Indoor building energy consumption monitoring and regulation system and method based on environmental data

Abstract

The application discloses an indoor building energy consumption monitoring and regulation system and method based on environmental data, and the method comprises the following steps: deploying intelligent power distribution nodes in each loop of a building distribution box, each intelligent power distribution node being connected to a corresponding loop power line and establishing an ad hoc network through power line carrier communication; judging the abnormal type of electrical parameters according to the collected data of the intelligent power distribution nodes; when the abnormal type of electrical parameters is harmonic pollution abnormality, requesting high-frequency waveform data of the same period from the topologically adjacent intelligent power distribution nodes through the ad hoc network; after receiving the return data of the adjacent intelligent power distribution nodes, analyzing the space-time propagation characteristics of the high-frequency waveform data to determine the abnormal nodes; sending a control signal to the abnormal nodes through the power line carrier; and the control signal is used for executing energy consumption regulation. The application relates to the technical field of indoor building energy consumption monitoring and intelligent regulation, and solves the technical problem of insufficient accuracy of energy consumption data monitoring in the prior art.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of energy consumption monitoring, and particularly relates to an indoor building energy consumption monitoring and regulation method based on environmental data. BACKGROUND

[0002] Indoor building energy consumption accounts for a considerable proportion of total social energy consumption, and its energy consumption status is directly related to the level of operating cost, the pros and cons of energy utilization efficiency, and the degree of influence on the environment. The existing indoor building energy consumption monitoring generally relies on wireless communication technologies such as Wi-Fi and ZigBee to realize data transmission and collaborative regulation. However, wireless communication signals are easily blocked by building walls, metal components and the like, resulting in poor communication stability between nodes, and often causing data transmission packet loss and delay, which makes the energy consumption data collection incomplete and the abnormal information reporting lag, seriously affecting the timely response to energy consumption abnormalities. Moreover, due to unstable communication, efficient collaboration between nodes is difficult to achieve, and abnormal positioning relies on isolated analysis of single node data, which not only has low positioning accuracy, but also causes the regulation command to be unable to quickly reach the target node, further reducing the timeliness of energy consumption abnormality processing. In addition, the existing method does not correct the collected electrical parameters for environmental adaptability, and the energy consumption metering data accuracy is insufficient due to the influence of temperature and humidity and line physical characteristics, which makes it difficult to support accurate energy consumption monitoring and regulation. SUMMARY

[0003] The present application aims to at least solve one of the technical problems existing in the prior art. To this end, the present application provides an indoor building energy consumption monitoring and regulation method based on environmental data, which is used to solve the technical problem of insufficient accuracy of energy consumption data monitoring in the prior art. The present application deploys intelligent power distribution nodes in each circuit of the building distribution box and establishes an ad hoc network using power line carrier communication, corrects electrical parameters combined with environmental data, collaboratively analyzes abnormal data to accurately locate abnormal nodes and executes targeted regulation, effectively improving the reliability and efficiency of indoor building energy consumption monitoring and regulation.

[0004] To achieve the above-mentioned purpose, the first aspect of the present application provides an indoor building energy consumption monitoring and regulation method based on environmental data, comprising:

[0005] Intelligent power distribution nodes are deployed in each circuit in the building distribution box, each intelligent power distribution node is connected to the power line of the corresponding circuit, and an ad hoc network is established through power line carrier communication;

[0006] The abnormal type of the electrical parameter is determined according to the collected data of the intelligent power distribution node;

[0007] When the electrical abnormal parameter type is harmonic pollution abnormality, the intelligent power distribution node requests high-frequency waveform data of the same period from the topologically adjacent intelligent power distribution nodes through the ad hoc network;

[0008] After receiving the returned data from neighboring smart distribution nodes, the smart distribution node analyzes the spatiotemporal propagation characteristics of the high-frequency waveform data to identify abnormal nodes.

[0009] Control signals are sent to abnormal nodes via power line carrier waves; these control signals are used to perform energy consumption regulation.

[0010] Based on the above technical solution, this invention first deploys intelligent power distribution nodes in each circuit within the building's power distribution box and establishes a self-organizing network using power line carrier communication. This allows for communication and collaboration between nodes based on the building's existing power lines, eliminating the need for additional communication lines and reducing the complexity and cost of system deployment. The self-organizing network also enhances network flexibility and stability, facilitating adaptation to indoor building environments of varying sizes and structures. Secondly, by analyzing the collected data from the intelligent power distribution nodes to determine the type of electrical parameter anomalies, accurate identification of energy consumption anomalies is achieved, providing a clear basis for subsequent control. Furthermore, when harmonic pollution anomalies occur, the self-organizing network requests high-frequency waveform data from neighboring nodes for the same time period and analyzes its spatiotemporal propagation characteristics to identify the abnormal node. This multi-node collaborative analysis improves the accuracy and reliability of anomaly location. Finally, by sending control signals to the abnormal node via power line carrier communication to execute energy consumption control, the timeliness and effectiveness of control command transmission are ensured, enabling rapid response and handling of energy consumption anomalies, thereby achieving efficient monitoring and precise control of indoor building energy consumption.

[0011] Furthermore, the establishment of an ad hoc network via power line carrier communication includes:

[0012] After each smart distribution node is powered on, it broadcasts a beacon frame containing its own identifier and the parent node's code;

[0013] The electrical distance D between nodes is calculated based on the carrier signal transmission delay Δt and the carrier propagation speed v: D = V × Δt;

[0014] Screening electrical distance ≤ D max The nodes construct a neighbor list; where D max The maximum adjacent distance is preset;

[0015] Based on the neighbor node table, the communication packet loss rate L, signal strength RSSI, and path hop count H of each node in the multi-hop path are calculated, and the dynamic routing weight W of each path is calculated; the multi-hop path refers to a path containing multiple nodes.

[0016] The paths are sorted in descending order according to the dynamic routing weight W, and the top N paths with the highest weights are retained to form a dynamic routing table. Each node uses the dynamic routing table to forward data and adjust the network topology, thus obtaining an ad hoc network.

[0017] Further, in the ad hoc network, when a dynamic routing weight W < W th , the communication is automatically switched to a backup relay node; wherein, W th represents a preset weight threshold, and the backup relay node is a pre-registered other neighboring node in the dynamic routing table which maintains communication connection with the current node.

[0018] Further, the calculation formula of the dynamic routing weight W is: ; wherein, RSSI min represents the minimum received signal strength, p represents the packet loss penalty factor, and k represents the hop attenuation coefficient.

[0019] Further, the ad hoc network includes a first channel and a second channel, the first channel is used for transmitting basic data, the basic data includes transmission temperature and humidity parameters, electrical parameters, line physical parameters and compensated energy consumption data; the second channel is based on the real-time collected noise spectrum and signal attenuation rate, and a preset frequency range of channel frequency band is adaptively allocated for the transmission of high-frequency waveform data, the transmission of high-frequency waveform data is segmented into multiple subcarriers for transmission by using orthogonal frequency division multiplexing technology, and forward error correction coding FEC technology is integrated in the transmission process to ensure transmission reliability.

[0020] Further, the intelligent power distribution node corrects the collected electrical parameters by using the obtained temperature and humidity parameters and the pre-stored line physical parameters, and outputs the compensated power data, including:

[0021] According to the temperature T, the current sampling value is linearly compensated to obtain a current compensation value;

[0022] According to the relative humidity RH, the power factor is corrected to obtain a compensated power factor;

[0023] Based on the wire diameter d and the material resistivity r, a line correction coefficient is calculated;

[0024] According to the product of the current compensation value, the corrected power factor and the line correction coefficient, the compensated power data is calculated.

[0025] Further, the calculation formula of the current compensation value I comp is: ; wherein, I raw represents the current sampling value, a represents the wire resistance temperature coefficient, T ref represents the reference temperature, and U represents the line voltage.

[0026] The calculation formula of the compensated power factor PF comp is: PF comp =F raw ×[1-β(RH-RH ref); wherein, β represents the humidity coefficient of the insulation material, RH ref represents the reference humidity;

[0027] The calculation formula of the line correction coefficient p is: p=(p ref ×d ref ²) / (r×d²); wherein, p ref represents the reference resistivity, d ref represents the reference line diameter.

