Electric control cabinet temperature control method based on simulation

Abstract

The invention relates to the technical field of temperature control, in particular to an electric control cabinet temperature control method based on simulation, and the method comprises the steps: obtaining the basic data of a target electric control cabinet; constructing a target three-dimensional model associated with a core component temperature tolerance threshold and a parameter stable temperature interval based on the basic data; fusing the target three-dimensional model and historical temperature information to generate a temperature monitoring area; dynamic temperature monitoring is realized based on a monitoring area, and adaptive temperature control adjustment is executed in combination with an association relationship between temperature rise and component parameters. According to the invention, through deep fusion of three-dimensional modeling, historical temperature analysis and component characteristics, precision of temperature monitoring and self-adaption of temperature control adjustment are realized, it is effectively guaranteed that core components operate in a safe temperature interval, temperature control reliability and operation stability of the electric control cabinet are significantly improved, and energy consumption and operation and maintenance cost are reduced.

Description

Technical Field

[0001] This invention relates to the field of temperature control technology, and in particular to a simulation-based method for temperature control of an electrical control cabinet. Background Technology

[0002] As a core device in industrial and power control systems, electrical control cabinets integrate numerous precision electronic components (such as controllers, relays, and power modules). The operational stability of these components directly determines the reliability of the entire control system. However, during long-term operation, the internal components of the electrical control cabinet continuously generate heat. Coupled with the enclosed space and limited ventilation, the internal temperature is easily elevated. If the temperature exceeds the tolerance threshold of the components, it can cause parameter drift, lifespan degradation, and in severe cases, even component burnout, leading to a shutdown of the entire control system. Therefore, precise temperature control and simulation monitoring of the electrical control cabinet are crucial.

[0003] Currently, temperature control and simulation technologies for electrical control cabinets have made some progress. Existing technologies are mainly divided into two categories: one is traditional temperature control technology, which involves placing fixed temperature sensors inside the cabinet to collect local temperature data in real time. When the temperature exceeds a preset threshold, it drives actuators such as cooling fans and air conditioners to perform overall cooling adjustment. The other is advanced simulation monitoring technology, which uses 3D modeling technology to construct a cabinet structure model, combines it with historical temperature data to generate a temperature distribution map, and then delineates the monitoring area to achieve targeted temperature monitoring and adjustment.

[0004] Although existing technologies have alleviated the temperature rise problem of electrical control cabinets to some extent, there are still many insurmountable technical pain points in practical applications: First, traditional temperature control technology adopts a fixed-point temperature measurement and passive cooling mode, with a fixed monitoring area, which cannot adapt to the hot spot migration problem caused by changes in cabinet load and external environmental fluctuations, resulting in insufficient targeted and flexible temperature control; Second, the existing advanced simulation technology's 3D models are mostly static models, which can only restore the physical structure of the cabinet and do not associate key data such as component temperature tolerance thresholds and component thermophysical properties, resulting in a large deviation between simulation results and actual operating conditions, and failing to provide a reliable basis for precise temperature control; Third, existing technologies are all passive control modes that respond after temperature rise, lacking the ability to predict temperature change trends, and the temperature control action has a significant lag, making it difficult to avoid the impact of temperature rise on the stability of component parameters in advance; Fourth, although some technologies have attempted to introduce simulation methods, they have not established a dynamic calibration mechanism for simulation accuracy. As the cabinet operates for longer periods, the thermal characteristics of components change, and the simulation accuracy will continue to decline, failing to guarantee the reliability of long-term monitoring.

[0005] Furthermore, as industrial control systems evolve towards intelligence and high precision, the integration of components within electrical control cabinets is continuously increasing, leading to increasingly stringent requirements for temperature control accuracy, response speed, and reliability. Existing technologies can no longer meet the integrated needs of precise simulation, predictive early warning, and adaptive temperature control for electrical control cabinets under complex operating conditions. There is an urgent need to develop a temperature control simulation method that integrates precise modeling, dynamic calibration, and predictive regulation to address the shortcomings of existing technologies, ensure the stable operation of components within the electrical control cabinet, and reduce the risk of system failure. Summary of the Invention

[0006] To at least solve one of the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: This invention provides a simulation-based method for temperature control of an electrical control cabinet, the method comprising the following steps: S1, Obtain the basic data of the target electrical control cabinet.

[0007] S2, construct a target three-dimensional model corresponding to the target electrical control cabinet based on the basic data, and associate the temperature tolerance threshold and parameter stability temperature range of the core components in the target electrical control cabinet with the target three-dimensional model.

[0008] S3, integrate the target 3D model with the historical temperature information of the target electrical control cabinet to generate the temperature monitoring area corresponding to the target electrical control cabinet.

[0009] S4. Based on the temperature monitoring area, the temperature of the corresponding area in the target electrical control cabinet is dynamically monitored, and adaptive temperature control is performed in combination with the correlation between temperature rise and component parameters.

[0010] The present invention has at least the following beneficial effects: This invention constructs a three-dimensional model relating the temperature tolerance thresholds of core components to their stable temperature ranges, and integrates historical temperature information to generate a temperature monitoring area that closely matches the actual temperature rise distribution. This achieves precise temperature monitoring and effectively overcomes the drawbacks of traditional temperature control methods, which often result in missed control of key areas and over-control of non-key areas due to overly generalized global management. Furthermore, this invention performs adaptive temperature control based on the actual temperature rise rate and component parameter stability requirements, dynamically matching the adjustment intensity (e.g., initiating refined cooling in advance when core components are nearing the upper limit of their stable parameter range). This ensures stable component operation while avoiding energy waste caused by excessive temperature control. Simultaneously, by optimizing the temperature control simulation model using real cabinet data and historical temperature information, this invention significantly improves the adaptability of temperature control simulation to actual operation. It can be upgraded based on the existing cabinet structure without adding complex hardware, effectively reducing the cost of technology implementation. Ultimately, this invention allows core components to continuously operate within a suitable temperature range, reducing the probability of high-temperature damage, and provides precise data for maintenance work, thereby extending the overall lifespan of the electrical control cabinet and reducing maintenance difficulty and workload.

[0011] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A flowchart illustrating a simulation-based temperature control method for an electrical control cabinet, as provided in an embodiment of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0016] It should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the steps as sequential processes, many of these steps can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the steps can be rearranged. A process can be terminated when its operation is complete, but it may also have additional steps not included in the figures. A process can correspond to a method, function, procedure, subroutine, subroutine, etc.

[0017] This invention provides a simulation-based method for temperature control of an electrical control cabinet, such as... Figure 1 As shown, the method includes the following steps: S1, Obtain the basic data of the target electrical control cabinet.

[0018] In this invention, the target electrical control cabinet refers to the specific electrical control cabinet for which the user intends to conduct temperature control simulation. The basic data specifically includes the following: Cabinet dimensional data: refers to the key dimensional parameters that can fully characterize the physical outline and internal cavity structure of the target electrical control cabinet. Specifically, it includes the length, width and height of the cabinet exterior, as well as the partition dimensions of each cavity inside the cabinet (such as the length × width × height of the main cavity and auxiliary cavity), cabinet door thickness, and the size and distribution spacing of ventilation openings / heat dissipation grilles, etc., to provide data support for constructing a scaled 3D model of the cabinet structure.

[0019] Functional components inside the cabinet: refers to the electronic components and assemblies inside the electrical control cabinet that realize core functions such as control, power distribution, and signal transmission. Specifically, these include controllers (such as PLCs and microcontrollers), power modules (such as IGBT modules), relays, contactors, fuses, terminal blocks, power modules, signal converters, etc. Their spatial location directly affects the temperature distribution inside the cabinet and is a key basis for delineating the temperature monitoring area.

[0020] Temperature sensor installation location data: refers to the precise spatial coordinate data of each temperature acquisition device deployed in the cabinet (based on a preset cabinet coordinate system, such as establishing a three-dimensional coordinate system with the lower left corner of the cabinet as the origin), including the specific coordinate values ​​of the sensor in the length, width, and height directions of the cabinet, and at the same time, the monitoring coverage area corresponding to the sensor is associated to ensure that the subsequent temperature data is accurately matched with the spatial position of the three-dimensional model.

