Electromechanical equipment dynamic tracking method based on multivariate heterogeneous data and edge calculation

Abstract

The invention belongs to the field of equipment operation and maintenance monitoring, and particularly relates to an electromechanical equipment dynamic tracking method based on multivariate heterogeneous data and edge computing, which comprises the following steps: presetting an equipment operation and maintenance monitoring network, and determining a distributed edge hierarchical control chain according to electromechanical equipment distribution and operation attribute data in the equipment operation and maintenance monitoring network, the distributed edge hierarchical control chain is constructed by combining a device distribution state, an operation load and a monitoring device correlation coefficient with a graph algorithm, and through combination of the distributed edge hierarchical control chain and monitoring data, a distributed enhanced control model is used to acquire a multi-level tracking control space and a bidirectional feedback control instruction. Uplink feedback and downlink electromechanical instruction control of the electromechanical equipment are realized; by means of distributed edge hierarchy control chain construction and model application, equipment data can be effectively integrated, accurate monitoring and control over electromechanical equipment are achieved, the equipment operation and maintenance efficiency is improved, and the equipment operation stability and reliability are enhanced.

Description

Technical Field

[0001] The present invention belongs to the field of equipment operation and maintenance monitoring, and in particular to a method for dynamic tracking of electromechanical equipment based on multi-dimensional heterogeneous data and edge computing. Background Art

[0002] The application of electromechanical equipment in various fields is becoming more and more extensive and important, such as manufacturing, energy industry, transportation, etc. The normal operation of electromechanical equipment is directly related to production efficiency, product quality and production safety. Under the background of Industry 4.0 and intelligent manufacturing, real-time and accurate monitoring and tracking of electromechanical equipment has become a key requirement to ensure stable production; however, the existing electromechanical equipment monitoring technology has many shortcomings and defects. On the one hand, most of the existing technologies monitor equipment from a single level. For large and complex electromechanical equipment, it contains multiple hierarchical structures such as equipment level and component level, but most monitoring methods only focus on tracking and monitoring equipment from the overall level, and it is difficult to go deep into the subtle component level, resulting in the inability to timely discover potential faults and performance degradation at the component level, and the irregular distribution of monitoring equipment leads to data confusion, which increases the difficulty of processing; on the other hand, for some widely distributed large-scale equipment, the current control and monitoring methods are too centralized, and data needs to be transmitted to the central control center for processing and analysis, which not only increases the burden of data transmission, but also causes delays in monitoring and feedback, and cannot respond to abnormal conditions of equipment in time. Summary of the invention

[0003] In view of the shortcomings of the prior art, the present invention proposes a dynamic tracking method for electromechanical equipment based on multi-heterogeneous data and edge computing, including: pre-setting an equipment operation and maintenance monitoring network, determining a distributed edge hierarchical control chain according to the distribution and operation attribute data of the electromechanical equipment in the equipment operation and maintenance monitoring network, wherein the distributed edge hierarchical control chain is constructed by combining the equipment distribution status, the operating load, and the correlation coefficient of the monitoring equipment with a graph algorithm, and the distributed edge hierarchical control chain is combined with the monitoring data to obtain a multi-level tracking control space and a two-way feedback control instruction using a distributed enhanced control model, thereby realizing the upward feedback and downward electromechanical instruction control of the electromechanical equipment; the present invention can effectively integrate equipment data with the help of distributed edge hierarchical control chain construction and model application, realize accurate monitoring and control of electromechanical equipment, improve equipment operation and maintenance efficiency, and enhance equipment operation stability and reliability.

[0004] To achieve the above object, the present invention provides the following technical solutions: The dynamic tracking method of electromechanical equipment based on multi-heterogeneous data and edge computing includes: Preset an equipment operation and maintenance monitoring network, and determine a distributed edge hierarchical control chain according to the distribution and operation attribute data of electromechanical equipment in the equipment operation and maintenance monitoring network; Through the distributed edge hierarchical control chain combined with the data monitored by the equipment operation and maintenance monitoring network, a multi-level tracking control space and a two-way feedback control instruction are obtained through a configured distributed enhanced control model; Based on multi-level tracking control space and bidirectional feedback control instructions, the electromechanical equipment is fed back upstream and the electromechanical instructions are controlled downstream; The distributed edge hierarchical control chain is constructed by combining the graph algorithm with the distribution status of electromechanical equipment, the operating load and the correlation coefficient of the corresponding monitoring equipment under each hierarchical control node in the equipment operation and maintenance monitoring network.

[0005] Specifically, the equipment operation and maintenance monitoring network includes a first-level directed electromechanical equipment simulation subnet and a second-level monitoring node chain; The steps of constructing the first-level directed electromechanical equipment simulation subnet include: Obtain the distribution location and attributes of electromechanical equipment and operation-related dynamic parameters, use each subcomponent in the electromechanical equipment as a node in the equipment operation and maintenance monitoring network, and obtain a first-level node set; The distribution location and attributes of the electromechanical equipment include the spatial coordinates, function labels and operation-related dynamic parameters corresponding to each subcomponent; the operation-related dynamic parameters include: the energy transmission path, control instruction transmission priority and electromagnetic coupling strength corresponding to each subcomponent; According to the energy transmission path, electromagnetic coupling strength, energy transmission efficiency and control instruction transmission priority among the dynamic parameters associated with the operation of each node of the electromechanical equipment, the directed association probability value between each node is obtained through the Bayesian algorithm; The directed association probability value between each node is used as the causal relationship verification factor between two nodes. The causal relationship between each node is verified through the causal analysis algorithm to obtain the causal node set corresponding to the first-level node set.

