CN116010114B - Equipment energy efficiency management and control system based on edge calculation - Google Patents

Abstract

The equipment energy efficiency management and control system based on edge calculation adopts an end layer, an edge layer and a cloud layer framework, wherein energy consumption equipment and energy consumption meters are uniformly distributed on the end layer, a central server is arranged on the cloud layer, and the edge layer comprises an energy data acquisition module, an energy consumption edge calculation module and an equipment management and control module. The equipment management and control module of the edge layer comprises an operation control sub-module, a fault management sub-module and an equipment ledger sub-module; the energy data acquisition module acquires energy consumption data of the energy consumption instrument; the energy consumption edge calculation module is used for carrying out energy consumption edge calculation according to the energy consumption data, and the calculation result is sent to the central server; the central server performs statistical analysis on the calculation result, performs energy efficiency control processing, establishes control description and corresponding control parameters of the energy consumption equipment, and maps the control description and corresponding control parameters to the operation control sub-module. The invention provides data computing capability at the edge layer, and ensures the instantaneity of energy efficiency management and control.

Description

Equipment energy efficiency management and control system based on edge computing

technical field

The invention belongs to the technical field of energy efficiency management and control of industrial equipment, is used to control energy consumption such as water, electricity and the like of industrial equipment, and is an equipment energy efficiency management and control system based on edge computing.

Background technique

The energy efficiency control of industrial equipment mainly collects various types of energy data, such as electricity, water, natural gas, industrial gas, cold heat, etc., analyzes energy consumption, and provides corresponding technical means to eliminate blind spots in energy consumption and reduce costs.

In the prior art, the main method for energy efficiency management and control of multiple industrial equipment is generally to collect the energy consumption data of a workshop or a factory area in a local area network, and the data is directly uploaded to the monitoring center through the data collection gateway, and the energy efficiency stored in the monitoring center The control center server manages, analyzes and displays the data. According to the observation data, experienced technicians intervene to control the energy-consuming equipment.

Obviously, the energy efficiency management and control system based on the above method has the following problems:

(1) The collected data is single, and the data of the main energy-consuming equipment is not collected and analyzed, but it is broadly collected for the workshop or factory area. (2) The data of energy consumption meters are uploaded in real time, the data volume and bandwidth consumption are large, and the storage operation and maintenance cost of the energy efficiency management and control center server is high. (3) There is little analysis work on the collected energy efficiency data, and the value of the data is not enough to guide and control the main energy-consuming equipment, relying heavily on the own experience of the technicians. (4) The data of various instruments are directly transmitted to the monitoring center, which has poor scalability and relatively complicated communication debugging. (5) The control programs and descriptions of major energy-consuming equipment are generally based on the designated brand controllers that come with the equipment. The interface is relatively closed, and it is relatively difficult to adjust or optimize the program. Replacement, it will be difficult for the operation and maintenance of the equipment.

Contents of the invention

In order to overcome the shortcomings of the above-mentioned prior art, the purpose of the present invention is to provide a device energy efficiency management and control system based on edge computing, which can provide complex data computing capabilities at the edge layer near the device end, and ensure the real-time performance of energy efficiency management and control of the IoT network. In order to improve the flexibility and scalability of the layout, reduce the server load; and further expand the scope of data collection to make the prediction more accurate.

In order to achieve the above object, the technical scheme adopted in the present invention is:

The equipment energy efficiency management and control system based on edge computing adopts the architecture of terminal layer, edge layer and cloud layer. Energy-consuming equipment and energy consumption meters are arranged on the terminal layer, the cloud layer is arranged with a central server, and the edge layer includes energy data. A collection module, an energy consumption edge computing module, and a device management and control module; the device management and control module includes an operation control sub-module, a fault management sub-module, and an equipment ledger sub-module;

The energy data collection module collects the energy consumption data of the energy consumption meter;

The energy consumption edge calculation module performs energy consumption edge calculation according to the energy consumption data, and sends the calculation result to the central server; the central server performs statistical analysis on the calculation result, performs energy efficiency control processing, and establishes energy The control description and corresponding control parameters of the consumption equipment are mapped to the operation control sub-module;

The operation control sub-module collects the parameter information of the energy-consuming equipment, and sends it to the fault management sub-module, and transmits the fault alarm information triggered according to the fault setting conditions to the equipment account sub-module, and passes through the equipment account sub-module The summary and ledger information are transmitted to the central server; the operation control sub-module controls the energy-consuming equipment to execute the control parameters.