[0021] S2, construct a target three-dimensional model corresponding to the target electrical control cabinet based on the basic data, and associate the temperature tolerance threshold and parameter stability temperature range of the core components in the target electrical control cabinet with the target three-dimensional model.

[0022] In this invention, the core components inside the cabinet refer to electronic components and assemblies that play a key supporting role in realizing the overall control function of the electrical control cabinet and are sensitive to temperature changes. Specifically, these include programmable logic controllers (PLCs), power modules such as IGBTs, relays, power supply modules, signal processing chips, etc. The working state of these components directly determines the operational reliability of the electrical control cabinet, and their performance is easily affected by temperature fluctuations. Therefore, they are the core protection objects for temperature monitoring and temperature control.

[0023] The temperature tolerance threshold refers to the highest critical temperature that the core component can withstand during short-term operation. Exceeding this threshold will cause irreversible damage (such as burnout or breakdown) or permanent performance failure of the component. This threshold is usually determined by the technical specifications provided by the component manufacturer. The parameter stability temperature range refers to the temperature range within which the various electrical parameters of the core component (such as resistance, output power, response speed, etc.) remain within the design standard range. Within this range, the component can stably perform its designed functions. If the temperature exceeds this range, even if the temperature tolerance threshold is not reached, it will cause component parameter drift, leading to problems such as decreased control accuracy and response delay in the electrical control cabinet.

[0024] By associating the above parameters with the target 3D model, the 3D model can not only restore the physical structure of the electrical control cabinet, but also accurately map the relationship between temperature changes and the working status of core components, providing a precise basis for subsequent delineation of temperature monitoring areas and execution of targeted temperature control adjustments based on the model.

[0025] Furthermore, S2 specifically includes the following sub-steps: S21. Based on the cabinet size data of the target electrical control cabinet, construct a three-dimensional model of the cabinet structure with the same scale as the actual cabinet. The three-dimensional model of the cabinet structure includes the core structural features of the cabinet, such as cabinet chamber partitions, cabinet doors, ventilation openings, heat dissipation grilles, cable holes, mounting brackets, etc., to completely restore the physical structure of the cabinet.

[0026] The specific construction process of the cabinet structure 3D model is as follows: First, import the cabinet dimension data obtained from S1 (including the cabinet's external length × width × height, internal chamber partition dimensions, cabinet door thickness, and other core dimensions) using professional 3D modeling software (such as SolidWorks, UG, Revit, etc.). Use this cabinet dimension data as a reference to draw the basic outline of the cabinet. Then, based on the core structural features of the cabinet, add 3D models of chamber partitions, cabinet doors, ventilation openings, heat dissipation grilles, cable holes, mounting brackets, and other structures one by one to the basic outline. After modeling is completed, use the software's scaling function to compare the deviation between the model size and the actual cabinet size, ensuring that the model scaling ratio is 1:1 and completely restoring the physical structural form of the cabinet. The cabinet structure 3D model contains the aforementioned core structural features of the cabinet and can accurately match the spatial layout of the actual cabinet.

[0027] S22. Based on the spatial location data of each functional component in the target electrical control cabinet, construct the three-dimensional model of each functional component one by one in the corresponding spatial area of ​​the cabinet structure three-dimensional model. The three-dimensional model of the functional component completely matches the actual component's shape, size, and material properties. In addition, the temperature tolerance threshold and parameter stable temperature range of the core components are marked in the three-dimensional model of the functional component.

[0028] The specific construction process of the 3D model of each functional component is as follows: First, for each functional component (such as controller, power module, relay, terminal block, etc.), detailed data such as its external dimensions and material properties are obtained individually; then, based on the spatial position data of each functional component obtained by S1, the precise installation area of ​​the component is located in the 3D model of the cabinet structure; subsequently, in this area, a 3D model that perfectly matches the external dimensions of the actual component is drawn using 3D modeling software, and the model is given the same material properties as the actual component (such as metal, plastic, etc., for subsequent thermal simulation analysis); at the same time, the temperature tolerance threshold and parameter stability temperature range corresponding to the core components are marked in the model.

[0029] S23. Based on the installation location data of the temperature sensors, construct a three-dimensional model of each temperature sensor at the corresponding installation point in the three-dimensional model of the cabinet structure. The three-dimensional model of the temperature sensor includes a detection range identifier to clearly define the temperature acquisition coverage area of ​​each sensor.

[0030] The specific construction process of the 3D model of each temperature sensor is as follows: First, based on the temperature sensor installation location data obtained by S1, accurately locate the installation point of each sensor (such as the side wall of the cabinet, the top of the chamber, the vicinity of the core components, etc.) in the coordinate system Ω of the 3D model of the cabinet structure (this coordinate system is established with the bottom left corner of the cabinet as the origin, the length direction of the cabinet as the x-axis, the width direction as the y-axis, and the height direction as the z-axis). Then, at this point, construct a 3D model of the sensor body according to the actual size and appearance of the temperature sensor (such as patch type, probe type, wall-mounted type, etc.). At the same time, add a detection range identifier to the model. The detection range identifier refers to the visual acquisition area identifier added around the 3D model of the temperature sensor (such as a conical range or a spherical range with the sensor as the vertex, etc.), and mark the specific parameters of the range (such as detection angle, effective acquisition distance, etc.) to intuitively reflect the temperature acquisition coverage area of ​​each sensor and ensure that the subsequent temperature monitoring area is accurately matched with the sensor acquisition range.

[0031] To ensure that the 3D models of functional components and temperature sensors accurately match the actual layout of the electrical control cabinet, precise coordinate mapping must be performed during the construction of the 3D models of functional components in S22 and the construction of the 3D models of temperature sensors in S23. First, for the i-th functional component (where i ranges from 1 to m1, and m1 is the total number of functional components in the cabinet), its design location data is PF. design (i)=(x i y i , z i For the j-th temperature sensor (j ranges from 1 to m2, where m2 is the total number of temperature sensors in the cabinet), its design location data is PS. design (j)=(x j y j , z j ); Secondly, the above design location data is mapped to the coordinate system Ω of the cabinet structure 3D model using the coordinate transformation function T(), to obtain the corresponding installation location within the model: for functional components, the mapped location is PF. model (i)=T(PF design (i)); For the temperature sensor, the mapped position is PS. model (j)=T(PS design (j)); Among them, the core function of the coordinate transformation function T() is to unify the origin, axis and scaling ratio of the design data and the model coordinate system, so as to ensure that the layout of each component and sensor in the three-dimensional model is completely consistent with the actual electrical control cabinet.

[0032] S24 aligns and merges the coordinates of the cabinet structure 3D model, the 3D models of each functional component, and the 3D models of each temperature sensor to generate a target 3D model that fully restores the physical structure of the target electrical control cabinet and the key temperature control related parameters.

[0033] The specific integration process is as follows: The first step is to confirm that all models to be merged have been unified to coordinate system Ω, that is, to complete the coordinate mapping of S22 and S23, and to ensure that the position reference of each model is consistent. The second step is to use the model alignment function of the 3D modeling software to accurately match the corresponding spatial positions of the 3D models of each functional component and the 3D model of the temperature sensor with the 3D model of the cabinet structure. For example, the functional components are matched with the cabinet mounting brackets, and the sensors are matched with the installation points to avoid positional deviations. The third step is to perform a model merging operation, which integrates the three types of models into a single model while retaining the independent attributes of each component, such as the material of functional components, the temperature control parameters of core components, and the detection range of sensors. The fourth step is to perform integrity verification on the fused model, checking for any missing, overlapping, or positional deviations, to ensure that the fused target 3D model can completely and accurately reproduce the physical structure of the actual electrical control cabinet and the correlation of key parameters related to temperature control.

[0034] S3, integrate the target 3D model with the historical temperature information of the target electrical control cabinet to generate the temperature monitoring area corresponding to the target electrical control cabinet.