[0028] Further, the judging the abnormal type of the electrical parameter according to the collected data of the intelligent power distribution node comprises:

[0029] Using the LSTM network model, the historical energy consumption data stored by the intelligent power distribution node and the corresponding calendar attributes, temperature data and humidity data are used to generate a reference power curve every hour;

[0030] According to the reference power curve, the deviation δ of the actual power value P(t) and the reference power value P base (t) is calculated, and the calculation formula of the deviation is: δ=[P(t)-P base (t)] / P base (t)×100%, t represents time, and P base (t)∈reference power curve;

[0031] The collected current waveform is subjected to Fourier transform, and the harmonic energy distribution entropy value E after the transform is calculated, and the calculation formula of the entropy value is: , n represents the harmonic order, N h represents the highest analysis harmonic order, and p n represents the proportion of the nth harmonic energy;

[0032] According to the deviation, the entropy value and the preset threshold range, whether the electrical parameter is abnormal is judged, and the electrical parameter abnormal type comprises device overload abnormality, device stop abnormality and harmonic pollution abnormality.

[0033] Further, the judging rule of the device overload abnormality is that the deviation is greater than a second deviation threshold, and the entropy value is greater than a first entropy threshold and less than a second entropy threshold;

[0034] The judging rule of the device stop abnormality is that the deviation is less than or equal to a first deviation threshold, and the entropy value is less than or equal to a first entropy threshold;

[0035] The judging rule of the harmonic pollution abnormality is that the deviation is greater than a first deviation threshold and less than or equal to a second deviation threshold, and the entropy value is greater than a third entropy threshold;

[0036] Wherein, the first deviation threshold < the second deviation threshold, the first entropy threshold < the second entropy threshold < the third entropy threshold.

[0037] Further, the high-frequency waveform data of the same period is, after detecting the node sending the trigger instruction containing the time stamp T event of the electrical parameter anomaly, the adjacent node backtracks the stored waveform data in the interval [T event -T pre , T event +T post ] based on T event , and the waveform data in the interval is obtained by wavelet transform and quantization encoding compression; wherein, T pre represents the preset forward buffering time, and T post represents the preset backward recording time.

[0038] Further, the analysis of the space-time propagation characteristics of the high-frequency waveform data comprises:

[0039] Taking the time stamp T event as the reference, the waveform time delay of each node is calculated by cross-correlation analysis;

[0040] The abnormal source positioning equation is constructed as: ; wherein, (x, y) is the abnormal source coordinate, τ i is the waveform time delay of the i-th intelligent power distribution node, (x i , y i ) is the coordinate of the i-th intelligent power distribution node, v is the carrier propagation speed, is the measurement error;

[0041] Based on the time difference positioning technology TDOA, an over-determined equation set is established: ; wherein, τ j is the waveform time delay of the j-th intelligent power distribution node, (x j , y j ) is the coordinate of the j-th intelligent power distribution node;

[0042] According to the Levenberg-Marquardt algorithm, the over-determined equation set is solved to obtain the abnormal source coordinate;

[0043] If the abnormal source coordinate is located within the loop range of the intelligent power distribution node i, the intelligent power distribution node i is marked as an abnormal node; otherwise, it is marked as a line crosstalk fault.

[0044] Further, the coordinate of the intelligent power distribution node is obtained by: when the intelligent power distribution node is deployed, the loop number and installation physical position of the intelligent power distribution node are converted into three-dimensional coordinates and pre-stored in the building information model (BIM) through the building coordination gateway.

[0045] Further, the energy consumption regulation includes:

[0046] If the electrical parameter abnormal type is the device overload abnormality, a hierarchical load instruction is sent, and the hierarchical load instruction is used to first close the non-key load device and then cut off the abnormal loop.

[0047] The non-key load device is a device with a priority lower than a preset level in a device priority list pre-configured by the device management platform, and the list is synchronized to each intelligent power distribution node through the initialization communication of the management platform and the intelligent power distribution node; and the abnormal loop is a specific loop determined based on the loop identification information and the abnormal node positioning result built in the intelligent power distribution node.

[0048] If the electrical parameter abnormal type is the harmonic pollution abnormality or the harmonic abnormal fluctuation, an active power filter (APF) is started, and a reverse harmonic current with an equal amplitude and an opposite phase to the harmonic current amplitude collected by the intelligent power distribution node is injected into the abnormal loop.

[0049] If the electrical parameter abnormal type is the device stop abnormality, a device identification and an installation position associated with the current intelligent power distribution node are pushed to the management platform.

[0050] The second aspect of the application provides an indoor building energy consumption monitoring and regulation system based on environmental data, which comprises a plurality of intelligent power distribution nodes, a building coordination gateway and a management platform.

[0051] The intelligent power distribution node (IPDU) is deployed in each power distribution box loop of a building and comprises:

[0052] An electric energy metering module is configured to collect electrical parameters of a power distribution loop, and the electrical parameters include voltage, current and power parameters.

[0053] A communication module is configured to transmit data through a power distribution line carrier wave to realize self-organizing network communication between IPDU nodes; the self-organizing network communication means that the IPDU nodes construct a mesh topology network through a power line, and the network supports point-to-point transmission of high-frequency waveform data and broadcast of collaborative control instructions.

[0054] An embedded processor is configured to execute a temperature and humidity compensation algorithm, and correct the collected electrical parameters by using the obtained temperature and humidity parameters and the pre-stored line physical parameters.

[0055] A wet temperature sensor is arranged to detect environmental parameters, including temperature and relative humidity, in real time at the installation position of the IPDU.

[0056] The building coordination gateway BCG is connected with each IPDU through a power line and is used to coordinate data interaction among distributed IPDU nodes and realize device-level energy consumption metering and abnormal positioning.

[0057] The management platform is in communication connection with the BCG and is used to realize three-dimensional visualization of indoor building energy consumption and perform energy consumption regulation and control.

[0058] Further, the building coordination gateway BCG comprises:

[0059] A carrier wave communication modulation and demodulation module is arranged to analyze ad hoc network data of the intelligent power distribution node IPDU and convert the ad hoc network data into a standard network protocol.

[0060] A data aggregation processing module is arranged to associate building space topology data of the IPDU node and generate a district energy consumption map; the district energy consumption map represents an energy consumption three-dimensional thermal map that fuses position information of the IPDU node and a building information model BIM model and dynamically displays energy consumption data of each floor.

[0061] An external communication interface is arranged to connect the management platform and a cloud platform; the cloud platform is used to store historical energy consumption data and perform power prediction according to the historical energy consumption data of the device by using an LSTM network model.

[0062] Compared with the prior art, the present application has the following beneficial effects:

[0063] In terms of energy consumption monitoring, the intelligent power distribution node is arranged in each circuit and an ad hoc network is established by using power line carrier wave communication, data transmission is realized by relying on the existing power line of the building, no additional wiring is needed, the deployment cost and complexity are reduced, and the stability and flexibility of network communication are ensured by dynamic routing weight adjustment and standby relay node switching, which is suitable for different building structures. At the same time, the intelligent power distribution node corrects electrical parameters in combination with temperature and humidity parameters and line physical parameters, effectively reduces the metering error caused by environmental factors, and improves the accuracy of energy consumption data; a reference power curve is generated based on an LSTM network model, abnormal types are judged in combination with bias and harmonic energy distribution entropy values, and multi-dimensional analysis makes abnormal identification more accurate.

[0064] The advantages of the method are obvious in energy consumption regulation efficiency and pertinence. For abnormal harmonic pollution, high-frequency waveform data of adjacent nodes are obtained through self-organizing network, and abnormal nodes are determined by combining time-space propagation characteristics and TDOA positioning algorithm, so as to realize accurate positioning of abnormal sources and avoid deviation of single node judgment. According to different abnormal types, hierarchical regulation strategies are adopted, such as closing non-critical load and cutting off circuit when equipment is overloaded, starting active filter to inject reverse harmonic current when harmonic abnormality occurs, and pushing maintenance alarm containing specific information when equipment stops, so that the regulation is more targeted, can quickly respond and handle energy consumption abnormalities, and improves the overall efficiency of indoor building energy consumption management. In addition, the building coordination gateway and the management platform generate a three-dimensional energy consumption atlas combined with the BIM model, realize the visualization monitoring of energy consumption, and facilitate the management personnel to intuitively master the energy consumption status, and further optimize the energy consumption management strategy. BRIEF DESCRIPTION OF DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0066] Figure 1 The flowchart of the indoor building energy consumption monitoring and regulation method based on environmental data provided by the present application is shown in the figure.

[0067] Figure 2 The flowchart of another indoor building energy consumption monitoring and regulation method based on environmental data provided by the present application is shown in the figure.

[0068] Figure 3 The flowchart of another indoor building energy consumption monitoring and regulation method based on environmental data provided by the present application is shown in the figure.

[0069] Figure 4 The framework diagram of the indoor building energy consumption monitoring and regulation system based on environmental data provided by the present application is shown in the figure.

