A power distribution cabinet heat dissipation control method and system based on temperature rise prediction
By introducing a heat dissipation capacity attenuation coefficient and real-time thermal resistance into the temperature rise prediction model and dynamically binding input variables, the prediction accuracy and anticipation of heat dissipation control of the distribution cabinet are achieved, solving the prediction deviation problem caused by the performance degradation of the heat dissipation device and ensuring the heat dissipation control effect during the equipment aging process.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TIANJIN HUAJIE POWER EQUIP MFG CO LTD
- Filing Date
- 2026-04-03
- Publication Date
- 2026-06-19
AI Technical Summary
Existing temperature rise prediction models do not take into account the performance degradation of heat dissipation devices, resulting in systematic inaccuracies in prediction results, and the timing of heat dissipation control actions continues to shift as the equipment ages.
By collecting the temperature of each heat source node in the distribution cabinet and the temperature and air volume of the inlet and outlet of the heat dissipation device, the heat dissipation capacity attenuation coefficient is calculated, the input variables of the temperature rise prediction model are dynamically bound, and real-time compensation and online updates are performed to ensure that the prediction results are consistent with the current physical state of the heat dissipation device.
It improves the predictive accuracy and anticipation of heat dissipation control in power distribution cabinets, ensuring the timing stability and cooling effect of heat dissipation control during equipment aging.
Smart Images
Figure CN121957196B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of thermal management technology for distribution cabinets, and in particular to a method and system for heat dissipation control of distribution cabinets based on temperature rise prediction. Background Technology
[0002] As the core equipment for power distribution and control in a power system, the switchgear contains key electrical components such as circuit breaker contacts, busbar connection points, and cable joints that continuously generate heat under rated operating conditions. The relatively enclosed nature of the cabinet structure makes it difficult for internal heat to dissipate naturally. Therefore, the timely and effective operation of the heat dissipation device is crucial for ensuring the safe and stable operation of the switchgear. Existing heat dissipation control methods for switchgear generally employ a passive triggering mechanism based on a fixed temperature threshold. This involves real-time monitoring of the cabinet's internal temperature using temperature sensors, triggering the heat dissipation device to initiate cooling when the monitored temperature exceeds a preset safety threshold. In this threshold-triggered mechanism, the activation of the heat dissipation device depends on the fact that the temperature has exceeded the limit. However, the thermal inertia time constant inside the switchgear is typically on the order of several minutes. There is a response time lag of 10 to 30 seconds between the heat dissipation device receiving the activation command and actually producing an effective cooling effect. This results in the heat dissipation control action always lagging behind the temperature exceeding the limit event, causing the cabinet temperature to continue to rise before the heat dissipation device completes its response, and the electrical components to endure over-temperature stress for an extended period.
[0003] To address the aforementioned response lag issue, existing technologies have introduced improved solutions that incorporate temperature rise prediction models. These models collect parameters such as current temperature, load current, and ambient temperature, and use these models to predict future temperature trends. When the predicted temperature exceeds a safety threshold, the cooling system is triggered in advance, advancing the cooling control timing before the actual temperature exceeds the limit. However, existing temperature rise prediction schemes suffer from a generally overlooked systemic flaw: these models use the rated cooling capacity of the cooling system as a fixed parameter in the prediction calculations, failing to account for the continuous degradation of cooling performance caused by factors such as dust accumulation in the air ducts, fan blade wear, and filter blockage during actual operation. As the cooling system's usage time increases, the discrepancy between its actual and rated cooling capacity widens. This leads to a systematic deviation between the future temperature sequence calculated by the prediction model based on the rated cooling capacity and the actual thermal state of the distribution cabinet, causing the prediction results to gradually become inaccurate. The error between the triggering timing of the cooling control action and the actual required lead time accumulates monotonically with equipment aging. Summary of the Invention
[0004] This application provides a heat dissipation control method and system for distribution cabinets based on temperature rise prediction. It solves the problems of systematic inaccuracy of prediction results and continuous shift of the timing of heat dissipation control action as the equipment ages due to the failure of existing temperature rise prediction models to consider the performance degradation of heat dissipation devices. It improves the prediction accuracy and timing stability of advanced control of distribution cabinet heat dissipation control throughout the entire life cycle of heat dissipation devices.
[0005] In a first aspect, this application provides a heat dissipation control method for distribution cabinets based on temperature rise prediction, the heat dissipation control method for distribution cabinets based on temperature rise prediction includes:
[0006] Step S1: Collect the node temperature of each heat source node in the power distribution cabinet, as well as the inlet temperature, outlet temperature and real-time air volume of the heat dissipation device, to obtain the thermal state dataset.
[0007] Step S2: Based on the inlet temperature, outlet temperature and real-time air volume in the thermal state data set, the real-time thermal resistance of the heat dissipation device is compared with the factory-calibrated reference thermal resistance to obtain the heat dissipation capacity attenuation coefficient.
[0008] Step S3: Use the highest node temperature in the thermal state dataset and the heat dissipation capacity attenuation coefficient as input variables for the temperature rise prediction model, so that the prediction result of the temperature rise prediction model is dynamically bound to the current physical attenuation state of the heat dissipation device, and obtain the future predicted temperature sequence.
[0009] Step S4: When a predicted temperature in the future predicted temperature sequence exceeds a safety threshold, the output power command of the heat dissipation device is proportionally compensated using the heat dissipation capacity attenuation coefficient. Before the temperature exceeds the limit, the heat dissipation device is triggered to perform a cooling action. The difference between the measured temperature and the corresponding predicted temperature in the future predicted temperature sequence is used to obtain the prediction execution deviation. The prediction execution deviation is used to drive the weight coefficient corresponding to the heat dissipation capacity attenuation coefficient in the temperature rise prediction model and the reference thermal resistance to be updated synchronously online.
[0010] Secondly, this application provides a heat dissipation control system for distribution cabinets based on temperature rise prediction, the heat dissipation control system for distribution cabinets based on temperature rise prediction includes:
[0011] The data acquisition module is used to collect the node temperature of each heat source node in the power distribution cabinet, as well as the inlet temperature, outlet temperature and real-time air volume of the heat dissipation device, to obtain a thermal state dataset.