[0035] In this invention, the historical temperature information specifically refers to the historical temperature dataset collected by each temperature sensor within the target electrical control cabinet during its entire lifecycle operation under full operating conditions. Each temperature sensor's historical temperature dataset contains complete temperature time-series data for four typical operating cycles: no-load cycle (no-load operation), rated load cycle (rated power output), peak load cycle (maximum load operation), and start-stop transition cycle (transition from start-up to stable operation, and from stable operation to shutdown), ensuring coverage of all core operating scenarios of the electrical control cabinet. Further, S3 specifically includes the following sub-steps: S31, combining the target 3D model with historical temperature information, generates an initial heatmap corresponding to the target 3D model, specifically including the following sub-steps: S311, the target three-dimensional model is uniformly divided along the length, width and height dimensions to form several target spatial grids with unique spatial coordinates. The target spatial grid is the smallest spatial unit for temperature analysis.

[0036] Specifically, the target 3D model is discretized in space into tiny cubic units of equal volume (i.e., voxel meshes). The outer envelope dimensions (Lx, Ly, Lz) of the target 3D model are defined as the maximum size range of the model along the x-axis (length direction), y-axis (width direction), and z-axis (height direction). The meshing precision Δ is defined as the side length of a single voxel mesh (unit: mm, set by the user according to temperature control accuracy requirements; the smaller Δ, the higher the temperature analysis accuracy). The spatial range corresponding to the mesh index (u, v, w) satisfies the following formula: x∈[u×Δ, (u+1)×Δ), y∈[v×Δ, (v+1)×Δ), z∈[w×Δ, (w+1)×Δ).

[0037] Where u, v, and w are all non-negative integers, and must satisfy 0 ≤ u < floor(Lx / △), 0 ≤ v < floor(Ly / △), 0 ≤ w < floor(Lz / △), and floor() is a floor function to ensure that the mesh completely covers the target 3D model. Each spatial mesh can be uniquely identified as g(u, v, w), and is uniformly denoted as the g-th target spatial mesh. The value of g is from 1 to n, where n is the total number of target spatial meshes, and n = floor(Lx / △) × floor(Ly / △) × floor(Lz / △).

[0038] S312, based on historical temperature information, obtains the thermal risk index of each target spatial grid.

[0039] In this invention, since the historical temperature information consists of discrete point data collected by temperature sensors inside the cabinet, which cannot directly cover all target spatial grids, a spatial interpolation algorithm (such as Kriging interpolation, inverse distance weighted interpolation, etc.) is first used to calculate the complete temperature time series data corresponding to each target spatial grid based on the known temperature time series data of each sensor. Then, based on the completed temperature data, the thermal risk index (HRI) of the g-th target spatial grid is calculated using a weighted summation formula. g Specifically, the following conditions must be met: HRI g =w1×F temp (g)+w2×F volatility (g)+w3×F d (g)+w4×F correlation (g)

[0040] Among them, F temp (g) represents the temperature level characteristic of the g-th target space grid, characterizing the degree to which the historical highest temperature of the g-th grid deviates from the safety benchmark. The calculation formula is: F temp (g)=(max(Th(g))-Ta) / (T0-Ta).

[0041] Where max(Th(g)) is the highest temperature in the historical temperature sequence collected by the temperature sensor associated with the g-th target spatial grid; Ta is the ambient temperature, defined as the normal external ambient temperature where the electrical control cabinet is located (default is 25℃, which can be calibrated according to the actual application scenario); T0 is the preset safe temperature threshold, defined as the highest allowable temperature for long-term stable operation of the core components inside the cabinet, determined by the component technical specifications, and must satisfy T0 > Ta to avoid the denominator being zero or negative. F temp The value of (g) ranges from [0,1]. The larger the value, the closer the grid temperature is to the safety threshold.

[0042] F volatility (g) represents the temperature fluctuation characteristics of the g-th target space grid, characterizing the stability of the historical temperature of the g-th grid. The calculation formula is: F volatility (g)=σ(Th(g)) / [Avg(Th(g))-Ta].

[0043] Where σ(Th(g)) and Avg(Th(g)) are the standard deviation and mean of the historical temperature sequence corresponding to the temperature sensor associated with the g-th target spatial grid, respectively. It should be noted that the denominator [Avg(Th(g))-Ta] > 0, meaning the average temperature is higher than the ambient temperature, consistent with the actual temperature rise characteristics of the electrical control cabinet during operation. volatility The larger the (g) value, the more drastic the temperature fluctuation of the grid, and the greater the impact on the stability of the components.

[0044] F d (g) represents the overheating duration characteristic of the g-th target spatial grid, characterizing the cumulative duration of the historical temperature exceeding the warning threshold of the g-th grid. The calculation formula is as follows: Where Th(g,t) is the temperature value of the temperature sensor associated with the g-th grid at the t-th historical data point; t is the index of the historical data point, ranging from 1 to Q, and Q is the number of historical data points, defined as the total number of historical temperature time-series data collected by a single temperature sensor, determined by the collection frequency and collection duration. For example, if the collection frequency is 1 time / minute and the collection duration is 100 hours, then Q=6000. Tw is the warning temperature threshold, defined as the early warning threshold value when the temperature is below the safe temperature threshold T0, where Tw<T0. It is set by the user according to the temperature control response requirements and is used to trigger the early warning mechanism in advance. I() is the indicator function, defined as I()=1 when the condition in parentheses is true; I()=0 when the condition is false; F d The value of (g) ranges from [0, 1]. The larger the value, the longer the cumulative time of the temperature exceeding the warning level in the grid, and the higher the risk.

[0045] F correlation(g) represents the spatial correlation characteristic of the g-th target spatial grid, characterizing the degree of temperature change linkage between the g-th grid and its neighboring grids. The calculation formula is: F correlation (g) = (1 / |N(g)|) × ∑ q∈N（g） ρ(Th(g), Th(q)).

[0046] Where N(g) is the set of neighboring grids of the g-th target spatial grid, defined as all target spatial grids that are spatially directly adjacent to the g-th grid (sharing a face, an edge, or a vertex); |N(g)| is the number of grids in the set of neighboring grids N(g) (i.e., the number of neighboring grids). ρ(Th(g), Th(q)) is the Pearson correlation coefficient, defined as a parameter used to quantify the linear correlation between the historical temperature sequences corresponding to the g-th and q-th target spatial grids, with a value range of [-1, 1]. The closer the value is to 1, the more consistent the temperature change trends of the two grids; Th(q) is the historical temperature sequence collected by the temperature sensor associated with the q-th target spatial grid. F correlation The value of (g) ranges from [-1, 1]. Here, the absolute value is mapped to [0, 1] (only the degree of linkage is considered, and positive and negative correlations are not distinguished). The larger the value, the stronger the temperature linkage between the grid and the surrounding area, and it needs to be considered as a whole area for temperature control.

[0047] w1 to w4 are the weight coefficients of each feature, and they satisfy w1+w2+w3+w4=1. The weight coefficients are set by the user according to the actual temperature control requirements. For example, the proportion of w1 and w3 can be increased in the area around the core components.

[0048] S313: Based on the thermal risk index of each target spatial grid, the target 3D model is rendered to generate an initial heatmap.

[0049] In this invention, the thermal risk index of all target spatial grids is first normalized to the [0, 1] interval to eliminate differences in the dimensions of different features; then, a mapping relationship between the normalized thermal risk index and the color gradient is established, for example: HRI g ∈[0,0.3] is mapped to blue, representing low risk; HRI g ∈(0.3, 0.7] is mapped to yellow, representing medium risk; HRI g ∈(0.7,1] is mapped to red, representing high risk; finally, based on this mapping relationship, the target 3D model is rendered grid by grid to generate an initial heat map that can intuitively reflect the temperature distribution and thermal risk level inside the cabinet.

[0050] S32, Based on the initial heat map, identify historical hotspot areas and delineate temperature monitoring areas after cluster optimization.

