Steam turbine valve flow fitting method and system

Abstract

The invention provides a steam turbine valve flow fitting method and system, and the method comprises the steps: collecting a current observation point vector related to the flow characteristics of asteam turbine valve, and carrying out the smoothing of an abnormal observation point in the current observation point vector; utilizing an observation point vector obtained by smoothing an abnormal observation point in the observation point vector to train to obtain a linear self-segmentation neural network model; and adjusting the valve opening degree of the neural network model, and fitting to obtain the valve flow corresponding to the valve opening degree. Based on the steam turbine valve flow fitting method, the invention further provides a steam turbine valve flow fitting system. According to the method, the fitting speed of the valve flow curve is increased, meanwhile, the method has the advantage of selection flexibility of the segment number of the piecewise function, rapid and accurate performance is shown on the fitting effect of the valve flow curve, and meanwhile, a fitting model is provided for the next step of valve flow characteristic optimization.

Description

technical field [0001] The invention belongs to the technical field of electric power systems, and in particular relates to a steam turbine valve flow fitting method and system. Background technique [0002] During the operation of thermal power units, a large amount of high-dimensional data is formed, which are stored centrally in the plant-level monitoring information system (Supervisory Information System, SIS) and distributed control system (Distributed Control System, DCS). Cooperative work provides a wealth of basic data. The valve flow curve characterizes the relationship between valve opening (or valve position command) and valve flow. Accurate identification of valve flow characteristics is of great significance to effectively avoid unqualified primary frequency modulation quality of thermal power units or abnormal unit coordination ability. If the steam turbine valve flow curve does not match the actual flow characteristics, then when the unit load changes and th...

AI Technical Summary

Problems solved by technology

[0006] Due to the need for test personnel to operate on site, there are four defects in the traditional steam turbine valve flow curve simulation method: (1) The valve flow characteristics of the steam turbine are obtained under the de-loading of the unit and the specific operating conditions of the boiler, but in the actual operation process, There will be dynamic changes in the working conditions of the boiler, and it is difficult to completely maintain the ideal test state. At the same time, the operating conditions of the steam turbine will also change dynamically, which will lead to deviations in the valve flow characteristic curve obtained by the traditional method; (2) in the traditional test The amount of data obtained is limited. This is because the different strokes of the valve opening are manually adjusted, and the corresponding working condition data is also relatively simple, which will lead to a low capacity of the fitting model. It is also very limited, and a large amount of data in SIS and DCS cannot be deeply excavated; (3) Since the traditional test requires a special person to go to the site for operation, and the test frequency is low, when the valve characteristics change, the original fitting model may It is no longer applicable; (4) In big data application scenarios, it is impossible to judge the data quality of all measuring points in real time based on expert experience alone, which leads to the input of abnormal points affecting the accuracy of valve characteristic fitting

Images