[0012] The analysis module is used to compare the real-time thermal resistance of the heat dissipation device with the factory-calibrated reference thermal resistance based on the inlet temperature, outlet temperature and real-time air volume in the thermal state data set, and obtain the heat dissipation capacity attenuation coefficient.
[0013] The binding module is used to take the highest node temperature in the thermal state dataset and the heat dissipation capacity decay coefficient as input variables of the temperature rise prediction model, so that the prediction result of the temperature rise prediction model is dynamically bound to the current physical decay state of the heat dissipation device to obtain the future predicted temperature sequence.
[0014] The compensation module is used to proportionally compensate the output power command of the heat dissipation device with the heat dissipation capacity attenuation coefficient when there is a predicted temperature in the future predicted temperature sequence that exceeds the safety threshold. Before the temperature exceeds the limit, the heat dissipation device is triggered to perform a cooling action. The difference between the measured temperature and the corresponding predicted temperature in the future predicted temperature sequence is used to obtain the prediction execution deviation. The prediction execution deviation is used to drive the weight coefficient corresponding to the heat dissipation capacity attenuation coefficient in the temperature rise prediction model and the reference thermal resistance to be updated synchronously online.
[0015] Thirdly, a heat dissipation control device for a distribution cabinet based on temperature rise prediction is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the heat dissipation control device for the distribution cabinet based on temperature rise prediction to execute the aforementioned heat dissipation control method for the distribution cabinet based on temperature rise prediction.
[0016] Fourthly, a computer-readable storage medium is provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform the above-described heat dissipation control method for power distribution cabinets based on temperature rise prediction.
[0017] The technical solution provided in this application synchronously collects node temperatures, inlet temperatures, outlet temperatures, and real-time airflow at each heat source node and the air inlet / outlet side of the heat dissipation device within the distribution cabinet, forming a timestamp-aligned thermal state dataset. This breaks through the data limitations of existing technologies that rely solely on single-point temperature monitoring, ensuring that the subsequent calculation of the heat dissipation capacity attenuation coefficient and the input to the temperature rise prediction model are both based on multi-dimensional raw data with strict time consistency. Furthermore, by comparing the real-time thermal resistance of the heat dissipation device with the factory-calibrated baseline thermal resistance, the heat dissipation capacity attenuation coefficient is obtained. This transforms the current physical degradation state of the heat dissipation device into a quantitative parameter that can directly participate in the prediction calculation. Unlike existing technologies that use a fixed rated heat dissipation capacity in prediction modeling, this approach ensures that the calculation benchmark of the prediction model is always anchored to the actual working state of the heat dissipation device rather than the theoretical design value. The heat dissipation capacity attenuation coefficient and the highest node temperature are used together as input variables for the temperature rise prediction model, and their cross term is introduced to capture the nonlinear amplification effect of heat dissipation performance degradation on the temperature rise rate. This dynamically binds the prediction results of the temperature rise prediction model to the current physical degradation state of the heat dissipation device, solving the fundamental problem of systematic prediction bias in existing prediction models as equipment ages.
[0018] At the control decision level, the output power command is proportionally compensated using a heat dissipation capacity attenuation coefficient. This allows the heat dissipation device to compensate for the decrease in actual heat dissipation capacity by increasing output power under performance degradation conditions, ensuring that the actual cooling effect of the heat dissipation control action is not weakened due to equipment aging. The prediction execution deviation is obtained by subtracting the measured temperature from the corresponding predicted temperature in the future predicted temperature sequence. This deviation drives the synchronous online update of the weight coefficient corresponding to the heat dissipation capacity attenuation coefficient and the reference thermal resistance in the temperature rise prediction model. This constructs a complete closed-loop path of prediction, execution, verification, and update, enabling the prediction model to dynamically self-correct according to the degree of heat dissipation device degradation. This fundamentally solves the defect of existing predictive control systems that operate in an open-loop inaccurate state for a long time throughout the entire life cycle of the heat dissipation device, and achieves continuous maintenance of the anti-correspondence and prediction accuracy of heat dissipation control during equipment aging. Attached Figure Description
[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0020] Figure 1 This is a schematic diagram of an embodiment of the heat dissipation control method for distribution cabinets based on temperature rise prediction in this application.
[0021] Figure 2 This is a schematic diagram of an embodiment of the heat dissipation control system for distribution cabinets based on temperature rise prediction in this application.
[0022] Figure 3 This is a schematic block diagram of the heat dissipation control device for the power distribution cabinet based on temperature rise prediction in an embodiment of the present invention. Detailed Implementation
[0023] This application provides a method and system for heat dissipation control of a power distribution cabinet based on temperature rise prediction. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.
[0024] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the heat dissipation control method for distribution cabinets based on temperature rise prediction in this application includes:
[0025] Step S1: Collect the node temperature of each heat source node in the power distribution cabinet, as well as the inlet temperature, outlet temperature and real-time air volume of the heat dissipation device, to obtain the thermal state dataset.
[0026] Specifically, the thermal state dataset refers to a multi-dimensional time-series data set formed by merging the node temperatures of each heat source node with the inlet and outlet temperatures of the heat dissipation device and the real-time airflow using timestamp alignment within the same sampling period. Each heat source node refers to one of five types of concentrated heat locations within the distribution cabinet: circuit breaker contacts, busbar connection points, cable joints, transformer winding surfaces, and cabinet outlets. These locations generate the most heat and have the fastest temperature rise rate under rated operating conditions, providing the most representative temperature input for temperature rise prediction. The sampling period is set to 5 seconds. This value is based on the fact that the thermal inertia time constant inside the distribution cabinet is typically on the order of several minutes; a 5-second sampling interval can capture temperature change trends without causing data redundancy due to overly frequent sampling.
[0027] Step S2: Based on the inlet temperature, outlet temperature and real-time air volume in the thermal state data set, the real-time thermal resistance of the heat dissipation device is compared with the factory-calibrated reference thermal resistance to obtain the heat dissipation capacity attenuation coefficient.
