanism analysis method is mainly used to establish soft sensoring model. The soft sensing results were as inputs of the mechanistic model of the boiler heating system to establish boiler performance analysis, calculation and prediction models.The realization of the boiler heating system based on the fluid network is discussed. At first, overall technical solutions and system architecture are identified, and main functions and key technologies are introduced. By use of the established boiler performance analysis and prediction system, thermal performance of a supercritical power plant boiler is calculated and predicted, and controlling the steam temperature by spraying water according to the predicted value. Monitoring and diagnostic methods of boiler fouling and slagging are provided, the economic of boiler fouling and soot blowing are analyzed, providing the best soot blowing cycle. At last, the fuzzy comprehensive evaluation method is used to optimize soot blowing judgment and soot blowing sequence.