An automatic measurement and control intelligent box-type transformer and its measurement and control system

Abstract

This invention relates to an intelligent prefabricated substation and its control system for automatic measurement and control. The intelligent prefabricated substation includes at least several sensors for collecting equipment parameters and a control module for controlling the operating status of the equipment. It also includes at least one IoT agent terminal. The IoT agent terminal determines abnormal conditions of the equipment based on the equipment parameters sent by the sensors and feeds back at least one corresponding control message to the control module based on the abnormal condition. In response to the control message, the control module sends control commands to the equipment related to the control message. This invention uses an IoT agent terminal in the intelligent prefabricated substation to perform edge computing processing on the data, enabling the IoT agent terminal to quickly handle abnormal conditions in the intelligent prefabricated substation and improving the efficiency of measurement and control.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent box-type transformers, and in particular to an automatic measurement and control intelligent box-type transformer and a measurement and control system thereof. Background Art

[0002] Box-type substation, also known as prefabricated substation or prefabricated substation. It is a kind of factory-prefabricated indoor and outdoor compact distribution equipment that is integrated with high-voltage switchgear, distribution transformer and low-voltage distribution device according to a certain wiring scheme. It is particularly suitable for urban network construction and transformation. It is a new type of substation that has emerged after civil substation. China's urban distribution network is constantly updated and transformed, and box-type substations are widely used.

[0003] For example, patent document CN106054704A discloses an intelligent box-type transformer measurement and control system with integrated communication management and optical fiber ring network communication, including: a central processing unit, a power module that transmits power to the central processing unit, a 485 or Ethernet communication module that is interactively connected to the central processing unit, an optical fiber ring network switch that transmits information to the central processing unit and can form a self-healing optical fiber ring network, an input signal and an output signal connected to the central processing unit, an indicator light component connected to the central processing unit, and a human-computer interaction module that interacts with the central processing unit. The present invention provides an intelligent box-type transformer measurement and control system with integrated communication management and optical fiber ring network communication, which can realize the functions of analog quantity collection of all information on the high and low voltage sides of the box-type transformer, input signal collection, local remote control, and event information upload; at the same time, it has a communication management function, and can communicate with the downstream equipment in real time to realize remote management and automated monitoring of the booster station.

[0004] For example, patent document CN110676701A discloses a multi-core intelligent box-type transformer measurement and control integrated device, including an upper cover, a bottom shell and a measurement and control system, the surface of the upper cover is fixedly connected with a display screen, an operation button and an indicator light, the display screen is fixedly connected at the center of the upper cover surface, the operation button is located below the display screen, the two sides of the display screen are respectively fixedly connected with indicator lights, the lower surface of the upper cover is fixedly connected with the bottom shell, the measurement and control system includes a microprocessor, a network module, a serial port module, a storage module, an I-stage AC signal acquisition module, a II-stage AC signal acquisition module, a protection module, a display module, an input module, an output module and a power supply module, and the lower surface of the bottom shell is fixedly connected with a connection interface, a USB interface, a network interface and a protection gate. The present invention adopts an integrated structural design, concentrates a variety of protection measurement and control, remote communication and local control functions, enhances the stability of the measurement and control device during operation, and reduces the difficulty of maintenance and operation.

[0005] For example, patent document CN110350490A discloses an intelligent integrated protection, measurement and control device for box-type transformers. By setting three-stage time-limited overcurrent protection for the transformer and setting the overcurrent time constant, overcurrent protection is performed on electrical equipment. After a grounding fault occurs in the current grounding, a zero-sequence time-limited protection module will cause a zero-sequence overcurrent, which is controlled by the main control module to be put in and out, and an alarm or tripping is selected. The low-voltage protection module collects three voltage values ​​through the low-voltage measurement unit. The current measurement unit monitors the current of the three lines and compares them with the low-voltage constant. When the three line voltages are all lower than the constant and there is current in any phase, the low-voltage protection action is turned on. The overvoltage protection module monitors the three lines. When the voltage between any phase is greater than the set voltage, the overvoltage protection is turned on. The normal operation of the box-type transformer is guaranteed by remote protection and measurement and control of the transformer current and voltage.

[0006] The above existing technologies only focus on the protection of current, voltage, information transmission and other aspects. In the existing technology, the life of the smart box transformer cannot be predicted, which requires a large amount of non-essential annual inspection and testing work, resulting in the shutdown of the smart box transformer. Not only that, the real-time performance in the existing technology, such as millisecond-level response, is difficult to achieve in the cloud. The current solution uses the transmission speed of light, which is already a limit. Even if it does not pass through some routers, switches, and network equipment calculations in the middle, and is only transmitted in the optical fiber, if the data center is not in the city, the round-trip transmission time may reach tens of milliseconds. When there are some scenarios that need to be responded within a hundred or even tens of milliseconds, there is no way to implement it in the cloud, and it needs to be implemented on the edge.