In one embodiment, the energy data acquisition module includes a data acquisition sub-module, a preprocessing sub-module and a data storage sub-module; the energy consumption meter includes a water meter, an electric meter and a gas meter;

The data collection sub-module collects the original energy consumption data of the energy consumption meter, and constructs a column vector of water consumption energy consumption data, a column vector of electricity consumption data and a column vector of gas consumption data. In each column vector, each The item represents the energy consumption within a time period, which is divided into multiple time periods every day;

The preprocessing sub-module respectively performs data cleaning, data transformation and data update on the column vectors of water energy consumption data, electricity consumption data column vectors and gas consumption data column vectors;

The data storage submodule is used for storing the preprocessing result of the preprocessing submodule.

In one embodiment, the data cleaning is to analyze whether the data of each period is an abnormal value, and if it is an abnormal value, the average value of adjacent bits is used to replace the abnormal value; the data transformation is to transform the original column vector into energy consumption prediction The calculated sequence to be processed; the data update, when performing online energy consumption prediction, assign the energy consumption values of each time period on the i+1 day to the energy consumption values of each time period on the i-th day.

In one embodiment, the energy consumption edge computing module includes an energy consumption multi-scale analysis submodule, an energy consumption alarm submodule and an energy consumption prediction submodule;

The energy consumption multi-scale analysis sub-module performs energy consumption analysis in two dimensions of spatial scale and time scale;

The energy consumption alarm sub-module sets a threshold percentage for water consumption according to a time scale, and if the water consumption exceeds the threshold compared with the previous time scale, an alarm is triggered; for gas consumption and electricity consumption , set the energy consumption data collection frequency, when the energy consumption fluctuation exceeds the set threshold, an alarm will be triggered;

The energy consumption prediction sub-module uses the moving average method to predict the water energy consumption and gas energy consumption of each energy-consuming equipment in a future time scale, and uses the long-short-term memory network prediction model to predict the electricity consumption in a future time scale.

In one embodiment, the central server obtains the long-short-term memory network prediction model through training with energy consumption data, and sends model parameters to the energy consumption prediction sub-module.

In one embodiment, the central server includes an energy consumption configuration monitoring submodule, an energy consumption prediction calculation submodule, and an energy efficiency control processing submodule;

The energy consumption configuration monitoring sub-module establishes an overall layout display interface for each energy consumption device in an industrial configuration mode according to the process flow, so as to monitor the energy consumption data of each energy consumption device;

The energy consumption prediction calculation sub-module uses the collected energy consumption data to train the long-short-term memory network to obtain model parameters, and sends the model parameters to the energy consumption prediction sub-module according to the update conditions, and performs Online energy consumption prediction;

The energy efficiency control processing sub-module uses a control programming language to control and describe energy-consuming equipment, and introduces production process parameters, process constraints, and energy consumption prediction values in a future time scale, and performs calculations with the lowest energy consumption cost as the optimization goal. Get the corresponding control parameters.

In one embodiment, the energy efficiency control processing sub-module takes the energy-consuming equipment as the control object, uses a top-down object-oriented programming method, and uses composite function blocks to establish the energy-consuming equipment control description corresponding to the energy-consuming data , use the configuration function to connect each compound function block, and form a control description based on the compound function block network, which is mapped to the operation control sub-module.

In one embodiment, the control description of the energy-consuming equipment is the parameter information of the energy-consuming equipment collected through the terminal layer, the mechanism of the energy-consuming equipment, the mapping relationship between the input and the output established by the control constraints, and the implementation code; The operation control sub-module provides the control description for the resources of the virtual container engine to describe the resource environment running at the edge layer.

In one embodiment, the edge layer includes a plurality of edge control devices, and each edge control device includes the energy data collection module, the energy consumption edge computing module and the equipment management and control module.

Compared with prior art, the beneficial effect of the present invention is:

The present invention utilizes the advantages of cloud computing and edge computing, and proposes an energy efficiency management and control system with a device-edge cloud structure, so that the energy efficiency management and control system is easy to deploy and easy to expand. Taking various energy consumption meters as the objects of data acquisition and perception, and the main energy consumption equipment as the controlled objects, through edge computing processing, it can provide complex data computing capabilities at the edge layer, greatly reducing the bandwidth requirements of uplink data, ensuring Energy efficiency controls the real-time nature of the IoT network, greatly reducing the storage requirements of the central server, reducing the investment and operation and maintenance costs of the cloud layer.

Description of drawings

Fig. 1 is a schematic structural diagram of the energy efficiency management and control system of the present invention.

Fig. 2 is a schematic flow chart of the energy efficiency management and control method of the present invention.

Fig. 3 is a schematic diagram showing the arrangement and association of multiple end layers, multiple edge layers and cloud layers in the present invention.

Detailed ways

The implementation of the present invention will be described in detail below in conjunction with the drawings and examples.

