CN114448094A - A data sharing system based on edge computing of intelligent service terminals in Taiwan area - Google Patents

Abstract

A data sharing system based on edge computing of intelligent service terminals in the Taiwan area. From a data-driven perspective, combined with container technology and APP development technology, it breaks through the data barriers between cloud centers and intelligent service terminals, and divides intelligent service terminals from bottom to top. Data access layer, data integration layer, edge data sharing layer, data mart layer, data application layer and cloud data storage layer, integrate various information resources in the power distribution station area on the edge side, reduce the pressure of cloud computing center bandwidth, slow down The computing resource requirements of intelligent service terminals, improve data processing capabilities, enable rapid response of diagnostic services, realize data collection and sharing, facilitate the realization of information collaboration, computing resource collaboration, and business collaboration on the cloud and edge sides, and realize dynamic adjustment of terminal services. It is more adaptable, enhances the ability to create data value, and improves the level of lean production and management of power grids.

Description

A data sharing system based on edge computing of intelligent service terminals in Taiwan area

technical field

The invention belongs to the technical field of power distribution automation, and in particular relates to a data sharing system based on edge computing of intelligent service terminals in station areas.

Background technique

The ubiquitous power Internet of Things applies modern information and communication methods such as artificial intelligence and mobile Internet to every link of the operation of the power system, so as to build an intelligent network that is interconnected between things, people and things, and people. . The power distribution Internet of Things is an important part of the ubiquitous power Internet of Things. At present, the construction mode of "cloud pipe edge terminal" is mainly adopted. The edge side is responsible for data collection and on-site processing to meet the latency requirements of real-time services. The intelligent service terminal deployed on the edge side acts as an edge node, which has the ability to exchange and purchase, and meets various information access requirements in the distribution station area. At the same time, the lightweight Linux container technology is deployed to flexibly configure various APPs according to various business scenarios. software. The cloud center is responsible for deep data mining and advanced business processing, and performs collaborative management of multiple intelligent fusion terminals to achieve cloud-side data collaboration, business collaboration, and computing resource collaboration.

With the in-depth application of advanced distribution Internet of Things technologies such as "cloud, big things, and intelligence", it can essentially boost the level of distribution network construction, operation and maintenance, and management, and achieve leapfrog development. Distribution Internet of Things is a new type of power network form produced by the deep integration of traditional industrial technology and Internet of Things technology. It can realize the comprehensive perception, data fusion and intelligent application of the distribution network, and meet the needs of lean and intelligent management of the distribution network. However, at present, the construction of the distribution Internet of things in my country is in its infancy, and there are the following key technical challenges: (1) the coverage of sensors is insufficient, and there are still monitoring blind spots in the power grid; (2) the data model and communication standards are not unified, and information islands are difficult to eliminate; (3) data analysis The degree of intelligence is low, and the value of power supply and consumption data has not been effectively tapped; (4) an effective collaboration model has not yet been established, and the multi-agent collaboration mechanism has not yet matured.

With the continuous increase of comprehensive energy equipment under the distribution station area, the intelligent service terminal has accumulated massive data generated by the power distribution and consumption links. Due to the complex processing logic, diverse structure, and long storage period of these data, the difficulty of data integration and sharing between cloud centers and edge nodes increases, which reduces the efficiency of cloud-edge collaboration and monitoring, operation and maintenance. Data sharing has become the primary consideration in the process of realizing centralized resource management and control, enhancing data application value and improving information construction in the power grid. At present, although power grid companies have built business data centers for different application requirements, in the scenario of typical business scenarios of power distribution IoT cloud-side collaborative governance platform area, a unified information model mechanism has not yet been established to achieve centralized management and control at the data level. Sharing and collaborative governance, therefore, it is necessary to build a complete data sharing system in the distribution Internet of Things.

SUMMARY OF THE INVENTION

In order to improve the data processing efficiency of the power Internet of Things cloud center and the intelligent service terminal, and realize the timely and refined diagnosis of abnormal working conditions in the distribution station area under the Internet of Things architecture, the invention discloses an edge computing based on the intelligent service terminal of the station area. The data sharing system, based on the "cloud-pipe-side-end" power distribution IoT system architecture, uses lightweight data transmission protocols, establishes internal and external data exchange specifications for intelligent service terminals, and builds a data sharing system based on edge computing of intelligent service terminals in the station area , and combined with the business scenarios of abnormal working conditions in the low-voltage station area, to complete the diagnosis case of abnormal working conditions in the distribution station area based on the data sharing system.