[0007] There are many sensors in the box-type transformer, which will generate a large amount of data. The backend server has a lot of data loss, and many frames are lost. Because the network is not stable enough and the bandwidth does not support the transmission of all information to the backend server, there are also network security issues and local response issues. Through edge computing of data, various devices can be reversed to restore the box-type transformer to the best operating state.

[0008] Therefore, the present invention relates to the research and development of intelligent box-type transformer edge Internet of Things gateway: the monitoring terminal is committed to solving the detection technology of box-type transformer and distribution network electrical equipment, and solves the problem of large uplink and downlink data volume and master station data volume in communication between the master station and the terminal through on-site monitoring, on-site processing and equipment preventive maintenance. At the same time, it considers whether it is possible to realize remote control of wireless private network through edge computing technology combined with 4G or 5G (currently the State Grid does not allow the use of wireless network to realize remote control), replacing existing data acquisition terminals such as TTU, FTU, DTU and RTU.

[0009] Not only that, how to quickly improve the processing efficiency of edge computing nodes is still a technical problem that has not been solved by existing technologies.

[0010] In addition, on the one hand, there are differences in understanding among those skilled in the art; on the other hand, the inventor studied a large number of documents and patents when making the present invention, but due to space limitations, not all details and contents are listed in detail. However, this does not mean that the present invention does not have the characteristics of these prior arts. On the contrary, the present invention already has all the characteristics of the prior art, and the applicant reserves the right to add relevant prior art to the background technology. Summary of the invention

[0011] In view of the deficiencies in the prior art, the present invention provides an intelligent box transformer with automatic measurement and control, which includes at least a number of sensors for collecting equipment parameters and a control module for controlling the operating status of the equipment, and also includes at least one Internet of Things agent terminal. The Internet of Things agent terminal determines the abnormal situation of the equipment based on the equipment parameters sent by the sensors, and feeds back corresponding at least one control information to the control module based on the abnormal situation. In response to the control information, the control module sends a control instruction to the equipment related to the control information.

[0012] Preferably, the IoT agent terminal evaluates the operating status information of the smart box transformer based on the device parameters, and sends the operating status information to at least one first edge agent node or at least one second edge agent node, and the first edge agent node or the second edge agent node feeds back at least one regulation suggestion to the IoT agent terminal based on the operating status information.

[0013] Preferably, the IoT proxy terminal can calculate necessary data according to data transmitted by the sensor, and when the data value of the necessary data exceeds a threshold, the IoT proxy terminal sends corresponding preset decision information to the control module.

[0014] Preferably, the IoT proxy terminal directly establishes a communication connection with the cloud server system, or the IoT proxy terminal establishes a communication connection with the cloud server through a first edge proxy node and / or at least one second edge proxy node, and in response to request information from the first edge proxy node or at least one second edge proxy node, the IoT proxy terminal sends data information corresponding to the request information to the first edge proxy node or at least one second edge proxy node.

[0015] Preferably, the geographical location of the first edge proxy node is close to the location of the cloud server system, the geographical location of the second edge proxy node is close to the geographical location range of the smart box transformer, the number of the first edge proxy nodes is less than the number of the second edge proxy nodes, and the number of the second edge proxy nodes is less than the number of the smart box transformers.

[0016] Preferably, the cloud server system comprehensively describes the operating status information of the smart box transformer based on the data association of power statistics information, loss analysis information and / or user equipment basic ledger information, thereby feeding back adjustment information to the IoT agent terminal based on the specific power problems determined by the analysis.

[0017] The present invention also provides a measurement and control system for an intelligent box transformer with automatic measurement and control, which at least includes an intelligent box transformer and a cloud server system. At least one Internet of Things agent terminal is arranged in the intelligent box transformer and / or near a geographical location. The Internet of Things agent terminal determines an abnormal condition of the equipment based on equipment parameters sent by several sensors in the intelligent box transformer, and feeds back corresponding at least one control information to the control module based on the abnormal condition. In response to the control information, the control module sends a control instruction to the equipment related to the control information. The Internet of Things agent terminal sends an alarm information to the cloud server system by direct communication and / or indirect communication based on the abnormal condition.

[0018] Preferably, the system also includes at least one first edge proxy node and / or at least one second edge proxy node, the geographical location of the first edge proxy node is close to the location of the cloud server system and its data task processing capability is better than the data task processing capability of the second edge proxy node, the geographical location of the second edge proxy node is close to the geographical location range of the smart box transformer, the number of the first edge proxy nodes is less than the number of the second edge proxy nodes, and the number of the second edge proxy nodes is less than the number of the smart box transformers.

[0019] Preferably, the IoT proxy terminal evaluates and selects at least one first edge proxy node and / or at least one second edge proxy node with a shorter predicted data processing time to establish a communication connection based on the physical distance parameter and the received data performance test parameter, and the IoT proxy terminal sends part of the data task to the first edge proxy node and / or the second edge proxy node. Preferably, the first edge proxy node and / or the second edge proxy node processes the data information based on the request information of the IoT proxy terminal, and / or

[0020] The first edge proxy node and / or the second edge proxy node sends alarm information to the cloud service system based on the data anomaly information.

