Research on Efficient Energy Consumption Algorithm for Ocean Oil and Gas IoT Based on Deep Learning

The management of offshore oil and gas is becoming more and more important. In this study, firstly, the data collected in the offshore oil and gas Internet of Things are prepared, including data cleaning and data standardization, to ensure the quality of input data. Next, a CNN model composed of convolution layer, aggregation layer and full connection layer is designed and constructed. These layers work together and are responsible for feature extraction and accurate prediction of future energy consumption trends. From the research results, the algorithm has obviously improved the energy saving rate and recall rate. The highest recall rate of CNN model can be as high as 95%, which is obviously better than the highest values of recurrent neural network (RNN) and long-term and short-term memory network (LSTM), which are 85% and 80% respectively. This proves the excellent ability of CNN model to identify abnormal energy consumption.