meters is physical inspection of meter dismantling within 8 years. The method leads to many issues including

high cost of testing, low sampling rate, unknown meter status huge waste of resources etc. Searching for non-

dismantling meter detection solution is necessary. Although the smart grid can be managed much better with

the increasing use of smart meters, the current standard brings many issues. To solve the problems like a huge

waste of resources, detecting inaccurate smart meters and targeting them for replacement must be done. Based

on the big data analysis of smart meters, abnormity can be predicted and diagnosed. For this purpose, the

method is based on Long Short-Term Memory (LSTM) and a modified Convolutional Neural Network (CNN)

to predict electricity usage patterns based on historical data. In this process, LSTM is used to fit the trend

prediction of smart meters, and recurrence plot is used to detect the abnormality of smart meter. Both LSTM

and recurrence plot method is the first time to be used in smart meter detection. In actual research, many

electricity power domain. With the success of the solution, the service time of a normal meter can be prolonged

by abnormity detection. This will lead to saving a lot of resources on smart meter applications.

The foundation of the smart grid is a swift two-way

intelligent communication network. The efficient

and safe operation of the power grid can be obtained

intelligent toll equipment. It can effectively measure

electric parameters and measure two-way electric

requirement of mandatory 8-year periodic rotation,

the annual replacement volume of smart meters in

China is about 80 million, [4].

entry of the smart grid. The smart meter has many

and processing system products etc., [5], [6] It is

estimated that nearly 1.2 billion smart meters will be

installed in the world by 2024 based on Wood

Mackenzie report. The penetration rate of smart

meters will be 60%, [7]. Currently, China has

become the largest consumer market for smart

meters in the world. Smart meters will also have

broad market demand along with the construction

focused on the research on new electromagnetic

measurement standards under big data. The purpose

of this study is to collect electricity data from the

electricity consumption area. The abnormal

information on the user's electricity meter can be

excavated based on the collected data. Then the

abnormal meters can be discovered, [8].

include failures or stealing electric leakage etc.

progress in China. The difference in user

consumption models under different time types is

obtained through analysis of electricity consumption

types of different users. The abnormal electricity

consumption can be detected by using support

vector machine theory. This method requires less

training time and does not need to classify artificial

abnormal data. It can effectively reduce the

including multi-algorithm combination strategy,

reinforcement learning strategy, optimized

scheduling algorithm, and prediction and dynamic

adjustment, providing a reference for achieving

more intelligent and efficient load balancing,

(2) Anomaly detection model construction: The use

of prediction models to learn normal behavior and

perform anomaly detection based on prediction

errors; (3) Abnormal score calculation: Calculate the

abnormal score for each data point based on

prediction error and data history; (4) Online

learning: When new data arrives, update the model

to adapt to the new data and continue learning, [15].

etc., [17], [18].

electricity meters are being replaced with smart

meters that can improve accuracy, efficiency, and

visibility in electrical energy consumption patterns

and measurements. However, in most of the

deployments, smart meters are only used to digitally

measure the energy usage of consumer premises and

smart meter detection based on artificial intelligence

is necessary for smart grid management in China.

Fig. 2: 2016.3.1-5.1 Error Distribution

Fig. 3: 2016.5.1-7.1 Error Distribution

4.1 Data Reliability Analysis

The reliability needs to be analyzed after getting the

data. Firstly different users at the community are

p.m. It is found that the current value at 3 a.m. is the

However, electricity consumption increased at 6

hottest season in summer. As temperatures rise,

users begin to use air-conditioning frequently which

result in electricity consumption increasing even at

night. It can explain why the current situation is

Fig. 5: Electricity Consumption of Same User

Fig. 6: Number of Users at Different Time of Same

4.3 Polynomial Fitting

By using the formula W=U*I*t and then summation

can be obtained. Thus the electricity value can be

months or 209 days. Different fitting figures can be

made by changing the number of the highest number

fitted graphs of the highest power of 4 and 8

respectively as shown in Figure 9 and Figure 10.

From 4 to 8, the fitting degree increased

the highest power from 8 to 10 is not obvious.

To realize smart grid management or abnormal

smart meters being replaced only, many deep

learning algorithms and models need to be analyzed.

Because nobody has done similar work before.

should be paid to training and data analysis, [32],

training set and validation set. But the different

The side-by-side split method is a mainstream way

of partitioning. The data set is segmented into

independent subsets. One part is for training and the

other is for verification.

more impressive. It is more consistent with the final

goal of the study. The future value is predicted with

historical value. However, this segmentation method

has its drawbacks. Because it needs to use data