Research On Performance Algorithm For Gas And Coal Co-firing Unit

With the transformation of China's economic development mode, energy saving becomes an important task for many high energy consumption enterprises. As the focus of energy saving industry, power industry will face greater pressure and more challenge. Iron and steel industry as the other energy saving key industry, electricity generation using by-product gas (mainly blast furnace gas, BFG) is an important measure for its achievement of energy consumption and environment protection indicators.Performance indicator is a measure of power plant units and their equipments running state, rapid and accurate calculation of the indicator can help operating personnel make adjustment timely to achieve the most optimal operation state of the unit. In this thesis we take five power units in an iron and steel plant for research, with reference to relevant technical standards at home and abroad, systematically summarize the methods of indicator calculation for different unit type includes CCGT unit, gas coal co-firing unit, gas firing unit. Targeting on large-scale utilization of low -calorific-value BFG, we participate in more than 50 field trials and more than 10 reference tests for the five power unit, calculate electricity load versus coal consumption for unit 0/4 and electricity load, gas blending ratio versus coal consumption in the form of two-variable function for unit 1/2/3 , which describes systematically performance of the unit and provides a theoretical basis for optimal allocation . Based on the test data, this paper analyses the economic impact on the boiler of gas coal co-firing unit for mixed burning BFG, improvement and optimization of various types of unit operation performance, energy conservation and gas emission reduction and so on, which provides basic data for the realization of 100% recycling of BFG and reducing coal consumption. Traditional method of performance calculation for gas coal co-firing units involves much offline data and the test is time-consuming and expensive, can not be calculated in real-time. Once a parameter measuring point is damaged, the performance calculation can not reflect the true state of the unit, thus traditional method is not suitable for online management of the energy. This Paper describes the simulation process of coal consumption for a gas coal co-firing unit according to BP neural network nonlinear mapping method. Through discussion the structure parameters of the neural network model such as input parameters, number of input parameters, number of learning sample set, number of hidden layer nodes, combination of learning rate and momentum factor, the paper establishes two optimal BP neural network models which suitable for simulation of turbine heat rate and boiler efficiency. Learning and training the mode use experimental data, the results show that selection of input parameters strongly associated with simulation object can get higher accuracy performance calculation; for the neural network model with about 15 input parameters, 20 hidden nodes can get optimal simulation results; can only use a big learning rate combined with a small momentum factor get a learning precision smoothly. Finally, the paper use the turbine heat rate and boiler efficiency optimal BP neural network model to calculate the gross coal consumption rate in actual test conditions, the mean square error is less than 0.030%, indicates that the model results can meet the needs of engineering precision and the model is of great importance for gas coal co-firing performance online calculation.The neural network model can ensure the coal consumption calculation accuracy, overcome the shortcomings of heat balance algorithm requires too many parameters and can not calculate real-timely, which provides a new way to realize online performance analysis for gas coal co-firing unit in steel plant.