[0051] The core objective of this step is to identify high-risk areas requiring focused management from the overall thermal risk distribution of the initial heat map, thereby improving the targeting and accuracy of subsequent temperature monitoring. This includes the following sub-steps: S321: Extract spatial grids whose heat risk index exceeds a preset risk threshold from the initial heat map to form a candidate hotspot unit set.

[0052] In this invention, the preset risk threshold is a critical value that distinguishes between regular risk areas and high-risk areas. Its value is determined based on the stable temperature range of the parameters of the core components in the cabinet (for example, it is set in combination with the thermal risk index corresponding to the critical temperature of the core component parameters drift). The value range is [0.5, 0.7] after normalization (which can be finely adjusted according to the actual temperature control accuracy requirements). During the extraction process, by traversing all target space grids in the initial heat map, grids with thermal risk indices greater than the preset risk threshold are selected. All grids that meet the conditions together constitute a candidate hot spot unit set, which is the basic data for subsequent hot spot area clustering.

[0053] S322 uses a density clustering algorithm to perform cluster analysis on the candidate hotspot unit set, and merges the spatially adjacent candidate hotspot units with a thermal risk index difference of less than a preset value into historical hotspot areas.

[0054] In this invention, the spatial adjacency criterion is that the target spatial grids corresponding to two candidate hotspot units are directly adjacent in space, i.e., sharing a face, an edge, or a vertex, thereby ensuring the spatial continuity of the clustered regions. The density-based clustering algorithm preferentially adopts the DBSCAN algorithm (density-based spatial clustering noise application algorithm). This algorithm does not require pre-setting the number of clusters and can adaptively identify spatially continuous hotspot clusters, accurately adapting to the irregular hotspot region morphology in this application scenario. The specific core parameter definitions, judgment rules, and clustering application steps are as follows: I. Definition of core parameters: Neighborhood radius ε: Based on the side length of the target space grid, the specific value is ε = grid side length × √3 (√3 is the square root of 3, this value can ensure that the neighborhood range can completely cover all adjacent grids of the current grid, accurately matching the aforementioned judgment criteria for spatial adjacency).

[0055] Minimum number of core points MinPts: set to 3, meaning that at least 3 candidate hotspot cells that meet the conditions are required to form a continuous region. This can effectively prevent a single grid from being misjudged as a hotspot region due to data errors or instantaneous temperature rise.

[0056] Preset value α: α is the critical value of the difference in the normalized heat risk index, ranging from 0.05 to 0.1, and can be adjusted according to the clustering accuracy requirements. The higher the clustering accuracy requirement, the smaller the value of α; in the case of normal accuracy, α can be set to 0.1, and the judgment criterion is: if the difference in the heat risk index of two candidate hotspot units is ≤ α, then it is determined to meet the clustering conditions.

[0057] II. Clustering Application Steps: Step 1: Traverse each target space grid g in the candidate hot spot cell set, define the neighborhood range with grid g as the center and ε as the radius, and filter out all candidate hot spot cells contained in the neighborhood.

[0058] Step 2: For each grid g, count the number of candidate hotspot cells in its neighborhood that satisfy the condition "difference between the thermal risk index of grid g and that of grid g is ≤ α". If the number is ≥ MinPts (i.e. ≥ 3), then grid g is determined to be a core point.

[0059] Step 3: Group mutually reachable core points (core points whose neighborhoods overlap and all units in the overlapping area satisfy the condition that the difference in thermal risk index is ≤ α) and non-core points in the neighborhood of core points (they themselves satisfy the clustering conditions, but the number of units in their neighborhoods that meet the requirements is < MinPts) into the same cluster.

[0060] Step 4: Repeat steps 1 to 3 until all grids in the candidate hotspot unit set have been traversed. Each cluster formed in the end is an independent historical hotspot region.

[0061] It should be noted that during the clustering process, if a candidate hotspot cell has no other cells in its neighborhood that meet the criteria (neither spatial adjacency nor the thermal risk index difference ≤ α), then the cell is considered an isolated noise cell and will not be included in the historical hotspot area (it will be uniformly removed in step S323). Through the above clustering process, the precise aggregation of scattered high-risk grids into continuous hotspot areas can be achieved, providing a clear regional basis for subsequent delineation of temperature monitoring areas and implementation of precise temperature control.

[0062] S323 smooths and optimizes the boundaries of historical hotspot areas, removes isolated noise units, and forms a complete temperature monitoring area.

[0063] In this invention, the isolated noise unit refers to a single or a few discrete target spatial grids that are not spatially adjacent to any historical hotspot area and whose thermal risk index is higher than the preset risk threshold but the difference is ≤0.05 (normalized value). It is usually caused by data error or local instantaneous temperature rise and is not a real continuous hotspot.

[0064] The specific removal method is as follows: Set a grid number threshold K (K≥3, which corresponds to the MinPts value in S322 to ensure consistent judgment criteria). If the number of grids in a discrete grid cluster (not adjacent to historical hotspot areas) is <K, it is determined to be an isolated noise unit. By traversing all candidate hotspot units not included in historical hotspot areas, all isolated noise units that meet the above conditions are filtered and removed, and valid continuous areas with a grid number ≥K are retained.

[0065] After removing isolated noise cells, for the jagged irregular shape of the boundaries of historical hotspot areas caused by mesh discretization, a morphological dilation-erosion algorithm or a Bézier curve fitting method is used for smoothing optimization. One of the two methods can be selected. Specific application details are as follows: (1) Morphological dilation-erosion algorithm: Select a cubic structural element that matches the size of the target space grid (the side length of the structural element = the side length of 1 grid). First, perform a dilation operation on the historical hot spot area (fill the small gaps in the boundary depression and connect the adjacent small gaps), and then perform an erosion operation (restore the main outline of the area and eliminate the jagged edges of the boundary). Through the combination of dilation-erosion operation, a region with a regular outline is obtained. (2) Bézier curve fitting: First, extract the coordinates of the boundary grid vertices of the historical hotspot area, sample the vertices in spatial order (sampling interval = 1 grid side length, balancing smoothness and accuracy), and then construct the Bézier curve with the sampling points as control points. Replace the original jagged boundary with a smooth curve, while ensuring that the fitted curve does not exceed the core range of the original hotspot area. The core purpose of smoothing optimization is to make the outline of the temperature monitoring area more regular, which facilitates the subsequent accurate matching of the temperature sensor's acquisition range and improves the targeting of temperature monitoring and the effectiveness of data acquisition.

[0066] After the above noise reduction and smoothing optimization processes, a complete temperature monitoring area with regular contours, no noise interference, and coverage of real and continuous hot spots is finally obtained.

[0067] S324 configures corresponding monitoring levels and differentiated early warning thresholds based on the thermal risk index of each temperature monitoring area.

[0068] The core logic of this step is that the higher the risk, the higher the monitoring level and the lower the warning threshold, ensuring that temperature control resources are accurately matched to high-risk areas. This is specifically achieved through the following two illustrative examples: Example 1: Classification based on fixed average threshold (suitable for scenarios with high temperature control accuracy requirements and stable monitoring area shape) In this embodiment, the average thermal risk index of each temperature monitoring area is used as the basis for level classification. The average thermal risk index is the arithmetic mean of the thermal risk indices of all target spatial grids within a single temperature monitoring area, and the calculation formula is as follows: AvgHRI(h) = (∑ g∈R(h) HRI g ) / M(h) In the formula, h is the unique identifier of the temperature monitoring area (1≤h≤D, where D is the total number of temperature monitoring areas), R(h) is the set of target spatial grids contained in the h-th monitoring area, and HRI g Let M(h) be the thermal risk index of the g-th grid in the region, and M(h) be the total number of grids in the region.