[0006] Specifically, the steps of constructing the first-level directed electromechanical equipment simulation subnet also include: Based on the directed association probability value between each node and the preset association degree interval, the directed association strength between each node is obtained; According to the nodes with causal relationships and the nodes without causal relationships in the first-level node set, the directed association strength between the nodes is divided into directed causal association strength and ordinary directed association strength; The directed causal association strength and the ordinary directed association strength are used as the directed connection relationship between each node and combined with the first-level node set and the corresponding position coordinates. The first-level directed electromechanical equipment simulation subnet is obtained through the hypergraph algorithm. At the same time, the directed connection relationship is annotated with the corresponding directed causal association strength and ordinary directed association strength through the automatic labeling algorithm.

[0007] Specifically, the steps of constructing the first-level directed electromechanical equipment simulation subnet also include: Based on the type and association strength of the directed connection relationship between the nodes of the first-level directed electromechanical equipment simulation subnet, a color mapping function of each type of directed association relationship and the corresponding association strength is constructed through a preset color mapping table and a kernel function in a support vector machine; The color mapping table includes color types and color level gradients; The color mapping function of each type of directed association relationship and the corresponding association strength is built into the directed connection relationship between the corresponding nodes; At the same time, the first-level directed electromechanical equipment simulation subnet after the built-in color mapping function is input into the three-dimensional simulation algorithm and combined with the HSL model. According to the type of directed connection relationship between each node and the change in correlation strength, the color of the directed connection relationship between each node of the first-level directed electromechanical equipment simulation subnet and the color corresponding brightness and chromaticity changes are simulated and verified by training to obtain a three-dimensional first-level directed electromechanical equipment simulation subnet.

[0008] Specifically, each connection relationship in the first-level directed electromechanical device simulation subnet has a built-in strength adjustment factor; the strength adjustment factor includes a positive strength adjustment factor and a negative strength adjustment factor; The intensity adjustment factor is obtained by fitting a multiple regression function based on the energy transfer delay change rate, the electromagnetic coupling intensity change rate and the control delay change rate, as well as the positive and negative signs corresponding to the energy transfer delay change rate, the electromagnetic coupling intensity change rate and the control delay change rate and the brightness and chromaticity of the color corresponding to the directed connection relationship.

[0009] Specifically, the steps for building the secondary monitoring node chain include: Obtaining the fault frequency corresponding to each node in the three-dimensional first-level directed electromechanical equipment simulation subnet and the corresponding associated fault nodes and the corresponding associated fault frequencies; Based on the acquired fault frequency corresponding to each node and the corresponding associated fault nodes and the corresponding associated fault frequencies, probability clustering is performed through an adaptive clustering algorithm combined with a Bayesian algorithm to obtain a set of associated fault nodes; According to the associated fault node set, the probability of the corresponding associated fault node failing is simulated by the Bayesian algorithm when a node in the corresponding associated fault node set fails, using the three-dimensional simulation algorithm in the three-dimensional first-level directed electromechanical equipment simulation subnet; Based on the probability of failure of the corresponding associated fault node and the preset associated fault accuracy threshold, probability clustering simulation training is performed to obtain the set of all associated fault nodes that meet the associated fault accuracy threshold.

[0010] Specifically, the steps of constructing the secondary monitoring node chain also include: Obtain the node with the largest data collection load for each associated fault node set as the load master node for the regional division of the corresponding associated fault node set, and obtain the regional division load master node set; Based on the directed causal association strength and common directed association strength of the corresponding mark of the directed connection relationship corresponding to the associated fault node set, the spatial distance between the edge nodes between the associated fault node sets, the functional consistency and the causal association strength and the maximum load processing energy corresponding to each tracking control node in the distributed edge hierarchical control chain, combined with the conditions of minimizing the directed association strength between the partitioned regions and maximizing the directed association strength between each associated fault node set in the region, the regional partition load main node set is divided into regions through a hypergraph partitioning algorithm to obtain a partitioned regional partition load main node subset space; Based on the divided area, the load main node subset space is divided into two parts. Combined with the intensity adjustment factor and the maximum load processing energy corresponding to each tracking control node in real time, the boundary range corresponding to each area divided load main node subset space is adjusted in real time through a three-dimensional simulation algorithm combined with a hypergraph segmentation algorithm.

[0011] Specifically, the steps of constructing the secondary monitoring node chain also include: A data monitoring point is set based on each completed regional partition load main node subset space to obtain a data monitoring point set, and a directed monitoring connection relationship is constructed based on the data transmission direction and control instruction execution priority order between the regional partition load main node subset spaces corresponding to the data monitoring point set; Based on the directed monitoring connection relationship and the data monitoring point set, a secondary monitoring node chain is obtained; Based on the secondary monitoring node chain, a random data collection index chain is constructed between each data monitoring point in the secondary monitoring node chain and each associated fault node set in the corresponding regional partition load main node subset space through a random algorithm combined with a discrimination algorithm, and data collection rules are set for the random data collection index chain corresponding to the same data monitoring point.

[0012] Specifically, the steps for building a distributed edge-level control chain include: A distributed federated tracking control model is constructed based on a distributed federated algorithm combined with a reinforcement learning algorithm, and an edge tracking control sub-model in the distributed federated tracking control model is deployed to each data monitoring point to obtain an edge control node set; Based on the correlation coupling and equipment subcomponent operation priority between the nodes corresponding to the load main node subset space in each area of ​​the data monitoring point set, an information sharing connection relationship corresponding to each two edge control nodes in the edge control node set is constructed; Based on the edge control node set and the information sharing connection relationship, a distributed edge hierarchical control chain is obtained.

[0013] Specifically, the data collection rules include: For all nodes with causal relationships in the associated fault node set in the main node subset space of the same area partition load and nodes with medium or above ordinary directed association strength, set the same data collection frequency; The random data collection index chain corresponding to each associated fault node set only randomly collects data from one node in the associated fault node set at a time, and marks the collected data with the label sequence number corresponding to the associated fault node set; Configure the intensity adjustment factor and the acquisition frequency fluctuation mapping function of the data acquisition frequency, and configure the acquisition frequency fluctuation mapping function to all random data acquisition index chains; When the current intensity adjustment factor changes, the data acquisition frequency corresponding to the random data acquisition index chain is adjusted in real time through the acquisition frequency fluctuation mapping function.