As shown in FIG. 1 , the present invention provides an edge computing-based device energy efficiency management and control system, which adopts a terminal layer, edge layer, and cloud layer architecture. Wherein, energy-consuming equipment and energy-consumption meters are arranged on the end layer, and a central server is arranged on the cloud layer, and the edge layer includes an energy data collection module, an energy consumption edge computing module, and a device management and control module; the device management and control The module includes fault management sub-module, equipment ledger sub-module and operation control sub-module. in:

The energy consumption data collection module communicates with the on-site energy consumption meter to collect the energy consumption data of the energy consumption meter;

The energy consumption edge calculation module performs energy consumption edge calculation according to the energy consumption data, and sends the calculation result to the central server; the central server performs statistical analysis on the calculation result, and performs energy efficiency control decision processing , establishing a control description of the energy-consuming equipment and corresponding control parameters, and mapping to the operation control sub-module, where the operation control sub-module controls the energy-consuming equipment to execute the control parameters.

The operation control sub-module collects the parameter information of the energy-consuming equipment, and sends it to the fault management sub-module, and transmits the fault alarm information triggered according to the fault setting conditions to the equipment account sub-module, and passes through the equipment account sub-module The summary and ledger information are transmitted to the central server together.

The present invention performs energy consumption prediction calculation and processing on the cloud layer based on historical energy consumption data, sends the calculated parameter information to the edge layer, and updates according to certain conditions at the same time, and uses the edge layer computing capability to collect energy consumption data at the edge layer. And carry out preprocessing and online energy consumption prediction, realize online prediction on the side near the device at the edge layer, and control the main energy-consuming devices at the end layer; and establish corresponding control descriptions on the cloud layer for the main energy-consuming devices, The equipment is arranged in different positions according to the site conditions, and the layers are connected in different ways. The energy efficiency management and control structure based on edge computing in the present invention can provide complex data calculation capabilities and ensure the real-time performance of the energy efficiency management and control IoT network.

In the present invention, energy-consuming equipment is arranged on the end layer, and energy-consuming instruments mainly include water meters, electric meters and gas meters, which are also arranged on the end layer. According to the device energy efficiency management and control system based on the cloud layer-edge layer-end layer Internet of Things structure of the present invention, its energy efficiency management and control method: energy consumption data such as water meters, electricity meters, and gas meters at the end layer are collected and sent to the edge layer for energy consumption After calculation, analysis and processing, it is then sent to the cloud center server. Referring to Figure 2, the specific process is as follows:

Energy consumption data collection: Collect energy consumption data such as water meters, electricity meters, and gas meters, including data collection, preprocessing, and data storage, and send them to the energy consumption edge computing module.

Edge computing of energy consumption: Use the collected energy consumption data to perform multi-scale analysis, energy consumption alarm and energy consumption prediction, and send the analysis results to the central server.

Equipment management and control: collect parameter information of major energy-consuming equipment, enter the operation control environment function to perform control and monitoring, and transmit it to the fault management sub-module, and transmit the fault alarm information triggered by the fault setting conditions to the equipment ledger sub-module. The summary of ledger sub-module is transmitted to the central server together with the ledger information.

In some embodiments of the present invention, in order to realize the above energy consumption data collection, the energy data collection module includes a data collection sub-module, a preprocessing sub-module and a data storage sub-module. The data acquisition sub-module communicates with the on-site energy consumption meter, collects the original energy consumption data of the energy consumption meter, and constructs the water consumption data column vector W(k), the electricity consumption data column vector E(k) and The gas consumption data column vector G(k) can be expressed as:

W(k) = [W(1,1), ..., W(1,24), W(2,1), ..., W(2,24), ..., W(N,1), ..., W (N,24)] ^T

E(k) = [E(1,1), ..., E(1,24), E(2,1), ..., E(2,24), ..., E(N,1), ..., E (N,24)] ^T

G(k) = [G(1,1), ..., G(1,24), G(2,1), ..., G(2,24), ..., G(N,1), ..., G (N,24)] ^T

In each column vector, each item represents the energy consumption in one period, for example, W(1,1) represents the water consumption in the first period on the first day, and W(N,1) represents the Nth The energy consumption of water in the first period of the day, W(N,24) represents the energy consumption of water in the 24th period of the Nth day, and N represents the number of days for collection, and each day is divided into 24 periods. Similarly, E(N,1) represents the energy consumption in the first period of the Nth day, and E(N,24) represents the energy consumption in the 24th period of the Nth day; G(N ,1) represents the energy consumption of gas in the first time period of the Nth day, and G(N,24) represents the energy consumption of gas in the 24th time period of the Nth day.

The present invention can perform protocol analysis on the data of the energy consumption meter to collect the data of the energy consumption meter with different protocols. For example, the collection perception support includes protocols such as Modbus TCP, Modbus RTU, and OPC.

The preprocessing sub-module performs data cleaning, data conversion and data processing on the water consumption data column vector W(k), electricity consumption data column vector E(k) and gas consumption data column vector G(k) respectively. renew. The preprocessed data can be sent and reported to the cloud layer, greatly reducing the upstream data bandwidth requirements.

