高级检索

    基于ARIMA-LSTM模型的MSWI过程CO2排放浓度多步预测

    Multi-step Prediction of CO2 Emission Concentration in MSWI Process Based on ARIMA-LSTM Model

      摘要
    • 摘要: 针对城市固废焚烧(municipal solid waste incineration, MSWI)过程CO2排放兼具线性趋势与非线性波动的复杂动态特性, 现有单一预测难以准确拟合的问题, 提出基于差分整合移动平均自回归-长短期记忆(autoregressive integrated moving average-long short-term memory, ARIMA-LSTM)模型的CO2排放浓度的多步预测方法。首先, 采用ARIMA算法构建线性主模型以进行CO2排放浓度预测; 然后, 以主模型的预测残差为真值, 采用LSTM算法构建非线性补偿模型; 最后, 将主模型和补偿模型的预测值进行组合得到超前多步的预测结果。基于北京某MSWI工厂的真实CO2数据集验证了所构建混合模型的有效性。

       

      Abstract: To address the problem that the CO2 emission of municipal solid waste incineration (MSWI) process has complex dynamic characteristics of linear trend and nonlinear fluctuation, and the existing single prediction model is difficult to accurately fit, a multi-step prediction method for CO2 emission concentration is proposed based on autoregressive integrated moving average-long short-term memory (ARIMA-LSTM) model. First, ARIMA algorithm was used to construct a linear master model to predict the CO2 emission prediction. Then, taking the prediction residual of the main model as the true value, LSTM algorithm was used to construct a nonlinear compensation model. Finally, the prediction values of the main model and the compensation model were combined to obtain the advanced multi-step prediction results. Based on the real CO2 dataset of a Beijing MSWI plant, the effectiveness of the constructed hybrid model was verified.

       

      • HTML全文
      • 参考文献(29)
      • 相关文章
      • 施引文献
    • 资源附件(0)

    /

    返回文章
    返回