The invention follows the principle of "centralized management and control, classified storage, and unified application" of the cloud center and the intelligent service terminal, and divides the intelligent service terminal into a data access layer, a data integration layer, an edge data sharing layer, a data mart layer, The data application layer and the cloud data storage layer break through the data barriers between the cloud center and smart service terminals, integrate various information resources in the distribution station area on the edge side, reduce the pressure on the bandwidth of the cloud computing center, and ease the computing resource requirements of smart service terminals. Improve data processing capabilities, enable diagnosis services to respond quickly, realize data collection and sharing, enhance data value creation capabilities, and improve the level of lean production management of power grids.

The present invention adopts the following technical solutions. The present invention provides a data sharing system based on the edge computing of intelligent service terminals in the station area, including: a data access layer, a data integration layer, an edge data sharing layer, a data mart layer, a data application layer and a cloud data storage layer. It is characterized by:

The data access layer includes a real-time data access module, an offline data access module, and an equipment parameter access module. The data access layer is connected to the intelligent service terminal, collecting and distributing transformers, branch boxes, household meters, charging piles, distributed Energy grid equipment operation data and equipment parameters. The collected power grid equipment operation data and equipment parameters are divided into real-time data, offline data, and equipment parameters, and then connected to the real-time data access module, offline data access module, and equipment parameter access respectively. module.

Preferably, the MQTT protocol is used to complete the sending and receiving operations of data between levels,

The MQTT protocol includes three roles: a message publisher, a message subscriber, and an MQTT server. The functional modules of the previous level serve as message publishers to send a data model with a unified format to the data center through the MQTT server. At the same time, the functions of the latter level As a message subscriber, the module configures the subscription interface information. After the MQTT server matches the subscription service information, it forwards the data model shared by the data center to the message subscriber to complete the data sharing process and realize the data interaction process.

The data integration layer includes a data traceability module, a multi-source collection data cleaning module, and a statistical calculation module; the entire data integration layer performs the multi-source collection data cleaning module through the MQTT protocol, and then sends JSON to the edge data sharing layer for hosting.

The edge data sharing layer includes a data center module, and the data center module is used to realize short-term data storage. In the data center module, the measurement class objects, statistics class objects, parameter class objects, and control class objects uploaded by the data integration layer are Objects are represented by a data model in a unified format, using the MQTT protocol to differentiate the data models in the form of interfaces, publish them by category and store them in the data center to realize data sharing and exchange.

The data mart layer sends the required data interface to the edge sharing layer, receives the data under the interface sent by the edge sharing layer, combines the received real-time data and feature data, and uploads it to the data application layer.

The functional module of the data application layer obtains the data model corresponding to the data mart layer according to the needs of the business, makes decision-making analysis on the business-oriented scenario, and realizes the localized computing and processing advantages of the intelligent business terminal in the Taiwan area to realize the edge measurement and localization. Abnormal working condition monitoring and rapid response management.

The cloud data storage layer is responsible for in-depth data mining and advanced business processing. The intelligent business terminal uploads the data set assembled based on the business function modules to the cloud data storage layer for storage, combined with the business real-time data of the intelligent business terminal, through machine learning and deep learning. The algorithm performs the classification and prediction of abnormal operating conditions.

The real-time data includes real-time environmental data monitored by sensors: temperature, humidity, and real-time electrical data collected by smart grid equipment: voltage, current, power quality, active power, reactive power;

The offline data includes the historical electrical and environmental data before the terminal that is not connected to the terminal that is artificially imported from the outside of the intelligent service terminal and the historical electrical and environmental data logs after the terminal is accessed and stored in the internal file system of the intelligent service terminal;

The device parameters include grid device parameter information and sensor parameter information.

