【摘要】 以某1 000 MW机组脱硫塔为例，提出基于LSTM的脱硫塔阻力预测模型。结合ISSA算法对LSTM超参数寻优，并利用对异常值更具鲁棒性的Huber函数作为损失函数，使脱硫系统阻力预测结果更优。改进后模型的平均绝对百分比误差为3.061%，决定系数为0.942，优于其他对比模型。更多还原

【Abstract】 Taking the desulfurization tower of a 1 000 MW unit as an example, a LSTM based desulfurization tower resistance prediction model is proposed. In combined with ISSA algorithm, LSTM hyperparameter is optimized, and Huber function, which is more robust to outliers, is used as the loss function, so that the resistance prediction result of desulfurization system is better. The average absolute percentage error of the improved model is 3.061%, with a determination coefficient is 0.942, which is superior to other comparison models.更多还原