Dual-intelligent-optimization control method for unit load

Abstract

A dual-intelligent-optimization control method for a unit load comprises the steps of firstly establishing a supercritical unit load characteristic nerve network prediction model for condensate throttling, training and verifying the prediction model by means of historical operation data of the unit, respectively predicting and optimizing the opening degree of a deoxygenator water level regulating valve and the opening degree of a stream turbine valve in a dynamic variable-load period and a constant-load period of the unit by the prediction model, and overlapping the deviation between each optimization result and an original control instruction on a corresponding control signal for performing optimization control on the unit. The dual-intelligent-optimization control method has functions of improving the load response speed of the steam turbine in the dynamic variable-load process, ensuring the deoxygenator water level in a safe operation range of the unit, furthermore ensuring high adjustment precision for the unit load in a condensate flow recovering process of the constant-load period, greatly improving deep peak load regulation capability of the unit, and ensuring long-term safe and reliable operation of a thermal power unit.

Description

technical field [0001] The invention relates to a dual intelligent optimization control method for unit load based on a load forecasting model and condensed water throttling, and belongs to the technical field of power generation. Background technique [0002] Large-capacity super (super) critical coal-fired units have gradually become the main unit of the power grid due to their obvious advantages in energy saving, consumption reduction and environmental pollution reduction. The continuous improvement of power supply quality indicators of the power grid requires that large-capacity super (super) critical units should have functions such as primary frequency regulation and automatic generation control (AGC). (It is hoped that the load change per minute can reach 1.5% to 2% of the rated load of the unit or even higher). Since the coordinated control system (CCS) needs to take into account the control deviation of load and main steam pressure, as well as the large inertia and...

AI Technical Summary

Problems solved by technology

[0006] (c) Condenser vacuum (back pressure), as an important cold-end parameter of the unit, is affected by various factors such as circulating water pump, vacuum pump, unit shaft seal operation, etc. Vacuum changes have a great influence on unit output and unit operation economy large impact, which is not considered in the above model;

[0007] (d) The high, medium and low pressure cylinders of large supercritical steam turbine units are all equipped with at least 8 stages of regenerative steam extraction, and the sudden changes in the amount of regenerative cycle extraction at all levels caused by various reasons during operation (such as high pressure heaters) decoupling, sudden change in condensate flow rate, etc.), all of which will have an impact on the work of the steam turbine in a short period of time, and the above model cannot reflect

[0008] That is to say, the existing simplified model ignores many factors, which makes the load prediction value output by the model not high in accuracy and poor in anti-interference ability, which affects the actual application effect of the model in engineering

Images