Journal
IEEE ACCESS
Volume 6, Issue -, Pages 23551-23560Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/ACCESS.2018.2825538
Keywords
Time series prediction; LSTM model; power equipment; industry Internet of Things
Categories
Funding
- Key Research Program of Shandong Province [2017GGX10140]
- Fundamental Research Funds for the Central Universities [15CX08015A]
- National Natural Science Foundation of China [61309024]
Ask authors/readers for more resources
Industrial Internet of Things (IoT) is producing massive data which are valuable for knowing running status of the underlying equipment. However, these data involve various operation events that span some time, which raise questions on how to model long memory of states, and how to predict the running status based on historical data accurately. This paper aims to develop a method of: (1) analyzing equipment working condition based on the sensed data; (2) building a prediction model for working status forecasting and designing a deep neural network model to predict equipment running data; and (3) improving the prediction accuracy by systematic feature engineering and optimal hyperparameter searching. We evaluate our method with real-world monitoring data collected from 33 sensors of a main pump in a power station for three months. The model achieves less root mean square error than that of autoregressive integrated moving average model. Our method is applicable to general IoT equipment for analyzing time series data and forecasting operation status.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available