Intelligent high-temperature power supply temperature monitoring method and system

Abstract

The invention relates to the technical field of power supply temperature abnormity identification, in particular to an intelligent high-temperature power supply temperature monitoring method and system. The method comprises the following steps: acquiring temperature data of a power supply module; performing temperature heat dissipation processing and temperature change processing according to the power supply module temperature data to obtain power supply module temperature heat dissipation data and power supply module temperature change data; performing power module temperature heat dissipation abnormity judgment according to the power module temperature heat dissipation data to obtain temperature heat dissipation abnormity data; performing power supply module temperature change abnormity judgment according to the power supply module temperature change data to obtain temperature change abnormity data; and performing temperature anomaly curve construction according to the temperature heat dissipation anomaly data and the temperature change anomaly data to obtain temperature anomaly curve data so as to perform high-temperature power supply temperature monitoring auxiliary operation. Based on temperature fluctuation analysis, heat transfer lag calculation and time sequence prediction, the accuracy of temperature anomaly judgment is improved, and false alarms and missing alarms are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power supply temperature anomaly recognition, and in particular to an intelligent high-temperature power supply temperature monitoring method and system. Background Art

[0002] Power modules operating in high temperature environments (such as industrial control systems, power electronic equipment, new energy battery management systems (BMS), server data centers, aerospace electronic equipment, etc.) have extremely strict requirements for temperature management. The temperature control of power modules not only affects their operating efficiency, but also directly determines their reliability and service life. However, existing temperature monitoring methods have obvious deficiencies in the following aspects. Existing methods usually adopt a fixed-interval temperature acquisition strategy, which cannot perceive rapid temperature fluctuations in real time, and lack dynamic analysis of temperature change trends. It is easy to miss the precursor stage of abnormalities, resulting in low heat dissipation management efficiency and delayed abnormality identification, which is difficult to meet the needs of complex working environments. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes an intelligent high-temperature power supply temperature monitoring method and system to solve at least one of the above technical problems.

[0004] The present application provides a method for monitoring temperature of an intelligent high-temperature power supply, comprising the following steps: Step S1: Acquire power module temperature data; Step S2: performing temperature heat dissipation processing and temperature change processing according to the temperature data of the power module to obtain the temperature heat dissipation data of the power module and the temperature change data of the power module; Step S3: judging abnormality of heat dissipation of the power module temperature according to the heat dissipation data of the power module temperature, and obtaining abnormal heat dissipation data of the power module temperature; Step S4: judging abnormal temperature change of the power module according to the temperature change data of the power module, and obtaining abnormal temperature change data; Step S5: constructing a temperature anomaly curve according to the temperature heat dissipation abnormality data and the temperature change abnormality data to obtain temperature anomaly curve data for performing auxiliary work of high-temperature power supply temperature monitoring.

[0005] In the present invention, the center + edge multi-point temperature measurement can more accurately reflect the overall temperature distribution of the power module and avoid the error of single-point collection. The thermal conductivity of the power module is calculated by the heat flux density model, and the parts with insufficient heat dissipation are identified to facilitate the optimization of the heat dissipation design. Combined with the temperature gradient analysis, the abnormal heat dissipation area is found to reduce the risk of local overheating. The heat dissipation performance is monitored in real time. When it is found that the temperature continues to rise but the heat dissipation power fails to increase accordingly, the system can immediately alarm to prevent equipment damage caused by poor heat dissipation. By using the rate of change analysis, it is possible to identify abnormal situations where the temperature rises or falls rapidly, and prevent equipment overheating or cooling failure caused by current fluctuations, load abnormalities and other factors. By comprehensively analyzing the temperature heat dissipation abnormality data and the temperature change abnormality data, a dynamic abnormality curve is constructed, which can not only identify the current abnormality, but also predict the future temperature abnormality, thereby achieving early warning and improving operation and maintenance efficiency.

[0006] Preferably, step S1 specifically includes: Step S11: controlling a temperature sensor preset at the center of the power module to collect the temperature of the power module to obtain the center temperature data of the power module; Step S12: controlling a temperature sensor preset at an edge of the power module to collect the temperature of the power module to obtain edge temperature data of the power module; Step S13: constructing a spatial matrix for the power module center temperature data and the power module edge temperature data to obtain the power module temperature data.

[0007] The present invention adopts a center + edge dual temperature sensor arrangement, which can simultaneously obtain temperature information of core components and peripheral areas, avoiding data errors caused by single-point monitoring. The center temperature data can reflect the heat load of the core components of the power supply (such as power modules and converters), and the edge temperature data can evaluate the efficiency of the heat dissipation system to ensure a comprehensive grasp of the temperature status. By constructing a spatial matrix model, the temperature gradient distribution can be analyzed more accurately, improving the reliability of temperature monitoring.

[0008] Preferably, step S2 specifically includes: Step S21: performing temperature heat dissipation rate processing according to the temperature data of the power module to obtain temperature heat dissipation rate data; Step S22: constructing a spatial gradient map of the temperature heat dissipation rate data to obtain the temperature heat dissipation data of the power module; Step S23: performing a center-edge heat accumulation effect simulation according to the power module temperature heat dissipation data and the power module temperature data to obtain the center-edge heat accumulation effect data; Step S24: Calculate the constant temperature change based on the center-edge heat accumulation effect data to obtain the power module temperature change data.