[0028] Specifically, the reference thermal resistance refers to the fixed reference value, measured in °C / W, which is the ratio of the inlet and outlet temperature difference to the actual heat dissipation power obtained through calibration testing under factory rated operating conditions. This value is stored in the equipment nameplate parameter database. The heat dissipation capacity attenuation coefficient is the ratio of the current real-time thermal resistance to the reference thermal resistance. This ratio is dimensionless; a value of 1 indicates that the heat dissipation device is in its rated performance state, while a value greater than 1 indicates that the heat dissipation performance has degraded, with larger values indicating greater degradation. The real-time heat dissipation power is obtained by multiplying the inlet and outlet temperature difference, real-time airflow, air density, and air specific heat capacity at constant pressure. The air density is taken as 1.205 kg / m³, and the air specific heat capacity at constant pressure is taken as 1005 joules / kg Celsius. These two physical properties are generally accepted values under standard atmospheric pressure.
[0029] Step S3: Use the highest node temperature and the heat dissipation capacity attenuation coefficient in the thermal state dataset as input variables for the temperature rise prediction model, so that the prediction results of the temperature rise prediction model are dynamically bound to the current physical attenuation state of the heat dissipation device, and obtain the future predicted temperature sequence.
[0030] Specifically, dynamic binding refers to using the heat dissipation capacity attenuation coefficient as one of the input variables of the temperature rise prediction model. This allows the prediction model to calculate the temperature based on the actual current heat dissipation capacity of the heat dissipation device in each sampling period, rather than based on a fixed rated heat dissipation capacity. This ensures that the prediction results adjust synchronously with changes in the physical state of the heat dissipation device. The temperature rise prediction model adopts a linear weighted structure. The input variables include the highest node temperature, the node temperature change rate, and the heat dissipation capacity attenuation coefficient. Each input variable corresponds to an independent weight coefficient. The weight coefficients are initially fitted offline using the least squares method and subsequently updated online in step S4 based on the prediction execution deviation. The prediction step size covers 12 sampling points from 5 to 60 seconds in the future, and the output is the future predicted temperature sequence.
[0031] Step S4: When a predicted temperature in the future predicted temperature sequence exceeds the safety threshold, the output power command of the heat dissipation device is proportionally compensated by the heat dissipation capacity attenuation coefficient. Before the temperature exceeds the limit, the heat dissipation device is triggered to perform a cooling action. The difference between the measured temperature and the corresponding predicted temperature in the future predicted temperature sequence is used to obtain the prediction execution deviation. The prediction execution deviation is used to drive the weight coefficient corresponding to the heat dissipation capacity attenuation coefficient and the reference thermal resistance in the temperature rise prediction model to be updated synchronously online.
[0032] Specifically, proportional compensation refers to the output power command being multiplied by a fixed compensation coefficient based on the amount by which the heat dissipation capacity attenuation coefficient exceeds the unit value, and then added to the rated power. The compensation coefficient is set to 1.2, ensuring that the output power command can still cover the actual required heat dissipation even when the heat dissipation device performance decreases by 20%. The trigger condition for synchronous online updates is that the absolute value of the prediction execution deviation exceeds a preset deviation tolerance threshold of 1 degree Celsius. If the value is below this threshold, the prediction is considered accurate, and no update is triggered. If the value is above this threshold, the weight coefficient corresponding to the heat dissipation capacity attenuation coefficient and the reference thermal resistance are simultaneously corrected. The correction direction is consistent with the sign direction of the prediction execution deviation, the correction step size is set to 0.005, and the learning rate is set to 0.01. All the above parameter values are set to ensure the convergence and stability of the update process.
[0033] In one specific embodiment, step S1 includes:
[0034] Temperature sensors are installed at the circuit breaker contacts, busbar connection points, cable joints, transformer winding surfaces, and air outlets of the distribution cabinet. The node temperature of each heat source node is collected at a preset sampling period to obtain the node temperature sequence.
[0035] Temperature sensors are arranged on the air inlet and air outlet sides of the heat dissipation device, and the air inlet temperature, air outlet temperature and real-time air volume are collected synchronously by a wind speed sensor at the same sampling period as the node temperature sequence to obtain the heat dissipation operation parameter sequence.
[0036] Based on the node temperature sequence and the heat dissipation operation parameter sequence, the thermal state dataset is obtained by synchronously merging them in a timestamp-aligned manner.
[0037] Specifically, the preset sampling period is set to 5 seconds. This value is based on the following: the thermal inertia time constant inside the distribution cabinet is typically on the order of several minutes; there is a response time lag of 10 to 30 seconds between the heat dissipation device receiving the control command and the actual cooling effect. A 5-second sampling interval can complete at least two data acquisitions within one response time lag period, ensuring that no nodes of temperature rise trend change are missed, while avoiding the impact on the real-time performance of control decisions due to excessive data processing per unit time caused by too short a sampling interval. The node temperature sequence refers to the time-series data set formed by arranging the temperature sensor output values of five types of heat source nodes—circuit breaker contacts, busbar connection points, cable joints, transformer winding surfaces, and cabinet air outlets—in the order of sampling time within each 5-second sampling period. These five locations are the parts of the distribution cabinet with the highest current density, highest contact resistance, and most concentrated heat generation under rated operating conditions. Placing sensors in these locations can cover the main heat sources inside the cabinet. The heat dissipation operation parameter sequence refers to the time-series data set formed by arranging the output values of the inlet side temperature sensor, outlet side temperature sensor, and wind speed sensor of the heat dissipation device in chronological order within a 5-second sampling period that is exactly the same as the node temperature sequence. This sequence has the same sampling period as the node temperature sequence, ensuring that the two sets of sequences have a point-to-point correspondence on the time axis.
[0038] Timestamp alignment refers to the synchronized triggering of all sensor acquisition actions of the node temperature sequence and the heat dissipation operation parameter sequence by the same hardware clock signal at the trigger time of each 5-second sampling cycle. Each acquisition record is marked with a timestamp of the trigger time when it is written to the cache. During the merging process, the timestamp is used as the index key to merge the data records with the same timestamp in the node temperature sequence and the heat dissipation operation parameter sequence into a multi-dimensional record. Each multi-dimensional record contains values of eight dimensions: circuit breaker contact temperature, busbar connection point temperature, cable joint temperature, transformer winding surface temperature, cabinet air outlet temperature, air inlet temperature, air outlet temperature, and real-time air volume. All dimension values correspond to the same sampling time, thereby forming a thermal state dataset. This eliminates the data mismatch problem caused by the inconsistent acquisition time of the two sets of sequences, ensuring that the calculation of real-time thermal resistance in step S2 and the input of the temperature rise prediction model in step S3 are based on the original data with strict time consistency.