[0069] The specific grading rules and parameter definitions are as follows: Monitoring level classification: Based on the average (normalized) heat risk index, it is divided into three levels, corresponding to two fixed average thresholds (H1, H2), and satisfying the preset risk threshold < H1 < H2 < 1. The higher the level number, the higher the importance (risk). Level 3 (High Importance / High Risk): Corresponds to AvgHRI(h)>H2. This level area is mostly the area where core components are concentrated (such as IGBT modules, PLC controllers and surrounding areas). Temperature control stability should be prioritized (e.g., shortening the monitoring sampling interval and starting redundant cooling equipment). Level 2 (Medium Importance / Regular High Risk): Corresponds to H1≤AvgHRI(h)≤H2, which is a regular high risk area (such as a power relay cluster area), and needs to be monitored as usual. Level 1 (Low Importance / Generally High Risk): Corresponds to the preset risk threshold < AvgHRI(h) < H1, which is a generally high-risk area (such as a dense area of ​​wiring terminals) and can be monitored at the normal frequency.

[0070] The average threshold is set based on the following: The values ​​of H1 and H2 are derived from the stable temperature range of the core components. For example, H2 = 0.8 (corresponding to the thermal risk index corresponding to the critical temperature of the core component parameter drift) and H1 = 0.6 (corresponding to the thermal risk index corresponding to the long-term stable operating temperature of the core component). These values ​​can be finely adjusted according to the temperature control accuracy requirements of the actual application scenario.

[0071] Example 2: Level classification based on dynamic step size (suitable for flexible scenarios with varying number of monitoring areas and dynamic fluctuations in operating conditions) In this embodiment, the monitoring level of each monitoring area is dynamically calculated by pre-setting a risk level division step size, and the specific rules are as follows: (1) Definition of core parameters: maxHRI: The maximum value among the average thermal risk indices of all D temperature monitoring areas (i.e., maxHRI=max(AvgHRI(1), AvgHRI(2), ..., AvgHRI(D))); HRI step : Preset risk level classification step size, normalized value, ranging from [0.2, 0.3]. The smaller the step size, the more refined the risk level classification. Example value: HRI step =0.25; roundup(): The roundup function rounds up, such as roundup(2.1)=3, roundup(3)=3.

[0072] (2) Formula for calculating monitoring level: The monitoring level L(h) of the h-th temperature monitoring zone is calculated as: L(h) = roundup(AvgHRI(h) / HRI) step ).

[0073] The grading mapping rule is as follows: the result is rounded up and mapped to three levels. The higher the grading number, the higher the importance (risk). For example, when HRI... step When L(h) = 0.25, L(h) = 4 corresponds to the third level (high importance), L(h) = 3 corresponds to the second level (medium importance), and L(h) = 2 corresponds to the first level (low importance) (if the roundup result > 4, it is uniformly classified as the third level; < 2, it is uniformly classified as the first level).

[0074] The core configuration logic for differentiated warning thresholds is as follows: the higher the importance (the higher the level number), the lower the warning threshold (the warning is triggered earlier, allowing more sufficient temperature control response time). The specific implementation is as follows: (1) Core mapping relationship: First, establish the mapping relationship between heat risk index and temperature. Then, use the reverse calculation formula to map the average heat risk index of each monitoring area to the corresponding benchmark temperature (as the core benchmark for the early warning threshold). The specific reverse calculation logic is as follows: Let the mean heat risk index of the h-th temperature monitoring area be AvgHRI(h), and the formula for calculating the corresponding baseline temperature Tbase(h) (warning threshold baseline) is as follows: Tbase(h)=Ta+AvgHRI(h)×(T0-Ta).

[0075] Derivation basis: directly based on the temperature level characteristic F in S312 temp The inverse solution of the forward calculation formula for (g) yields (F) temp (g) is the only core feature in the thermal risk index that is directly linearly related to temperature); F has been defined above. temp(g) = (max(Th(g)) - Ta) / (T0 - Ta) (positive direction: temperature → thermal risk characteristics). Since AvgHRI(h) is the mean of the thermal risk index in the region, it can be used to approximate F. temp (g) The core temperature correlation logic is obtained by inverse solving (reverse: thermal risk index → ​​temperature). Furthermore, to improve mapping accuracy, a weighted correction inverse formula can be used. Tbase(h) = Ta + [w1 × AvgF] temp (h)+w2×AvgF volatility (h)×k2+w3×AvgF d (h)×k3+w4×AvgF correlation [(h)×k4]×(T0-Ta).

[0076] Among them, AvgF temp (h), AvgF volatility (h), AvgF d (h) and AvgF correlation (h) represent the F values ​​within the h-th temperature monitoring area. temp (h), F volatility (h), F d (h) and F correlation The mean value of k2 to k4 is the correction coefficient, with a value range of [0.1, 0.3], which is obtained from experimental calibration and is used to weaken the influence of indirect features such as fluctuations and duration on temperature mapping.

[0077] (2) Threshold calculation formula: The warning threshold for the h-th temperature monitoring area is Talert(h) = Tbase(h) - β × (L(h) - 1).

[0078] Where β is the threshold increment coefficient, in ℃, with a value range of [2,5]. For example, β=3℃, which means that for each monitoring level increase, the warning threshold decreases by β℃.

[0079] Example: Given parameters: Ta=25℃, T0=75℃ (safe temperature of core components), AvgHRI(h)=0.8 (high risk) for a certain monitoring area, L(h)=3 (third level, high importance), β=3℃.

[0080] Step 1: Back-calculate the reference temperature Tbase(h) = 25 + 0.8 × (75 - 25) = 65℃; Step 2: Calculate the final warning threshold Talert(h) = 65 - 3 × (3 - 1) = 59℃; The results show that the warning threshold for this high-risk area is 59℃ (lower than the baseline temperature of 65℃), which achieves early warning and fully meets the core requirement of prioritizing the control of high-importance areas.

[0081] S4. Based on the temperature monitoring area, the temperature of the corresponding area in the target electrical control cabinet is dynamically monitored, and adaptive temperature control is performed in combination with the correlation between temperature rise and component parameters.

[0082] The core objective of this step is to accurately trigger early warnings and temperature control adjustments through real-time temperature monitoring and threshold comparison, ensuring that the core components of the target electrical control cabinet operate within a safe temperature range. S4 specifically includes the following sub-steps: S41, according to the preset sampling period, collect the current temperature data of the temperature sensor in each temperature monitoring area, and obtain the current area temperature of each temperature monitoring area based on the collected current temperature data.

[0083] The preset sampling period is strongly correlated with the monitoring level defined in S324. The higher the level (the higher the importance / risk), the shorter the sampling period, ensuring that temperature changes in high-risk areas can be captured in a timely manner. The specific values ​​are: Level 3 (high importance) sampling period T1 = 1~3s, Level 2 (medium importance) sampling period T2 = 5~10s, Level 1 (low importance) sampling period T3 = 15~30s (which can be finely adjusted according to actual temperature control response requirements).

[0084] For each temperature monitoring area, real-time temperature data from all temperature sensors within that area is collected (if multiple sensors are deployed in a single area, all sensor data must be collected to avoid missing local temperatures); the data is collected in real time through sensor communication interfaces (such as RS485, Modbus protocol) and stored in a temporary data buffer.

[0085] All collected current temperature data are denoised (using a moving average filtering algorithm with a filter window size of 3 consecutive sampling points) to remove abnormal jump data (if the difference between a sample value and its two adjacent sample values ​​is greater than 5℃, it is determined to be an outlier and replaced by the average of the two adjacent sample values) to ensure the accuracy of subsequent comparisons.

[0086] In this invention, the current temperature in each temperature monitoring area can be the average or maximum value of all current temperature data in each temperature monitoring area: the average value is suitable for normal scenarios and is used to reflect the overall temperature level of the area; the maximum value is suitable for areas with dense core components and is used to accurately capture local extreme high temperatures and avoid overheating omissions.

[0087] S42 compares the current temperature with the warning threshold of the corresponding temperature monitoring area and calculates the temperature deviation value and temperature change rate.

