Fault management technology of power communication room based on the Internet of Things technology

With the continuous development of society, the construction of smart grids has gradually begun to be valued. Smart grids can liberate more labor and enable managers to manage the grid quickly and easily. The power communication network occupies an irreplaceable position in the smart grid. However, it is inevitable that various equipment failures will occur. Therefore, it is necessary to advance the time of fault handling through fault management, then reduce the fault harmfulness. In this paper, with the support of the Internet of Things technology, a monitoring system of the power communication room is constructed based on the configuration status and application environment of equipment, and an intelligent analysis method for fault location of the power communication room based on active detection is proposed. As a result, the reliable and stable operation of the power communication room can be guaranteed by using passive management and active management.

content of other three modules can be remotely grasped by managers through the APP monitoring, and it removes the working environment restriction of monitoring personnel, but does not involve monitoring content expanded, so the following only design the monitoring content of the environment, security and equipment modules according to actual needs.

Environment monitoring module
The environment monitoring module can remotely control the start and stop of the air conditioner by detecting the environment indexes such as temperature and humidity in the power communication room. The monitoring server is connected to the environmental monitoring equipment through the network interface to realize the monitoring of the environment [5]. The specific monitoring content is shown in Table 1.

Security monitoring module
The security monitoring module realizes the video monitoring, personnel access monitoring and illegal intrusion monitoring through the monitoring camera and access control system of the power communication room. The specific monitoring content is shown in Table 2.

Equipment monitoring module
The monitoring of the equipment monitoring module includes power supply, DC, UPS, switches, etc. In this module, the IOT technology is used to identify the status of power supply, DC and UPS, fault state, and the port fault status are identified through the switch command, the equipment running status are monitored through the cabinet temperature, finally the status of the main equipment are summarized, and the equipment faults are detected in time through the remote online monitoring, then the maintenance time is shorten [6]. The specific monitoring content is shown in Table 3.

Internet of Things technology in the power communication room
The Internet of Things (IOT) is a network to realize intelligent identification, positioning, tracking, monitoring and management by connecting any item to the Internet according to agreed protocols through information sensing equipment such as RFID, infrared sensors, global positioning system, laser scanner and so on. In terms of its essence, it is still the Internet that we are familiar with, except that the participating objects are no longer groups composed only of people, and the objects also participate in them, thus it has the functions of real-time tracking and positioning, real-time data collection, etc [7]. The Internet of Things has a three-layer structure, from the bottom to the top: the perception layer, the network layer and the application layer [8]. The lower layer serves as the basis of the upper layer, which in turn realizes the functions of collecting data, transmitting data and processing data. The traditional Internet of Things technology can collect data， and issue alarms after data processing to remind managers to take corresponding measures. However, there is a problem when it is used in the power communication room. Passively waiting for the alarm information is not a good measure to deal with failures, because the time left for the management personnel to deal with the fault after receiving the alarm information is not enough, we can take active detection to find the fault, so as to weaken the fault at an earlier stage. In order to achieve this function, a signal transmitter and a signal receiver can be added to the traditional sensor to realize the functions of sending a detection signal and receiving a detection signal, then transmit the data back to the control center through the traditional signal transmission channel of the Internet of Things, finally, on the basis of these data, the management personnel locate the fault and take countermeasures at an earlier stage.
In the application of the Internet of Things in the power communication room, the reliability and integrity of the data are the key to ensuring the effectiveness of the subsequent management work, so stable data transmission is particularly critical. In addition, a variety of sensors are used in the power communication room, correspondingly, various types of data need to be transmitted, so that there is a problem of mutual data interference. Therefore, a transmission method that can stably transmit data and can effectively avoid mutual influence among data is required [9]. The multi-carrier aggregation transmission technology is adopted in the power communication room, which is an orthogonal frequency division multiplexing technology, it can divide the channel into multiple orthogonal channels, and can decompose a high-speed data into low-speed parallel data, then modulates these data onto the channel to achieve information transmission. Orthogonal signals can be separated at the receiving end to avoid mutual interference among the various channels. Since the channel-related bandwidth is greater than the signal transmission bandwidth of the sub-channels, each sub-channel can be treated as a flat fading, so the interference among the various signals is eliminated.

Fault location method of the power communication room based on active detection
The monitoring system of the power communication room ensures that the fault can be quickly resolved after the alarm is issued, but there are problems with the way of passively waiting for the alarm information [10]. Therefore, with the support of the Internet of Things technology, this paper proposes a fault location method of the power communication room based on active detection.