[0039] In one specific embodiment, step S2 involves comparing the real-time thermal resistance of the heat dissipation device with the factory-calibrated reference thermal resistance to obtain the heat dissipation capacity attenuation coefficient, including:
[0040] The inlet and outlet temperature difference is obtained by subtracting the inlet and outlet temperatures from the thermal state dataset.
[0041] The real-time heat dissipation power is obtained by multiplying the temperature difference between the inlet and outlet air with the real-time air volume, air density, and specific heat capacity of air at constant pressure.
[0042] The real-time thermal resistance is obtained by comparing the temperature difference between the inlet and outlet air vents with the real-time heat dissipation power.
[0043] The heat dissipation capacity attenuation coefficient is obtained by comparing the real-time thermal resistance with the factory-calibrated reference thermal resistance.
[0044] Specifically, the inlet and outlet temperature difference refers to the difference between the outlet temperature and the inlet temperature of the heat dissipation device. The calculation direction is outlet temperature minus inlet temperature. This difference reflects the increase in airflow temperature after the heat dissipation device removes heat from the cabinet; a larger difference indicates that more heat is removed in a single airflow cycle. Real-time heat dissipation power refers to the actual rate at which the heat dissipation device transfers heat to the outside at the current sampling moment. It is obtained by multiplying the inlet and outlet temperature difference, real-time airflow, air density, and air specific heat capacity at constant pressure. The air density is taken as 1.205 kg / m³, and the air specific heat capacity at constant pressure is taken as 1005 joules / kg Celsius. These two physical properties are generally accepted values under standard atmospheric pressure and a temperature of 20 degrees Celsius. Within the normal operating environment of the distribution cabinet, the variation range should not exceed 2%, and using fixed values does not affect the calculation accuracy. Real-time thermal resistance refers to the ability of a heat dissipation device to convert a unit of heat dissipation power into a temperature difference between the inlet and outlet at the current sampling time. It is obtained by dividing the temperature difference between the inlet and outlet by the real-time heat dissipation power, and the unit is degrees Celsius per watt. The larger the value, the greater the temperature difference generated by the heat dissipation device under the same heat dissipation power, that is, the weaker the heat dissipation capacity.
[0045] The reference thermal resistance (RTR) is a fixed reference value, measured in degrees Celsius per watt, representing the ratio of the inlet and outlet temperature difference to the actual heat dissipation power, obtained through calibration testing under factory rated operating conditions. It is determined by the manufacturer during factory inspection under standard testing conditions and stored in the equipment nameplate parameter database as a benchmark for judging whether the heat dissipation device's performance has degraded. The heat dissipation capacity attenuation coefficient is the ratio of the real-time thermal resistance to the reference thermal resistance. This ratio is dimensionless. When the heat dissipation capacity attenuation coefficient equals 1, it indicates that the current heat dissipation capacity of the heat dissipation device is consistent with its factory rated state. When the heat dissipation capacity attenuation coefficient is greater than 1, it indicates that the heat dissipation performance has degraded due to factors such as dust accumulation in the air duct, fan blade wear, or filter blockage. The higher the value, the greater the degree of degradation. This coefficient is updated synchronously with the thermal state dataset every 5-second sampling cycle to ensure that it always reflects the current physical state of the heat dissipation device.
[0046] In one specific embodiment, in step S3, the highest node temperature in the thermal state dataset and the heat dissipation capacity attenuation coefficient are used together as input variables for the temperature rise prediction model, so that the prediction result of the temperature rise prediction model is dynamically bound to the current physical attenuation state of the heat dissipation device, to obtain the future predicted temperature sequence, including:
[0047] The highest node temperature is obtained by filtering the node temperatures of each heat source node in the thermal state dataset.
[0048] The node temperature change rate is obtained by subtracting the highest node temperature at the current sampling time from the highest node temperature at the adjacent sampling time.
[0049] Using the highest node temperature, node temperature change rate, and heat dissipation capacity attenuation coefficient corresponding to 720 consecutive sampling points in the thermal state dataset as training samples, and taking the minimum mean square error as the objective function, the least squares method is used to perform offline fitting on the bias weights, the weights corresponding to the highest node temperature, the weights corresponding to the node temperature change rate, the weights corresponding to the heat dissipation capacity attenuation coefficient, the weights corresponding to the cross term of the highest node temperature and the heat dissipation capacity attenuation coefficient, and the weights corresponding to the historical highest node temperature in the temperature rise prediction model to obtain the initial weight coefficient set.
[0050] The highest node temperature, the rate of change of node temperature, and the heat dissipation capacity attenuation coefficient are input into the temperature rise prediction model. The model is then used to progressively calculate the future predicted temperature sequence by summing six items: the predicted temperature at the current moment equals the product of the bias weight, the highest node temperature and its corresponding weight, the rate of change of node temperature and its corresponding weight, the heat dissipation capacity attenuation coefficient and its corresponding weight, the highest node temperature and the heat dissipation capacity attenuation coefficient and their corresponding weights, and the historical highest node temperature and its corresponding weights.
[0051] Specifically, the temperature rise prediction model adopts a linear weighted structure, consisting of six input nodes, six corresponding weight coefficients, and one bias weight. The six input nodes correspond to the highest node temperature, the node temperature change rate, the heat dissipation capacity decay coefficient, the interaction term between the highest node temperature and the heat dissipation capacity decay coefficient, the historical highest node temperature, and the bias term, respectively. The output node is the predicted temperature value at a future prediction step. Introducing the interaction term between the highest node temperature and the heat dissipation capacity decay coefficient as an independent input node is a key structural design feature that distinguishes this scheme from existing temperature rise prediction models. This interaction term captures the nonlinear amplification effect of heat dissipation device performance degradation on the temperature rise rate of high-temperature nodes. That is, at the same highest node temperature level, the larger the heat dissipation capacity decay coefficient, the more significantly the temperature rise rate is amplified. Simply using these two as independent input nodes cannot reflect this coupling relationship. The historical highest node temperature refers to the highest node temperature at the previous sampling time. Introducing this variable enables the model to perceive the temporal inertia of temperature changes and reflect the continuity of temperature variations. The node temperature change rate refers to the difference between the highest node temperature at the current sampling time and the highest node temperature at the previous sampling time. It reflects the instantaneous rate of temperature increase or decrease. A positive difference indicates that the temperature is on an upward trend, and a negative difference indicates that the temperature is on a downward trend. This variable, together with the historical highest node temperature, provides the model with information on the temporal direction of temperature.
