The Application Research Of Neural Network For Coal Consumption Calculation In Power Station

In recent years, with the rapid development of china’s economic and the fast improving of people’s living standards, the energy demand and the contradiction between energy supply and demand is going to grow and increasing, all this make the energy-saving issue become the peoples’ major concern. The electric power industry is an important support for economic and social development, especially the thermal power generation which consumpt large energy is an important areas of taking energy conservation.Coal consumption is an important indicator of the energy efficiency of the thermal power plant. Current power plant coal consumption calculation is based on the balance of positive energy balance method and counter-balance method, the calculation requires a large number of late amendments, can not reflect the real situation of the power plant coal consumption good. In recent years, the rapid development of the neural network modeling of complex conditions calculated to carry out similar power plant coal consumption provides a new way of thinking.In this paper, We research the way to forecast coal consumption based on neural network. We present a new unburned carbon of the fly ash from utility boilers forecast and a new coal consumption forecast based on BP neural network, also present a novel coal consumption forecast based on Extreme Learning Machine, and combined with the actual working conditions of two different coal consumption calculation model performance comparison and analysis. The main research content and innovation as follows:1. A new unburned carbon of the fly ash from utility boilers forecast based on BP neural network.Based on BP neural network forecast that unburned carbon of the fly ash from utility boilers. The method based on BP neural network theory, combined with the characteristics of the generation of the power plant boiler fly ash carbon content, and take full account of the impact of network training speed and various parameters of the boiler on the carbon content of fly ash from the boiler fuel characteristics, three load and excess air ratio to select the appropriate parameters as the input of the training network, simplifying the calculation of parameters of the fly ash carbon content; automation Qunce limited liability company by the formation of the Yunnan Power Plant Information Monitoring System the the300MW boiler system history library part for the BP neural network training and validation data. Simulation results show that the BP neural networks used in power plant boiler fly ash carbon content to be able to give full play to the BP neural network nonlinear approximation ability to overcome the traditional method requires accurate data and computational models of defects is calculated The important factors of the carbon content of fly ash, power plant coal consumption provides a new way of thinking.2. The application Research of BP neural network based on genetic algorithm for coal consumption forecast in power plant.We design a new BP neural network based on genetic algorithm and present a novel way to forecast coal consumption based on this new BP neural network.Combined with the impact of the power plant coal consumption factors, instead of the traditional coal consumption calculation model based on anti-balancing algorithm based on BP neural network nonlinear approximation ability of BP neural network model. BP neural network is difficult to determine the characteristics of the model structure and initial weights, taking into account the design based on genetic algorithm BP neural network based on genetic algorithm optimization search capabilities; taking into account the impact of the recent samples on the the postorder forecast amount The fact that BP neural network design based on genetic algorithm. Simulation results show that the power plant energy balance test data can be established on the Matlab simulation platform coal consumption calculation based on BP neural network model, predict the results of the overall coal consumption in the model significantly more realistic than the original coal consumption calculation model calculations.3. A novel coal consumption forecast based on Extreme Learning Machine.We present a novel coal consumption forecast based on Extreme Learning Machine.The actual working conditions of the combined power plant coal consumption to calculate the real-time requirements, based on the Extreme Learning Machine forecast power plant coal consumption calculation method, based on the Extreme Learning Machine step to complete the training of thinking. Calculation model based on BP neural network to establish the coal consumption of coal consumption calculation model established by this method in training speed of obvious advantages, the ability of the coal consumption of non-linear approximation is not weaker than the BP neural network for power plant coal consumption calculation provides a more viable approach.