[0014] Specifically, the steps of obtaining the multi-level tracking control space and the bidirectional feedback control instructions include: Based on the three-dimensional first-level directed electromechanical equipment simulation subnet, the electromechanical equipment operation simulation is performed, and at the same time, data is collected from the nodes corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet through the dynamic collection frequency configured by the second-level monitoring node chain combined with the collection frequency fluctuation mapping function to obtain a simulation tracking data set space; The corresponding area in the simulated tracking data set space is divided into the data set obtained by the load main node subset space, and synchronously input to the corresponding edge control node. The running status of the corresponding collection node and the associated fault node set is judged by the configured discrimination algorithm, and the multi-level tracking control space is obtained and the label serial number of the collection node corresponding to the associated fault node set is marked; Based on the corresponding acquisition nodes in the multi-level tracking control space and the judgment accuracy of the fault operation status of the associated fault node set, combined with the preset judgment accuracy threshold, the intensity adjustment factor and the dynamic acquisition frequency are adjusted through a three-dimensional simulation algorithm, and the overall electromechanical equipment operation status tracking process corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet, the second-level monitoring node chain and the distributed edge level control chain is subjected to cyclic simulation training; When the judgment accuracy of the corresponding acquisition node and the fault operation status of the associated fault node set is greater than a preset judgment accuracy threshold, the judgment result corresponding to each edge control node is output.

[0015] Specifically, the steps of obtaining the multi-level tracking control space and the bidirectional feedback control instructions also include: According to the judgment result corresponding to each edge control node, a corresponding bidirectional feedback control instruction is generated through the edge tracking control sub-model configured for the corresponding edge control node; Sharing the bidirectional feedback control instruction along the direction of the distributed edge hierarchical control chain according to the priority of the corresponding edge control node in the distributed edge hierarchical control chain; At the same time, the bidirectional feedback control instruction is transmitted along the random data acquisition index chain to the nodes in the associated fault node set corresponding to the identified fault for associated control and fault warning, and the corresponding fault warning is fed back and uploaded to the edge warning interface. At the same time, according to the mapping relationship between the fault severity and the intensity adjustment factor, the color type and color depth of the directed connection relationship corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet are dynamically adjusted and displayed in real time.

[0016] Specifically, the step of determining the operating status of the corresponding collection node and the associated fault node set includes: If the corresponding acquisition node is judged to have a fault, the associated fault node is secondary judged by using the Bayesian algorithm and the associated fault accuracy threshold, and the fault judgment result and location information corresponding to the acquisition node and the associated fault node that is secondary judged to have a fault are output; If the corresponding collection node determines that there is no fault but there is a fault in the associated fault node set, the location information and fault result corresponding to the collection node with fault are obtained by using the corresponding Bayesian algorithm and the associated fault accuracy threshold between the nodes in the associated fault node set.

[0017] Compared with the prior art, the present invention has the following beneficial effects: In view of the deficiencies of the prior art, the present invention obtains detailed properties of each sub-component of electromechanical equipment through multivariate heterogeneous data, constructs a first-level directed electromechanical equipment simulation subnet, and determines node associations and causal relationships with the help of Bayesian algorithms, causal analysis algorithms, etc., thereby overcoming the difficulty of in-depth component-level monitoring in the prior art. At the same time, a distributed edge hierarchical control chain is built using a distributed federation algorithm and a reinforcement learning algorithm, and the edge tracking control sub-model is deployed to the data monitoring point, solving the problems of heavy burden of centralized control data transmission and delayed monitoring feedback. The secondary monitoring node chain performs regional division and data collection based on fault frequency and correlation conditions, combines strength adjustment factors and dynamic collection frequency, and cooperates with three-dimensional simulation algorithm cyclic training to achieve accurate identification of the equipment's operating status, generate and transmit two-way feedback control instructions, and realize associated control and fault warning, which comprehensively improves the accuracy and real-time performance of monitoring and control of large and complex electromechanical equipment, effectively avoids potential faults, and reduces the difficulty of data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1This is a flow chart of a method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to an embodiment of the present invention; Figure 2 This is an architecture diagram of a three-dimensional first-level directed electromechanical device simulation subnet according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In the automated production lines of large factories, numerous electromechanical devices are widely distributed and have complex structures, including multi-layer structures such as equipment level and component level. Existing monitoring methods are difficult to deeply monitor subtle components, and centralized control leads to data transmission delays, making it impossible to detect component failures and abnormal responses in a timely manner. For this reason, please refer to Figure 1 , an embodiment provided by the present invention: a method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing, the steps comprising: S1. Preset an equipment operation and maintenance monitoring network, and determine a distributed edge hierarchical control chain according to the distribution and operation attribute data of electromechanical equipment in the equipment operation and maintenance monitoring network; S2. Acquire multi-level tracking control space and bidirectional feedback control instructions through the distributed edge hierarchical control chain combined with the data monitored by the equipment operation and maintenance monitoring network and through the configured distributed enhanced control model; S3, based on the multi-level tracking control space and bidirectional feedback control instructions, the electromechanical equipment is fed back and the electromechanical instructions are controlled downward; The distributed edge hierarchical control chain is constructed by combining the graph algorithm with the distribution status of electromechanical equipment, the operating load and the correlation coefficient of the corresponding monitoring equipment under each hierarchical control node in the equipment operation and maintenance monitoring network.

[0020] For further information, see Figure 2 , the equipment operation and maintenance monitoring network in this embodiment includes a first-level directed electromechanical equipment simulation subnet and a second-level monitoring node chain.