[0021] Beneficial technical effects of the present invention:

[0022] (1) By setting up an IoT proxy terminal that can provide feedback on abnormal conditions of the smart box transformer, the response speed of the measurement and control system to the alarm of the smart box transformer can be improved;

[0023] (2) By setting up edge proxy nodes with different processing capabilities in a hierarchical manner to undertake the data processing tasks of the IoT proxy terminals, the pressure of data processing can be further reduced, and abnormal situations cannot be sent to the cloud service system as soon as possible.

[0024] (3) The cloud service system can associate the overall information of the smart box-type transformer with the national power parameter data, non-power parameter data and / or user equipment basic ledger information to comprehensively describe the operating status information of the smart box-type transformer, so that the measurement and control status of the smart box-type transformer can be more specific, which is more conducive to timely detection of measurement and control abnormalities in the smart box-type transformer. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic diagram of the logic module of the edge computing node of the present invention. DETAILED DESCRIPTION

[0026] The following is a detailed description with reference to the accompanying drawings.

[0027] Edge computing refers to the use of an open platform that integrates network, computing, storage, and application core capabilities to provide the nearest service on the side close to the source of objects or data. Its applications are initiated on the edge side, resulting in faster network service responses, meeting the basic needs of the industry in terms of real-time business, application intelligence, security and privacy protection, etc. Edge computing is between physical entities and industrial connections, or on top of physical entities. Cloud computing can still access historical data of edge computing.

[0028] Different from the limitations of RTU / measurement and control sampling frequency and transmission mode, SCADA information is generally refreshed every 4 to 5 seconds, that is, the SCADA system represents the system status of the power grid several seconds ago. The edge computing in the present invention refers to the calculation program completed by the edge area close to the data source. Edge computing is a distributed computing mode that inputs real-time small-scale data and outputs judgment result information for regulating electrical equipment to a stable state. Edge computing sinks most of the computing process at the edge close to the device, and a small part of the computing process is performed through the distant cloud, which can enhance the self-healing ability of the power system when the communication link is interrupted by force majeure.

[0029] like Figure 1As shown, the edge computing network is provided with several computing proxy nodes. The intelligent box transformer can process and receive feedback information through the proximal computing proxy node to achieve faster network response. The cloud server of the edge computing center can coordinate and specify control strategies for several intelligent box transformers. However, in the prior art, an intelligent box transformer basically communicates with the nearby computing proxy nodes in a fixed manner. The operating state of the intelligent box transformer is changing, and some intelligent box transformers need to process less tasks or even shut down. The amount of tasks that some intelligent box transformers need to process increases, and the processing speed of the computing proxy nodes connected to them remains unchanged, which will lead to delayed feedback of the processing results of some processing tasks. Similarly, some processing computing nodes are busy with tasks, and some processing computing nodes are idle, which makes it impossible to maximize the efficiency of processing computing nodes in edge computing. Alternatively, an intelligent box transformer is directly connected to the cloud server through the communication network for task processing, which makes the cloud server or cloud server system busy with task processing, resulting in delayed feedback of task processing results, which is not conducive to the timely processing of abnormalities of the intelligent box transformer. Therefore, how to improve the task processing efficiency of the intelligent box transformer with automatic measurement and control and its measurement and control system during edge computing is a technical problem that has not been solved.

[0030] The present invention provides an intelligent box-type transformer with automatic measurement and control and its measurement and control system, which can also be called an intelligent box-type transformer data processing method, system and device based on edge computing.

[0031] like Figure 1 As shown, several smart box transformers 40 are distributed in different geographical locations. At least one IoT proxy terminal for receiving and sending data is provided in the smart box transformer 40. Several edge proxy nodes close to the edge geographical locations of the smart box transformer are provided between the IoT proxy terminal and the cloud server 10. The edge proxy nodes can be provided with exactly the same processing modules, so that the task processing from each edge geographical node is equivalent. The edge processing nodes can be set to processing modules of different levels, so that the types and levels of tasks that the edge proxy nodes can process are different, so that some edge proxy nodes can process urgent data tasks, and some edge proxy nodes are capable of processing non-urgent data tasks.

[0032] The edge agent nodes of the present invention include at least two types: a first edge agent node 20 and a second edge agent node 30 .

[0033] The data capacity of the first edge proxy node 20 is close to that of the cloud server, and can handle a large number of urgent data processing tasks and a large number of non-urgent data processing tasks. The first edge proxy node can establish a communication connection with the cloud service system through a communication network. Preferably, the geographical location distribution of the first edge proxy nodes is close to that of the cloud server system, and the number of the first edge proxy nodes is in the hundreds.