The data traceability module draws on the idea of serialized storage, reorders the real-time data and offline data of each device collected by the data access layer, attaches the performance parameters of the device itself, and then stores it in the respective device according to the size of the timestamp. table;

The multi-source pooled data cleaning module performs data cleaning based on the time series data reordered by the data traceability module, and applies the related methods of data missing value filling and outlier processing in data mining to further supplement and complete the data and improve data accuracy;

The statistical calculation module artificially extracts business characteristics from the time domain and the frequency domain from the data tables organized by the data traceability module and the multi-source aggregated data cleaning module according to the common knowledge of experts: abnormal duration and abnormal occurrence frequency, Relevant business characteristics are obtained through statistical learning algorithm calculation.

The data center module includes a data format unification unit, a data short-term storage unit, and a data classification management unit, wherein:

The data format unification unit names each data transmitted from the data integration layer with a unified rule, annotates the data type, the data generation time, together with the value of the data, to form a JSON string;

The data short-term storage unit is to centrally store the JSON strings of non-real-time data after the data format is unified, and sort the data by time;

The data classification management unit is to combine the real-time data and the historical data in the storage unit to give an interface for the data mart layer to call.

The beneficial effect of the present invention is that, compared with the prior art, the key point of the present application is to start from the data-driven point of view, combine the container technology, and follow the principle of "centralized management and control, classified storage, and unified application" of the cloud center and the intelligent business terminal, The intelligent business terminal is divided into data access layer, data integration layer, edge data sharing layer, data mart layer, data application layer and cloud data storage layer from bottom to top, so as to break through the data barriers between the cloud center and the intelligent business terminal. The side integrates various information resources in the distribution station area, reduces the pressure on the bandwidth of the cloud computing center, reduces the computing resource demand of the intelligent service terminal, improves the data processing capability, enables the diagnosis service to respond quickly, realizes the collection and sharing of data, and enhances the creation of data value. capacity, and improve the lean level of power grid production management.

Description of drawings

Figure 1 shows a data sharing system based on edge computing of intelligent service terminals in the station area.

Figure 2 is a flow chart of the data access layer.

Figure 3 is a flow chart of the data integration layer.

Figure 4 is a flow chart of data interaction.

Figure 5 shows the workflow of the data mart.

Figure 6 shows the results of the three-phase unbalanced data mart.

FIG. 7 is a diagram showing the results of diagnosis of abnormal three-phase unbalance.

Detailed ways

The present application will be further described below with reference to the accompanying drawings. The following examples are only used to more clearly illustrate the technical solutions of the present invention, and cannot be used to limit the protection scope of the present application.

The terms mentioned in the present invention are explained as follows:

1. MQTT (Message Queuing Telemetry Transport, Message Queuing Telemetry Transport Protocol), is a "lightweight" communication protocol based on the publish/subscribe mode, which is built on the TCP/IP protocol and is developed by IBM Released in 1999. The biggest advantage of MQTT is that it can provide real-time and reliable message services for connecting remote devices with very little code and limited bandwidth. As a low-overhead, low-bandwidth instant messaging protocol, it has a wide range of applications in the Internet of Things, small devices, and mobile applications. It has the following main characteristics:

(1) Use the publish/subscribe message mode to provide one-to-many message publishing and decouple the application. This is very similar to XMPP, but the information redundancy of MQTT is much less than that of XMPP, because XMPP uses XML formatted text to pass data.

(2) Message transmission that shields the payload content.

(3) Use TCP/IP to provide network connection. The mainstream MQTT is based on TCP connection for data push, but there is also a UDP-based version called MQTT-SN. The advantages and disadvantages of these two versions are naturally different due to different connection methods.

(4) There are three kinds of message publishing service quality:

"At most once", message publishing relies entirely on the underlying TCP/IP network. Message loss or duplication can occur. This level can be used in the case of environmental sensor data, it doesn't matter if one read record is lost, because there will be a second transmission soon.

"At least once" ensures that messages arrive, but message duplication may occur.

"Only Once" ensures that the message arrives once. This level can be used in some more demanding billing systems. In billing systems, duplicate or lost messages can lead to incorrect results.

(5) Small transmission, small overhead (fixed-length header is 2 bytes), and protocol exchanges are minimized to reduce network traffic. This makes it very suitable for "In the field of the Internet of Things, the communication between sensors and servers, the collection of information, the computing power and bandwidth of embedded devices are relatively weak, and the use of this protocol to transmit messages is perfect.