[0052] The offline fitting process uses 720 consecutive sampling points from the thermal state dataset as training samples. These 720 sampling points correspond to a 60-minute historical data window. The window length is determined based on the fact that, under normal operating conditions, the complete temperature rise process caused by the load current switching from low load to full load in a distribution cabinet typically takes 30 to 60 minutes. The 720 sampling points can fully cover at least one temperature rise process, ensuring that the training samples contain sufficient information on dynamic temperature changes. The least squares method is executed as follows: The six input values and corresponding measured temperature values of each record in the 720 training samples are substituted into a system of linear equations. Six weight coefficients and bias weights are used as variables to be solved. The objective function is to minimize the sum of the squares of the differences between the predicted and measured temperatures of all training samples. By solving the normal equations, the weight coefficient combination that minimizes the objective function is obtained, i.e., the initial weight coefficient set. The prediction step size is set from 1 to 12, corresponding to temperature predictions for 12 time points from 5 to 60 seconds in the future. Each prediction step size corresponds to an independent linear weighted summation operation. The six types of input values at the current time are substituted into the weight coefficient combination of the corresponding prediction step size to progressively calculate the temperature values at the 12 prediction time points, which are then arranged in chronological order to form a future predicted temperature sequence. Dynamic binding means that the heat dissipation capacity attenuation coefficient is updated synchronously with the thermal state dataset in each 5-second sampling period and directly used as the input value of the temperature rise prediction model in the current sampling period to participate in the weighted summation operation. This ensures that the performance parameters of the heat dissipation device used in each prediction operation are the measured values under the current physical conditions, rather than fixed rated design values.
[0053] In one specific embodiment, in step S4, when a predicted temperature in the future predicted temperature sequence exceeds a safety threshold, the output power command of the heat dissipation device is proportionally compensated using a heat dissipation capacity attenuation coefficient, triggering the heat dissipation device to perform a cooling action before the temperature exceeds the limit, including:
[0054] By iterating through the future predicted temperature sequence, the prediction step size corresponding to the first time the predicted temperature exceeds the safety threshold is selected, and the minimum over-limit prediction step size is obtained.
[0055] Based on the heat dissipation capacity attenuation coefficient and rated power, the output power command is calculated according to the rule that the output power command is equal to the sum of the rated power and the heat dissipation capacity attenuation coefficient minus the unit value, multiplied by the attenuation compensation coefficient, and then multiplied by the rated power.
[0056] Before the time window corresponding to the minimum over-limit prediction step size is reached, the output power command is converted into a PWM signal and sent to the heat dissipation device drive controller to trigger the heat dissipation device to perform cooling action.
[0057] Specifically, the safety threshold is determined based on the lowest rated temperature resistance of the electrical components in the distribution cabinet, set at 70 degrees Celsius. This value is referenced to the upper limit of the long-term allowable operating temperature of key electrical components such as circuit breaker contacts and busbar connection points. Operating below this temperature ensures that the insulation material does not undergo accelerated aging. Traversing the future predicted temperature sequence means starting from the predicted temperature with a prediction step size of 1, comparing each predicted temperature in ascending order of prediction step size with the safety threshold of 70 degrees Celsius. The prediction step size corresponding to the first predicted temperature exceeding 70 degrees Celsius is the minimum over-limit prediction step size. Multiplying this step size by the sampling period of 5 seconds gives the remaining time window from the current moment until the predicted temperature first exceeds the limit. The heat dissipation device must start up and reach the target speed within this time window to achieve cooling intervention before the actual temperature exceeds the limit. The calculation rule for the output power command is as follows: subtract the unit value of 1 from the heat dissipation capacity attenuation coefficient to obtain the attenuation range. Multiply the attenuation range by the attenuation compensation coefficient of 1.2 and then by the rated power to obtain the compensation increment. Add the compensation increment to the rated power to obtain the output power command. The attenuation compensation coefficient of 1.2 is based on the fact that under the condition that the performance attenuation range of the heat dissipation device does not exceed 20%, a compensation coefficient of 1.2 times can ensure that the actual heat dissipation power is not lower than the rated level. The upper limit of the output power command is set to 1.5 times the rated power. When the upper limit is exceeded, the upper limit value shall be executed to prevent the heat dissipation device from being damaged by overload operation of the motor.
[0058] PWM (Pulse Width Modulation) signal controls the average input voltage of the cooling fan motor by adjusting the proportion of high-level pulses within a single cycle, thereby controlling the fan speed. The ratio of the output power command to the rated power is the duty cycle of the PWM signal. When the duty cycle is 1, the fan operates at the rated speed. When the duty cycle is greater than 1, the PWM carrier frequency is increased to achieve operation above the rated speed. This conversion relationship is executed by the power-duty cycle mapping table built into the cooling device driver controller, which is preset at the factory. The time window corresponding to the minimum over-limit prediction step size refers to the remaining time between the current sampling time and the predicted first over-limit time. During this time, the PWM signal must complete transmission and drive the fan to reach the target speed. The response time of the fan from rest to the target speed is usually no more than 3 seconds, while the minimum time window corresponding to the minimum over-limit prediction step size is at least 5 seconds, ensuring that the fan completes its response before the predicted over-limit time arrives.
[0059] In one specific embodiment, step S4, subtracting the measured temperature from the corresponding predicted temperature in the future predicted temperature sequence to obtain the prediction execution deviation, includes:
[0060] After the heat dissipation device performs the cooling action, when the time window corresponding to the minimum over-limit prediction step size has passed, the node temperature of each heat source node is re-acquired, and the maximum value of the node temperature of each heat source node is filtered to obtain the actual measured temperature after execution.