[0034] The data capacity of the second edge proxy node 30 is less than that of the first edge proxy node 20, and it can handle common, large-scale urgent data processing tasks. The second edge proxy node can establish a communication connection with the cloud service system through a communication network. Preferably, the geographical location distribution of the second edge proxy node is close to the geographical location of the smart box transformer. The number of the second edge proxy nodes is in the thousands or even tens of thousands. That is, the number of the second edge proxy nodes is ten times or even a hundred times the number of the first proxy nodes.

[0035] In the case of a non-urgent data task that cannot be processed received by the second edge proxy node, the second edge proxy node forwards the data task to the first edge proxy node or the cloud server system. The first edge proxy node or the cloud server system processes the data task and feeds back the processing result to the IoT proxy terminal of the corresponding smart box transformer.

[0036] Preferably, when the data tasks are not busy, the second edge proxy node can receive urgent data tasks from the IoT proxy terminal and / or the first edge proxy node.

[0037] Figure 1 The first edge proxy node and the second edge proxy node in the figure only indicate the geographical distribution trend, and do not represent the actual setting location of the edge proxy node. In the actual setting environment, the smart box transformer can be set at any location where humans live. Therefore, there may be several smart box transformers distributed around the cloud server system 10, the first edge proxy node 20, and the second edge proxy node 30. The first edge proxy node 20 and the second edge proxy node 30 may receive data task requests from the IoT proxy terminal of the adjacent smart box transformer to process some data tasks as quickly as possible.

[0038] The automatic measurement and control intelligent box transformer of the present invention at least includes a box transformer body, a plurality of sensors, a plurality of IoT proxy terminals and a cloud server. A plurality of sensors establish information connection with at least one IoT proxy terminal. At least one IoT proxy terminal establishes information connection with the cloud server through a plurality of edge proxy nodes. The IoT proxy terminal is arranged at the edge of the data source, and judges the stable state of at least one electrical device based on the data input by the sensor. The IoT proxy terminal sends the stable state information of at least one electrical device to a distant cloud server. Preferably, the IoT proxy terminal can send control information to the corresponding electrical device based on the stable state of the electrical device. The decision information includes control information and alarm information for the electrical device. The control information includes adjustment control information, start-stop control information, and the like. For example, the control information is: automatic gear adjustment information of the transformer, automatic control information of the reactive power compensation device, automatic control information of the air conditioner, automatic control information of the industrial dehumidifier, automatic start-stop information of the box transformer exhaust fan, and the like. The alarm information is, for example, the triggering over-limit alarm information when the power data is abnormal.

[0039] For example, the temperature sensor transmits the uploaded temperature data to the IoT proxy terminal. After the IoT proxy terminal determines that the temperature has reached the alarm threshold, the IoT proxy terminal sends a startup message to the air conditioner to achieve the reduction. At the same time, the IoT proxy terminal sends abnormal information, control information and alarm information to the adjacent first edge proxy node and / or second edge proxy node. The first edge proxy node and / or the second edge proxy node or even the cloud server system evaluates whether the control information of the IoT proxy terminal is comprehensive based on the abnormal information, control information and alarm information. In the case that the control information fed back by the IoT proxy terminal is incomplete, the first edge proxy node and / or the second edge proxy node or even the cloud server system can continue to send supplementary control information to the IoT proxy terminal, so that the IoT proxy terminal can adjust the equipment operation status in the smart box transformer by supplementary means, so that the abnormal operation in the smart box transformer disappears quickly and the operation efficiency is improved.

[0040] The IoT proxy terminal includes one or more of a dedicated integrated chip, a processor, a server, and a server group.

[0041] The first edge proxy node and / or the second edge proxy node may be one or more of a dedicated integrated chip, a processor, a server, and a server group.

[0042] The data collected by several sensors for the transformer include at least: ambient temperature and humidity, noise, oil sample monitoring, insulating oil medium, initial value difference of winding short-circuit impedance, C2H2 content and aldehyde content.

[0043] The data collected by the sensors on the line at least include the temperature of the hardware and the sag of the line.

[0044] The data collected by several sensors for high and low voltage cabinets include at least: high and low voltage cabinet temperature, circuit breaker mechanical characteristics, opening and closing circuit resistance, SF6 micro-water content, insulation resistance to ground, service life and opening DC resistance.

[0045] The data collected by several sensors for the comprehensive protection device include at least: comprehensive protection device temperature, AD sampling consistency, secondary circuit integrity, communication interface status detection, insulation resistance, equipment temperature, message delay and sampling message packet loss rate.

[0046] The data collected by the sensors for the reactive power compensation device at least include: compensation device temperature and compensation efficiency.

[0047] The data collected by several sensors for power cables include at least: cable temperature, partial discharge parameters and insulation performance.

[0048] The data collected by the sensors for the environment include at least: ambient humidity and ambient temperature.

[0049] Preferably, the cloud server system trains and predicts the service life of electrical equipment based on AI self-learning and / or neural network algorithms.

[0050] Preferably, the sensor uploads all data to the IoT proxy terminal according to a preset program. The IoT proxy terminal uploads the processed necessary data to the cloud server system based on the edge computing method. The monitoring personnel can view the necessary data that is meaningful to the operation and maintenance of the smart box transformer through the cloud server system, which is conducive to the monitoring personnel to monitor and understand the work of the smart box transformer.

