混煤煤灰软化温度的实验研究与预测

Experimental Research and Forecast of the Softening Temperature of Blended Coal Ash

从某电厂700 MW机组锅炉取混煤灰样,在智能灰熔点测定仪上采用角锥法对其进行软化温度测定。实验结果表明,混煤煤灰软化温度与掺混比呈非线性规律。采用径向基神经网络(RBFNN)在MATLAB环境下建立了混煤软化温度的智能预测模型。为检验模型的预测效果,以实验的8个混煤煤灰作为受验样本,应用该网络模型对其软化温度进行预测。预测表明:RBFNN模型的预测结果与实验结果吻合良好,二者的最大相对误差为3.79%,平均相对误差为1.56%,预测效果远远优于线性预测模型。

Samples of blended coal ash were taken from the boiler of a power plant 700MW unit.Their softening temperature was measured by a pyramid method on an intelligent ash-melting point measuring device.Experimental results indicate that the softening temperature of blended coal ash and blend/mixture ratio assume a nonlinear relationship.Through the use of a radial-based function neural network(RBFNN) an intelligent forecasting model for the blended-coal softening temperature was set up under MATLAB environment.To verify the forecast effectiveness of the model,with 8 blended coal ash samples under test serving as samples to be examined a forecast of their softening temperature was conducted using the above-mentioned RBFNN-based model.The results of the forecast indicate that the forecast results of the RBFNN model agree well with those of experiments.The maximum relative error between the above two results is 3.79% with the average relative error being 1.56%.The effectiveness of the forecast has been found to be by far superior to that of a linear forecast model.