In practical applications, if there are failures such as abnormal communication sensitivity and communication signal interference, data collection problems will occur, which will affect the accuracy of subsequent data calculations. For example, in this embodiment, the data cleaning is to analyze whether the data of each period is an abnormal value by using methods such as the Raida criterion, and if it is an abnormal value, the average value of adjacent bits is used to replace the abnormal value.

The data transformation transforms the original column vector into a sequence to be processed for energy consumption prediction calculation, namely:

W'(k) = [W(1), W(2), …, W(N)] ^T

E'(k) = [E(1), E(2), …, E(N)] ^T

G'(k) = [G(1), G(2), ..., G(N)] ^T

Here, W(1) represents the sequence of W(1,1), ..., W(1,24), and W(N) represents the sequence of W(N,1), ..., W(N,24) ;Similarly, use E(1) to represent the sequence of E(1,1), ..., E(1,24), and use E(N) to represent the sequence of E(N,1), ..., E(N,24) Sequence; use G(1) to represent the sequence of G(1,1), ..., G(1,24), and use G(N) to represent the sequence of G(N,1), ..., G(N,24).

The data update is to continuously update the input data sequence when performing online prediction of energy consumption. Specifically, assign the energy consumption values of each time period on the i+1 day to each energy consumption value of each time period on the i-th day. consumption value.

The data storage submodule is used for storing the preprocessing result of the preprocessing submodule.

In an embodiment of the present invention, the energy consumption edge computing module includes an energy consumption multi-scale analysis submodule, an energy consumption warning submodule and an energy consumption prediction submodule.

The energy consumption multi-scale analysis sub-module performs energy consumption analysis in two dimensions of space scale and time scale, wherein the space scale and time scale can be customized as required. Among them, the spatial scale is generally defined as four levels of production line, workshop, factory building and park, and the time scale is generally defined as four levels of day, month, quarter and year. For different energy media, select physical space and different time scales for analysis and comparison.

Typically, for water energy consumption, monthly and annual statistical analysis is carried out according to the plant; for gas energy consumption and electric energy consumption, monthly and quarterly statistical analysis is carried out according to the four spatial scales of production line, workshop, factory building and park.

On the basis of multi-scale analysis of energy consumption, energy consumption alarms can be carried out, and historical energy consumption alarms and real-time energy consumption alarms can be performed for energy consumption data of different spatial and time scales. On the basis of the 24-hour statistical analysis of the production line, monitor the real-time fluctuation of gas and water consumption in the production line, set the energy consumption fluctuation percentage within the time threshold, and trigger the energy consumption alarm based on the threshold.

Specifically, the energy consumption alarm sub-module sets a threshold percentage for water consumption according to a time scale such as monthly and annual water consumption, and if the water consumption exceeds the threshold compared with the previous time scale, an alarm is triggered to remind Operation and maintenance personnel pay attention to and analyze the possibility of energy consumption changes in a targeted manner, and optimize and improve auxiliary energy consumption. The typical threshold percentage is generally 10%. Obviously, the early warning of water consumption belongs to the historical energy consumption alarm.

For gas consumption and electricity consumption, set the energy consumption data collection frequency, and trigger an alarm when the energy consumption fluctuation exceeds the set threshold. Obviously, gas consumption and electricity consumption belong to real-time alarms. In the actual park energy consumption monitoring, the energy consumption data collection frequency can be set to 30 seconds, and the alarm will be triggered when the energy consumption fluctuation exceeds a certain threshold. The energy consumption fluctuation threshold is generally set 15%. Similarly, energy consumption alarm information is stored, and the reasons for analysis are recorded to provide decision-making basis for subsequent alarm analysis and optimization.

The energy consumption prediction sub-module uses the moving average method to predict the water energy consumption and gas energy consumption of each energy-consuming device on a time scale in the future, that is, the average value of the observed values of the past K moments is used as the prediction of the future moment. Use the long short-term memory network prediction model to predict the electricity consumption and energy consumption of a time scale in the future.

For example, the Long short-term memory network (Long short-term memory, LSTM) adopts a network structure of a single hidden layer, including an input layer, an LSTM layer, and an output layer. The LSTM layer has 10 LSTM units, and uses an effective Adam version of stochastic gradient descent Fit the model, using a sigmoid activation function, and optimize with a mean squared error loss function. In one embodiment, the input is the power consumption of the past 30 days, and the output is the predicted power consumption of the next day, and the training sequence is generated by using a sliding window method. If the power consumption data of 20 days is used to predict the next For the power consumption of a day, the training window size is set to 20, and the label window size is 1, which is equivalent to 1 data after every 20 data predictions, and the window moves to the right by one bit, and the next 20 data are used to predict The last 1 data, until the entire sample data is traversed. Set 60% of them, that is, 6 sequences as the training set, and the rest as the test set. Set 1 batch, iterate 5 rounds for training; perform single-step prediction on the trained model, denormalize the prediction results, and output the predicted power consumption of the next day.

