摘要
Abstract
To address the challenges of delayed energy consumption data acquisition and fine-grained man-agement in discrete manufacturing workshops,this paper proposes an energy consumption prediction and optimi-zation method based on a collaborative"cloud-edge-end"Industrial Internet of Things(IIoT)architecture.Firstly,a three-layer IIoT system comprising the perception,network,and application layers is designed,le-veraging edge computing gateways to achieve real-time collection and preprocessing of multi-source heteroge-neous data.Subsequently,a hybrid VMD-CNN-BiLSTM prediction model is developed by integrating the ad-vantages of variational mode decomposition(VMD),convolutional neural networks(CNN),and bidirectional long short-term memory networks(BiLSTM)to enhance prediction accuracy.VMD is employed to reduce noise,while CNN and BiLSTM are combined to capture the multi-dimensional spatial features and temporal de-pendencies of energy consumption data,respectively.Empirical analysis conducted in an automotive stamping workshop demonstrates that the proposed model achieves a mean absolute percentage error(MAPE)of 3.42%,representing reductions of 5.8%and 2.1%compared to SVM and LSTM models,respectively.In practical ap-plication,the workshop's monthly electricity consumption decreased by 8.5%,resulting in an approximately 14.2%reduction in electricity costs and a monthly CO2 emission reduction of approximately 3.8 tons,thereby validating the energy-saving and carbon reduction efficacy of the proposed method.关键词
工业物联网/大数据/节能降耗/深度学习/VMD-CNN-BiLSTM模型/离散制造/能耗预测Key words
industrial internet of things(IIoT)/big data Analytics/energy saving optimization/deep Learning/VMD-CNN-BiLSTM model/discrete manufacturing/energy consumption分类
信息技术与安全科学