[0070] Figure 5 The framework diagram of another indoor building energy consumption monitoring and regulation system based on environmental data provided by the present application is shown in the figure.

[0071] Figure 6 The framework diagram of another indoor building energy consumption monitoring and regulation system based on environmental data provided by the present application is shown in the figure. DETAILED DESCRIPTION

[0072] The technical solutions of the present application will be described clearly and completely below in connection with the embodiments. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of the present application.

[0073] To solve the technical problem that the networking in the prior art is vulnerable to the shielding of building walls and has poor communication stability, an indoor building energy consumption monitoring and regulation method based on environmental data is provided, and the method comprises:

[0074] Intelligent power distribution nodes are deployed in each circuit in a building distribution box, and each intelligent power distribution node is connected to a corresponding circuit power line and establishes an ad hoc network through power line carrier communication.

[0075] The abnormal type of the electrical parameter is determined according to the collected data of the intelligent power distribution node.

[0076] When the electrical abnormal parameter type is harmonic pollution abnormality, high-frequency waveform data of the same period is requested from the topologically adjacent intelligent power distribution nodes through the ad hoc network.

[0077] After receiving the returned data of the adjacent intelligent power distribution nodes, the spatiotemporal propagation characteristics of the high-frequency waveform data are analyzed to determine the abnormal node.

[0078] A control signal is sent to the abnormal node through the power line carrier, and the control signal is used to perform energy consumption regulation.

[0079] Based on this, the present application relies on the existing power lines of the building to realize communication without additional line laying, which can improve the communication stability of indoor building energy consumption monitoring, and through the collaborative analysis of multiple nodes, the accurate identification and regulation of energy consumption abnormalities are realized.

[0080] As shown in Figure 1 The indoor building energy consumption monitoring and regulation method based on environmental data provided by the embodiments of the present application comprises:

[0081] S1, intelligent power distribution nodes are deployed in each circuit in a building distribution box, and each intelligent power distribution node is connected to a corresponding circuit power line and establishes an ad hoc network through power line carrier communication.

[0082] The intelligent power distribution node is a device with electrical parameter collection, data communication and basic control functions, which is used to monitor the power operation state of the circuit in real time and participate in network collaboration. The ad hoc network is a distributed network formed by multiple intelligent power distribution nodes through power line carrier communication, which is used for data transmission and instruction interaction between nodes.

[0083] In some implementations, the smart power distribution node can include an electric energy metering unit, a PLC (Power Line Communication, PLC) communication unit, and a microprocessor, which realizes data transmission while obtaining power supply through the power line; the ad hoc network can adopt common topological structures such as star type, tree type, or Mesh, and the nodes complete identity authentication and network access through a preset communication protocol (such as a PLC protocol conforming to the IEEE 1901 standard).

[0084] It should be noted that the number of deployed smart power distribution nodes can be adjusted according to the number of building power distribution circuits and the monitoring accuracy requirement, and a single node can correspond to one or more electric equipment circuits.

[0085] For example, in a three-story building, the lighting, air conditioning, socket, and other circuits in the power distribution box on each floor are respectively deployed with smart power distribution nodes, the nodes are connected through the power line to form an ad hoc network covering the entire building, and cross-floor data interaction is realized.

[0086] S2, determining the abnormal type of the electrical parameter according to the collected data of the smart power distribution node.

[0087] Among them, the collected data of the smart power distribution node includes electrical parameters such as voltage, current, and power, and can include auxiliary parameters such as environmental temperature and humidity; the abnormal type of the electrical parameter includes harmonic pollution abnormality, and can also include device overload, current / voltage exceeding standard, and other common power abnormalities.

[0088] In some implementations, whether the electrical parameter is abnormal and the abnormal type can be determined by comparing the collected real-time electrical parameter with a preset reference value (such as a device rated parameter, a historical normal operation parameter range), or using a Fourier Transform analysis method to analyze the harmonic component.

[0089] It should be noted that the reference value can be dynamically adjusted according to the type of building electrical equipment, the operation period, and other factors to adapt to the normal operation range in different scenarios.

[0090] For example, the smart power distribution node collects current data in real time, and if it is found through Fourier Transform that the amplitude ratio of 3rd and 5th harmonics exceeds a preset threshold, it is determined that there is a harmonic pollution abnormality.

[0091] S3, when the electrical abnormal parameter type is harmonic pollution abnormality, requesting high-frequency waveform data of the same period from the topologically adjacent smart power distribution nodes through the ad hoc network.

[0092] Among them, the high-frequency waveform data refers to continuous sampling data of voltage or current waveform containing high-frequency harmonic component, which is used to analyze the propagation path and source of the harmonic, and is obtained by real-time collection of the high-frequency sampling module of the smart power distribution node on the circuit power signal.

[0093] In some implementations, the "same period" can be determined according to the time of abnormal occurrence, for example, a period of 1-5 seconds before and after the time of abnormal triggering; the request instruction is sent through broadcast or point-to-point communication of the ad hoc network, and the adjacent nodes extract the stored waveform data of the corresponding period according to the time range in the instruction.

[0094] It should be noted that the storage of high-frequency waveform data can adopt a circular buffer mode to ensure that the key period data can be retrieved when an abnormality occurs.

[0095] For example, node A detects a harmonic pollution abnormality, sends a request to the three nearest adjacent nodes, and requests to return the current high-frequency waveform data from 2 seconds before to 3 seconds after the time of abnormal occurrence (e.g., 10:00:00).

[0096] S4, analyze the space-time propagation characteristics of the high-frequency waveform data to determine the abnormal node.

[0097] The space-time propagation characteristics of the high-frequency waveform data reflect the propagation time difference of the harmonic waveform between different nodes, the degree of waveform distortion, and other changes with time and space, which can be used to locate the position of the harmonic source.

[0098] In some implementations, the space-time propagation characteristics can be analyzed by comparing the phase difference, amplitude decay law, or calculating the time difference of the waveform arriving at each node of the same period high-frequency waveform collected by different nodes; the determination of the abnormal node can be combined with the physical location distribution of the node to determine the intelligent power distribution node corresponding to the loop where the harmonic source is located.

[0099] It should be noted that the physical location of the node can be determined by the pre-configured installation address information (e.g., "1st floor east power distribution box 3rd loop").

[0100] For example, by analyzing the high-frequency waveforms returned by nodes A, B, and C, it is found that the harmonic waveform first appears at node B, and then propagates to nodes A and C in turn, and the harmonic amplitude of node B is the largest, and accordingly node B is determined as the abnormal node.

[0101] S5, send an energy consumption control signal to the abnormal node through power line carrier.

[0102] The energy consumption control signal is used to adjust the power consumption state of the loop where the abnormal node is located to eliminate or reduce harmonic pollution, for example, to control the start of harmonic control equipment, adjust the operating power of electrical equipment, etc.

[0103] In some implementations, the control signal can be transmitted in digital code form through the power line carrier, and the abnormal node receives and analyzes the signal and drives the execution unit (such as a relay, a harmonic filter) to act.

[0104] It should be noted that the control signal can contain a check code to ensure that the signal is not tampered with or misexecuted during transmission.

[0105] For example, the control signal sent to the abnormal node instructs it to start the built-in harmonic filter to filter out the main harmonic component and reduce the harmonic pollution of the loop.

[0106] Based on the above technical solutions, the indoor building energy consumption monitoring and control method based on environmental data provided in the application establishes an ad hoc network through power line carrier communication, avoiding the influence of wireless communication being blocked by walls and improving communication stability; abnormality is located by using multi-node cooperative analysis of high-frequency waveform data, improving the accuracy of energy consumption anomaly monitoring; data transmission and control are realized relying on existing power lines, reducing system deployment cost and being suitable for various indoor building scenarios.

[0107] In a possible implementation form of the embodiment of the application, in combination with Figure 1 As shown in Figure 2 The above S1 can be implemented through the following S101, S102, S103 and S104, which will be described in detail below:

[0108] S101, after each intelligent power distribution node is powered on, a beacon frame containing its own identifier and parent node code is broadcasted.

[0109] The beacon frame is a communication frame used by the intelligent power distribution node for network identity identification and topology discovery, the own identifier is a unique number of the node, such as a MAC address or a device serial number, and the parent node code is an upper node identifier associated when the node initially accesses the network (which can be set as a default value when not accessed).

[0110] In some implementation forms, the broadcast period of the beacon frame can be set to 1-5 seconds, and the broadcast power is adjusted according to the communication range requirement of the node to ensure that the adjacent nodes can stably receive; the parent node code can be determined automatically through the initialization process after the node is powered on, and the node with the highest signal strength is preferentially selected as the initial parent node.

