| In the background of the tension of current energy situation and the big adjustment of energy structure, it has become an increasingly important research topic for thermal power units to achieve energy saving and improve economic benefit deeply. Due to adopting the traditional PID control,the cold end system of power plant is often difficult to keep the condenser vacuum often near the best vacuum operation, result in the operation technology and the unit economy remains to be further improved. The current coal-fired boiler flue gas waste heat utilization project, often because of low temperature corrosion and adopting the conventional PID control, smoke temperature cannot be reduced to wet desulphurization temperature near the best efficiency, lead to waste heat recovery failed to maximize. So aiming at the remaining problems of cold end optimization and boiler flue gas waste heat utilization in the thermal power plants, some analysis, studies and solutions are given in this paper,in order to give viewpoints and theoretical supports what have a better energy saving and emission reduction as well as the economic benefits of units. For this purpose,this paper focuses on the research and discussion from the following two aspects:First of all, the supercritical unit cold circulating water system belongs to typical large lag and large inertia〠large couplingã€many factors affect the system, PID control is difficult to guarantee the good control quality and load adaptability, etc. Conventional PID control is often less consideration, regulate the cycle of water and slow response, low control precision, at the same time the adaptive ability is poor, which makes the cold end system tend to deviate from the condenser in the operation of the best vacuum state, the economic benefit is often poor, severely restricts the implementation of energy saving of the unit. Therefore, On the basis of analyzing many influence factors of the cold end system, to determine the load and the inlet temperature of circulating water, circulating water level and condenser heat temperature, circulation pump frequency as input of neural network model, and design optimization control scheme of circulating water as well as the comprehensive study the current main related state parameter, and with the difference between the unit output increment and circulating water pump power consumption increment as training signal, using advanced ant colony algorithm in network training, real-time parallel to calculate the Optimal circulating water quantity, timely adjusting loop water condenser is always near the best vacuum operation. The simulation results show that, the optimal control can achieve control of the best circulating water more quickly and accurately, and has a good robustness and better ability adapting to the changing conditions and environment, and good energy saving and unit economic benefit.Secondly, at present, the recovery waste heat failed to maximize, as it was mainly limited the low temperature corrosion. The conventional PID control have some deficiencies, such as the lag of temperature control in the condensate system, the single input of the traditional PID algorithm, the inadaptability to the different kinds of coal and the poor robustness, making low temperature corrosion uncontrollable and inevitable, and the flue gas temperature reducing is restricted. For low temperature corrosion problems in the field of boiler flue gas waste heat utilization, based on the theory of corrosion speed limited studies the problem how to depth recovery of waste heat from flue gas. To select the ultra supercritical boiler which install the low temperature economizer system and work near the design exhaust temperature as the research object, in order to ensure safe depth recovery of flue gas waste heat at the same time, the comprehensive technical and economic consideration, puts forward the low temperature economizer outlet smoke temperature as far as possible close to the thinking of the flue gas acid dew point in real time. In order to achieve the precise control of low temperature economizer outlet smoke temperature drop, by optimizing the structure of neural network, with hybrid dynamic recursive neural network as control model, design the ACO-MDRNN smoke temperature optimization control scheme, and the comprehensive learning unit load, import and export of low temperature economizer feedwater temperature, low temperature economizer major parameters, with real-time flue gas acid dew point as low temperature economizer outlet smoke temperature set point. At the same time, the mutative scale chaos search strategy was integrated into the ant colony algorithm, the mutative scale chaos ant colony optimization algorithm as network training algorithm, real-time calculated control smoke temperature drop of best condensation water, to timely adjust the flue gas outlet temperature of low temperature economizer, maximize energy efficiency. Proved by the simulation and experimental research, the optimal control scheme can guarantee to avoid low temperature corrosion or corrosion in limited cases, the recovery of flue gas surplus heat to maximize, also greatly improve the load adaptability. |