(6) A mechanism for notifying all parties involved of client abnormal interruption using Last Will and Testament features.

2. mosquitto server

Mosquitto is an open source message broker software that implements the message push protocol MQTT v3.1. It provides a lightweight, publishable/subscribeable message push mode, making short message communication between devices simple. , can be ported to Linux virtual machines and Linux development boards, such as low-power sensors, mobile phones, embedded computers, microcontrollers and other mobile devices that are widely used now. A typical application case is the remote monitoring and automation implemented at home by Andy Stanford-ClarkMosquitto (one of the founders of the MQTT protocol). The Mosquitto project also provides a C library for implementing MQTT clients, as well as the very popular mosquitto_pub and mosquitto_sub line of command MQTT clients.

Aiming at the shortcomings of the existing data sharing system in the distribution station area, from the perspective of data driving, combined with the container technology and APP development technology, the present invention proposes a data sharing system based on the edge computing of the intelligent service terminal in the station area. The system follows the cloud The principle of "centralized management and control, classified storage, and unified application" of the center and intelligent business terminals ensures simplicity and generality, high efficiency and controllability, and low power consumption, reduces the nesting level of the shared system, reduces redundant information description, and facilitates rapid development and formation. Minimize the occupancy of communication resources and realize fast data upload and delivery.

Embodiment 1 of the present invention provides a data sharing system based on edge computing of an intelligent service terminal in a station area. As shown in Figure 1, from bottom to top, it specifically includes data access layer, data integration layer, edge data sharing layer, data mart layer, data application layer and cloud data storage layer.

1. The data access layer includes a real-time data access module, an offline data access module, and an equipment parameter access module. The data access layer is connected to intelligent service terminals, collecting distribution transformers, branch boxes, household meters, charging piles, distributed Energy grid equipment operation data and equipment parameters. The collected power grid equipment operation data and equipment parameters are divided into real-time data, offline data, and equipment parameters, and then connected to the real-time data access module, offline data access module, and equipment parameter access respectively. module.

Among them, real-time data refers to real-time environmental data monitored by sensors, such as temperature, humidity, etc., and real-time electrical data collected by smart grid devices such as multi-function meters, such as voltage, current, power quality, active power, reactive power, etc., Offline data refers to the historical electrical and environmental data before the terminal that is not connected to the terminal that is manually imported from the outside of the smart service terminal, and the historical electrical and environmental data logs after the terminal is connected to the internal file system of the smart service terminal. Device parameters refer to grid device parameter information and sensor parameter information; the data access layer uploads real-time data, offline data, and device parameters to the data integration layer, and the data access layer process is shown in Figure 2.

2. The data integration layer includes a data traceability module, a multi-source collection data cleaning module, and a statistical calculation module; the data traceability module draws on the idea of serialized storage and reorders the real-time data and offline data of each device collected by the data access layer. Attach the performance parameters of the device itself, and then store them in their respective tables according to the size of the timestamp; the multi-source aggregated data cleaning module performs data cleaning based on the time series data reordered by the data traceability module, and applies data missing values in data mining. The related methods of filling and outlier processing, such as mean filling, etc., further complement the data and improve the accuracy of the data; the statistical calculation module is based on the common knowledge of experts, in the data traceability module and multi-source collection data cleaning module. In the table, business features are artificially extracted from the time domain and frequency domain, such as abnormal duration and abnormal occurrence frequency, etc., and relevant business features are calculated through statistical learning algorithms; finally, the data integration layer traces the source and collects the cleaned data through the MQTT protocol. The measurement data (real-time data) in the table, the parameters of the device itself (parameter data) and the characteristic data (statistical data) after the statistical calculation module are sent together with JSON to the edge data sharing layer for hosting, the data integration layer process As shown in Figure 3.

As the level of processing raw data, business system data is extracted mainly through multi-source collection of data, preliminary cleaning, and real-time calculation. For business system data based on time scale, business system data extraction needs to be completed through data traceability and statistical calculation.

3. The edge data sharing layer includes a data center module, which is used to realize short-term data storage. In the data center module, the measurement objects, statistical objects, parameter objects, The control class object is represented by a data model in a unified format, and adopts the MQTT protocol, a lightweight communication protocol based on publish/subscribe.