[0111] It should be noted that the beacon frame needs to contain a check field (such as a CRC check code) to avoid identification information errors caused by signal interference.

[0112] For example, after a certain intelligent power distribution node is powered on, it broadcasts a beacon frame every 2 seconds, which contains its own MAC address “IPDU-001” and parent node code “NULL”, and “NULL” indicates that it is not associated with a parent node. After being received by the adjacent nodes, the existence and initial state of the node can be identified.

[0113] S102, the electrical distance D between nodes is calculated according to the carrier signal transmission delay Δt and the carrier propagation speed V, and nodes with an electrical distance ≤D maxThe node constructs a neighbor node table.

[0114] wherein the electrical distance D is a virtual distance between nodes calculated based on communication signal transmission characteristics, and the calculation formula is D=V×Δt; V is the propagation speed of the carrier in the power line, and is usually taken as 2×10 8 m / s; D max is a preset maximum adjacent distance, used to limit the range of the neighbor node.

[0115] In some implementations, the carrier signal transmission delay Δt is measured through a time synchronization mechanism between nodes (such as the IEEE 1588 precise time protocol); D max The value of D needs to be determined in combination with the length of the building power distribution line and the signal attenuation characteristics. For a general residential building, it is taken as 50-100 m, and for a large commercial building, it is taken as 100-200 m.

[0116] It should be noted that the electrical distance is not the physical distance, but an index reflecting the communication reachability between nodes. Nodes with the same physical distance may have different electrical distances due to differences in line noise.

[0117] For example, the carrier signal transmission delay Δt between node A and node B is 0.2 μs, calculated at V=2×10 8 m / s, the electrical distance D=40 m, and if D max is set to 50 m, then node B is included in the neighbor node table of node A.

[0118] S103, based on the neighbor node table, a dynamic routing weight W of a multi-hop path is calculated, the first N paths are reserved after being sorted in descending order of the weight to form a dynamic routing table, and data forwarding between nodes and network topology adjustment are realized.

[0119] wherein the multi-hop path refers to a communication path containing multiple nodes; the dynamic routing weight W is used to evaluate the communication quality of the path, and the calculation formula is ; RSSI is the received signal strength, RSSI min is the minimum received signal strength, L is the communication packet loss rate, ρ is the packet loss penalty factor, H is the path hop count, and k is the hop count attenuation coefficient.

[0120] In some implementations, RSSI min is taken as -100 dBm, which is the minimum receiving sensitivity of common power line carrier communication, the default value of ρ is 1.2, and k is taken as 0.3-0.5 according to the network size; the value of N is determined according to the number of nodes, and is generally taken as 3-5 to ensure path redundancy.

[0121] For example, the RSSI of a certain multi-hop path is -70 dBm, L=0.05, H=2, and W≈(-70 / -100)×1 / (0.05 1.2) x e -0.4×2 ≈0.7 x 63.1 x 0.45 ≈ 20.0, if one of the top 3 paths with the highest weight is selected, it is included in the dynamic routing table.

[0122] S104, when the dynamic routing weight W < W th , automatically switch to the backup relay node for communication.

[0123] wherein, W th is a preset weight threshold value, used to determine whether the current path communication quality meets the standard; the backup relay node is another adjacent node pre-registered in the dynamic routing table and maintaining communication connection with the current node.

[0124] In some implementations, the value of W th is determined according to historical communication quality data, and is usually 60%-70% of the average weight of the normal path; the priority of the backup relay node is ranked according to its dynamic routing weight with the current node, to ensure the optimal communication quality after switching.

[0125] It should be noted that the switching process needs to use a seamless switching mechanism (such as pre-establishing the connection state of the backup path) to avoid data transmission interruption.

[0126] For example, the dynamic routing weight of a certain path is W = 15, and W th is set to 20, because W < W th , the system automatically switches to the backup relay node with a weight of 18 in the dynamic routing table, to ensure the stable transmission of data.

[0127] Based on the above technical solution, through the steps of broadcasting the beacon frame, calculating the electrical distance, constructing the dynamic routing table, and automatically switching the backup node, the self-organizing network based on power line carrier communication is constructed, which can adapt to the network topology change, guarantee the stability and reliability of data transmission between nodes, and provide efficient communication support for subsequent energy consumption monitoring and regulation.

[0128] In a possible implementation of the embodiment of the application, the above S2 can be implemented through the following S201, S202 and S203, which are specifically described as follows:

[0129] S201, using a long short-term memory network (LSTM) model, based on the historical energy consumption data stored by the intelligent power distribution node and the corresponding calendar attributes, temperature data, and humidity data, to generate a baseline power curve every hour.

[0130] The historical energy consumption data includes hourly power values in the past period (e.g., 3 months); the calendar attributes include weekdays / holidays, seasons, time periods (e.g., peak / flat / valley periods); and the reference power curve is a curve of reference power values corresponding to different time points in a normal operating state.

[0131] In some implementations, an LSTM model (Long Short-Term Memory, LSTM) needs to be pre-trained and then applied to actual power prediction. The pre-training process can include:

[0132] First, collect the historical energy consumption data stored by the intelligent power distribution node, the calendar attributes of the corresponding period, the environmental temperature data, and the relative humidity data to form an original data set; the calendar attributes include weekdays and holidays, the historical energy consumption data is collected every half hour, and then aggregated into hourly average power values, the environmental temperature data and the relative humidity data are both average temperatures per hour as the corresponding data for that hour. The original data set obtained in this way is time series data, the interval between data and data is 1 hour, and the data of each time point includes: the average power of the hour, the average temperature of the hour, the average humidity of the hour, and the calendar attributes corresponding to the hour.

[0133] Next, the original data set is cleaned to remove abnormal values caused by equipment failure and data transmission errors; the calendar attributes are converted into numerical features, such as marking weekdays as 1 and holidays as 0; and then all feature data is normalized to map the data to the [0, 1] interval to improve model convergence speed.

[0134] An LSTM model is constructed based on a deep learning algorithm, and the input data of the model is determined to be a feature vector containing "past 24-hour hourly power values + current time calendar attributes + current time temperature and humidity"; and the output data is the reference power prediction value for the next hour at the current time.

[0135] Then, the preprocessed data set is divided into a training set, a validation set, and a test set in a ratio of 7:2:1; the mean squared error (MSE) is used as the loss function, the Adam optimizer, the learning rate is initially set to 0.001, and the iteration is trained for 100-300 rounds. After each round, the model performance is evaluated using the validation set, and when the validation set loss does not decrease for 10 consecutive rounds, the training is stopped and the optimal model parameters are saved.

[0136] The optimal model is evaluated using the test set. If the average absolute error (MAE) is ≤5% and the root mean square error (RMSE) is ≤8%, the model passes the validation; otherwise, the model structure needs to be adjusted, such as increasing the number of LSTM layer neurons or expanding the data set and retraining.

[0137] Finally, the trained model is deployed to the embedded processor of the intelligent power distribution node or the cloud server, real-time receives the feature data (power in the past 24 hours, calendar attributes, temperature and humidity) at the current time, outputs the reference power value every hour, and forms a continuous reference power curve.

[0138] S202, calculate the deviation δ of the actual power value and the reference power value, and the harmonic energy distribution entropy value E.

[0139] The calculation formula of the deviation δ is δ = [P(t) - P base (t)] / P base (t) x 100%, P(t) is the actual power value, P base (t) is the reference power value corresponding to the time t in the reference power curve.

[0140] The calculation formula of the harmonic energy distribution entropy value E is: , n is the harmonic order, N h is the highest analysis harmonic order (usually 30 times), p n is the proportion of the nth harmonic energy to the total harmonic energy, which is obtained by Fourier transform on the collected current waveform.

[0141] It should be pointed out that the deviation δ is used to reflect the deviation degree of the actual energy consumption from the normal state, and the entropy value E is used to reflect the disorder of the harmonic component. Through comprehensive analysis and threshold comparison of the two features, the specific abnormal type of the electrical parameter can be obtained, and the different abnormal conditions such as equipment overload, stop and harmonic pollution can be accurately distinguished.

[0142] For example, the actual power P(t) = 90kW at a certain time, the corresponding reference power P base (t) = 80kW, then the deviation δ = (90-80) / 80 x 100% = 12.5%; the Fourier transform gives that the harmonic energy proportions of 3 times and 5 times are 0.3 and 0.2 respectively, and the total proportion of the remaining harmonics is 0.5, so the entropy value E = -(0.3 x log20.3 + 0.2 x log20.2 + 0.5 x log20.5) ≈ 1.489.