[0051] The IoT agent terminal can identify the necessary data of electrical equipment on-site based on edge computing methods, so that it can send decision information to related electrical equipment in a timely manner, making the abnormal operating status in the smart box transformer disappear quickly.

[0052] Some data types and decision information for on-site judgment by the IoT agent terminal are shown in Table 1. The over-limit refers to the situation where the value exceeds the threshold.

[0053] The IoT proxy terminal can calculate the necessary data based on the data transmitted by the sensor, and when the data value of the necessary data exceeds the threshold, the IoT proxy terminal will send the corresponding preset decision information to the corresponding electrical equipment. With this setting, the data transmission distance between the IoT proxy terminal and the sensor is short, which can reduce data loss during the transmission process. The IoT proxy terminal pre-screens and calculates the necessary data, and sends the corresponding decision information in time when abnormal data is found. There is no need to wait for the cloud server system to comprehensively analyze and obtain the decision information, so it can save time to obtain some decision information and improve the adjustment efficiency of the smart box transformer.

[0054] Preferably, the IoT agent terminal can send the necessary data obtained to the first edge agent node and / or the second edge agent node, and the first edge agent node and / or the second edge agent node analyze the operating status of each device in the smart box transformer.

[0055] There are several smart box transformers and several IoT proxy terminals distributed in a city. An IoT proxy terminal may receive the equipment parameters of one smart box transformer alone, or may receive the equipment parameters of multiple smart box transformers. There may be multiple second edge proxy nodes or first edge proxy nodes within the communication range of an IoT proxy terminal. In the process of data processing, some edge proxy nodes are bound to be busy and some edge proxy nodes are idle. Due to the geographical distance between the IoT proxy terminal and the first edge proxy node and / or the second edge proxy node, the IoT proxy terminal needs to evaluate and select the first edge proxy node and / or the second edge proxy node that can quickly feedback the results, so as to be able to timely measure and control each device in the smart box transformer.

[0056] The IoT proxy terminal evaluates and selects at least one first edge proxy node and / or at least one second edge proxy node with a shorter predicted data processing time to establish a communication connection based on the physical distance parameter and the received data performance test parameter. The IoT proxy terminal sends part of the data task to the first edge proxy node and / or the second edge proxy node.

[0057] The IoT proxy terminal sends a performance test allocation request to at least one first edge proxy node and / or at least one second edge proxy node within the communication range. In response to the received performance test allocation request, the first edge proxy node and / or at least one second edge proxy node object to the current performance test result to the IoT proxy terminal.

[0058] The IoT proxy terminal predicts the information feedback time and / or information feedback speed of the first edge proxy node and / or at least one second edge proxy node based on the physical distance information and the performance test results, and selects the first edge proxy node and / or at least one second edge proxy node with a shorter predicted information feedback time to send the data task.

[0059] Preferably, the performance test request and the performance test result contain timestamp information, so that the IoT proxy terminal can evaluate the time required for physical transmission of information and whether the information transmission of the connected communication network is congested.

[0060] Preferably, the cloud service system, the first edge proxy node, the second edge proxy node and the IoT proxy terminal store a location index table containing geographic location information of all first edge proxy nodes and second edge proxy nodes.

[0061] Preferably, the first edge proxy node can periodically send non-urgent data results to the cloud server system, so that the cloud server system can analyze the operation status of the smart box transformers within the regional range.

[0062] When the IoT agent terminal predicts the information feedback time and / or information feedback speed of the first edge agent node and / or at least one second edge agent node based on the physical distance information and performance test results, part of the prediction time will undoubtedly be consumed, thereby causing a delay in the decision to select the first edge agent node and / or at least one second edge agent node.

[0063] In order to reduce the delay of the IoT proxy terminal in selecting the first edge proxy node and / or at least one second edge proxy node, that is, to improve the selection efficiency of the IoT proxy terminal in selecting the first edge proxy node and / or at least one second edge proxy node, the processor in the IoT proxy terminal is configured as follows:

[0064] Automatically optimize based on physical distance information and performance test results to make a decision. The decision is to determine one or two first edge proxy nodes and / or at least one second edge proxy node.

[0065] Preferably, the processor in the IoT proxy terminal is configured to generate an optimization decision for selecting the first edge proxy node and / or at least one second edge proxy node through a reinforcement learning algorithm to achieve self-optimization. The basic process of reinforcement learning is a Markov decision process. The Markov decision process can be represented by a four-tuple {s, m, p, r} consisting of state s, selection object m, state transition probability p, and state transition reward or return r. For a discrete-time Markov decision process, the set of states and selection objects is called the state space S and the selection space A. Specifically represented as state s i ∈S,a i ∈A. According to the action selected in step t, the state is determined by probability P(s t+1 ,s t ,a t ) from s t Transfer to t+1 At the same time as the state transfer, the decision-making subject receives an immediate reward R(s t+1 ,s t ,a t ). In the above expression, s t Represents the state at time t. t It is represented as the selected object at time t. The accumulated reward at the end of the above process is:

[0066] G t =R t +γR t+1 +γ 2 R t+2 +…+γ k R t+k =∑ k＝0 γ k R t+k

[0067] R in the formula t is the accumulated reward within time t. γ is the discount factor, which ranges from 0 to 1. The discount factor is used to reduce the reward weight corresponding to the long-term decision. The ultimate goal of the decision is to maximize the accumulated reward while reaching the target state.

