Research And Application Of Energy Data Processing Technology For Workshop Equipement Based On Edge Computing

Energy plays an important role in discrete manufacturing enterprises.The energy data of manufacturing equipment has a significant reference value for improving the production efficiency of enterprises and controlling production costs.The equipment energy management system based on the Cloud Computing mode is widely used in discrete manufacturing workshops,However,the rapid increase of workshop manufacturing equipment and the way of collecting data in seconds exist,which makes the amount of energy data show a trend of massive growth.In this case,based on the processing method of device energy data directly to the cloud in the Cloud Computing mode,problems such as network transmission congestion and increased data pool storage pressure are caused.At the same time,the production tasks of the discrete manufacturing workshop are implemented according to the production arrangement.And,some the production arrangement of the workshop does not take into account the energy consumption of the equipment.And some irrational arrangements are leading to the higher energy consumption of the equipment,which boosts the production costs of the enterprise.Energy consumption of equipment is essential for the rational production arrangement in the actual production process.However,the energy consumption of equipment is usually discovered rather slowly.Therefore,it is a significant demand for the energy management of enterprises to forecast the energy consumption of equipment in advance.Therefore,Edge Computing,Data Compression,and Time Series Prediction methods are proposed to solve the above problems,with an auto parts enterprise in Guangdong Province as a research object.It can guarantee the healthy operation of energy management and production tasks in the manufacturing workshop,and realize the rational utilization of energy and the improvement of production efficiency of the enterprise.The main work of the thesis includes:First,this thesis analyzes the specific needs according to the actual production workshop situation,designs and implements an energy data acquisition system based on Edge Computing.The system is comprised of the device energy data source layer,the edge data processing layer,and the cloud platform data processing layer.The edge data processing layer is the core layer.The edge gateway of this layer completes the device information configuration,data communication,and data processing function requirements.The data transmission function is completed by the Kafka cluster.In short,the system realizes real-time and accurate energy data collection of all manufacturing equipment,and provides a stable data source for edge data processing.Second,this thesis proposes a Swinging Door Trending compression algorithm based on anomaly detection and dynamic tolerance adjustment regarding excessive energy data transmission and storage pressure.First of all,according to the optimal anomaly detection parameters selected by the sliding window anomaly detection algorithm based on state change,abnormality detection is performed on energy data to improve the accuracy of energy data.In additon,the data compression process is performed in combination with the expected error of the scene demand and the actual data fluctuation state,according to the Swinging Door Trending compression algorithm based on dynamic adjustment of tolerance,to reduce the amount of data and relieve the pressure of transmission and storage.Lastly,the data of the enterprise’s actual key energy monitoring equipment is used to test and verify,which shows that the algorithm has specific feasibility.Third,this thesis proposes an energy consumption prediction algorithm based on the ARIMA model,combines the characteristics of energy consumption data with nonlinear parts,and proposes a combined model based on the ARIMA-SVR energy consumption prediction algorithm in response to the needs of enterprises to use the predicted value of equipment energy consumption to arrange production tasks.The two algorithms are validated for the energy consumption data of the same critical energy monitoring equipment.The experimental results show that the combined model has a better effect on energy consumption prediction.It shows that the model can meet the demand of equipment energy consumption prediction for short-term production tasks.Fourth,this thesis designs and implements an edge platform for energy management at the edge data processing layer to verify the algorithm’s rationality and effectiveness.The data compression algorithm and energy consumption prediction algorithm are deployed on the edge platform of energy management.Furthermore,the energy data can be compressed and predicted in real-time at the edge to meet the realtime business requirements in practical application scenarios,which shows that the solution in this thesis has certain practicability in solving the two major problems.