[0143] S203, judging the electrical parameter abnormal type according to the deviation δ, the entropy value E and the preset threshold range.

[0144] The electrical parameter abnormal type includes equipment overload abnormality, equipment stop abnormality and harmonic pollution abnormality.

[0145] In some implementations, the judgment rule of the equipment overload abnormality is δ>the second deviation threshold value, and the first entropy value threshold value

[0146] The judgment rule of the device stoppage anomaly is that δ≤the first deviation threshold value and E≤the first entropy value threshold value.

[0147] The judgment rule of the harmonic pollution anomaly is that the first deviation threshold value < δ≤the second deviation threshold value and E>the third entropy value threshold value.

[0148] And, the first deviation threshold value < the second deviation threshold value, the first entropy value threshold value < the second entropy value threshold value < the third entropy value threshold value.

[0149] It should be noted that the threshold values of the present application are set through a large amount of experimental data and device operation characteristic analysis, and are as follows: the first deviation threshold value δ1 is -30%, that is, when the power drops by more than 30%, it usually indicates that the device is completely stopped or a serious fault occurs, thereby causing the motor to stop, so that below this value triggers the stoppage anomaly; the second deviation threshold value δ2 is +15%, that is, through experiments, it is found that more than 15% power rise will cause line overheating risk; the first entropy value threshold value E1 is 0.5, because the harmonic energy distribution is concentrated when the device is normally operated (E≤0.5 indicates that the harmonic component is simple, which meets the stoppage characteristics; the second entropy value threshold value E2 is 1.0, because the overloaded device is usually accompanied by an increase in specific harmonic, so that the E value is in the range of 0.6-0.9, and therefore 1.0 is taken as a safe upper limit; the third entropy value threshold value E3 is 1.5, because typical harmonic pollution will cause wideband harmonic energy dispersion, and laboratory data shows that the probability of E>1.5 is 95%.

[0150] It should be noted that when δ or E is at the threshold boundary (such as δ=the second deviation threshold value), the trend judgment of a plurality of consecutive sampling periods (such as 3 consecutive periods exceeding the threshold value) can be combined to avoid misjudgment.

[0151] Based on the above technical solution, the dynamic reference curve is generated through the LSTM model, and the multi-dimensional judgment of the deviation and the harmonic entropy value is combined, so that the accurate identification of the electrical parameter anomaly type is realized, which not only considers the deviation of the total energy consumption, but also takes into account the harmonic pollution and other power quality problems, and the threshold value can be dynamically adapted to different scenes, thereby improving the robustness and applicability of the anomaly judgment.

[0152] In a possible implementation manner of the embodiment of the present application, the S3 is specifically implemented by combining the S301, the S302 and the S303. Figure 1 As shown in the S301, the S302 and the S303, the S301, the S302 and the S303 can be combined to implement the S3. Figure 3 The S301, the S302 and the S303 are specifically implemented as follows.

[0153] The S301, the intelligent power distribution node detecting the harmonic pollution anomaly generates a trigger instruction containing a time stamp T event , and broadcasts the instruction to the intelligent power distribution nodes topologically adjacent through the ad hoc network.

[0154] The time stamp T eventTo detect the time of harmonic pollution anomaly; trigger instruction is used to inform the adjacent node to return the high frequency waveform data of a specific period, including request node identification, anomaly type and timestamp information.

[0155] In some implementations, the trigger instruction adopts a fixed data frame format, such as frame header + instruction type + timestamp + request node ID + check bit, and is transmitted to the adjacent nodes through the first channel of the ad hoc network; the judgment basis of the adjacent nodes is the adjacent node table of the node, i.e. the electrical distance ≤D max nodes.

[0156] It should be noted that the broadcast range of the trigger instruction can be limited to nodes within 2-3 hop paths to avoid network congestion, and the hop limit can be set by the path hop count H in the dynamic routing table.

[0157] For example, node A detects a harmonic pollution anomaly at 10:05:23, generates a trigger instruction containing timestamp T event =10:05:23, and broadcasts it to nodes B, C and D recorded in the adjacent node table through the ad hoc network, and the instruction explicitly requests to return the high frequency waveform data related to the timestamp.

[0158] S302, after receiving the trigger instruction, the adjacent intelligent power distribution node parses the timestamp T event , and determines that the time interval of the high frequency waveform data to be traced back is [T event -T pre , T event +T post ].

[0159] Wherein, T pre is the preset forward buffering time, i.e. the data duration before the anomaly occurs, and T post is the preset backward recording time, i.e. the data duration after the anomaly occurs; the time interval is used to ensure the complete waveform characteristics before and after the harmonic pollution occurs.

[0160] In some implementations, the values of T pre and T post are determined according to the harmonic propagation speed and the device response time: for ordinary building power distribution systems, T pre can be set to 1-3 seconds, and T post can be set to 2-5 seconds to capture the complete process from the generation to the diffusion of harmonics; the determination of the time interval ensures the consistency of the time reference of each node through the built-in clock synchronization mechanism of the node (such as time synchronization with the building coordination gateway BCG).

[0161] It should be noted that if the adjacent node does not store the high frequency waveform data of the time interval (such as the node just powers on or buffer overflow), it needs to return a data missing response to the request node to avoid the request node waiting timeout.

[0162] For example, after receiving the trigger instruction, the neighboring node B analyzes T event =10:05:23, if T pre =2 seconds, T post =3 seconds, it is determined that the time interval to be traced back is the high-frequency waveform data from 10:05:21 to 10:05:26.

[0163] S303, the adjacent intelligent power distribution node returns to the requesting node through the second channel of the ad hoc network after wavelet transform and quantization encoding compression processing of the high-frequency waveform data in the time interval.

[0164] Among them, wavelet transform is used to extract high-frequency coefficients containing harmonic characteristics and ignore weak coefficients dominated by noise, and quantization encoding is used to compress the retained coefficients to reduce data transmission volume; the second channel is a channel specially for transmitting high-frequency waveform data, and its working mechanism is based on real-time acquisition of noise spectrum and signal attenuation rate to achieve dynamic adaptation:

[0165] First, the channel will continuously monitor the noise distribution and signal attenuation in the current transmission environment, and adaptively allocate the optimal channel frequency band for the high-frequency waveform data from the preset frequency range, so as to avoid the influence on transmission quality caused by interference to fixed frequency band;

[0166] In the data transmission stage, the Orthogonal Frequency Division Multiplexing (OFDM) technology is used to divide the high-frequency waveform data processed by wavelet transform and quantization encoding into multiple parallel subcarrier signals, and the data transmission efficiency is improved by transmitting multiple subcarriers at the same time, and the influence of single carrier interference is reduced;

[0167] In order to further guarantee the reliability, forward error correction coding (FEC) technology is integrated into the transmission process, which can automatically detect and correct part of the errors in the transmission process by adding redundant check information in the data, so as to reduce the retransmission demand caused by data loss or error code.

[0168] In some implementations, the operation process of the second channel is as follows:

[0169] (1) Channel frequency band allocation: the second channel real-time acquires the noise spectrum and signal attenuation rate of the current transmission path, and selects the available frequency band with low noise and small attenuation from the preset frequency range (such as 300kHz-1MHz), and allocates a dedicated transmission frequency band for the high-frequency waveform data;

[0170] (2) Data segmentation and modulation: the compressed high-frequency waveform data is segmented into multiple sub-data streams according to the requirements of the Orthogonal Frequency Division Multiplexing technology, and is modulated onto different subcarriers, and the orthogonality between each subcarrier is maintained to avoid mutual interference;

[0171] (3) Error correction coding and transmission: FEC error correction coding is added to the modulated subcarrier signal to form a complete transmission frame, which is sent to the target node through the allocated frequency band;

[0172] (4) Reception and decoding: after receiving the signal through the second channel, the receiving node first corrects the signal using the FEC technology, and then restores each subcarrier signal to complete high-frequency waveform data through orthogonal frequency demodulation, finally completes the accurate reception of data.

[0173] Therefore, through the above mechanism, the second channel can provide efficient and reliable transmission support for high-frequency waveform data in a complex power line transmission environment, ensuring the data integrity and timeliness when multiple nodes cooperate to analyze harmonic pollution anomalies.

[0174] In some implementations, the wavelet transform can use db4 wavelet basis function, and the decomposition layer is set to 3-5 layers to balance the feature extraction accuracy and computational complexity; uniform quantization is used for quantization coding, and the quantization step is adaptively adjusted according to the waveform amplitude range, such as 0.01A when the amplitude range is ±10A; point-to-point communication is used for data transmission, and the highest weight path in the dynamic routing table is used for transmission.