In this embodiment, the long-short-term memory network prediction model can be trained by the central server using energy consumption data, that is, the model training process is executed on the cloud layer, and the model parameters are delivered to the energy consumption prediction sub-module. For example, the model parameters of the cloud layer are delivered to the edge layer through the standard MQTT (Message Queuing Telemetry Transport) protocol, and calculations and processing are performed in the edge layer to realize the online prediction function near the device.

The equipment management and control module has the functions of operation control, fault management and equipment ledger for the main energy-consuming equipment that needs to collect data at the end layer. In particular, the equipment management and control module will establish the control description of the main energy-consuming equipment, and use the corresponding composite function Block code packaging is mapped to an operation control environment that deploys Docker containers, where Docker containers are open source application container engines. The control description of the main energy-consuming equipment established here refers to the input-output mapping relationship and implementation code established through the parameter information of the main energy-consuming equipment collected by the end layer, the mechanism of the equipment, and the control constraints.

Specifically, in the sub-module of the equipment ledger, records are kept of the equipment name, model specification, purchase date, service life, depreciation period, asset number, usage status of the user department, etc., and the data can be synchronized to cloud storage. The data synchronization period of ledger information can be set freely, generally 1 year.

In the fault management sub-module, equipment fault information is stored, fault types are classified, and statistical analysis is performed. At the same time, typical fault matching is targeted for troubleshooting and solutions, and operation and maintenance personnel are guided to troubleshoot. With the occurrence, storage, classification and input of corresponding solutions for new faults, there is a complete fault knowledge base for different major energy-consuming devices. Provide fault alarm signals for the three most frequently occurring faults of different energy-consuming equipment.

The operation control sub-module provides energy efficiency control to describe the resource environment that can run at the edge layer, and is generally a virtual container engine resource, which is arranged at the edge layer.

The cloud layer mainly implements functions such as energy consumption configuration monitoring, energy consumption prediction calculation, and energy efficiency control processing. Energy consumption configuration monitoring utilizes different types of components in the cloud layer to monitor, overview, data query, and build energy flow diagrams for the energy consumption of water, electricity, gas energy and custom energy consumption equipment in the entire park and each workshop. The energy efficiency management and control decision-making process is based on the control description of the main energy-consuming equipment in the IEC61499 standard, object-oriented programming, and event-driven, forming a composite function block code.

In some embodiments of the present invention, the central server includes an energy consumption configuration monitoring submodule, an energy consumption prediction calculation submodule, and an energy efficiency control processing submodule.

The energy consumption configuration monitoring sub-module establishes an overall layout display interface for each energy consumption device in an industrial configuration mode according to the process flow, so as to monitor the energy consumption data of each energy consumption device;

The energy consumption prediction calculation sub-module uses the collected energy consumption data to train the long-short-term memory network to obtain model parameters, and sends the model parameters to the energy consumption prediction sub-module according to the update conditions, and performs Online energy consumption forecasting.

The energy efficiency control processing sub-module uses a control programming language to control and describe energy-consuming equipment, and introduces production process parameters, process constraints, and energy consumption prediction values in a future time scale, and performs calculations with the lowest energy consumption cost as the optimization goal. Get the corresponding control parameters. The device management and control module performs control on the parameters of the main energy-consuming devices at the end layer according to the optimization target.

The control description of the main energy-consuming equipment established in the cloud layer of the present invention is based on the IEC61499 standard, object-oriented programming, event-driven, and forms composite function block codes. Describe and generate composite function block codes for the control of energy-consuming devices established in the cloud layer, package and map them to the device management and control module, and control the main energy-consuming devices connected to the end layer; energy consumption data collection and energy consumption edge computing in the form of applications run.

Specifically, the energy efficiency control processing sub-module takes the energy-consuming equipment as the control object, uses the top-down object-oriented programming method conforming to IEC61499, and uses the composite function block to establish the energy-consuming equipment control description corresponding to the energy-consuming data , use the configuration function to connect each compound function block, and form a control description based on the compound function block network, which is mapped to the operation control sub-module.

Composite function blocks follow the modular design paradigm. In the IEC61499 standard, function block instances can be combined according to a certain logic to form a function block network with specific functions, and form a compound function block type that can be reused through encapsulation. The present invention can use code blocks that satisfy equipment control logic and execution relationship established by structured text, ladder diagram, continuous function flow chart language and the like. For the control description established for the main energy-consuming equipment, the control description is issued in the form of a composite function block through a wired connection, and through the mapping function of the integrated development environment software. In this embodiment, the control description of the main energy-consuming devices is constructed on the cloud layer, and the control description is delivered to the application container engine resources at the edge layer.