[0068] The processor in the IoT proxy terminal performs training and learning based on the amount of data to be processed, physical distance information and performance test results, as well as the information feedback time and / or information feedback speed of the first edge proxy node and / or at least one second edge proxy node, and makes optimization decisions with the shortest information feedback time and / or the maximum information feedback speed value as the optimization goal.

[0069] The processor in the IoT proxy terminal constructs a state space S based on the amount of data to be processed collected by the sensor. The processor constructs a selection space A based on the optimization decision it makes, and selects physical distance information and performance test results as the state space S. Preferably, attributes such as ambient temperature that affect transmission efficiency can also be included in the state space S.

[0070] Since there are n first edge proxy nodes and / or second edge proxy nodes, the mathematical representation of the state space S is a 2×n dimensional matrix. The matrix includes an n-dimensional column vector of physical distance information and an n-dimensional column vector of performance test results. Preferably, the n-dimensional column vector of physical distance information is o1, o2, …, o n The n-dimensional column vector of the performance test results is q1, q2, …, q n .

[0071] The processor in the IoT proxy terminal determines that the selected selection space variable is the number and the preferred order of the first edge proxy node and / or the second edge proxy node.

[0072] Since there are n communicable first edge proxy nodes and / or second edge proxy nodes, the selection state of the first edge proxy node and / or the second edge proxy node is m. i , the preferred order is a i It should be noted that the setting of the selection space A needs to meet the requirements of completeness and legality.

[0073] Preferably, completeness is the possibility that the selected space A can provide for achieving the expected goal. For example, by selecting a first edge proxy node to complete the information feedback time goal

[0074] Preferably, the selection space also meets the legality. In practical applications, not all executed selections are valid in any state. For example, selecting the first edge proxy node and / or the second edge proxy node in a fault state or a closed state is an invalid selection.

[0075] Based on the above reasons, the decision made by the processor in the IoT agent terminal is a low-dimensional output of an integer, which can make effective choices simply and efficiently, thereby shortening the information feedback time.

[0076] Preferably, the selection state c of the first edge proxy node and / or the second edge proxy node i The characteristic quantity is:

[0077]

[0078] c iThe value of is 1, 0, -1. When the processor analyzes and obtains a conclusion of 1, 0, -1 for a first edge proxy node and / or a second edge proxy node, it indicates the next step of selection, irrelevant, or non-selection.

[0079] The processor in the IoT proxy terminal selects the first edge proxy node and / or the second edge proxy node whose predicted feedback time is greater than the actual feedback time for priority selection based on the historical predicted feedback time and the actual feedback time. The more times the predicted feedback time is greater than the actual feedback time, the higher the priority.

[0080] Preferably, the processor in the IoT proxy terminal constructs a function regarding the reward based on the actual feedback time after the selection result of the first edge proxy node and / or the second edge proxy node.

[0081] The maximum value of the reward function should be equivalent to the optimization target. Preferably, the reward function is a function of the selected object m executed by the IoT agent terminal and the environmental state s collected by the sensor. The reward function R(m,s) is as follows:

[0082]

[0083] The processor in the IoT proxy terminal generates at least one selection information in the optimization decision by determining the state transition probability according to the following steps:

[0084] In the learning round, the sensor module explores the environment based on the Boltzmann distribution strategy;

[0085] After the learning round is over, the selection object is determined based on the ε-greedy strategy.

[0086] The Boltzmann distribution strategy is as follows:

[0087]

[0088] p(m t |s t ) is the state transition probability. Specifically, p(m t |s t ) indicates that the IoT agent terminal is in state s t Use selection object m t probability. is the value function of the chosen state. is the state t You can choose A subset of . Belongs to a subset of the selection space A. According to the previous state s t-1 The corresponding choice is m t-1Determine. The Boltzmann distribution strategy is that the choice with a large value function has a high probability of being selected. The choice with a small value function has a low probability of being selected. In practical applications, it is not practical to apply any series of input choices to the system, so the derivation of input values ​​should be limited. Therefore, the state s t The selected object should be located in the selection m t-1 Nearby, by collection indicates, rather than from Here, the temperature coefficient τ (τ>0) is used to control the exploration / exploitation characteristics. High temperature leads to more exploration and more random selection. Low temperature leads to more exploitation and more greedy selection. In other words, a low temperature coefficient will cause the agent to choose the option with the highest value or the highest corresponding reward. In addition, during each round of learning, the temperature coefficient will be updated according to the following update rule.