[0175] It should be noted that the compressed high-frequency waveform data needs to include timestamp information to ensure that the requesting node can perform time alignment analysis on the data returned by different nodes.

[0176] For example, node B performs db4 wavelet transform (3-layer decomposition) on the current waveform data from 10:05:21 to 10:05:26, retains the high-frequency coefficients corresponding to the 1st to 5th harmonics, and compresses the data to 30% of the original data size after quantization coding, and transmits it back to node A through the adaptive frequency band (such as 300kHz) of the second channel. The accurate timestamp of each sampling point is attached in the data frame.

[0177] Based on the above technical solution, through the precise broadcast of the trigger instruction, the clear definition of the time interval, and the compression and transmission of the high-frequency waveform data, the efficient acquisition of high-frequency waveform data in the same time period is ensured, which not only guarantees the integrity and timeliness of the data, but also improves the transmission efficiency through compression and dedicated channel, and lays a reliable data foundation for subsequent analysis of harmonic propagation characteristics and positioning of abnormal nodes.

[0178] In one possible implementation of the embodiments of the present application, the above S4 specifically includes the following S401 to S403:

[0179] S401, taking timestamp T event as the reference, calculate the waveform time delay τ i of each intelligent power distribution node through cross-correlation analysis.

[0180] wherein, waveform time delay τ i is the time difference of high-frequency waveform propagation from the abnormal source to the i-th intelligent power distribution node; the cross-correlation analysis is to determine the time offset of the waveform reaching each node by calculating the similarity between the high-frequency waveform data collected by different nodes, so as to obtain the time delay value.

[0181] In some implementations, the time window of the cross-correlation analysis can be set as T pre +T post , that is, consistent with the time interval of the high-frequency waveform data, and the sliding step is set as 1 / 10 of the sampling period, such as 10 kHz of sampling frequency, the step is 10 μs, so as to improve the time delay calculation accuracy; for the waveform data with greater noise interference, filtering preprocessing can be performed first, such as using 50 Hz notch filter to remove power frequency interference.

[0182] S402, based on the waveform time delay, an abnormal source positioning equation and an over-determined equation set are constructed, and the Levenberg-Marquardt algorithm is used for solving, so as to obtain the abnormal source coordinates (x, y).

[0183] Firstly, the abnormal source positioning equation is constructed: ; (x, y) is the abnormal source coordinates, (xi, yi) is the coordinates of the i-th intelligent power distribution node, v is the carrier propagation speed, , and the measurement error is

[0184] Then, the over-determined equation set is established based on the time difference positioning technology TDOA: ; wherein, τ j is the waveform time delay of the j-th node, (x j , y j ) is the coordinates of the j-th node; finally, the Levenberg-Marquardt algorithm is used to solve the nonlinear equation set, and the optimal abnormal source coordinates are obtained.

[0185] In some implementations, the carrier propagation speed v is the propagation speed of electromagnetic wave in the power line, which is usually 2*10^8 m / s. The line material (such as copper cable, aluminum cable) can be fine-tuned; the initial iteration step of the Levenberg-Marquardt algorithm is set as 0.1, the initial value of the damping factor is set as 0.01, and the iteration is stopped when the residual sum of squares is less than 10^-6.

[0186] For example, the coordinates and corresponding time delays of nodes A (10, 20), B (30, 40), and C (20, 50) are selected to construct an over-determined equation set of 3 equations, and after 15 times of iteration by the Levenberg-Marquardt algorithm, the abnormal source coordinates are solved as (25, 35).

[0187] S403, match the abnormal source coordinate (x, y) with the loop range of the smart power distribution node to determine the abnormal node or line crosstalk fault.

[0188] The loop range of the smart power distribution node refers to the physical coverage area of the power distribution loop monitored by the node in the building space, which can be predefined by associating the loop topology in the building information model (BIM) with the coordinate range, such as a node corresponding to "3rd floor 302 room air conditioning loop", whose coordinate range is x∈[20, 30], y∈[30, 40].

[0189] In some implementations, the coordinate boundary of the loop range can be determined by the wall in the BIM model and the position of the power distribution box. If the abnormal source coordinate falls within the loop range of a node, the node is marked as an abnormal node. If the coordinate does not fall within the loop range of any node, it is determined to be a line crosstalk fault, i.e. the harmonics are coupled to the non-target loop through the line.

[0190] It should be noted that when the abnormal source coordinate is close to the loop range boundary of multiple nodes, the harmonic amplitude of each node can be combined to assist in determining the abnormal node.

[0191] For example, the abnormal source coordinate (25, 35) falls within the loop range of node B, i.e. x∈[25, 35], y∈[30, 45], so node B is marked as an abnormal node. If the coordinate is (15, 15) and does not fall within the loop range of any node, it is determined to be a line crosstalk fault.

[0192] Based on the above technical solution, the waveform delay is obtained by cross-correlation analysis, the positioning equation set is constructed to solve the coordinate, and the abnormal node is matched with the loop range to realize accurate positioning of the harmonic pollution abnormal source, which considers the time-space propagation characteristics of high-frequency waveforms and improves the positioning reliability by combining building physical space information, providing a clear target object for subsequent energy consumption regulation.

[0193] In one possible implementation of the embodiments of the present application, S5 specifically includes the following S501 to S503:

[0194] S501, generate corresponding energy consumption regulation instructions according to the electrical parameter abnormal type of the abnormal node.

[0195] The energy consumption regulation instruction is an instruction for controlling the abnormal node to adjust the power consumption state. The content is determined according to the abnormal type. If it is a harmonic pollution abnormality, the instruction contains the parameters of starting an active power filter (APF) (such as target harmonic number and compensation current amplitude). If it is a device overload abnormality, the instruction contains the steps of hierarchical unloading (such as the sequence of shutting down non-critical loads). If it is a device stop abnormality, the instruction contains maintenance alarm information (such as device identification and location).

[0196] In some implementations, the instruction adopts a standardized data format (such as frame header + abnormal type code + regulation parameter + check code). The abnormal type code is represented by 1 byte (such as 01H for harmonic pollution, 02H for overload, and 03H for stop). The value of the regulation parameter is dynamically adjusted according to the abnormal degree. For example, the compensation depth of the APF is set according to the harmonic current amplitude ratio when there is harmonic pollution (the compensation depth is set to 100% when the ratio is > 20%).

[0197] It should be noted that the instruction needs to contain a timestamp and a sending node identifier to ensure that the abnormal node can verify the timeliness and legitimacy of the instruction, and to avoid receiving expired or fake instructions.

[0198] For example, for a node with harmonic pollution abnormality, the generated regulation instruction has an abnormal type code of 01H, a regulation parameter of “target harmonic number: 3, 5, 7; compensation current amplitude: 15A”, and is accompanied by a sending timestamp and the identifier of node A.

[0199] S502, transmit the energy consumption regulation instruction to the abnormal node through power line carrier transmission through the first channel of the ad hoc network.

[0200] The first channel is a channel for transmitting basic data, is suitable for transmitting small amount of data information such as instructions, and adopts an anti-interference modulation method (such as binary phase shift keying BPSK) to ensure transmission reliability. When transmitting through power line carrier, the instruction signal is modulated and loaded onto the power line, and is forwarded to the abnormal node through the path with the highest weight in the dynamic routing table.

[0201] In some implementations, the instruction transmission adopts an acknowledgement mechanism, that is, the abnormal node needs to return an acknowledgement frame after receiving the instruction. The acknowledgement frame contains the instruction check result. If the sending node does not receive the acknowledgement within a preset time (such as 1 second), the instruction is automatically retransmitted. The transmission power can be dynamically adjusted according to the electrical distance between nodes, that is, the farther the distance, the higher the power, and the upper limit is not more than 30 dBm.

[0202] It should be noted that when the first channel is congested, the instruction can be temporarily transmitted through the second channel. The switching trigger condition is that the real-time packet loss rate of the first channel is > 5%.

[0203] For example, node A transmits the control instruction through the path with the highest weight in the ad hoc network (node A→node C→abnormal node B), adopts the BPSK modulation mode, and transmits the power of 20 dBm. After receiving the instruction, the abnormal node B returns an acknowledgement frame, and node A confirms that the instruction is delivered.

[0204] In S503, the abnormal node receives and parses the energy consumption control instruction, performs the corresponding energy consumption control operation, and feeds back the operation result.