Fault location mechanism of the power communication room based on active detection
Aiming at the fault location of the power communication room, this paper divides it into three stages, namely fuzzy fault detection, detailed fault detection and fault judgment. Among them, the fuzzy fault detection stage determines the initial detection set, the detailed fault detection stage determines the detailed detection set on the basis of it to ensure the integrity of the detection data, finally the fault judgment stage uses the results of the detection set of the first two stages to determine the fault position.
As the beginning of the entire location method, the fuzzy fault detection involves the installation position of the detection signal transmitting device. This device can be installed in the location of the sensor of the Internet of Things, and the transmitting device is required to transmit detection signals to all nodes in the network. In the fuzzy fault detection stage, the detection set from each detection signal transmitter to all other network nodes is used to form the initial detection set T 0 . Secondly, the greedy algorithm is used to solve the minimum detection set that can cover the entire network to form the fuzzy detection set T 1 ; Finally, according to the initial detection set, the detection signal transmitter is operated and the return signal is fed back to the control center.
The fuzzy detection set cannot provide enough data for fault location, so the detailed fault detection stage is needed to efficiently obtain more detailed detection results. In the detailed fault detection stage, the candidate detection set T 2 is first established, which is obtained by subtracting the fuzzy detection set T 1 from the initial detection set T 0 ; secondly, the detection value of each detection in the candidate detection set is calculated and sorted, and the largest detection t 1 is selected, and add the detection t 1 to the detailed probe set T 3 and delete it in the candidate detection set; then add the detection t 1 to the network to interact, and update the value of the detection in the candidate detection set, among them, the submodular of the detection value function can be used to reduce the scale of updating detection value. Then, sort the candidate detection set again, and select the largest detection t 2 , then repeat the above operation. Perform the above loop process until the detection cost reaches the maximum value or the candidate detection set is empty, and the final detailed detection set T 3 is obtained. Finally, according to the detailed detection set, the detection signal transmitter is operated and the return signal is fed back to the control center.
As the final loop, the fault judgment stage determines the faulty equipment and the faulty line set based on the detection results of the fuzzy detection set and the detailed detection set to ensure that the fault set should be able to explain all the fault information in the detection result. Among them, the detection results of the fuzzy detection set and the detailed detection set can be represented as the detection fault bipartite graph shown in Figure.

Minimum detection set algorithm
The minimum detection set algorithm aims to select as few detections as possible from the initial detection set, so that the path through all equipment and lines. Because there is no isolated equipment in the network, all equipment is connected by lines, so it is only necessary to ensure that all lines are covered, and accordingly all equipment will also be covered [11]. Based on the greedy algorithm, this paper takes the following steps to find the minimum detection set: 1) Firstly, set a set Y to store all the lines covered by the fuzzy detection set, which is initially empty.
2) Set a variable J for each detection, indicating the number of lines in the detection that are different from the lines in Y, and initially it is the number of all lines in the detection.
3) Select the detection with the largest value J, count it into the fuzzy detection set, and delete it from the initial detection set. 4) Update the value J of each detection in the initial detection set. If all lines in the network are not fully covered by the set Y, go to step 3, otherwise, stop the calculation and obtain the final fuzzy detection set.

Detection value and its submodular
At each step in the selection process of the detailed detection set, the detection that contributes the most to the fault location needs to be selected, so the detection value is taken to measure it. Specifically, the detection value can be divided into two aspects, one is the reduction degree of network uncertainty, and the other is the number of important nodes in the detection.
The uncertainty function of the network is not unique, but it has the same feature, that is, it is obtained based on the state of the network X, and it is set to H(X). It is easy to know that the process of detection is the process of continuously reducing network uncertainty, and the detection set is set to T, so it is obviously H(X|T) less than H(X). In order to measure the degree of each detection on reducing network uncertainty, a gain value G(t i ) is introduced, as shown in Equation 1.
Therefore, combining the number of important nodes of each detection N(t i ), the detection value of each detection P(t i ) can be obtained, as shown in Equation 2.
Among them,  and  respectively indicate the importance of the two aspects in the detection value, which can be adjusted according to the needs of different networks.
In the detailed detection stage, each update of the detection value is not a small amount of calculation. However, the fault location has high requirements for timeliness. Therefore, the submodular of the detection value function can be used to obviously reduce the amount of calculation. Specifically, when the detection value function satisfies the submodular requirement, after a detection enters the network for interaction, the detection value of all other detections will decrease. Therefore, if the detection with the largest detection value after the update still has higher detection value than the detection value of other detections before the update, the detection is the selection of next loop, and there is no need to update the detection value of other detections.

Minimum fault algorithm
The minimum fault algorithm is the core algorithm of fault judgement, and it aims to explain all the fault information obtained by fault detection with the minimum faults [12]. The specific steps are as follows: 1) Firstly, O is used to indicate the fault area, K is used to indicate the connection line between the fault area and the non-fault area, and V(i) and L(i) are used to indicate the candidate set of the faulty equipment and faulty line, respectively.
2) The detection set in the fault state is extracted from all the detection results to form the fault area O.
3) Check each line in the network. If one end belongs to O and the other end does not belong to O, the line is the connecting line between the fault area and non-fault area, and count the line to K. 4) Use a |k| bit binary number to enumerate the states of each connected line at the same time. There are 2 |k| kinds of combinations of different states of the connected line, so that each combination is further analyzed when i=0, 1, ..., 2 |k| -1. 5) For each combination i, there are |k| connected lines, then further analysis is performed for each connected line: if the code of a connected line in combination i is 1, it means that there is fault on the connection line, then Count this connection line into L(i); if the code of the connection line in combination i is 0, it means that there is no fault on the connection line, but that there is a fault on the end of the connection line that locates in the fault area, then count the end into V(i) after confirming that the end is not in V(i).
6) After analyzing all combinations, find the combination whose |V(i)|+|L(i)| is smallest, and treat this combination as the final fault set.