The data center module further includes a data format unification unit, a data short-term storage unit, and a data classification management unit. The data format unification unit names each data transmitted from the data integration layer with a unified rule, and marks the data type, the data generation time, and the value of the data. , form a JSON string; the data short-term storage unit is to store the JSON strings of non-real-time data centrally after the data format is unified, and the data is sorted by time; the data classification management unit is to store the real-time data and historical data in the storage unit. The data combination is given an interface for the data mart layer to call; for example, the real-time voltage, current, active power, reactive power, temperature, humidity, etc., composed of raw data JSON strings, are integrated as the basic general interface that needs to be called for real-time research and judgment services. ;A JSON string of overload-related characteristic data such as A-phase overload rate, B-phase overload rate, C-phase overload rate, A-phase overload accumulation time, B-phase overload accumulation time, C-phase overload accumulation time, etc. Integration, as a characteristic interface that needs to be called by overloaded services.

The system uses the MQTT protocol to complete the data transmission and reception operations between various levels, and each level is installed with one or more functional modules for processing data services. The protocol has three main roles, namely message publisher, message subscriber and MQTT server. The function module at the previous level, as a message publisher, sends a data model with a unified format to the data center through the MQTT server. At the same time, the function module at the latter level, as a message subscriber, configures the subscription interface information. After the MQTT server matches the subscription service information, The data model shared by the data center is forwarded to the message subscribers to complete the data sharing process. The data interaction process is shown in Figure 4.

4. The data mart layer sends the required data interface to the edge sharing layer, and receives the data under the interface sent by the edge sharing layer. The data mart layer includes the reactive power compensation data mart module, the leakage detection data mart module, etc. , the reactive power compensation data mart sends a basic general interface for subscribing to real-time research and judgment services to the edge data sharing layer, and receives the data contained in the interface through the MQTT protocol, including power factor, A-phase active power, B-phase active power, and C-phase active power; At the same time, subscribe to the feature interface required by the reactive power compensation service, and receive the data of the feature interface, including freezing forward active energy for 15 minutes, freezing forward reactive energy for 15 minutes, freezing reverse active energy for 15 minutes, and freezing reverse power for 15 minutes. Reactive power and other characteristics, and then combine the above real-time data and characteristic data, and upload them to the reactive power compensation module of the data application layer. In the same way, the leakage detection data mart sends the basic general interface for subscribing to real-time research and judgment services to the edge data sharing layer, and receives the data contained in the interface through the MQTT protocol, including three-phase voltage, three-phase current, total power loss, and no-load power loss. ;At the same time, subscribe to the feature interface required by the leakage protection detection business, and receive the data of the feature interface, including the leakage protection time, leakage protection enable and other characteristics; then combine the above real-time data and feature data and upload it to the leakage protection detection of the data application layer module.

The data mart is oriented to specific edge services, analyzes the data characteristics that can characterize the service, dynamically builds the required data model, and flexibly loads data based on the data cluster of the data center. The data mart layer is more inclined to the tactical requirements of the business, and the goal is to meet the immediacy and efficiency of the analysis of specific applications in the intelligent fusion terminal. The workflow of the data mart layer is shown in Figure 5.

5. Data application layer

The functional module of the data application layer obtains the data model corresponding to the data mart layer according to the business needs, makes decision-making analysis on the business-oriented scenarios, and realizes the edge detection and local abnormal work based on the localized computing and processing advantages of the intelligent business terminal in the Taiwan area. Condition monitoring and rapid response governance.

6. Cloud data storage layer

The cloud center is responsible for in-depth data mining and advanced business processing. The intelligent business terminal uploads data sets assembled based on business function modules to the cloud center for storage. Combined with the business real-time data of the intelligent business terminal, machine learning and deep learning algorithms are used to detect abnormal conditions. Classification and prediction.

In the second embodiment of the present invention, an intelligent service terminal is used as the edge terminal, mosquitto is used as the agent in the MQTT protocol, Huawei cloud is selected as the cloud, and the Huawei IoT platform is selected to realize the communication between the edge terminal and the cloud, and the device_id of the created product is 5fed8387aaafca02dba396b2_123456 , the present invention specifically takes the typical business of abnormal diagnosis of three-phase unbalance of distribution transformers in the station area as an example, and realizes the rapid response of the business by building a data sharing system.