[0205] The parsing process of the abnormal node includes verifying the integrity of the instruction by the check code and verifying the legality by the verification of the sending node identifier. After parsing, the driving execution unit (such as an active power filter APF, a relay, and a communication module) is driven to act. The operation result feedback includes the execution state (success / failure) and the current parameter (such as the harmonic suppression rate after the APF is started).

[0206] In some implementations, the energy consumption control includes:

[0207] If the abnormality is harmonic pollution, the execution operation includes: starting an active filter (APF), collecting current data of the abnormal loop in real time, extracting harmonic current components therefrom, generating a reverse harmonic current with the same amplitude and opposite phase as the harmonic current and injecting the reverse harmonic current into the loop, dynamically adjusting the reverse harmonic current through a closed-loop control mechanism (control period ≤10 ms), and ensuring that the harmonic suppression rate reaches 90% or more to eliminate harmonic pollution.

[0208] If the abnormality is device overload, the execution operation includes: according to the device priority list pre-configured by the device management platform and synchronized to the intelligent power distribution node, sequentially shutting down non-critical load devices (i.e., devices with a priority lower than a preset level) in order from low to high priority, and after each non-critical load is shut down, detecting whether the actual power of the loop returns to the normal range of the baseline power curve every 2 seconds, if not, the next non-critical load is shut down, and this process continues until the overload state is removed. If the overload state still exists after all non-critical loads are shut down, the abnormal loop is cut off to avoid overheating and damage to the line. The abnormal loop is determined by the loop identification information built-in the intelligent power distribution node and the positioning result of the abnormal node.

[0209] If the abnormality is device stop, the execution operation includes: the intelligent power distribution node pushes the device identifier (such as device number, model) associated with itself and the pre-stored installation physical location (such as “X layer X area power distribution box X loop associated air conditioning device”) to the management platform through the building coordination gateway BCG. After receiving, the management platform automatically triggers a maintenance alarm, generates a maintenance work order and assigns it to the relevant maintenance personnel, so as to timely investigate the cause of the device stop and restore operation.

[0210] It should be noted that if an operation fails (such as APF failure) during execution, the abnormal node needs to immediately send an alarm message to request manual intervention.

[0211] For example, after the abnormal node B parses the instruction and starts the APF, the 3rd and 5th harmonic currents are suppressed from 20A to 1.5A within 5 seconds, and the harmonic suppression rate reaches 92.5%. Subsequently, the node A is fed back with "execution success, current harmonic suppression rate 92.5%".

[0212] Based on the above technical solution, by generating targeted regulation instructions, reliable instruction transmission and accurate operation execution, efficient regulation of different types of electrical abnormalities is realized, which not only guarantees the safety and timeliness of instruction transmission, but also ensures the regulation effect through closed-loop control and result feedback, effectively improving the dynamic management capability of indoor building energy consumption.

[0213] Since the change of environmental temperature and humidity will affect the line resistance and device power factor, and the difference of line physical parameters will also cause energy consumption metering deviation, in order to realize the accuracy of indoor building energy consumption monitoring, in a possible implementation manner of the embodiment of the present application, the intelligent power distribution node will utilize the temperature and humidity compensation algorithm to correct the collected electrical parameters by the obtained temperature and humidity parameters and the pre-stored line physical parameters, and output the compensated power data, specifically including:

[0214] (1) Linearly compensate the current sampling value according to the temperature T to obtain the current compensation value I comp , and the calculation formula is

[0215] , wherein I raw is the current sampling value, a is the temperature coefficient of wire resistance, T ref is the reference temperature, which is 25°C by default, and can also be adjusted according to the average temperature environment of the building location, R ref is the line resistance at the reference temperature, and U is the line voltage.

[0216] (2) Correct the power factor according to the relative humidity RH to obtain the compensated power factor PF comp , and the calculation formula is PF comp = F raw × [1-β(RH-RH ref )], wherein P Fraw is the original power factor, β is the humidity coefficient of the insulating material, RH ref is the reference humidity, which is 50%RH by default, and can be set according to the average environmental humidity in the building;

[0217] (3) Calculate the line correction coefficient p based on the wire diameter d and the material resistivity r, and the calculation formula is p=(ρ ref ×d ref2) / (r x d2), where p ref is the reference resistivity, d ref is the reference line diameter, the default value is 1.5 mm2, which can be set according to the standard line diameter specification of the building power distribution line;

[0218] (4) The compensated power data is calculated according to the product of the current compensation value, the corrected power factor and the line correction coefficient.

[0219] In some implementations, the default a value of the copper conductor is 0.00393 / ℃, and the default a value of the aluminum conductor is 0.00429 / ℃, which is obtained by querying the physical property parameter table of the conductor material; the default β value of the polyvinyl chloride insulation material is 0.0015 / %RH, and the default β value of the rubber insulation material is 0.002 / %RH, which is obtained by statistical humidity characteristic experimental data of the insulation material; the default p ref value of copper is , and the default p ref value of aluminum is , which is obtained by querying the standard resistivity parameter manual of the conductor material.

[0220] Figures 4-6 The indoor building energy consumption monitoring and regulation method based on environmental data provided by the embodiments of the present application can be applied to an indoor building energy consumption monitoring and regulation system based on environmental data as shown in Figure 4 , which includes a plurality of intelligent power distribution nodes (IPDU), a building coordination gateway (BCG) and a management platform.

[0221] The intelligent power distribution node (IPDU) is the core monitoring and execution unit of the system, which is deployed in the loop of each power distribution box of the building, and each IPDU includes:

[0222] An electric energy metering module for collecting electrical parameters such as voltage, current, power (including power factor, harmonic distortion rate) of the power distribution loop;

[0223] A power line carrier communication module for realizing self-organizing network communication between nodes through the power distribution line, constructing a Mesh topology network, supporting point-to-point transmission of high-frequency waveform data and broadcasting of collaborative control instructions;

[0224] An embedded processor for executing temperature and humidity compensation algorithms, correcting the collected electrical parameters by using temperature and humidity parameters and pre-stored line physical parameters, and outputting compensated power data;

[0225] A temperature and humidity sensor for detecting environmental parameters such as temperature and relative humidity at the installation position of the IPDU in real time.

[0226] The building coordination gateway (BCG) is connected with each IPDU through a power line, is responsible for coordinating data interaction between distributed nodes, and comprises:

[0227] A carrier communication modulation and demodulation module is used for parsing ad hoc network data of the IPDU and converting the ad hoc network data into a standard network protocol.

[0228] A data aggregation processing module is used for associating building space topology data, generating a district energy consumption three-dimensional thermal map that integrates IPDU position information and a building information model (BIM), and dynamically displaying energy consumption data of each floor.

[0229] An external communication interface is used for connecting a management platform and a cloud platform, the cloud platform can store historical energy consumption data and use an LSTM network model to perform power prediction according to the historical energy consumption data of the equipment.

[0230] The management platform is in communication connection with the BCG, is used for realizing three-dimensional visual monitoring of indoor building energy consumption, and performing energy consumption regulation and control operations, such as receiving abnormal information uploaded by the BCG, issuing regulation and control instructions, and the like.

[0231] In some implementation manners, each IPDU is connected with each other through a power line carrier communication module, forms an ad hoc network covering each loop of the building, and realizes collaborative interaction of data and instructions; the BCG is connected with all IPDUs through a power line, aggregates node data and distributes control instructions; the management platform is in communication with the BCG through a wired or wireless manner, realizes remote management and monitoring of the entire system.

[0232] In some implementation manners, the intelligent power distribution node (IPDU) is arranged in a corresponding loop in a power distribution box of each floor according to the number of building power distribution loops, is directly connected to a loop power line to obtain power supply and transmit data; the building coordination gateway (BCG) can be installed in a building power distribution master control box or a weak current room, forms a communication link with each floor IPDU through a power line; the management platform can be deployed in a building management room (a local server) or a cloud, and establishes a stable connection with the BCG through a network.

[0233] Some data in the above formula are calculated by removing the dimension and taking the numerical value, the formula is obtained by software simulation of a large amount of collected data to obtain a formula closest to the real situation; the preset parameters and the preset threshold in the formula are set by a person skilled in the art according to the actual situation or are obtained by a large amount of data simulation.