[0021] Furthermore, the steps of constructing the first-level directed electromechanical device simulation subnet in this embodiment include: Obtain the distribution location and attributes of electromechanical equipment and operation-related dynamic parameters, use each subcomponent in the electromechanical equipment as a node in the equipment operation and maintenance monitoring network, and obtain a first-level node set; The distribution location and attributes of the electromechanical equipment include the spatial coordinates, function labels and operation-related dynamic parameters corresponding to each subcomponent; the operation-related dynamic parameters include: the energy transmission path, control instruction transmission priority and electromagnetic coupling strength corresponding to each subcomponent; In this embodiment, the air-conditioning control system in a shopping mall is taken as an example, and each sub-component in each air-conditioning and ventilation system, such as fans, compressors, valves, sensors, etc., is taken as a node to obtain its spatial coordinates (specific location in the equipment room or pipeline), functional labels (such as refrigeration, ventilation, monitoring, etc.) and operation-related dynamic parameters (energy transmission path such as electrical energy to mechanical energy to drive the fan, control instruction transmission priority, i.e. which control instructions are executed first, and electromagnetic coupling strength mainly for electromagnetic interference between electrical components).

[0022] According to the energy transmission path, electromagnetic coupling strength, energy transmission efficiency and control instruction transmission priority among the dynamic parameters associated with the operation of each node of the electromechanical equipment, the directed association probability value between each node is obtained through the Bayesian algorithm; The directed association probability value between each node is used as the causal relationship verification factor between two nodes. The causal relationship between each node is verified through the causal analysis algorithm to obtain the causal node set corresponding to the first-level node set.

[0023] Based on the directed association probability value between each node and the preset association degree interval, the directed association strength between each node is obtained; According to the nodes with causal relationships and the nodes without causal relationships in the first-level node set, the directed association strength between the nodes is divided into directed causal association strength and ordinary directed association strength; The directed causal association strength and the ordinary directed association strength are used as the directed connection relationship between each node and combined with the first-level node set and the corresponding position coordinates. The first-level directed electromechanical equipment simulation subnet is obtained through the hypergraph algorithm. At the same time, the directed connection relationship is annotated with the corresponding directed causal association strength and ordinary directed association strength through the automatic labeling algorithm.

[0024] Based on the type and association strength of the directed connection relationship between the nodes of the first-level directed electromechanical equipment simulation subnet, a color mapping function of each type of directed association relationship and the corresponding association strength is constructed through a preset color mapping table and a kernel function in a support vector machine; The color mapping table includes color types and color level gradients; further, the color types and color level gradients in this embodiment are customized by those skilled in the art according to usage habits.

[0025] The color mapping function of each type of directed association relationship and the corresponding association strength is built into the directed connection relationship between the corresponding nodes; At the same time, the first-level directed electromechanical equipment simulation subnet after the built-in color mapping function is input into the three-dimensional simulation algorithm and combined with the HSL model. According to the type of directed connection relationship between each node and the change in correlation strength, the color of the directed connection relationship between each node of the first-level directed electromechanical equipment simulation subnet and the color corresponding brightness and chromaticity changes are simulated and verified by training to obtain a three-dimensional first-level directed electromechanical equipment simulation subnet.

[0026] Furthermore, in this embodiment, a strength adjustment factor is built into each connection relationship in the first-level directed electromechanical device simulation subnet; the strength adjustment factor includes a positive strength adjustment factor and a negative strength adjustment factor; The intensity adjustment factor in this embodiment is obtained by fitting a multivariate regression function based on the energy transfer delay change rate, the electromagnetic coupling intensity change rate and the control delay change rate, as well as the positive and negative signs corresponding to the energy transfer delay change rate, the electromagnetic coupling intensity change rate and the control delay change rate, and the brightness and chromaticity of the color corresponding to the directed connection relationship.

[0027] Furthermore, in this embodiment, when constructing a first-level directed electromechanical equipment simulation subnet in the equipment operation and maintenance monitoring network, the process obtains electromechanical equipment attributes from multiple dimensions, providing a basis for fully understanding the equipment status. By using the Bayesian algorithm and causal analysis algorithm, the directed association probability value and causal relationship are determined through the node operation associated dynamic parameters. Compared with traditional methods, it can deeply explore the potential causal relationship between components, accurately locate the source of the fault, and prevent the fault in advance.

[0028] This process obtains the directed association strength based on the association probability value combined with the preset association degree interval, further divides the directed causal association strength and the ordinary directed association strength, and constructs and annotates the simulated subnet with the help of the hypergraph algorithm and the automatic annotation algorithm to clearly show the connection relationship and association degree of each component, so that the operation and maintenance personnel can quickly locate the key nodes and improve the efficiency of troubleshooting; secondly, the mapping function is constructed and built-in based on the color mapping table and the support vector machine kernel function, and the simulation verification training is carried out in combination with the three-dimensional simulation algorithm and the HSL model to convert the abstract association relationship and strength into intuitive color, brightness and chromaticity changes, so that the operation and maintenance personnel can intuitively understand the dynamics of the equipment and find abnormalities in time; thirdly, the strength adjustment factor is obtained by fitting the multivariate regression function from the energy transfer delay change rate, the electromagnetic coupling strength change rate and the control delay change rate, and the directed connection relationship strength can be dynamically adjusted according to the real-time operation status of the equipment, so that the simulated subnet is more in line with the actual operation situation, which greatly enhances the real-time and accuracy of monitoring, and provides a strong guarantee for the stable operation and efficient operation and maintenance of the equipment.