1. Functional module design of data access layer

The intelligent service terminal is connected to the key equipment of the distribution network such as distribution transformers, branch boxes, household meters, charging piles, distributed energy sources, etc., and performs data collection, monitoring and management on them. The business is complex and the data is massive and heterogeneous. Real-time data (sensor data and smart grid equipment data), offline data (external timing import data and data stored in the internal file system) and equipment parameters (grid equipment parameter information and sensor parameter information), so the present invention includes 104 protocol, Modbus The intelligent business terminal communication acquisition board interface communication protocol including the protocol, designed the data access layer function module, provides the terminal to the basic communication analog acquisition function, and the collected raw data is released to the data integration layer in JSON format through the mosquitto agent. The model is shown in Table 3.

Table 3 Data release format of data access layer

2. Functional module design of data integration layer

The data integration layer, as the layer for processing raw data, mainly extracts business system data through multi-source collection of data, preliminary cleaning, and real-time calculation. The time scale-based business system data needs to be traced through data and statistical calculation. For data calculation, the data integration layer obtains all the features required by the upper-layer application function module business, including the features required for three-phase unbalanced abnormal diagnosis, as follows:

①Three-phase voltage unbalance ε

Where: U _max is the maximum phase current, U _min is the minimum phase current.

② Three-phase unbalance duration D _tpu

In the formula: D _tpu-1 is the three-phase unbalance duration of the distribution transformer at the previous moment, and ε is the three-phase unbalance degree of the distribution transformer at the current moment;

③ Three-phase unbalance frequency F _tpu

In the formula: D _tpu is the three-phase unbalanced duration of the distribution transformer at the i-th time, D _tpu(i-1) is the three-phase unbalanced duration of the distribution transformer at the i-1-th time, and k _i is the i-th time. The three-phase unbalanced event occurs in the distribution transformer, and t is the statistical time scale.

④ Three-phase unbalance cumulative time C _tpu

In the formula: m _i is the three-phase unbalanced state of the distribution transformer at the i-th moment, and t is the statistical time scale.

After the functional modules of the data integration layer calculate all the features, they are published to the edge data sharing layer in JSON format through the mosquitto agent. The JSON model is shown in Table 4.

Table 4 Data release format of data integration layer

3. Functional module design of edge data sharing layer

The data center is a functional module for short-term data storage. In order to reduce the nesting level of the data sharing system, the measurement class objects, statistics class objects, parameter class objects, and control class objects of the data integration layer are in a unified format. Characterization, using the MQTT protocol, a lightweight communication protocol based on publish/subscribe, to publish and store the model in the data center to realize data sharing and exchange.

This architecture uses the MQTT protocol to complete the data transmission and reception operations between various levels, and each level is installed with one or more functional modules for processing data services. The protocol has three main roles, namely message publisher, message subscriber and MQTT server. The APP at the previous level, as the message publisher, sends the data model in a unified format to the data center through the MQTT server. At the same time, the function module at the latter level, as the message subscriber, configures the subscription interface information. After the MQTT server matches the subscription service information, the The data model shared by the data center is forwarded to the message subscribers to complete the data sharing process. The data interaction process is shown in Figure 4.

4. Data mart layer functional module design

The data mart is oriented to specific edge services, analyzes the data characteristics that can characterize the service, dynamically builds the required data model, and flexibly loads data based on the data cluster of the data center. The data mart layer is more inclined to the tactical requirements of the business, and the goal is to meet the immediacy and efficiency of the analysis of specific applications in the intelligent fusion terminal. The workflow of the data mart layer is shown in Figure 5. For the three-phase unbalanced abnormal diagnosis scenario, there is a corresponding three-phase unbalanced data mart function module, which is used to extract the features required for three-phase unbalanced abnormal diagnosis from the features stored in the edge data sharing layer, as shown in Figure 6. After obtaining the required characteristics, it is published to the three-phase unbalance abnormality diagnosis function module of the data application layer in JSON format through the mosquitto agent. The JSON model is shown in Table 5.