[0061] The difference between the measured temperature after execution and the predicted temperature corresponding to the smallest out-of-limit prediction step size in the future predicted temperature sequence is used to obtain the prediction execution deviation.
[0062] Specifically, the timing starts at the moment when the heat dissipation device drive controller receives the PWM signal and confirms the fan start. From this moment, after a time length equal to the minimum over-limit prediction step multiplied by the sampling period of 5 seconds, the temperature of each heat source node is re-acquired. The re-acquisition operation is completely consistent with the acquisition method in step S1. The output values of the temperature sensors for five types of heat source nodes—circuit breaker contacts, busbar connection points, cable joints, transformer winding surfaces, and cabinet air outlets—are read. The maximum value of the five node temperature values is taken to obtain the measured temperature after execution. This operation is consistent with the selection rule for the highest node temperature in step S3, ensuring that the measured temperature after execution corresponds completely with the predicted temperature in the future predicted temperature sequence in terms of data scope. The predicted temperature corresponding to the minimum over-limit prediction step is selected as the comparison benchmark because the control command of the heat dissipation device is triggered based on the predicted temperature exceeding the safety threshold at this prediction step. The difference between the measured temperature after the same time window and the predicted temperature directly reflects the accuracy of the prediction at the corresponding time node.
[0063] The prediction execution deviation is obtained by subtracting the predicted temperature corresponding to the smallest out-of-limit prediction step in the future predicted temperature sequence from the actual measured temperature after execution. A positive difference indicates that the actual temperature is higher than the predicted temperature, which means that the temperature rise prediction model underestimates the actual temperature rise and the correction amount of the heat dissipation capacity attenuation coefficient in step S3 is insufficient. The corresponding weight coefficient needs to be updated in the direction of increasing. A negative difference indicates that the actual temperature is lower than the predicted temperature, which means that the temperature rise prediction model overestimates the actual temperature rise. The corresponding weight coefficient needs to be updated in the direction of decreasing. When the absolute value of the difference does not exceed the preset deviation tolerance threshold of 1 degree Celsius, it is determined that the prediction result matches the measured result and does not trigger the subsequent online update process in step S4. The prediction execution deviation is passed to the next stage as the input to drive the online update.
[0064] In one specific embodiment, step S4 involves synchronously updating the weight coefficients corresponding to the heat dissipation capacity attenuation coefficient and the reference thermal resistance in the temperature rise prediction model based on the prediction execution deviation, including:
[0065] When the absolute value of the prediction execution deviation exceeds the preset deviation tolerance threshold, the product of the prediction execution deviation and the heat dissipation capacity attenuation coefficient is multiplied by the preset learning rate to update the weight coefficients corresponding to the heat dissipation capacity attenuation coefficient in the temperature rise prediction model in the gradient descent direction, and the updated weight coefficients are obtained.
[0066] Based on the sign direction of the predicted execution deviation and the difference between the current heat dissipation capacity attenuation coefficient and the heat dissipation capacity attenuation coefficient of the previous sampling period, the reference thermal resistance is corrected in the same direction with a preset correction step size to obtain the updated reference thermal resistance.
[0067] The updated weighting coefficients and the updated baseline thermal resistance are written back to the temperature rise prediction model parameter storage area and the baseline thermal resistance storage area, respectively, and will take effect in the next sampling period.
[0068] Specifically, the preset deviation tolerance threshold is set to 1 degree Celsius. This value is based on the fact that the measurement accuracy of temperature sensors in the distribution cabinet is typically ±0.5 degrees Celsius. A tolerance threshold of 1 degree Celsius is higher than the sensor's measurement error range, ensuring that the prediction execution deviation that triggers online updates originates from model prediction deviation rather than sensor measurement error. The preset learning rate is set to 0.01. This value is based on the fact that an excessively large learning rate will cause the weight coefficients to oscillate during the update process and fail to converge, while an excessively small learning rate will cause the weight coefficients to update too slowly and fail to keep up with the performance degradation of the heat dissipation device. A learning rate of 0.01 ensures that the weight coefficients can effectively converge within 10 sampling periods under the condition that the performance degradation rate of the heat dissipation device is typically 3% to 5% increase in thermal resistance per 100 hours. The update rule for the gradient descent direction is as follows: the updated weight coefficient is equal to the original weight coefficient corresponding to the heat dissipation capacity attenuation coefficient minus the product of the prediction execution deviation and the heat dissipation capacity attenuation coefficient, and then multiplied by the learning rate of 0.01 to obtain the correction amount. When the prediction execution deviation is positive, the correction amount is positive, the updated weight coefficient decreases, and the contribution of the heat dissipation capacity attenuation coefficient to the predicted temperature in the next sampling period decreases; when the prediction execution deviation is negative, the correction amount is negative, the updated weight coefficient increases, and the contribution of the heat dissipation capacity attenuation coefficient to the predicted temperature increases.
[0069] The preset correction step size is set to 0.005. This value is set based on the following: the correction range of the reference thermal resistance must be less than the upper limit of the actual change range of the thermal resistance in a single sampling period of the heat dissipation device. Under normal dust accumulation rate, the change in thermal resistance of the heat dissipation device in a single sampling period usually does not exceed 0.5% of the reference thermal resistance. The correction step size of 0.005 matches this change range. The specific meaning of the same-direction correction is: a logical AND operation is performed between the sign of the predicted execution deviation and the sign of the difference between the current heat dissipation capacity attenuation coefficient and the heat dissipation capacity attenuation coefficient of the previous sampling period. When the signs are the same, the updated reference thermal resistance is equal to the original reference thermal resistance plus the product of the correction step size of 0.005 and the original reference thermal resistance. When the signs are different, the updated reference thermal resistance is equal to the original reference thermal resistance minus the product of the correction step size of 0.005 and the original reference thermal resistance, ensuring that the correction direction of the reference thermal resistance is consistent with the actual degradation trend of the heat dissipation device. The temperature rise prediction model parameter storage area refers to the address range in the controller's built-in storage unit specifically used to store all weight coefficients of the temperature rise prediction model. The reference thermal resistance storage area refers to the address range in the same storage unit specifically used to store the current value of the reference thermal resistance. The write-back operation is performed after all data processing steps of the current sampling period are completed. After the write-back is completed, the new parameter values will take effect in step S2 of the next 5-second sampling period.