[0029] Furthermore, the steps of constructing the secondary monitoring node chain in this embodiment include: Obtaining the fault frequency corresponding to each node in the three-dimensional first-level directed electromechanical equipment simulation subnet and the corresponding associated fault nodes and the corresponding associated fault frequencies; Based on the acquired fault frequency corresponding to each node and the corresponding associated fault nodes and the corresponding associated fault frequencies, probability clustering is performed through an adaptive clustering algorithm combined with a Bayesian algorithm to obtain a set of associated fault nodes; According to the associated fault node set, a three-dimensional simulation algorithm in a three-dimensional first-level directed electromechanical equipment simulation subnet is used to simulate the failure of a node in the corresponding associated fault node set, and the probability of the corresponding associated fault node failing is obtained by using a Bayesian algorithm; Based on the probability of failure of the corresponding associated fault node and the preset associated fault accuracy threshold, probability clustering simulation training is performed to obtain the set of all associated fault nodes that meet the associated fault accuracy threshold.

[0030] Obtain the node with the largest data collection load for each associated fault node set as the load master node for the regional division of the corresponding associated fault node set, and obtain the regional division load master node set; Based on the directed causal association strength and common directed association strength of the corresponding mark of the directed connection relationship corresponding to the associated fault node set, the spatial distance between the edge nodes between the associated fault node sets, the functional consistency and the causal association strength and the maximum load processing energy corresponding to each tracking control node in the distributed edge hierarchical control chain, combined with the conditions of minimizing the directed association strength between the partitioned regions and maximizing the directed association strength between each associated fault node set in the region, the regional partition load main node set is divided into regions through a hypergraph partitioning algorithm to obtain a partitioned regional partition load main node subset space; Based on the partitioned regional load main node subset space, combined with the strength adjustment factor and the maximum load processing energy corresponding to each tracking control node in real time, the boundary range corresponding to each regional load main node subset space is adjusted in real time through a three-dimensional simulation algorithm combined with a hypergraph partitioning algorithm; Furthermore, in this embodiment, when the strength adjustment factor is a positive strength adjustment factor, the boundary range corresponding to the corresponding regional division load main node subset space becomes larger at the data processing load range of the corresponding data monitoring point, and vice versa.

[0031] A data monitoring point is set based on each completed regional partition load main node subset space to obtain a data monitoring point set, and a directed monitoring connection relationship is constructed based on the data transmission direction and control instruction execution priority order between the regional partition load main node subset spaces corresponding to the data monitoring point set; Based on the directed monitoring connection relationship and the data monitoring point set, a secondary monitoring node chain is obtained; Based on the secondary monitoring node chain, a random data collection index chain is constructed between each data monitoring point in the secondary monitoring node chain and each associated fault node set in the corresponding regional partition load main node subset space through a random algorithm combined with a discrimination algorithm, and data collection rules are set for the random data collection index chain corresponding to the same data monitoring point.

[0032] Furthermore, the data collection rules in this embodiment include: For all nodes with causal relationships in the associated fault node set in the main node subset space of the same area partition load and nodes with medium or above ordinary directed association strength, set the same data collection frequency; The random data collection index chain corresponding to each associated fault node set only randomly collects data from one node in the associated fault node set at a time, and marks the collected data with the label sequence number corresponding to the associated fault node set; Configure the intensity adjustment factor and the acquisition frequency fluctuation mapping function of the data acquisition frequency, and configure the acquisition frequency fluctuation mapping function to all random data acquisition index chains; When the current intensity adjustment factor changes, the data acquisition frequency corresponding to the random data acquisition index chain is adjusted in real time through the acquisition frequency fluctuation mapping function.

[0033] Furthermore, in this embodiment, when the change corresponding to the intensity adjustment factor exceeds the normal change threshold, the corresponding acquisition frequency also increases continuously.

[0034] Furthermore, the steps of constructing the distributed edge level control chain in this embodiment include: A distributed federated tracking control model is constructed based on a distributed federated algorithm combined with a reinforcement learning algorithm, and an edge tracking control sub-model in the distributed federated tracking control model is deployed to each data monitoring point to obtain an edge control node set; Based on the correlation coupling and equipment subcomponent operation priority between the nodes corresponding to the load main node subset space in each area of ​​the data monitoring point set, an information sharing connection relationship corresponding to each two edge control nodes in the edge control node set is constructed; Based on the edge control node set and the information sharing connection relationship, a distributed edge hierarchical control chain is obtained.

[0035] Furthermore, the steps of acquiring the multi-level tracking control space and the bidirectional feedback control instructions in this embodiment include: Based on the three-dimensional first-level directed electromechanical equipment simulation subnet, the electromechanical equipment operation simulation is performed, and at the same time, data is collected from the nodes corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet through the dynamic collection frequency configured by the second-level monitoring node chain combined with the collection frequency fluctuation mapping function to obtain a simulation tracking data set space; The corresponding area in the simulated tracking data set space is divided into the data set obtained by the load main node subset space, and synchronously input to the corresponding edge control node. The running status of the corresponding collection node and the associated fault node set is judged by the configured discrimination algorithm, and the multi-level tracking control space is obtained and the label serial number of the collection node corresponding to the associated fault node set is marked; Furthermore, the step of determining the operating status of the corresponding collection node and the associated fault node set in this embodiment includes: If the corresponding acquisition node is judged to have a fault, the associated fault node is secondary judged by using the Bayesian algorithm and the associated fault accuracy threshold, and the fault judgment result and location information corresponding to the acquisition node and the associated fault node that is secondary judged to have a fault are output; If the corresponding collection node determines that there is no fault but there is a fault in the associated fault node set, the location information and fault result corresponding to the collection node with fault are obtained by using the corresponding Bayesian algorithm and the associated fault accuracy threshold between the nodes in the associated fault node set.