[0089]

[0090] Where ζ is the learning rate. The value range of ζ is between 0 and 1. ζ determines the rate at which the IoT agent terminal goes from exploring the environment to acquiring experience knowledge for utilization.

[0091] At the end of learning, the strategy is completely greedy. In each state, the IoT agent terminal will choose the object with the largest value function value. The value of the learning rate ζ should allow the IoT agent terminal to have enough exploration / utilization time.

[0092] Preferably, the ε-greedy strategy is:

[0093]

[0094] In the formula, ε is the greed factor. The value range of ε is between 0 and 1. The greed factor represents the probability that the IoT agent terminal randomly selects a selection object to obtain a reward. argmaxQ(s,m) represents the selection object with the largest value function under the environment state s.

[0095] Preferably, the IoT proxy terminal or the processor therein is configured to make optimization decisions in the following manner:

[0096] Construct a value function about the environment state and the choice, and record different environment states and choice objects to construct a value table. The value represents a discrete record of the value function. Preferably, the value function is a quadratic function group about the data processing volume and information feedback time. For example, the data processing volume is -l(xm) 2 +n. The settings of the three coefficients l, m, and n at least satisfy the requirement that the data processing volume is positive in half of the data processing cycle.

[0097] Preferably, when the value function converges and the optimization decision of the IoT proxy terminal does not make the environmental state reach the optimization target, or when the value function converges and the system is not damaged, the IoT proxy terminal is configured to obtain the first selection object under the corresponding environmental state based on the ε-greedy greedy strategy. Preferably, the first selection object is obtained based on the ε-greedy greedy strategy. Preferably, the first selection object is randomly selected. The ε-greedy greedy strategy enables the IoT proxy terminal to select the selection object corresponding to the maximum value of the value function in the later stage of learning and training, but there is still a certain probability that ε randomly selects a selection object to obtain a reward.

[0098] Preferably, the cloud server system also establishes information connection with the power statistics module, the loss analysis module and the user equipment basic ledger information system.

[0099] The cloud server system associates the data of power statistics, loss analysis and user equipment basic ledger information system to comprehensively describe the status of a specific site or specific equipment, thereby analyzing specific power problems.

[0100] Preferably, when the cloud server system determines that the loss value of a specific site or specific equipment exceeds the limit, it determines that the specific site or specific equipment has an abnormal loss point. The cloud server system eliminates the abnormal loss to achieve energy saving and consumption reduction.

[0101] The present invention is described by taking a method for eliminating abnormal loss points of a transformer as an example.

[0102] The cloud server system calculates the TOC values ​​of various candidate transformers based on the transformer comprehensive energy efficiency cost method.

[0103] The expression for calculating the TOC value of various candidate transformers is:

[0104] TOC=CI+(A×P O +B×P K ) / 1000

[0105] Among them, TOC represents the comprehensive energy efficiency cost of the transformer (yuan), CI represents the initial cost of the transformer equipment (yuan) and the purchase price of the transformer; P O Indicates the rated no-load loss of the transformer (W), P K represents the rated load loss of the transformer (W), A represents the equivalent initial cost of the unit no-load loss of the transformer (yuan / KW), and B represents the equivalent initial cost of the unit load loss of the transformer (yuan / KW).

[0106] Without considering the additional loss costs of the power supply network, the average annual incremental costs, and assuming that the load rate of the transformer remains unchanged during the emergency trial period, the expression of the transformer comprehensive energy efficiency cost TOC value is simplified to:

[0107]

[0108] Among them, K pv ={1-1 / (1+i) n} / i.

[0109] E represents the average electricity price of transformer users (yuan / kWh); n represents the economic service life of the transformer, which is generally 20 years; K pv H represents the present value coefficient of the expenses for consecutive n years with a discount rate of i; i represents the annual discount rate, which is not lower than the bank loan interest rate of the same period; py Indicates the annual energized hours of the transformer, usually 8760h; τ indicates the annual maximum load loss hours (h); β o Indicates the initial load factor of the transformer, per unit value.

[0110] When making a decision to update an existing old transformer, the TOC value of the existing transformer that can continue to operate is compared with the TOC value of the new transformer to be updated within the same period. If the TOC value of the existing transformer is smaller, it should not be replaced. Otherwise, the transformer should be replaced as soon as possible.

[0111] The initial cost of the existing transformer is: CI = V O0 -V OT / (1+i) T .

[0112] Among them, V O0 V represents the saleable price of the existing transformer in the decision year, that is, the current value of the transformer; OT It represents the residual value of the existing transformer at the end of the Tth year, that is, the net residual value, which is equivalent to the scrap recycling price (appropriate). T represents the number of years that the existing transformer can continue to operate.