[0070] The above describes the heat dissipation control method for distribution cabinets based on temperature rise prediction in the embodiments of this application. The following describes the heat dissipation control system for distribution cabinets based on temperature rise prediction in the embodiments of this application. Please refer to [link / reference]. Figure 2 One embodiment of the heat dissipation control system for distribution cabinets based on temperature rise prediction in this application includes:
[0071] The data acquisition module is used to collect the node temperature of each heat source node in the power distribution cabinet, as well as the inlet temperature, outlet temperature and real-time air volume of the heat dissipation device, to obtain a thermal state dataset.
[0072] The analysis module is used to compare the real-time thermal resistance of the heat dissipation device with the factory-calibrated reference thermal resistance based on the inlet temperature, outlet temperature and real-time air volume in the thermal state data set, and obtain the heat dissipation capacity attenuation coefficient.
[0073] The binding module is used to take the highest node temperature in the thermal state dataset and the heat dissipation capacity decay coefficient as input variables of the temperature rise prediction model, so that the prediction result of the temperature rise prediction model is dynamically bound to the current physical decay state of the heat dissipation device to obtain the future predicted temperature sequence.
[0074] The compensation module is used to proportionally compensate the output power command of the heat dissipation device with the heat dissipation capacity attenuation coefficient when there is a predicted temperature in the future predicted temperature sequence that exceeds the safety threshold. Before the temperature exceeds the limit, the heat dissipation device is triggered to perform a cooling action. The difference between the measured temperature and the corresponding predicted temperature in the future predicted temperature sequence is used to obtain the prediction execution deviation. The prediction execution deviation is used to drive the weight coefficient corresponding to the heat dissipation capacity attenuation coefficient in the temperature rise prediction model and the reference thermal resistance to be updated synchronously online.
[0075] above Figure 2 The heat dissipation control system for distribution cabinets based on temperature rise prediction in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The heat dissipation control device for distribution cabinets based on temperature rise prediction in this embodiment of the invention will be described in detail from the perspective of hardware processing.
[0076] Reference Figure 3 This invention also provides a heat dissipation control device for power distribution cabinets based on temperature rise prediction. This heat dissipation control device can be a server, and its internal structure can be as follows: Figure 3 As shown, the temperature rise prediction-based power distribution cabinet heat dissipation control device includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor, designed as a computer, provides computing and control capabilities. The memory of the temperature rise prediction-based power distribution cabinet heat dissipation control device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the temperature rise prediction-based power distribution cabinet heat dissipation control device stores the data corresponding to this embodiment. The network interface of the temperature rise prediction-based power distribution cabinet heat dissipation control device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.
[0077] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the heat dissipation control device for the distribution cabinet based on temperature rise prediction applied thereto.
[0078] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the distribution cabinet heat dissipation control method based on temperature rise prediction.
[0079] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0080] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a temperature rise prediction-based power distribution cabinet heat dissipation control device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0081] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A heat dissipation control method for distribution cabinets based on temperature rise prediction, characterized in that, The method includes: Step S1: Collect the node temperature of each heat source node in the power distribution cabinet, as well as the inlet temperature, outlet temperature and real-time air volume of the heat dissipation device, to obtain the thermal state dataset. Step S2: Based on the inlet temperature, outlet temperature and real-time air volume in the thermal state data set, the real-time thermal resistance of the heat dissipation device is compared with the factory-calibrated reference thermal resistance to obtain the heat dissipation capacity attenuation coefficient. Step S3: Use the highest node temperature in the thermal state dataset and the heat dissipation capacity attenuation coefficient as input variables for the temperature rise prediction model, so that the prediction result of the temperature rise prediction model is dynamically bound to the current physical attenuation state of the heat dissipation device, and obtain the future predicted temperature sequence. Step S4: When a predicted temperature in the future predicted temperature sequence exceeds a safety threshold, the output power command of the heat dissipation device is proportionally compensated using the heat dissipation capacity attenuation coefficient. Before the temperature exceeds the limit, the heat dissipation device is triggered to perform a cooling action. The difference between the measured temperature and the corresponding predicted temperature in the future predicted temperature sequence is used to obtain the prediction execution deviation. The prediction execution deviation is used to drive the weight coefficient corresponding to the heat dissipation capacity attenuation coefficient in the temperature rise prediction model and the reference thermal resistance to be updated synchronously online.
2. The heat dissipation control method for distribution cabinets based on temperature rise prediction according to claim 1, characterized in that, Step S1 includes: Temperature sensors are installed at the circuit breaker contacts, busbar connection points, cable joints, transformer winding surfaces, and air outlets of the distribution cabinet. The node temperature of each heat source node is collected at a preset sampling period to obtain the node temperature sequence. Temperature sensors are arranged on the air inlet and air outlet sides of the heat dissipation device, and the air inlet temperature, air outlet temperature and real-time air volume are collected synchronously by a wind speed sensor at the same sampling period as the node temperature sequence to obtain the heat dissipation operation parameter sequence. Based on the node temperature sequence and the heat dissipation operation parameter sequence, they are synchronously merged in a timestamp-aligned manner to obtain a thermal state dataset.
3. The heat dissipation control method for distribution cabinets based on temperature rise prediction according to claim 1, characterized in that, In step S2, the real-time thermal resistance of the heat dissipation device is compared with the factory-calibrated reference thermal resistance to obtain the heat dissipation capacity attenuation coefficient, including: The inlet and outlet temperature difference is obtained by subtracting the inlet and outlet temperatures from the thermal state dataset. The real-time heat dissipation power is obtained by multiplying the temperature difference between the inlet and outlet air with the real-time air volume, air density, and air specific heat capacity at constant pressure. The real-time thermal resistance is obtained by comparing the temperature difference between the air inlet and outlet with the real-time heat dissipation power. The heat dissipation capacity attenuation coefficient is obtained by comparing the real-time thermal resistance with the factory-calibrated reference thermal resistance.