In some embodiments of the present invention, a park-level overall distributed control system description is established, with a single main energy-consuming device as the controlled object, including each main energy-consuming device that collects data from the end layer, based on the IEC61499 standard, using the second The class equipment model establishes the control description of the main energy-consuming equipment that collects data from the end layer. The control description of the cloud arrangement uses 4diac as the integrated development environment, and installs the Eclipse 4diac IDE software on the cloud center server. Among them, Eclipse 4diac is an open source project of IEC61499 distributed control system. It is mainly composed of two parts: development environment IDE and runtime Forte. The runtime Forte software is installed in the edge control device. The development environment IDE is usually placed on the edge server or cloud server. The advantage of this control description is that, based on the standard IEC61499 standard language, it will form the control description of each major energy-consuming device at the end layer, which can be reused later and does not rely on a single brand.

The equipment management and control module maps the control description of energy-consuming equipment established by the cloud layer and the corresponding control parameters to the operation control sub-module of the edge layer, and the energy consumption data collection and energy consumption edge computing modules run in the edge layer in the form of applications.

In the embodiment of the present invention, the edge layer includes a plurality of edge control devices, and each edge control device includes the energy data acquisition module, the energy consumption edge calculation module and the equipment management and control module, as shown in FIG. 3 . Collect water, electricity, gas and other energy consumption data at the edge layer and calculate energy consumption at the edge, and manage and control major energy-consuming devices at the edge layer.

In the prior art, the data collection is directly from the equipment to the central server, there is no edge layer in the middle, and it must go to the general service center during debugging, which is not conducive to expansion. In this embodiment, each edge control device independently includes an energy data collection module, an energy consumption edge calculation module, and a device management and control module, and its hardware is also the hardware part of each edge layer. In order to achieve effective collection, an edge control device can be arranged in each workshop or each group of energy-consuming equipment, and it is preferably arranged in a place close to the energy-consuming equipment and energy-consuming meters in the workshop. For example, it can be installed in the existing In a control cabinet or a separately designated electrical cabinet. The hardware of the cloud layer in this embodiment includes a central server, which can be a public central server or a private central server, and can generally be arranged in the central control room of the park.

In this embodiment, the devices in each layer are arranged in different positions according to the site conditions, and the layers are connected in different ways to realize communication interaction. Its system implementation includes the following steps:

Step 1: arrange the hardware of the terminal layer, edge layer and cloud layer; the terminal layer, the hardware includes the energy consumption meters and energy consumption equipment of each workshop or each group of energy consumption equipment; the edge layer is mainly a plurality of edge control device; each workshop or each group of energy-consuming equipment arranges at least one edge control device; the cloud layer, the hardware includes a central server. The grouping of energy-consuming equipment is generally based on physical distance.

Step 1.1, install various energy consumption meters on the end layer and ensure that each meter has a Modbus RTU or ModbusTCP protocol port; identify and clarify the communication interface type of the key energy consumption equipment.

Step 1.2, install the edge control device near the main energy-consuming equipment. It is recommended to install it in the electric control cabinet of the main energy-consuming equipment or install a separate cabinet next to it. One edge control device is arranged in one workshop. In particular, for equipment groups such as boilers, air-conditioning units, and air compressors, one edge control device is arranged separately.

Step 1.3, in the central server of the cloud layer, a private cloud can be arranged or a public cloud server can be rented directly, and the cloud server must have a public network IP.

Step 2: Connect and debug the hardware of the terminal layer, edge layer and cloud layer;

The communication debugging is to connect the edge control device and the hardware of the end layer through Modbus RTU communication, or use the LoRa networking mode to connect; connect the edge control device and the cloud layer through 4G or 5G connection, WiFi connection, wired connection any form of connection.

Step 3: Test the data flow of communication between layers;

The data flow test of the communication is to test separately:

The communication between the data of each energy consumption meter at the end layer and the edge layer is normal and stable;

The communication between the parameters of each energy-consuming device at the end layer and the edge layer is normal and stable;

And, the data of the energy consumption meter and the parameters of the energy consumption equipment communicate normally and stably from the edge layer to the cloud layer. Set the packet loss rate. If the packet loss rate meets the set value, the test is successful. Otherwise, it needs to be checked.

Step 4: Configure the edge layer; the configuration includes: according to the working conditions and collection requirements, configure the parameter type and collection frequency of the energy consumption equipment to be collected, and the data type and collection frequency of the energy consumption meter to be collected to said edge control device;

The above configuration work is set through the debugging tool that comes with the edge control device. After the above configuration, the data of the energy consumption meter is collected, preprocessed and stored; the equipment management and control module sends the parameters of the energy consumption equipment to the operation control sub-module for operation and sets the management fault information, and further controls the equipment Account information is processed.

Step 5: In the central server of the cloud layer, use the data uploaded from the edge layer to perform energy consumption prediction calculations, train the prediction model, and perform energy efficiency control processing on major energy-consuming devices.

