Skip to content
Go to file

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Meter Detection

Detecting malfunctional smart meters based on electricity usage and targeting them for replacement can save significant resources. For this purpose, we developed a novel deep-learning method for malfunctional smart meter detection based on long short-term memory (LSTM) and a modified convolutional neural network (CNN). Our method uses LSTM to predict the reading of a master meter based on data collected from submeters. If the predicted value is significantly different from master meter reading data over a period of time, the diagnosis part will be activated, classifying every submeter to identify the malfunctional submeter based on CNN. We propose a time series-recurrence plot (TS-RP) CNN, by combining the sequential raw data of electricity and its recurrence plots in the phase space as dual input branches of CNN.

For more details, please refer to the paper.

If you are using our work in your research, please cite us as

       author = {{Liu}, Ming and {Liu}, Dongpeng and {Sun}, Guangyu and {Zhao}, Yi and
         {Wang}, Duolin and {Liu}, Fangxing and {Fang}, Xiang and {He}, Qing and
         {Xu}, Dong},
        title = "{Detection of Malfunctioning Smart Electricity Meter}",
      journal = {arXiv e-prints},
     keywords = {Electrical Engineering and Systems Science - Signal Processing, Computer Science - Machine Learning, Statistics - Machine Learning},
         year = "2019",
        month = "Jul",
          eid = {arXiv:1907.11377},
        pages = {arXiv:1907.11377},
archivePrefix = {arXiv},
       eprint = {1907.11377},
 primaryClass = {eess.SP},
       adsurl = {},
      adsnote = {Provided by the SAO/NASA Astrophysics Data System}

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.



Explanations for each file


Our raw data is in fodler sitaiqu including the usage(kilowatt_everyday_2year.xlsx), the current(electriccurrent_hours_2year.xlsx) and the voltage(voltage_hours_2year.xlsx).

Data processing is accomplished in

Residential Area’s Error Prediction Task will generate the input for lstm. is used to compare the result in different sequence length. Hence, in order to exlude the contingency, we choose to predict 10 times for each sequence length in and

The comparision of classical methods is accomplished in

Malfunction-injected Residential Area Detection Task

We generated our data of residential area with malfunctional meters in

The detection task is finished in

Malfunctional Submeter Classification Task

We generated our data in, which imported and

The classification task is accomplished in

To test the performance of different proportions of malfunctional meters, we did some comparision in


Detection of Malfunctional Smart Electricity Meters Based on Deep Learning of Electricity Usage Data (under review)




No releases published


No packages published