4. The heat dissipation control method for distribution cabinets based on temperature rise prediction according to claim 1, characterized in that, In step S3, the highest node temperature in the thermal state dataset and the heat dissipation capacity attenuation coefficient are used together as input variables for the temperature rise prediction model. This dynamically binds the prediction results of the temperature rise prediction model with the current physical attenuation state of the heat dissipation device, resulting in a future predicted temperature sequence, including: The highest node temperature is obtained by filtering the node temperatures of each heat source node in the thermal state dataset. The node temperature change rate is obtained by subtracting the highest node temperature at the current sampling time from the highest node temperature at the adjacent sampling time. Using the highest node temperature, the node temperature change rate, and the heat dissipation capacity attenuation coefficient corresponding to 720 consecutive sampling points in the thermal state dataset as training samples, with the minimum mean square error as the objective function, the least squares method is used to perform offline fitting on the bias weights, the weights corresponding to the highest node temperature, the weights corresponding to the node temperature change rate, the weights corresponding to the heat dissipation capacity attenuation coefficient, the weights corresponding to the cross term of the highest node temperature and the heat dissipation capacity attenuation coefficient, and the weights corresponding to the historical highest node temperature in the temperature rise prediction model to obtain the initial weight coefficient group. The highest node temperature, the node temperature change rate, and the heat dissipation capacity attenuation coefficient are input into the temperature rise prediction model. The model is then used to progressively calculate the future predicted temperature sequence by summing six items: the current predicted temperature equals the product of the bias weight, the highest node temperature and its corresponding weight, the node temperature change rate and its corresponding weight, the heat dissipation capacity attenuation coefficient and its corresponding weight, the highest node temperature and the heat dissipation capacity attenuation coefficient and their corresponding weights, and the historical highest node temperature and its corresponding weights.
5. The heat dissipation control method for distribution cabinets based on temperature rise prediction according to claim 1, characterized in that, In step S4, when a predicted temperature in the future predicted temperature sequence exceeds a safety threshold, the output power command of the heat dissipation device is proportionally compensated using the heat dissipation capacity attenuation coefficient, triggering the heat dissipation device to perform a cooling action before the temperature exceeds the limit, including: The predicted temperature sequence is traversed, and the prediction step size corresponding to the first time the predicted temperature exceeds the safety threshold is selected to obtain the minimum over-limit prediction step size. Based on the heat dissipation capacity attenuation coefficient and rated power, the output power command is calculated according to the rule that the output power command is equal to the sum of the rated power and the heat dissipation capacity attenuation coefficient minus the unit value, multiplied by the attenuation compensation coefficient, and then multiplied by the rated power. Before the time window corresponding to the minimum over-limit prediction step size is reached, the output power command is converted into a PWM signal and sent to the heat dissipation device drive controller to trigger the heat dissipation device to perform cooling action.
6. The heat dissipation control method for distribution cabinets based on temperature rise prediction according to claim 5, characterized in that, In step S4, the difference between the measured temperature and the corresponding predicted temperature in the future predicted temperature sequence is used to obtain the prediction execution deviation, including: After the heat dissipation device performs the cooling action, when the time window corresponding to the minimum over-limit prediction step size has passed, the node temperature of each heat source node is re-acquired, and the maximum value of the node temperature of each heat source node is filtered to obtain the actual measured temperature after execution. The difference between the measured temperature after execution and the predicted temperature corresponding to the minimum over-limit prediction step size in the future predicted temperature sequence is used to obtain the prediction execution deviation.
7. The heat dissipation control method for distribution cabinets based on temperature rise prediction according to claim 6, characterized in that, In step S4, the weighting coefficients corresponding to the heat dissipation capacity attenuation coefficient and the reference thermal resistance in the temperature rise prediction model are synchronously updated online using the prediction execution deviation, including: When the absolute value of the prediction execution deviation exceeds the preset deviation tolerance threshold, the product of the prediction execution deviation and the heat dissipation capacity attenuation coefficient is multiplied by the preset learning rate to update the weight coefficients corresponding to the heat dissipation capacity attenuation coefficient in the temperature rise prediction model in the gradient descent direction, so as to obtain the updated weight coefficients. Based on the sign direction of the predicted execution deviation and the difference between the current heat dissipation capacity attenuation coefficient and the heat dissipation capacity attenuation coefficient of the previous sampling period, the reference thermal resistance is corrected in the same direction with a preset correction step size to obtain the updated reference thermal resistance. The updated weighting coefficients and the updated reference thermal resistance are written back to the temperature rise prediction model parameter storage area and the reference thermal resistance storage area, respectively, and will take effect in the next sampling period.
8. A heat dissipation control system for a distribution cabinet based on temperature rise prediction, characterized in that, For implementing the distribution cabinet heat dissipation control method based on temperature rise prediction as described in any one of claims 1-7, the distribution cabinet heat dissipation control system based on temperature rise prediction includes: The data acquisition module is used to collect the node temperature of each heat source node in the power distribution cabinet, as well as the inlet temperature, outlet temperature and real-time air volume of the heat dissipation device, to obtain a thermal state dataset. The analysis module is used to compare the real-time thermal resistance of the heat dissipation device with the factory-calibrated reference thermal resistance based on the inlet temperature, outlet temperature and real-time air volume in the thermal state data set, and obtain the heat dissipation capacity attenuation coefficient. The binding module is used to take the highest node temperature in the thermal state dataset and the heat dissipation capacity decay coefficient as input variables of the temperature rise prediction model, so that the prediction result of the temperature rise prediction model is dynamically bound to the current physical decay state of the heat dissipation device to obtain the future predicted temperature sequence. The compensation module is used to compensate the output power command of the heat dissipation device proportionally with the heat dissipation capacity attenuation coefficient when there is a predicted temperature in the future predicted temperature sequence that exceeds the safety threshold. Before the temperature exceeds the limit, the heat dissipation device is triggered to perform a cooling action. The difference between the measured temperature and the corresponding predicted temperature in the future predicted temperature sequence is used to obtain the prediction execution deviation. The prediction execution deviation is used to drive the weight coefficient corresponding to the heat dissipation capacity attenuation coefficient in the temperature rise prediction model and the reference thermal resistance to be updated synchronously online.
9. A heat dissipation control device for a distribution cabinet based on temperature rise prediction, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the heat dissipation control method for distribution cabinets based on temperature rise prediction as described in any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the heat dissipation control method for distribution cabinets based on temperature rise prediction as described in any one of claims 1 to 7.