[0088] In this invention, the temperature deviation value characterizes the degree of deviation between the current area temperature and the warning threshold. The calculation formula is: ΔT(h) = Tcurr(h) - Talert(h), where ΔT(h) is the temperature deviation value (unit: °C) of the h-th temperature monitoring area; Tcurr(h) is the current area temperature of the h-th area; Talert(h) is the warning threshold of the h-th area; ΔT(h) > 0 indicates that the current area temperature exceeds the warning threshold, and ΔT(h) ≤ 0 indicates that it does not exceed the warning threshold; the larger the absolute value of the deviation value, the more significant the deviation from the warning threshold.

[0089] The rate of temperature change characterizes how fast the temperature changes in a region and is used to predict the upward or downward trend of temperature. The calculation formula is: RT(h)=[Tcurr(h)-Tprev(h)] / Tsample(h), where RT(h) is the rate of temperature change of the h-th temperature monitoring region (unit: ℃ / s); Tcurr(h) is the current temperature of the h-th region in the current sampling period; Tprev(h) is the temperature of the h-th region in the previous sampling period; Tsample(h) is the preset sampling period corresponding to the h-th region; RT(h)>0 indicates that the temperature is rising, and the larger the value, the faster the rise; RT(h)≤0 indicates that the temperature is stable or is falling.

[0090] S43, based on the comparison results, determine whether each temperature monitoring area has triggered a temperature warning or alarm status.

[0091] Among them, the alarm status is a more urgent temperature anomaly than the warning status, corresponding to a temperature closer to the safe temperature threshold T0 of the core components. The specific judgment conditions and status management rules are as follows: No state triggered: When ΔT(h)≤△T1 (i.e. ≤0℃) and |RT(h)|≤△RT1, it indicates that the current temperature in the area has not exceeded the warning threshold, and the temperature change is gradual, which meets the temperature requirements for stable operation of core components. It is judged as normal temperature and no state is triggered. Among them, △T1 is the first preset temperature deviation threshold, which is 0℃. It is set based on the critical benchmark of the warning threshold and is used to distinguish the boundary between not exceeding the warning threshold and exceeding the warning threshold. △RT1 is the first preset temperature change rate threshold, which is 0.1~0.2℃ / s (0.2℃ / s in the example). It is set based on the stable temperature fluctuation range under normal operating conditions and is used to determine whether the temperature is in a stable state.

[0092] Temperature warning status is triggered when any of the following conditions are met: ① ΔT(h)∈(△T1, △T2], indicating that the current temperature slightly exceeds the warning threshold and the change is gradual; ② ΔT(h)≤△T1 but RT(h)>△RT1, indicating that the current temperature does not exceed the warning threshold, but the rate of increase is relatively fast, and there is a risk of exceeding the threshold; △T2 is the second preset temperature deviation threshold, with a value range of 0.8~1.2℃ / s (1℃ / s in the example), which is set based on the critical value of rapid temperature rise that the core components can withstand, and is used to determine whether there is a risk of rapid overheating.

[0093] Temperature alarm state is triggered when any of the following conditions are met: ① ΔT(h) > ΔT2, indicating that the current temperature significantly exceeds the warning threshold; ② Tcurr(h) > (T0-Tc), indicating that the current temperature is close to the core component's safe temperature threshold T0, reserving emergency response space, where Tc is a preset temperature value, for example, 2℃; ③ RT(h) > ΔRT2, indicating that the temperature is rising rapidly and may exceed the warning threshold or even the safe threshold in a short time (e.g., within a few seconds), requiring emergency cooling; ΔRT2 is the second preset temperature change rate threshold, with a value of 3℃, set based on the core component's temperature tolerance buffer zone, used to distinguish between slightly exceeding the warning threshold and significantly exceeding the warning threshold risk levels; Once a warning or alarm state is triggered, the state is immediately locked and continuously monitored to avoid frequent state switching due to instantaneous temperature fluctuations; The lock can only be released and updated to normal state after the following release conditions are met and a preset number of sampling cycles are maintained (e.g., 3, the sampling cycle follows the Tsample(h) of the corresponding area of ​​S41): The conditions for easing restrictions are: ΔT(h)≤ΔT1 (i.e., ≤0℃) and |RT(h)|≤ΔRT1.

[0094] The release condition is completely consistent with the judgment condition for not triggering any state, ensuring the logical consistency of state switching; the continuous requirement of 3 sampling cycles can filter out erroneous release cases due to instantaneous temperature drop, improving the stability of state judgment.

[0095] S44, if any temperature monitoring area triggers the temperature warning or alarm state, then according to the monitoring level of the temperature monitoring area, a corresponding temperature control adjustment command is generated, and the cooling actuator in the target electrical control cabinet is driven to perform local or overall cooling adjustment.

[0096] In this invention, the cooling actuator is defined as the equipment used for cooling within the electrical control cabinet. Based on their scope of action, cooling mechanisms are divided into localized cooling mechanisms and overall cooling mechanisms. Specific classifications and applicable scenarios are as follows: 1. Localized cooling mechanisms: such as zone-specific cooling fans, miniature water-cooling modules, heat sinks, etc., corresponding to one or more adjacent temperature monitoring areas, adapting to the need for "precise cooling of local hot spots", which can avoid energy waste caused by global cooling; 2. Overall cooling mechanism: such as cabinet-wide fans, air conditioning cooling system, etc., covering the entire target electrical control cabinet, adapting to scenarios where multiple areas are heated simultaneously or high-importance areas are cooled urgently, ensuring that the overall temperature of the cabinet is controllable.

[0097] The core of temperature control commands is determining the operating parameters of the cooling actuator (such as fan speed and cooling power). Two implementation methods are provided below to suit different temperature control accuracies and engineering cost requirements: Example 1: Precise Parameter Calculation Based on PID Control Algorithm This embodiment uses a PID (Proportional-Integral-Derivative) control algorithm to calculate the specific operating parameters of the cooling actuator, achieving precise closed-loop temperature regulation. The standard form of the algorithm and the parameter definitions are as follows: .

[0098] Where u(t) is the control output, i.e., the executable parameters of the cooling actuator (such as fan speed setpoint, cooling module power setpoint, etc.). K p The proportionality coefficient is used to quickly respond to the current temperature deviation, and the adjustment intensity is positively correlated with the magnitude of the deviation; e(t) is the temperature deviation, e(t) = Tset - TCk(t), where TCk(t) is the current area temperature (completely consistent with Tcurr(h) defined in S41 and S42 above, i.e., the pre-processed area average temperature or maximum value); Tset is the dynamically set target temperature, determined according to the monitoring level and warning / alarm status (for example, under the third level + alarm status, Tset is set to T0-5℃; under the second level + warning status, Tset is set to Talert(h) to ensure that the target temperature is adapted to the risk level); K d K is the differential coefficient used to predict temperature change trends, suppress temperature fluctuations, and improve regulation stability. i This is the integral coefficient, used to eliminate static temperature deviation and avoid temperature drift after long-term operation.

[0099] In this invention, K p K i K d The value of K is strongly correlated with the monitoring level and trigger status; the higher the level (the higher the risk), the higher the value. p K i A larger value for K ensures a faster response; d The value is dynamically fine-tuned based on the rate of temperature change (the greater the rate of temperature change, the higher the K value). dThe larger the value, the stronger the trend prediction capability. Specific examples of values ​​are as follows (which can be calibrated according to the actual temperature control characteristics of the cabinet): Level 3 + Alarm Status: K p =5.0, K i =0.8, K d =1.2; Level 2 + Warning Status: K p =3.5, K i =0.5, K d =0.8; Level 1 + Warning Status: K p =2.0, K i =0.3, K d =0.5.

[0100] Level 3 + Alarm Status: Tset = T0 - 5℃; Level 2 + Warning Status: Tset = Talert(h).