Step 6: Based on the set update conditions, send the model parameters obtained by the cloud layer training to the edge layer, and send the control description and optimized control parameters processed by the cloud layer energy efficiency control to the edge layer, and at the same time, control the description in the running control Submodule processing.

Conditions for updating the processing results of energy consumption prediction calculations include: 1) Initialization, that is, the device is powered on again. 2), based on the time threshold T, which is generally set to 1 month and can be adjusted dynamically. If any of the above conditions is met, the update condition is satisfied, and the cloud layer sends the energy consumption prediction model parameters to the edge layer. Similarly, the corresponding control parameter update conditions are: 1), initialization, that is, the device is powered on again. 2), process constraints change. 3), the controlled equipment has been replaced. If any of the above conditions are met, the corresponding control parameters can be updated to the edge layer.

Further, in the embodiment of the present invention, in order to better realize equipment energy efficiency management and control, the following configuration is also performed:

First, the operation control environment is configured at the edge layer, and on this basis, the control description for the main energy-consuming equipment in the end layer is constructed and mapped to the edge control device. Specifically include the following steps:

S1. For major energy-consuming equipment such as air compressors and boiler equipment, use the IEC61499 function block model as a model to encapsulate the logical relationship into reusable function block types. The writing and implementation of logical relationship can use the traditional sequential function diagram, ladder diagram, function block diagram, instruction list and structured text, and the structured text is preferred.

S2. Connect the corresponding function block instances according to the event flow and data flow specified by the application model, and construct the application described by the complete park-level overall distributed control system in the form of function block network. At this time, the application does not contain any hardware configuration information , focusing on the functional design and verification of system control descriptions.

S3. Under the framework of the development environment IDE system mode, configure and connect the edge control device first, and then configure and run the function blocks in the control application on an edge control device through its own mapping mechanism.

In a typical embodiment, a certain brand of air compressor in the air compressor workshop is used as the controlled object, and the control description of the air compressor is written and established using structured text in the development environment IDE. On this basis, the modified, adjusted and established The control description of other air compressors is packaged into functional blocks that can be reused; according to the actual layout of the network structure of the air compressor workshop, the functional blocks corresponding to the event flow and data stream are connected to form a function block network, and the air compressor workshop is established. The overall distributed control is described and mapped to the edge control device arranged in the air compressor workshop for corresponding control operations. The establishment of control descriptions in other major workshops is similar to that of the air compressor workshop, and finally forms the overall distributed control system description at the park level. In this example, the constructed control description is stored in the central server.

Second, set fault management functions at the edge layer. Specifically include the following steps:

S1. Define the faults of major energy-consuming equipment such as air compressors and boilers, set up and classify faults in multiple dimensions, generally according to alarm frequency and impact of alarms on production.

S2. Set the fault alarm triggering conditions for the classified faults of different devices, and set the alarm information push form.

S3. According to the processing of the equipment after the fault information is pushed, save it and establish a fault knowledge base.

Finally, enter and store the device ledger information. First establish the original information of the equipment, equipment name, model specification, purchase date, service life, depreciation period, asset number, use status of the department, etc., and then dynamically update and store the equipment information. The original information of the equipment is updated every year. For equipment failure Information, once the equipment fails, the failure information and the corresponding disposal will be saved to form a historical failure database of each equipment, which is convenient for subsequent employees to maintain the equipment.

After the edge layer is processed by the energy consumption data acquisition module, the data flow enters the energy consumption edge computing module for processing, and the data is transmitted to the cloud layer through the MQTT protocol, and the computing resources of the cloud layer are used for energy consumption prediction calculation and processing, and the calculated parameters are downloaded Go to the edge layer to realize online energy consumption prediction. In one embodiment of the present invention, the calculated parameters of the energy consumption prediction of the cloud layer are sent to the edge layer edge control device through the standard MQTT protocol, so as to realize the function of online prediction at the near-device end.

Through the above scheme, the present invention adopts the terminal-edge cloud structure, and installs and arranges edge control devices near the main energy-consuming equipment in the workshop. It has three functions of data collection, energy consumption edge computing, and equipment control. It communicates with the monitoring center server and uploads the processed data to the central server in the cloud layer. At the same time, the central server uses its sufficient computing resources to predict, calculate, control and process energy consumption Send the predicted and calculated parameters, control description, etc. to the edge control device, and control the sent equipment parameters such as major energy-consuming equipment.