Table 5 Data Mart Layer Data Publishing Format

5. Data application layer functional module design

Each business function module of the data application layer obtains the data model corresponding to the data mart layer according to its own business needs, makes decision analysis on the business-oriented scenarios, and realizes the edge measurement on the spot based on the localized computing and processing advantages of the intelligent business terminal in the Taiwan area. Abnormal working condition status monitoring and rapid response management. In this example, a three-phase unbalance abnormality diagnosis function module is designed in the data application layer. According to the interactive mechanism, the model parameters of the cloud, that is, the threshold value of each level, are subscribed to, and a three-phase unbalance abnormality diagnosis model is established, and then the current data sharing mechanism is used to extract the current Characteristic data, according to the diagnosis rules, the current abnormal degree of three-phase unbalance is diagnosed on the spot. The abnormal degree is measured by the square sum error. The square sum error d _k is used to measure the deviation between the observed value and the real value. The formula is

In the formula: k is the degree series, the number k=4 in this paper, n is the number of features of this degree, and in this example, the number of features n=4.

Compare the square sum errors of all degree thresholds, and select the degree threshold with the smallest square sum error. The degree represented by the degree threshold is the abnormal degree of three-phase imbalance in the current scene. The level is used to indicate the abnormal degree of three-phase imbalance in the current station area, which is divided into 4 levels (normal, attention, abnormal, serious), which are respectively expressed as 1, 2, 3, and 4. The calculation formula is as follows:

The diagnosis result is shown in Figure 7. From the above example, it can be seen that the abnormal degree of three-phase imbalance in the current station area is 1, which is a normal state. The cloud operation and maintenance personnel receive the information that the current station area is in a normal state and confirm that the station area is operating normally; Diagnosis according to the above method can reduce the computing resource demand of the intelligent service terminal, improve the data processing capability, and enable the diagnosis service to respond quickly, thereby providing guarantee for reducing operation and maintenance costs and improving the reliability of power distribution equipment.

6. Cloud data storage layer

The cloud center is responsible for in-depth data mining and advanced business processing. The intelligent business terminal uploads the data set assembled based on the business APP to the cloud center for storage, and combines the real-time business data of the intelligent business terminal to classify abnormal conditions through machine learning and deep learning algorithms. and forecast.

The beneficial effect of the present invention is that, compared with the prior art, the present invention can follow the "centralized management and control, classified storage, and unified application" of the cloud center and the intelligent service terminal under the "cloud-pipe-side-terminal" power distribution Internet of Things system architecture. The principle of intelligent business terminal is divided into data access layer, data integration layer, edge data sharing layer, data mart layer, data application layer and cloud data storage layer from bottom to top, so as to break through the data barriers between cloud center and intelligent business terminal. Integrate various information resources in the power distribution station area on the edge side, reduce the pressure on the bandwidth of the cloud computing center, reduce the computing resource demand of the intelligent service terminal, improve the data processing capability, enable the diagnosis service to respond quickly, realize data collection and sharing, improve The ability to create data value and improve the lean level of power grid production management.

The applicant of the present invention has described and described the embodiments of the present invention in detail with reference to the accompanying drawings, but those skilled in the art should understand that the above embodiments are only preferred embodiments of the present invention, and the detailed description is only to help readers better Rather, any improvement or modification based on the spirit of the present invention should fall within the protection scope of the present invention.

Claims (10)