[0234] Working principle of the application:

[0235] By deploying intelligent power distribution nodes in each loop of the building distribution box, a self-organizing network is established using power line carrier communication to realize data transmission and cooperation between nodes; the intelligent power distribution node collects electrical parameters and environmental temperature and humidity data, combines with pre-stored line physical parameter correction data to generate compensated power data; a reference power curve is generated based on historical data and environmental parameters, the deviation of actual power from reference power and harmonic energy distribution entropy value are analyzed to determine the type of electrical parameter anomaly; if it is a harmonic pollution anomaly, high-frequency waveform data of adjacent nodes in the same period is requested to analyze its space-time propagation characteristics to determine the abnormal node; finally, a control signal is sent to the abnormal node through the power line carrier, and corresponding energy consumption regulation and control operations are performed according to the type of anomaly, such as starting an active filter, hierarchical unloading load or pushing a maintenance alarm, so as to realize the monitoring and regulation of indoor building energy consumption.

[0236] The above embodiments are only used to illustrate the technical method of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical method of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical method of the present application.

Claims

1. A method for monitoring and controlling indoor building energy consumption based on environmental data, characterized in that, Intelligent power distribution nodes are deployed in each circuit within the building's power distribution box. Each intelligent power distribution node is connected to the power line of the corresponding circuit and establishes an ad hoc network through power line carrier communication. Determine the type of electrical parameter anomaly based on the data collected from the intelligent power distribution nodes; When the electrical anomaly parameter type is harmonic pollution anomaly, the smart distribution node requests high-frequency waveform data of the same time period from the topologically adjacent smart distribution nodes through the self-organizing network; After receiving the returned data from neighboring smart distribution nodes, the smart distribution node analyzes the spatiotemporal propagation characteristics of the high-frequency waveform data to identify abnormal nodes. Control signals are sent to abnormal nodes via power line carrier waves; these control signals are used to perform energy consumption regulation; wherein... The high-frequency waveform data for the same time period is: data sent at the node that detects an abnormal electrical parameter, including a timestamp T. event After the trigger command is executed, neighboring nodes backtrack the stored data in T format. event Based on [ Tevent -T pre ,T event +T post The waveform data within the interval, wherein the waveform data within the interval is obtained through wavelet transform and quantization encoding compression; wherein, T pre T represents the preset forward buffer time. post Indicates the preset backward recording time; The analysis of the spatiotemporal propagation characteristics of high-frequency waveform data includes: Based on the timestamp Tevent, the waveform delay of each node is calculated through cross-correlation analysis; The anomaly source localization equation is constructed as follows: Where (x,y) are the coordinates of the anomaly source, and τ i Let (x) be the waveform delay of the i-th smart distribution node. i ,y i Let be the coordinates of the i-th smart distribution node, and v be the carrier propagation speed. For measurement error; Establishing an overdetermined system of equations based on Time Difference-of-Origin (TDOA) positioning technology: ; where τ j Let (x) be the waveform delay of the j-th smart distribution node. j ,y j Let ) be the coordinates of the j-th smart distribution node; The coordinates of the anomaly source were obtained by solving the overdetermined system of equations using the Levenberg-Marquardt algorithm. If the coordinates of the abnormal source are within the loop range of smart distribution node i, then smart distribution node i is marked as an abnormal node; otherwise, it is marked as a line crosstalk fault.

2. The method for monitoring and controlling indoor building energy consumption based on environmental data according to claim 1, characterized in that, The establishment of an ad hoc network via power line carrier communication includes: After each smart distribution node is powered on, it broadcasts a beacon frame containing its own identifier and the parent node's code; The electrical distance D between nodes is calculated based on the carrier signal transmission delay Δt and the carrier propagation speed v: D = V × Δt; Screening electrical distance ≤ D max The nodes construct a neighbor list; where D max The maximum adjacent distance is preset; Based on the neighbor node table, the communication packet loss rate L, signal strength RSSI, and path hop count H of each node in the multi-hop path are calculated, and the dynamic routing weight W of each path is calculated; the multi-hop path refers to a path containing multiple nodes. The paths are sorted in descending order according to the dynamic routing weight W, and the top N paths with the highest weights are retained to form a dynamic routing table. Each node uses the dynamic routing table to forward data and adjust the network topology, thus obtaining an ad hoc network.

3. The method for monitoring and controlling indoor building energy consumption based on environmental data according to claim 2, characterized in that, In the self-organizing network, when the dynamic routing weight W < W th Automatically switches to the backup relay node for communication; among which, W th This indicates a preset weight threshold. The backup relay node is another neighboring node that is pre-registered in the dynamic routing table and maintains a communication connection with the current node.

4. The method for monitoring and controlling indoor building energy consumption based on environmental data according to claim 1, characterized in that, The self-organizing network includes a first channel and a second channel. The first channel is used to transmit basic data, which includes temperature and humidity parameters, electrical parameters, line physical parameters, and compensated energy consumption data. The second channel adaptively allocates a channel band within a preset frequency range for the transmitted high-frequency waveform data based on real-time collected noise spectrum and signal attenuation rate. It uses orthogonal frequency division multiplexing technology to divide the transmitted high-frequency waveform data into multiple subcarriers for transmission, and incorporates forward error correction coding (FEC) technology into the transmission process to ensure transmission reliability.

5. The method for monitoring and controlling indoor building energy consumption based on environmental data according to claim 1, characterized in that, The step of determining the abnormal type of electrical parameters based on the collected data from the smart power distribution node includes: Using an LSTM network model, a baseline power curve for each hour is generated based on historical energy consumption data stored in smart power distribution nodes, as well as corresponding calendar attributes, temperature data, and humidity data. Based on the reference power curve, calculate the actual power value P(t) and the reference power value P. base The deviation δ of (t) is calculated as: δ=[P(t)-P base (t)] / P base (t)×100%, where t represents time, and P base (t)∈Reference Power Curve; Perform a Fourier transform on the acquired current waveform and calculate the entropy value E of the transformed harmonic energy distribution. The formula for calculating the entropy value is: , n represents the harmonic order, N h p represents the highest harmonic order analyzed. n This indicates the proportion of energy of the nth harmonic; The electrical parameters are determined to be abnormal based on the deviation, the entropy value, and the preset threshold range. The abnormal electrical parameters include equipment overload abnormality, equipment shutdown abnormality, and harmonic pollution abnormality.

6. The method for monitoring and controlling indoor building energy consumption based on environmental data according to claim 1, characterized in that, The intelligent power distribution node uses the acquired temperature and humidity parameters and pre-stored line physical parameters to correct the collected electrical parameters, and outputs compensated power data, including: The current compensation value is obtained by linearly compensating the current sampling value based on the temperature T; The power factor is corrected based on the relative humidity (RH) to obtain the compensated power factor. Calculate the line correction coefficient based on wire diameter d and material resistivity r; The compensated power data is calculated by multiplying the current compensation value, the corrected power factor, and the line correction coefficient.

7. An indoor building energy consumption monitoring and control system based on environmental data, applied to the indoor building energy consumption monitoring and control method based on environmental data as described in any one of claims 1-6, characterized in that, This includes multiple smart power distribution nodes, building coordination gateways, and a management platform. The intelligent power distribution node IPDU is deployed in the circuits of each power distribution box in the building, and includes: An energy metering module is used to collect electrical parameters of the power distribution circuit, including voltage, current and power parameters; The communication module is used to transmit data via power line carrier waves to realize self-organizing network communication between IPDU nodes; the self-organizing network communication means that the IPDU nodes build a mesh topology network through power lines, and the network supports point-to-point transmission of high-frequency waveform data and broadcasting of cooperative control commands; An embedded processor is used to execute a temperature and humidity compensation algorithm, which uses the acquired temperature and humidity parameters and pre-stored circuit physical parameters to correct the collected electrical parameters. A humidity and temperature sensor is used to detect environmental parameters at the IPDU installation location in real time, including temperature and relative humidity. The building coordination gateway (BCG) is connected to each IPDU via power lines to coordinate data interaction between distributed IPDU nodes, enabling device-level energy consumption metering and anomaly location. The management platform is connected to BCG to realize three-dimensional visualization of indoor building energy consumption and to perform energy consumption control.

8. The indoor building energy consumption monitoring and control system based on environmental data according to claim 7, characterized in that, The Building Coordination Gateway (BCG) includes: The carrier communication modulation and demodulation module is used to parse the self-organizing network data of the intelligent power distribution node IPDU and convert the self-organizing network data into standard network protocols; The data aggregation and processing module is used to associate IPDU nodes with building space topology data to generate regional energy consumption maps. The regional energy consumption map represents a three-dimensional heat map of energy consumption that integrates the location information of IPDU nodes and the building information model (BIM) model, and dynamically displays the energy consumption data of each floor. An external communication interface is used to connect the management platform and the cloud platform; the cloud platform is used to store historical energy consumption data and use the LSTM network model to perform power prediction based on the device's historical energy consumption data.

Images