[0101] Example 2: Hierarchical Adjustment Intensity Matching Based on Monitoring Level and Status Type This embodiment directly matches the adjustment intensity of the cooling actuator by pre-setting the correspondence between monitoring level and status type, adapting to scenarios with simple engineering implementation and cost sensitivity. The specific rules are as follows: Level 3 (High Importance) + Warning Status: Activate the local cooling mechanism in this area and adjust it to medium power (e.g., adjust the cooling fan speed to 70% of the rated speed). Level 3 (High Importance) + Alarm Status: Simultaneously activate the local cooling mechanism (high power, such as fan speed 100%) and the overall cabinet cooling mechanism (medium power); if the temperature deviation value does not decrease within the first preset time, such as 30 seconds, then activate the redundant cooling equipment; Level 2 (Medium Importance) + Warning Status: Activate the local cooling mechanism in this area and adjust it to medium power. Level 2 (Medium Importance) + Alarm Status: Activate the local cooling mechanism for this area (high power); if the temperature deviation does not decrease within the second preset time, such as 1 minute, then activate the overall cooling mechanism (low power). Level 1 (Low Importance) + Warning Status: Activate the local cooling mechanism in this area and adjust it to low power. Level 1 (Low Importance) + Alarm Status: Activate the local cooling mechanism in this area and adjust it to medium power.

[0102] Low-power refers to the basic operating power of the cooling actuator, with the lowest output intensity. It is suitable for cooling needs in low-priority areas and non-emergency scenarios, and its core purpose is to reduce energy consumption while meeting basic temperature control requirements. Taking a cooling fan as an example, low-power corresponds to 40%~50% of the rated speed (the example value is 45% of the rated speed); for other cooling mechanisms such as refrigeration modules, low-power can correspond to 30%~40% of the rated cooling power (the specific value can be calibrated according to the temperature control characteristics of the electrical control cabinet).

[0103] The generated temperature control commands are converted into electrical signals (such as PWM signals to adjust fan speed) that can be recognized by the actuators through a controller (such as a PLC), driving the corresponding cooling actuators to run. At the same time, temperature data of the area is continuously collected during the adjustment process, and the trend of temperature deviation is calculated. If the temperature deviation continues to increase after adjustment (without decreasing within 5 sampling periods), the adjustment strategy is automatically upgraded (such as upgrading from low power to high power, or upgrading from local cooling only to local + overall cooling) to ensure that the cooling effect covers the risk changes.

[0104] Furthermore, S2 also includes the following sub-steps: S25 assigns thermophysical property parameters to each component within the target electrical control cabinet to the target 3D model, constructing a digital twin model of the target electrical control cabinet. The target 3D model fully reproduces the cabinet's dimensions, internal component installation locations, cooling actuator layout, and other physical structures, providing a geometric framework for the construction of the digital twin model.

[0105] In this invention, thermophysical properties include, but are not limited to, the thermal conductivity (unit: W / (m·K)), specific heat capacity (unit: J / (kg·K)), density (unit: kg / m³), and surface emissivity of each component, as well as the rated heat output (unit: W) of the core heat-generating components. The official parameters specified in the component technical specifications are preferred. Parameters not explicitly stated in the specifications are obtained through experimental testing (e.g., using the hot-wire method to test the thermal conductivity of the cabinet material) to ensure the accuracy of the model construction.

[0106] In this invention, based on the actual component layout of the target electrical control cabinet, corresponding parameters are precisely assigned to each component entity in the 3D model (such as IGBT modules, cooling fans, cabinet shells, etc.) to ensure that the thermal characteristics of the model and the physical entity are consistent. After the parameters are assigned, two key finishing operations need to be completed: First, establish the association mapping between the physical entity and the model, clarify the one-to-one correspondence between the temperature sensors, cooling actuators, and other equipment in the target electrical control cabinet and the corresponding components in the 3D model, to ensure that subsequent real-time temperature data can be accurately transmitted to the model and model correction commands can be accurately matched to the physical devices; Second, perform initial calibration, collect the initial temperature data of each area of ​​the target electrical control cabinet under normal operating conditions, compare it with the simulation temperature data of the corresponding area in the 3D model, and fine-tune the relevant thermophysical property parameters to make the initial state of the digital twin model consistent with the thermal characteristics of the physical entity, laying the foundation for model calibration in the subsequent S4 dynamic monitoring process.

[0107] Furthermore, during the dynamic monitoring process of S4, at each preset calibration cycle, the current temperature data of each temperature monitoring area is compared with the simulated temperature data of the digital twin model to calculate the temperature deviation value, and the thermophysical property parameters of the digital twin model are dynamically corrected through the PID algorithm.

[0108] In this invention, the preset calibration cycle is strongly correlated with the level of the temperature monitoring area. The higher the level (the higher the risk), the shorter the calibration cycle, ensuring the model accuracy in high-risk areas. Specifically, the values ​​are: Level 3 (high importance) calibration cycle Tcal1 = 30~60min, Level 2 (medium importance) Tcal2 = 60~120min, Level 1 (low importance) Tcal3 = 120~240min (which can be finely adjusted according to actual temperature control accuracy requirements). Collect the current temperature data Tcurr(h) of each temperature monitoring area, and simultaneously acquire the simulation temperature data Tsim(h) of the corresponding area of ​​the digital twin model, and calculate the calibration temperature deviation value ΔT between the two. cal (h)=|Tcurr(h)-Tsim(h)|; when ΔT cal (h) When the preset allowable deviation threshold ΔTallow (range 0.5~1℃, example ΔTallow=0.8℃) is reached, the PID control algorithm is activated to dynamically correct the thermophysical property parameters of the corresponding component in the digital twin model; the core parameter definition of the PID correction algorithm is as follows: Control output u cal Correction amount for thermophysical property parameters (such as the correction increment / decrease of thermal conductivity). Deviation e cal (t): ΔTallow-ΔT cal (h) (The goal is to keep the calibration deviation within the allowable range); proportionality coefficient Kpcal Integral coefficient K ical Differential coefficient K dcal : Dynamically set according to parameter type (e.g., when correcting thermal conductivity, K) pcal =0.3, K ical =0.1、K dcal =0.05), to ensure a smooth and undulating correction process.

[0109] After parameter correction, the simulated temperature Tsim'(h) of the corresponding region is obtained by resimulating using a digital twin model, and the deviation ΔT from the current region temperature Tcurr(h) is recalculated. cal '(h), if ΔT cal If (h) ≤ ΔTallow, the calibration is complete; if it still exceeds the threshold, repeat the correction process until the deviation meets the requirements.

[0110] By constructing a digital twin model in S25 and dynamically calibrating it in S4, the following can be achieved: ① Using simulated temperature to assist in judging the validity of real-time temperature data (e.g., if the deviation between real-time temperature and simulated temperature is too large, it indicates that the sensor may be faulty); ② Predicting temperature change trends based on accurate models and adjusting temperature control strategies in advance to improve the initiative of temperature control; ③ In long-term operation, dynamic correction can offset the changes in thermal characteristics caused by component aging, ensuring the long-term stability of the temperature control system.

[0111] Furthermore, step S4 also includes the following sub-steps: S45 preprocesses the current area temperature, historical temperature data, and real-time workload data of the target electrical control cabinet obtained in S41, providing standardized and time-series input data for subsequent LSTM model prediction.

[0112] In this invention, historical temperature data is selected from the most recent 1 to 2 hours, and the sampling interval is consistent with the preset sampling period of S41 (e.g., the third-level area is extracted at 1 to 3 seconds interval) to ensure the continuity of time sequence characteristics; real-time workload data includes the input current and output power of the target electrical control cabinet, which are collected in real time through the cabinet's current sensor and power sensor (the sampling frequency is synchronized with the temperature data).

[0113] The preprocessing operation involves standardizing the three types of data. Min-Max standardization can be used to map the data to the [0,1] interval, eliminating the dimensional differences between different data types (such as the difference between temperature units in °C and current units in A), thus improving the accuracy of model training and prediction. Simultaneously, input samples are constructed with a time step of 60 (i.e., every 60 consecutive sampling points form a complete time-series input sequence), ensuring that the LSTM model can effectively capture the temporal dependency between temperature and load data, laying the foundation for accurate prediction of temperature change trends.