[0036] Based on the corresponding acquisition nodes in the multi-level tracking control space and the fault operation state judgment accuracy of the associated fault node set combined with the preset judgment accuracy threshold, the intensity adjustment factor and the dynamic acquisition frequency are adjusted through a three-dimensional simulation algorithm, and the overall electromechanical equipment operation state tracking process corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet, the second-level monitoring node chain and the distributed edge level control chain is subjected to cyclic simulation training; When the accuracy of the judgment of the corresponding acquisition node and the fault operation status of the associated fault node set is greater than the preset judgment accuracy threshold, the judgment result corresponding to each edge control node is output; According to the judgment result corresponding to each edge control node, a corresponding bidirectional feedback control instruction is generated through the edge tracking control sub-model configured for the corresponding edge control node; Sharing the bidirectional feedback control instruction along the direction of the distributed edge hierarchical control chain according to the priority of the corresponding edge control node in the distributed edge hierarchical control chain; At the same time, the bidirectional feedback control instruction is transmitted along the random data acquisition index chain to the nodes in the associated fault node set corresponding to the identified fault for associated control and fault warning, and the corresponding fault warning is fed back and uploaded to the edge warning interface. At the same time, according to the mapping relationship between the fault severity and the intensity adjustment factor, the color type and color depth of the directed connection relationship corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet are dynamically adjusted and displayed in real time.

[0037] Furthermore, when a subcomponent corresponding to a node fails, the degree of its impact on the associated subcomponents will change, that is, the association strength between the corresponding nodes will change. When the association strength changes, the color type mapping and color depth mapping corresponding to the directed connection relationship between the corresponding nodes will change in real time. Through this visual change, the operating status of the corresponding subcomponent can be better tracked and monitored, and it can be determined whether there is any abnormality in the corresponding subcomponent.

[0038] Taking the above-mentioned air conditioning and ventilation system as an example, the corresponding data collection and status determination process includes: Based on the three-dimensional first-level directed electromechanical equipment simulation subnet, the operation of the air conditioning and ventilation system is simulated. At the same time, the second-level monitoring node chain collects data from the simulation subnet nodes in combination with the dynamic collection frequency to obtain a simulated tracking data set space. The data set of the corresponding area-divided load main node subset space is input into the edge control node, and the operation status of the collection node and the associated fault node set is judged by the discrimination algorithm, and the label sequence number is marked to obtain a multi-level tracking control space. For example, if the data of a temperature sensor collection node is found to be abnormal, the associated air conditioning unit and other nodes are secondary discriminated through the Bayesian algorithm and the associated fault accuracy threshold.

[0039] The training and instruction generation loop includes: According to the discrimination accuracy and the preset discrimination accuracy threshold, the intensity adjustment factor and dynamic acquisition frequency are adjusted through the three-dimensional simulation algorithm, and the entire system is subjected to cyclic simulation training. When the discrimination accuracy is greater than the preset discrimination accuracy threshold, the edge control node discrimination result is output. According to the result, a two-way feedback control instruction is generated through the edge tracking control sub-model, shared along the distributed edge hierarchical control chain, and transmitted to the fault node through the random data acquisition index chain for associated control and fault warning, and fed back to the edge warning interface. At the same time, according to the mapping relationship between the severity of the fault and the intensity adjustment factor, the color type and depth of the directed connection relationship of the three-dimensional first-level directed electromechanical equipment simulation subnet are dynamically adjusted and displayed. For example, when the air-conditioning compressor fails seriously, the color of the corresponding connection relationship turns red and the depth deepens, which intuitively presents the fault status.

[0040] Furthermore, in this embodiment, from the perspective of data collection and analysis, the adaptive clustering algorithm and the Bayesian algorithm are used to perform probability clustering on the fault frequency and the associated fault information, which can accurately identify the associated fault node set, and then combine the three-dimensional simulation algorithm and the Bayesian algorithm to obtain the fault probability and perform simulation training to ensure that the obtained associated fault node set has a high accuracy rate, laying the foundation for subsequent accurate monitoring. By establishing an association between the intensity adjustment factor and the spatial boundary range of the regional division load main node subset and the data acquisition frequency, the monitoring range and data acquisition frequency can be adjusted in real time according to changes in energy transfer, electromagnetic coupling, and control delay. When the intensity adjustment factor changes exceed the normal threshold, the acquisition frequency is dynamically adjusted to ensure the pertinence and timeliness of data acquisition.

[0041] In terms of monitoring network construction, based on the hypergraph partitioning algorithm, the regional load main node set is divided according to the directed association strength, spatial distance, functional consistency and maximum load processing energy, and the boundary range is adjusted in real time, which can effectively optimize the monitoring area and balance the data processing load. Construct a secondary monitoring node chain and a random data collection index chain, set reasonable data collection rules, set different collection frequencies for nodes with different association strengths, randomly collect and mark label numbers, and adjust the collection frequency in real time with the collection frequency fluctuation mapping function, which not only ensures the comprehensiveness of the data, but also avoids data redundancy and improves monitoring efficiency.

[0042] For equipment status tracking and control, the simulated tracking data set space is obtained through the three-dimensional first-level directed electromechanical equipment simulation subnet and the second-level monitoring node chain. The discriminant algorithm and Bayesian algorithm are used to discriminate the operating status of the collection node and the associated fault node set, and a multi-level tracking control space is constructed. Then, the intensity adjustment factor and dynamic collection frequency are optimized through cyclic simulation training to ensure the accuracy of the discrimination. When the discrimination accuracy reaches the standard, a two-way feedback control instruction is generated, shared along the distributed edge hierarchical control chain and transmitted to the fault node for associated control and fault warning. At the same time, the warning is fed back to the edge warning interface to achieve all-round monitoring and timely response of the equipment. In addition, according to the mapping relationship between the severity of the fault and the intensity adjustment factor, the color type and depth of the directed connection relationship of the three-dimensional first-level directed electromechanical equipment simulation subnet are dynamically adjusted and displayed, and the color mapping is changed in real time based on the change of the node association strength, so that the operation and maintenance personnel can track the operating status of the sub-components more intuitively and accurately through visualization means, and timely judge whether the sub-components are abnormal, which greatly improves the efficiency and accuracy of equipment operation and maintenance, and effectively ensures the stable operation of the equipment.