[0113] Table 1 Decision examples of IoT agent terminals

[0114]

[0115]

[0116] Table 2 Example of electricity collection statistics and display of smart meters

[0117]

[0118]

[0119] It should be noted that the above specific embodiments are exemplary, and those skilled in the art can come up with various solutions inspired by the disclosure of the present invention, and these solutions also belong to the disclosure scope of the present invention and fall within the protection scope of the present invention. Those skilled in the art should understand that the present invention description and its drawings are illustrative and do not constitute a limitation of the claims. The protection scope of the present invention is defined by the claims and their equivalents.

[0120] The present invention specification contains multiple inventive concepts, and the applicant reserves the right to file a divisional application based on each inventive concept. The present invention specification contains multiple inventive concepts, such as "preferably", "according to a preferred embodiment" or "optionally" all indicate that the corresponding paragraph discloses an independent concept, and the applicant reserves the right to file a divisional application based on each inventive concept.

Claims

1. An automatic measurement and control intelligent box-type transformer, comprising at least a number of sensors for collecting equipment parameters and a control module for controlling the operating status of the equipment, characterized in that: Also includes at least one IoT agent terminal, The IoT proxy terminal determines an abnormal condition of the device based on the device parameters sent by the sensor, and feeds back corresponding at least one control information to the control module based on the abnormal condition. In response to the control information, the control module sends a control instruction to a device related to the control information, the IoT agent terminal evaluates the operating status information of the smart box transformer based on the device parameters, and sends the operating status information to at least one first edge agent node or at least one second edge agent node. The first edge proxy node or the second edge proxy node feeds back at least one control suggestion to the IoT proxy terminal based on the operation status information, The processor in the IoT proxy terminal is configured as follows: An optimization decision for selecting a first edge proxy node and / or at least one second edge proxy node is generated through a Markov decision process to achieve self-optimization, and training and learning are performed based on the amount of data to be processed, physical distance information, and performance test results, as well as the information feedback time and / or information feedback speed of the first edge proxy node and / or at least one second edge proxy node, and an optimization decision is made in a manner that the information feedback time is shortest and / or the information feedback speed value is maximized as the optimization target, Determine the selected selection space variable as the number and order of the first edge proxy node and / or the second edge proxy node, and select the first edge proxy node and / or the second edge proxy node whose predicted feedback time is greater than the actual feedback time for priority selection according to the historical results of the predicted feedback time and the actual feedback time, The state transition probability determined according to the following steps generates at least one selection information in the optimization decision: In the learning round, the sensor module explores the environment based on the Boltzmann distribution strategy; After the learning round is over, the selection object is determined based on the ε-greedy greedy strategy; The geographical location of the first edge proxy node is close to the location of the cloud server system and its data task processing capability is better than that of the second edge proxy node. The geographical location of the second edge proxy node is close to the geographical location range of the smart box transformer. The number of the first edge proxy nodes is less than the number of the second edge proxy nodes, and the number of the second edge proxy nodes is less than the number of the intelligent box transformers.

2. The automatic measurement and control intelligent box-type transformer according to claim 1 is characterized in that: The IoT proxy terminal directly establishes a communication connection with the cloud server system, or the IoT proxy terminal establishes a communication connection with the cloud server through the first edge proxy node and / or at least one second edge proxy node, In response to the request information of the first edge proxy node or the at least one second edge proxy node, the IoT proxy terminal sends data information corresponding to the request information to the first edge proxy node or the at least one second edge proxy node.

3. The automatic measurement and control intelligent box-type transformer according to claim 2 is characterized in that: The cloud server system comprehensively describes the operating status information of the smart box transformer based on the data association of power parameter data, non-power parameter data and / or user equipment basic ledger information, thereby feeding back adjustment information to the IoT agent terminal based on the specific power problems determined by analysis.

4. A measurement and control system for an intelligent box-type transformer using the automatic measurement and control according to any one of claims 1 to 3, characterized in that: At least including intelligent box-type transformer and cloud server system, At least one IoT proxy terminal is arranged in the smart box-type transformer and / or near the geographical location. The IoT proxy terminal determines the abnormality of the equipment based on the equipment parameters sent by several sensors in the smart box-type transformer, and feeds back corresponding at least one control information to the control module based on the abnormality. In response to the control information, the control module sends a control instruction to a device related to the control information, The IoT proxy terminal sends an alarm message to the cloud server system in a direct communication and / or indirect communication manner based on the abnormal situation.

5. The automatic measurement and control system for intelligent box-type transformer according to claim 4 is characterized in that: The IoT proxy terminal evaluates and selects at least one first edge proxy node and / or a node with a shorter predicted data processing time based on the physical distance parameter and the received data performance test parameter. or at least one second edge proxy node establishes a communication connection, The IoT proxy terminal sends part of the data task to the first edge proxy node and / or the second edge proxy node.

6. The automatic measurement and control system for intelligent box-type transformer according to claim 5 is characterized in that: The first edge proxy node and / or the second edge proxy node processes data information based on the request information of the IoT proxy terminal, and / or The first edge proxy node and / or the second edge proxy node sends alarm information to the cloud server system based on the data anomaly information.

Images