Claims (7)

1. The equipment energy efficiency management and control system based on edge computing adopts the architecture of terminal layer, edge layer and cloud layer. Energy-consuming equipment and energy consumption meters are arranged on the terminal layer, the cloud layer is arranged with a central server, and the edge layer includes An energy data collection module, an energy consumption edge computing module, and a device control module; the device control module includes a device ledger sub-module, a fault management sub-module, and an operation control sub-module; The energy data collection module collects the energy consumption data of the energy consumption meter; The energy consumption edge calculation module performs energy consumption edge calculation according to the energy consumption data, and sends the calculation result to the central server; the central server performs statistical analysis on the calculation result, performs energy efficiency control processing, and establishes energy The control description and corresponding control parameters of the consumption equipment are mapped to the operation control sub-module; The operation control sub-module collects the parameter information of the energy-consuming equipment, and sends it to the fault management sub-module, and transmits the fault alarm information triggered according to the fault setting conditions to the equipment account sub-module, and passes through the equipment account sub-module The summary and ledger information are transmitted to the central server together; The energy data acquisition module includes a data acquisition sub-module, a preprocessing sub-module and a data storage sub-module; the energy consumption meter includes a water meter, an electric meter and a gas meter; The data collection sub-module collects the original energy consumption data of the energy consumption meter, and constructs a column vector of water consumption energy consumption data, a column vector of electricity consumption data and a column vector of gas consumption data. In each column vector, each The item represents the energy consumption within a time period, which is divided into multiple time periods every day; The preprocessing sub-module respectively performs data cleaning, data transformation and data update on the column vectors of water energy consumption data, electricity consumption data column vectors and gas consumption data column vectors; The data storage submodule is used to store the preprocessing result of the preprocessing submodule; It is characterized in that the data cleaning is to analyze whether the data in each time period is an abnormal value, and if it is an abnormal value, the average value of adjacent bits is used to replace the abnormal value; the data transformation is to transform the original column vector into the energy consumption prediction calculation Sequence to be processed; the data is updated. When performing online energy consumption prediction, the energy consumption values of each time period on the i+1 day are assigned to the energy consumption values of each time period on the i-th day. 2. The device energy efficiency management and control system based on edge computing according to claim 1, wherein the energy consumption edge computing module includes an energy consumption multi-scale analysis submodule, an energy consumption alarm submodule and an energy consumption prediction submodule; The energy consumption multi-scale analysis sub-module performs energy consumption analysis in two dimensions of spatial scale and time scale; The energy consumption alarm sub-module sets a threshold percentage for water consumption according to a time scale, and if the water consumption exceeds the threshold compared with the previous time scale, an alarm is triggered; for gas consumption and electricity consumption , set the energy consumption data collection frequency, when the energy consumption fluctuation exceeds the set threshold, an alarm will be triggered; The energy consumption prediction sub-module uses the moving average method to predict the water energy consumption and gas energy consumption of each energy-consuming equipment in a future time scale, and uses the long-short-term memory network prediction model to predict the electricity consumption in a future time scale. 3. The device energy efficiency management and control system based on edge computing according to claim 2, wherein the central server uses energy consumption data training to obtain the long-short-term memory network prediction model, and sends model parameters to the Energy consumption prediction sub-module. 4. The device energy efficiency management and control system based on edge computing according to claim 2 or 3, wherein the central server includes an energy consumption configuration monitoring submodule, an energy consumption prediction calculation submodule, and an energy efficiency control processing submodule; The energy consumption configuration monitoring sub-module establishes an overall layout display interface for each energy consumption equipment in an industrial configuration mode according to the process flow, so as to monitor the energy consumption data of each energy consumption equipment; The energy consumption prediction calculation sub-module uses the collected energy consumption data to train the long-short-term memory network to obtain model parameters, and sends the model parameters to the energy consumption prediction sub-module according to the update conditions, and performs Online energy consumption prediction; The energy efficiency control processing sub-module uses a control programming language to control and describe energy-consuming equipment, and introduces production process parameters, process constraints, and energy consumption prediction values in a future time scale, and performs calculations with the lowest energy consumption cost as the optimization goal. Get the corresponding control parameters. 5. The device energy efficiency management and control system based on edge computing according to claim 4, characterized in that, the energy efficiency control processing sub-module takes energy-consuming devices as control objects, uses a top-down object-oriented programming method, and utilizes The compound function block establishes the energy-consuming device control description corresponding to the energy consumption data, connects each compound function block by using the configuration function, forms a control description based on the compound function block network, and maps it to the operation control sub-module. 6. The equipment energy efficiency management and control system based on edge computing according to claim 1, wherein the control description of the energy-consuming equipment is the parameter information of the energy-consuming equipment and the mechanism of the energy-consuming equipment collected through the terminal layer, The mapping relationship between input and output established by control constraints and the implementation code; the operation control sub-module is a virtual container engine resource, which provides the resource environment for the control description to run at the edge layer. 7. The device energy efficiency management and control system based on edge computing according to claim 1, wherein the edge layer includes a plurality of edge control devices, and each edge control device includes the energy data acquisition module, energy consumption edge Computing module and device management and control module.

Images