[0114] S46: The preprocessed data is input into a pre-trained LSTM (Long Short-Term Memory) deep learning model to predict the temperature change trend of each temperature monitoring area within a preset time period. Based on the prediction results, it is determined whether an early warning condition is triggered. If triggered, a pre-adjustment command is generated to drive the cooling actuator to avoid the risk of abnormal temperature from the source. The specific implementation is as follows: In this invention, the LSTM adopts a deep learning architecture of input layer - 2 LSTM hidden layers - fully connected output layer; wherein, the input layer has a dimension of 3 (corresponding to three types of standardized data: temperature, current, and power respectively), the number of neurons in the 2 LSTM hidden layers is set to 64 and 32 respectively, the output layer has a dimension of N (N is the total number of temperature monitoring areas in the target electrical control cabinet), and finally outputs the temperature sequence of each monitoring area within a preset time period in the future; Training requirements: Pre-training should be performed based on the historical dataset of the target electrical control cabinet under all operating conditions. The dataset should cover typical scenarios such as normal operating conditions, light overload, and peak load, and include temperature-load time series data for the corresponding scenarios. The training objective is to minimize the mean square error (MSE) between the predicted temperature and the actual temperature. After the model is trained, it should be verified on the test set and meet the temperature control accuracy requirement of prediction error ≤ 0.5℃ before it can be put into use.

[0115] In this invention, the preset time period is 5 to 10 minutes, dynamically adapted according to the level of the temperature monitoring area (5 minutes for the third-level high-importance area to allow for more sufficient pre-cooling response time; 10 minutes for the first and second-level areas to balance the control effect and energy consumption). In this invention, the early warning trigger condition is: if the prediction result shows that in the temperature sequence of any temperature monitoring area within a preset time period, there are consecutive preset time periods, such as 2 minutes or more, with temperature values ​​≥ the early warning threshold Talert(h) of that area, then it is determined that predictive prevention and control needs to be initiated.

[0116] In this invention, the core logic of the pre-adjustment command is to drive the local cooling mechanism of the corresponding temperature monitoring area to perform low-power pre-cooling. The low power is 40% to 50% of the rated power, so as to achieve pre-cooling while avoiding energy waste caused by high-power operation.

[0117] In this invention, if the prediction results show that multiple adjacent temperature monitoring areas will trigger the warning conditions, the local cooling mechanism of the corresponding adjacent area will be activated for coordinated pre-cooling to ensure full coverage of potential high-temperature risk areas.

[0118] This invention also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being configured to perform the method described in this invention.

[0119] This invention also provides a computer-readable storage medium storing computer-executable instructions for performing the methods described in this invention.

[0120] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0121] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A simulation-based method for temperature control of an electrical control cabinet, characterized in that, The method includes the following steps: S1, Obtain the basic data of the target electrical control cabinet; S2, construct a target three-dimensional model corresponding to the target electrical control cabinet based on the basic data, and associate the temperature tolerance threshold and parameter stable temperature range of the core components in the target electrical control cabinet with the target three-dimensional model; S3, integrate the target 3D model with the historical temperature information of the target electrical control cabinet to generate the temperature monitoring area corresponding to the target electrical control cabinet; S4. Based on the temperature monitoring area, the temperature of the corresponding area in the target electrical control cabinet is dynamically monitored, and adaptive temperature control is performed in combination with the correlation between temperature rise and component parameters.

2. The method according to claim 1, characterized in that, The basic data includes: the cabinet dimensions of the target electrical control cabinet, the spatial location data of each functional component inside the cabinet, and the installation location data of several temperature sensors deployed inside the cabinet.

3. The method according to claim 2, characterized in that, S2 specifically includes the following sub-steps: S21. Based on the cabinet size data of the target electrical control cabinet, construct a three-dimensional model of the cabinet structure with the same scale as the actual cabinet. The three-dimensional model of the cabinet structure includes the core structural features of the cabinet. S22. Based on the spatial location data of each functional component in the cabinet, construct a three-dimensional model of each functional component in the corresponding spatial area of ​​the three-dimensional model of the cabinet structure. The three-dimensional model of the functional component matches the actual component's external dimensions and material properties, and marks the temperature tolerance threshold and parameter stable temperature range of the core components. S23, Based on the installation location data of the temperature sensors, construct the three-dimensional model of each temperature sensor at the corresponding installation point in the three-dimensional model of the cabinet structure. The three-dimensional model of the temperature sensor includes the detection range identifier. S24 aligns and merges the coordinates of the cabinet structure 3D model, the 3D models of each functional component, and the 3D models of each temperature sensor to generate the target 3D model.

4. The method according to claim 1, characterized in that, The historical temperature information includes a full-condition historical temperature dataset collected by each temperature sensor in the cabinet; the historical temperature dataset of any temperature sensor includes temperature time-series data for no-load, rated load, peak load, and start-stop transition cycles.

5. The method according to claim 1, characterized in that, S3 specifically includes the following sub-steps: S31, combine the target 3D model with historical temperature information to generate an initial heat map corresponding to the target 3D model; S32, Based on the initial heat map, identify historical hotspot areas and delineate temperature monitoring areas after cluster optimization.

6. The method according to claim 5, characterized in that, S31 specifically includes the following sub-steps: S311, the target three-dimensional model is uniformly divided along the length, width and height dimensions to form several target spatial grids with unique spatial coordinates, and the target spatial grid is the smallest spatial unit for temperature analysis; S312, based on historical temperature information, obtains the thermal risk index of each target spatial grid; S313: Based on the thermal risk index of each target spatial grid, the target 3D model is rendered to generate an initial heatmap.

7. The method according to claim 6, characterized in that, Thermal Risk Index (HRI) of Target Space Grid g The following conditions must be met: HRI g =w1×F temp (g)+w2×F volatility (g)+w3×F d (g)+w4×F correlation (g); Among them, F temp (g) represents the temperature level characteristic of the g-th target space grid, where g ranges from 1 to n, and n is the total number of target space grids. volatility (g) represents the temperature fluctuation characteristics of the g-th target space grid, F d (g) represents the overheating duration characteristic of the g-th target space grid, F correlation (g) represents the spatial correlation feature of the g-th target spatial grid, where w1 to w4 are the weight coefficients of each feature, and w1+w2+w3+w4=1.

8. The method according to claim 5, characterized in that, S32 specifically includes the following sub-steps: S321, extract spatial grids whose heat risk index exceeds the preset risk threshold from the initial heat map to form a candidate hotspot unit set; S322, a density clustering algorithm is used to perform cluster analysis on the candidate hot spot unit set, and the candidate hot spot units that are spatially adjacent and have a difference in heat risk index less than a preset value are merged into historical hot spot areas; S323 smooths and optimizes the boundaries of historical hotspot areas, removes isolated noise units, and forms a complete temperature monitoring area. S324 configures corresponding monitoring levels and differentiated early warning thresholds based on the thermal risk index of each temperature monitoring area.

9. The method according to claim 8, characterized in that, S4 specifically includes the following sub-steps: S41, according to the preset sampling period, collect the current temperature data of the temperature sensor in each temperature monitoring area, and obtain the current area temperature of each temperature monitoring area based on the collected current temperature data. S42, compare the current temperature of the area with the warning threshold of the corresponding temperature monitoring area, and calculate the temperature deviation value and temperature change rate; S43, determine whether each temperature monitoring area has triggered a temperature warning or alarm state based on the comparison results; S44, if any temperature monitoring area triggers the temperature warning or alarm state, then according to the monitoring level of the temperature monitoring area, a corresponding temperature control adjustment command is generated, and the cooling actuator in the target electrical control cabinet is driven to perform local or overall cooling adjustment.

10. The method according to claim 1, characterized in that, S2 also includes the following sub-steps: S25 assigns thermophysical property parameters to each component inside the target electrical control cabinet to the target 3D model, and constructs a digital twin model of the target electrical control cabinet; during the dynamic monitoring process in S4, at each preset calibration cycle, the current temperature data of each temperature monitoring area is compared with the simulated temperature data of the digital twin model, the temperature deviation value is calculated, and the thermophysical property parameters of the digital twin model are dynamically corrected through the PID algorithm.

Images