1. a data sharing system based on the edge computing of intelligent service terminals in the station area, comprising: data access layer, data integration layer, edge data sharing layer, data mart layer, data application layer and cloud data storage layer, it is characterized in that : The data access layer includes a real-time data access module, an offline data access module, and an equipment parameter access module. The data access layer is connected to the intelligent service terminal, collecting and distributing transformers, branch boxes, household meters, charging piles, distributed Energy grid equipment operation data and equipment parameters. The collected power grid equipment operation data and equipment parameters are divided into real-time data, offline data, and equipment parameters, and then connected to the real-time data access module, offline data access module, and equipment parameter access respectively. module. 2. a kind of data sharing system based on edge computing of intelligent service terminal in station area according to claim 1, is characterized in that: The MQTT protocol is used to complete the transmission and reception of data between various levels. The MQTT protocol includes three roles: a message publisher, a message subscriber, and an MQTT server. The functional modules of the previous level serve as message publishers to send a data model with a unified format to the data center through the MQTT server. At the same time, the functions of the latter level As a message subscriber, the module configures the subscription interface information. After the MQTT server matches the subscription service information, it forwards the data model shared by the data center to the message subscriber to complete the data sharing process and realize the data interaction process. 3. a kind of data sharing system based on edge computing of intelligent service terminal in station area according to claim 2, is characterized in that: The data integration layer includes a data traceability module, a multi-source collection data cleaning module, and a statistical calculation module; the entire data integration layer performs the multi-source collection data cleaning module through the MQTT protocol, and then sends JSON to the edge data sharing layer for hosting. 4. a kind of data sharing system based on edge computing of intelligent service terminal in station area according to claim 3, is characterized in that: The edge data sharing layer includes a data center module, and the data center module is used to realize short-term data storage. In the data center module, the measurement class objects, statistics class objects, parameter class objects, and control class objects uploaded by the data integration layer are Objects are represented by a data model in a unified format, using the MQTT protocol to differentiate the data models in the form of interfaces, publish them by category and store them in the data center to realize data sharing and exchange. 5. a kind of data sharing system based on edge computing of intelligent service terminal in station area according to claim 4, is characterized in that: The data mart layer sends the required data interface to the edge sharing layer, receives the data under the interface sent by the edge sharing layer, combines the received real-time data and feature data, and uploads it to the data application layer. 6. a kind of data sharing system based on edge computing of intelligent service terminal in station area according to claim 5, is characterized in that: The functional module of the data application layer obtains the data model corresponding to the data mart layer according to the needs of the business, makes decision-making analysis on the business-oriented scenario, and realizes the localized computing and processing advantages of the intelligent business terminal in the Taiwan area to realize the edge measurement and localization. Abnormal working condition monitoring and rapid response management. 7. a kind of data sharing system based on edge computing of intelligent service terminal in station area according to claim 6, is characterized in that: The cloud data storage layer is responsible for in-depth data mining and advanced business processing. The intelligent business terminal uploads the data set assembled based on the business function modules to the cloud data storage layer for storage, combined with the business real-time data of the intelligent business terminal, through machine learning and deep learning. The algorithm performs the classification and prediction of abnormal operating conditions. 8. a kind of data sharing system based on the edge computing of the intelligent service terminal in the station area according to claim 7, is characterized in that: in the data access layer, The real-time data includes real-time environmental data monitored by sensors: temperature, humidity, and real-time electrical data collected by smart grid equipment: voltage, current, power quality, active power, reactive power; The offline data includes the historical electrical and environmental data before the terminal that is not connected to the terminal that is artificially imported from the outside of the intelligent service terminal and the historical electrical and environmental data logs after the terminal is accessed and stored in the internal file system of the intelligent service terminal; The device parameters include grid device parameter information and sensor parameter information. 9. a kind of data sharing system based on edge computing of intelligent service terminal in station area according to claim 8, is characterized in that: The data traceability module draws on the idea of serialized storage, reorders the real-time data and offline data of each device collected by the data access layer, attaches the performance parameters of the device itself, and then stores it in the respective device according to the size of the timestamp. table; The multi-source pooled data cleaning module performs data cleaning based on the time series data reordered by the data traceability module, and applies the related methods of data missing value filling and outlier processing in data mining to further supplement and complete the data and improve data accuracy; The statistical calculation module artificially extracts business characteristics from the time domain and frequency domain from the data tables organized by the data traceability module and the multi-source aggregated data cleaning module according to the common knowledge of experts: abnormal duration and abnormal occurrence frequency. The statistical learning algorithm calculates the relevant business characteristics. 10. a kind of data sharing system based on edge computing of intelligent service terminal in station area according to claim 9, is characterized in that: The data center module includes a data format unification unit, a data short-term storage unit, and a data classification management unit, wherein: The data format unified unit names each data transmitted from the data integration layer with a unified rule, and marks the data type, the data generation time, and the value of the data to form a JSON string; The data short-term storage unit is to centrally store the JSON strings of non-real-time data after the data format is unified, and sort the data by time; The data classification management unit is to combine the real-time data and the historical data in the storage unit to give an interface for the data mart layer to call.

Images