[0043] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are within the protection of the present invention.

Claims

1. A dynamic tracking method for electromechanical equipment based on multi-heterogeneous data and edge computing, characterized in that: include: Preset an equipment operation and maintenance monitoring network, and determine a distributed edge hierarchical control chain according to the distribution and operation attribute data of electromechanical equipment in the equipment operation and maintenance monitoring network; Through the distributed edge hierarchical control chain combined with the data monitored by the equipment operation and maintenance monitoring network, a multi-level tracking control space and a two-way feedback control instruction are obtained through a configured distributed enhanced control model; Based on multi-level tracking control space and bidirectional feedback control instructions, the electromechanical equipment is fed back upstream and the electromechanical instructions are controlled downstream; The distributed edge hierarchical control chain is constructed by combining the graph algorithm with the distribution status of electromechanical equipment, the operating load and the correlation coefficient of the corresponding monitoring equipment under each hierarchical control node in the equipment operation and maintenance monitoring network.

2. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 1, characterized in that: The equipment operation and maintenance monitoring network includes a first-level directed electromechanical equipment simulation subnet and a second-level monitoring node chain; The steps of constructing the first-level directed electromechanical equipment simulation subnet include: Obtain the distribution location and attributes of electromechanical equipment and operation-related dynamic parameters, use each subcomponent in the electromechanical equipment as a node in the equipment operation and maintenance monitoring network, and obtain a first-level node set; The distribution location and attributes of the electromechanical equipment include the spatial coordinates, function labels and operation-related dynamic parameters corresponding to each subcomponent; the operation-related dynamic parameters include: the energy transmission path, control instruction transmission priority and electromagnetic coupling strength corresponding to each subcomponent; According to the energy transmission path, electromagnetic coupling strength, energy transmission efficiency and control instruction transmission priority among the dynamic parameters associated with the operation of each node of the electromechanical equipment, the directed association probability value between each node is obtained through the Bayesian algorithm; The directed association probability value between each node is used as the causal relationship verification factor between two nodes. The causal relationship between each node is verified through the causal analysis algorithm to obtain the causal node set corresponding to the first-level node set.

3. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 2, characterized in that: The step of constructing the first-level directed electromechanical equipment simulation subnet also includes: Based on the directed association probability value between each node and the preset association degree interval, the directed association strength between each node is obtained; According to the nodes with causal relationships and the nodes without causal relationships in the first-level node set, the directed association strength between the nodes is divided into directed causal association strength and ordinary directed association strength; The directed causal association strength and the ordinary directed association strength are used as the directed connection relationship between each node and combined with the first-level node set and the corresponding position coordinates. The first-level directed electromechanical equipment simulation subnet is obtained through the hypergraph algorithm. At the same time, the directed connection relationship is annotated with the corresponding directed causal association strength and ordinary directed association strength through the automatic labeling algorithm.

4. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 3, characterized in that: The step of constructing the first-level directed electromechanical equipment simulation subnet also includes: Based on the type and association strength of the directed connection relationship between the nodes of the first-level directed electromechanical equipment simulation subnet, a color mapping function of each type of directed association relationship and the corresponding association strength is constructed through a preset color mapping table and a kernel function in a support vector machine; The color mapping table includes color types and color level gradients; The color mapping function of each type of directed association relationship and the corresponding association strength is built into the directed connection relationship between the corresponding nodes; At the same time, the first-level directed electromechanical equipment simulation subnet after the built-in color mapping function is input into the three-dimensional simulation algorithm and combined with the HSL model. According to the type of directed connection relationship between each node and the change in correlation strength, the color of the directed connection relationship between each node of the first-level directed electromechanical equipment simulation subnet and the color corresponding brightness and chromaticity changes are simulated and verified by training to obtain a three-dimensional first-level directed electromechanical equipment simulation subnet.

5. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 4, characterized in that: A strength adjustment factor is built into each connection relationship in the first-level directed electromechanical device simulation subnet; the strength adjustment factor includes a positive strength adjustment factor and a negative strength adjustment factor; The intensity adjustment factor is obtained by fitting a multiple regression function based on the energy transfer delay change rate, the electromagnetic coupling intensity change rate and the control delay change rate, as well as the positive and negative signs corresponding to the energy transfer delay change rate, the electromagnetic coupling intensity change rate and the control delay change rate and the brightness and chromaticity of the color corresponding to the directed connection relationship.

6. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 5, characterized in that: The steps of constructing the secondary monitoring node chain include: Obtaining the fault frequency corresponding to each node in the three-dimensional first-level directed electromechanical equipment simulation subnet and the corresponding associated fault nodes and the corresponding associated fault frequencies; Based on the acquired fault frequency corresponding to each node and the corresponding associated fault nodes and the corresponding associated fault frequencies, probability clustering is performed through an adaptive clustering algorithm combined with a Bayesian algorithm to obtain a set of associated fault nodes; According to the associated fault node set, the probability of the corresponding associated fault node failing is simulated by the Bayesian algorithm when a node in the corresponding associated fault node set fails, using the three-dimensional simulation algorithm in the three-dimensional first-level directed electromechanical equipment simulation subnet; Based on the probability of failure of the corresponding associated fault node and the preset associated fault accuracy threshold, probability clustering simulation training is performed to obtain the set of all associated fault nodes that meet the associated fault accuracy threshold.

7. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 6, characterized in that: The step of constructing the secondary monitoring node chain also includes: Obtain the node with the largest data collection load for each associated fault node set as the load master node for the regional division of the corresponding associated fault node set, and obtain the regional division load master node set; Based on the directed causal association strength and common directed association strength of the corresponding mark of the directed connection relationship corresponding to the associated fault node set, the spatial distance between the edge nodes between the associated fault node sets, the functional consistency and the causal association strength and the maximum load processing energy corresponding to each tracking control node in the distributed edge hierarchical control chain, combined with the conditions of minimizing the directed association strength between the partitioned regions and maximizing the directed association strength between each associated fault node set in the region, the regional partition load main node set is divided into regions through a hypergraph partitioning algorithm to obtain a partitioned regional partition load main node subset space; Based on the divided area, the load main node subset space is divided into two parts. Combined with the intensity adjustment factor and the maximum load processing energy corresponding to each tracking control node in real time, the boundary range corresponding to each area divided load main node subset space is adjusted in real time through a three-dimensional simulation algorithm combined with a hypergraph segmentation algorithm.

8. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 7, characterized in that: The step of constructing the secondary monitoring node chain also includes: A data monitoring point is set based on each completed regional partition load main node subset space to obtain a data monitoring point set, and a directed monitoring connection relationship is constructed based on the data transmission direction and control instruction execution priority order between the regional partition load main node subset spaces corresponding to the data monitoring point set; Based on the directed monitoring connection relationship and the data monitoring point set, a secondary monitoring node chain is obtained; Based on the secondary monitoring node chain, a random data collection index chain is constructed between each data monitoring point in the secondary monitoring node chain and each associated fault node set in the corresponding regional partition load main node subset space through a random algorithm combined with a discrimination algorithm, and data collection rules are set for the random data collection index chain corresponding to the same data monitoring point.

9. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 8, characterized in that: The steps of constructing the distributed edge level control chain include: A distributed federated tracking control model is constructed based on a distributed federated algorithm combined with a reinforcement learning algorithm, and an edge tracking control sub-model in the distributed federated tracking control model is deployed to each data monitoring point to obtain an edge control node set; Based on the correlation coupling and equipment subcomponent operation priority between the nodes corresponding to the load main node subset space in each area of ​​the data monitoring point set, an information sharing connection relationship corresponding to each two edge control nodes in the edge control node set is constructed; Based on the edge control node set and the information sharing connection relationship, a distributed edge hierarchical control chain is obtained.

10. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 9, characterized in that: The data collection rules include: For all nodes with causal relationships in the associated fault node set in the main node subset space of the same area partition load and nodes with medium or above ordinary directed association strength, set the same data collection frequency; The random data collection index chain corresponding to each associated fault node set only randomly collects data from one node in the associated fault node set at a time, and marks the collected data with the label sequence number corresponding to the associated fault node set; Configure the intensity adjustment factor and the acquisition frequency fluctuation mapping function of the data acquisition frequency, and configure the acquisition frequency fluctuation mapping function to all random data acquisition index chains; When the current intensity adjustment factor changes, the data acquisition frequency corresponding to the random data acquisition index chain is adjusted in real time through the acquisition frequency fluctuation mapping function.

11. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 10, characterized in that: The steps of obtaining the multi-level tracking control space and the bidirectional feedback control instructions include: Based on the three-dimensional first-level directed electromechanical equipment simulation subnet, the electromechanical equipment operation simulation is performed, and at the same time, data is collected from the nodes corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet through the dynamic collection frequency configured by the second-level monitoring node chain combined with the collection frequency fluctuation mapping function to obtain a simulation tracking data set space; The corresponding area in the simulated tracking data set space is divided into the data set obtained by the load main node subset space, and synchronously input to the corresponding edge control node. The running status of the corresponding collection node and the associated fault node set is judged by the configured discrimination algorithm, and the multi-level tracking control space is obtained and the label serial number of the collection node corresponding to the associated fault node set is marked; Based on the corresponding acquisition nodes in the multi-level tracking control space and the judgment accuracy of the fault operation status of the associated fault node set, combined with the preset judgment accuracy threshold, the intensity adjustment factor and the dynamic acquisition frequency are adjusted through a three-dimensional simulation algorithm, and the overall electromechanical equipment operation status tracking process corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet, the second-level monitoring node chain and the distributed edge level control chain is subjected to cyclic simulation training; When the judgment accuracy of the corresponding acquisition node and the fault operation status of the associated fault node set is greater than a preset judgment accuracy threshold, the judgment result corresponding to each edge control node is output.

12. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 11, characterized in that: The step of acquiring the multi-level tracking control space and the bidirectional feedback control instructions also includes: According to the judgment result corresponding to each edge control node, a corresponding bidirectional feedback control instruction is generated through the edge tracking control sub-model configured for the corresponding edge control node; Sharing the bidirectional feedback control instruction along the direction of the distributed edge hierarchical control chain according to the priority of the corresponding edge control node in the distributed edge hierarchical control chain; At the same time, the bidirectional feedback control instruction is transmitted along the random data acquisition index chain to the nodes in the associated fault node set corresponding to the identified fault for associated control and fault warning, and the corresponding fault warning is fed back and uploaded to the edge warning interface. At the same time, according to the mapping relationship between the fault severity and the intensity adjustment factor, the color type and color depth of the directed connection relationship corresponding to the three-dimensional first-level directed electromechanical equipment simulation subnet are dynamically adjusted and displayed in real time.

13. The method for dynamic tracking of electromechanical equipment based on multi-heterogeneous data and edge computing according to claim 11, characterized in that: The step of determining the operating status of the corresponding acquisition node and the associated fault node set comprises: If the corresponding acquisition node is judged to have a fault, the associated fault node is secondary judged by using the Bayesian algorithm and the associated fault accuracy threshold, and the fault judgment result and location information corresponding to the acquisition node and the associated fault node that is secondary judged to have a fault are output; If the corresponding collection node determines that there is no fault but there is a fault in the associated fault node set, the location information and fault result corresponding to the collection node with fault are obtained by using the corresponding Bayesian algorithm and the associated fault accuracy threshold between the nodes in the associated fault node set.

Images