[0009] In the present invention, the traditional method usually only monitors the temperature value, while the present method dynamically evaluates the heat dissipation performance by calculating the temperature heat dissipation rate (i.e., the rate at which the temperature decreases over time). The sliding window calculation is used to smooth the heat dissipation rate in different time periods, identify the trend change of heat dissipation efficiency, and improve the stability of the data. If the heat dissipation rate decreases abnormally (for example, the temperature is at a high level for a long time and cannot be effectively reduced), the system will warn of heat dissipation failures in advance, such as heat dissipation fan failure, cooling system efficiency reduction, etc. Based on the temperature heat dissipation rate data, a spatial gradient map is constructed to analyze the heat dissipation capacity of different regions. The heat conduction equation is used to calculate the heat flow and form a temperature distribution map to more accurately identify areas of local overheating or insufficient heat dissipation. Combined with heat flow modeling, the heat accumulation inside the power module is simulated to predict whether there is a risk of continuous overheating. Compared with the traditional static temperature control strategy, this method can predict the temperature change trend, provide intelligent temperature control decisions, and improve the efficiency of temperature control management.

[0010] Preferably, step S3 specifically includes: Step S31: judging the abnormality of heat dissipation of material properties according to the heat dissipation data of the power module temperature, and obtaining abnormal heat dissipation data of material properties; Step S32: Correct the heat dissipation effect of ambient temperature and humidity according to the material property heat dissipation abnormality data and the power module temperature heat dissipation data to obtain ambient temperature and humidity heat dissipation correction data; Step S33: Perform heat transfer medium interaction processing according to the ambient temperature and humidity heat dissipation correction data to obtain temperature heat dissipation abnormality data.

[0011] In the present invention, the actual heat dissipation rate is calculated and compared with the theoretical heat dissipation rate. If the deviation is too large, it means that the material is aged, damaged or contaminated (such as dust accumulation on the heat sink). The present invention introduces temperature and humidity correction to ensure that the heat dissipation anomaly can still be accurately evaluated under different environmental conditions. Humidity affects the heat capacity and thermal conductivity of the air. Excessive humidity reduces the heat dissipation capacity of the air and even causes condensation to affect the power module. The correction of the heat dissipation effect of the ambient temperature and humidity can improve the dynamic adaptability. Different heat transfer media (such as air, liquid coolant) have different heat dissipation characteristics. A deeper degree of heat dissipation anomaly judgment is performed based on the interaction of the heat transfer media.

[0012] Preferably, step S4 is specifically: Step S41: judging abnormal temperature change of material properties according to the temperature change data of the power module, and obtaining abnormal temperature change data of the material; Step S42: Perform temperature, humidity and heat dissipation correction according to the abnormal temperature change data of the material to obtain the abnormal temperature change data.

[0013] The present invention improves the accuracy of abnormality detection based on parameters such as material thermal expansion coefficient and thermal conductivity, and adopts temperature and humidity correction to improve the accuracy of temperature anomaly detection under different environmental conditions.

[0014] Preferably, step S5 is specifically: Step S51: constructing an abnormal curve according to the temperature heat dissipation abnormal data and the temperature change abnormal data to obtain abnormal curve data; Step S52: Acquire standard power module temperature data; Step S53: constructing a temperature curve according to the standard power module temperature data to obtain temperature curve data; Step S54: extracting similarities and differences based on the abnormal curve data and the temperature curve data to obtain the temperature abnormal curve data to perform auxiliary work for high-temperature power supply temperature monitoring.

[0015] In the present invention, the temperature data of the power module under normal working conditions is selected to construct a standard curve. Hierarchical modeling is adopted to generate corresponding standard temperature curves for different ambient temperatures and different load conditions, so as to make a more accurate comparison. Historical data fitting (such as polynomial regression and Bayesian optimization) is adopted to improve the accuracy of the standard curve. As a reference benchmark for temperature anomaly analysis, it can be used to evaluate whether the current temperature curve is normal. Combined with the dynamic comparison mechanism, the appropriate standard curve is selected under different working conditions to improve the accuracy of anomaly detection. Traditional methods can often only provide a judgment of "whether it is abnormal", while this method can accurately analyze the type, severity and possible causes of the anomaly by extracting similarities and differences in features.

[0016] Preferably, the temperature heat dissipation rate processing is specifically as follows: Performing heat dissipation rate processing of the original area temperature according to the temperature data of the power module to obtain first temperature heat dissipation rate data; Perform cross-region temperature heat dissipation rate processing according to the power module temperature data to obtain cross-region temperature heat dissipation rate data; The heat flow simulation of the heat dissipation medium is performed according to the cross-region temperature heat dissipation rate data to obtain the second temperature heat dissipation rate data.

[0017] In the present invention, traditional heat dissipation monitoring mainly focuses on overall temperature changes, while the present invention performs independent heat dissipation rate analysis on local areas of the power module (such as power modules, heat sinks, air ducts, etc.) to improve the accuracy of local heat dissipation monitoring. For example, the power components of the power module change temperature faster than the heat sink area. Analysis of the heat dissipation rate in this area can identify hot spots in advance. Traditional methods are difficult to accurately analyze the flow of heat between different areas, and this method can provide more accurate reflection of cross-regional temperature heat dissipation data by calculating the heat dissipation rate across regions. If the temperature drop rate in a certain area is slower than that in the adjacent area, it means that the heat in the area is retained, and it is judged that the heat dissipation design needs to be optimized, providing a more refined data perspective from this dimension. Based on thermal flow simulation (such as CFD computational fluid dynamics simulation), by simulating the heat flow of the heat dissipation medium, the heat transfer path can be more accurately optimized and the heat dissipation efficiency can be improved.

[0018] Preferably, the simulation of the center-edge heat collection effect is specifically as follows: Extract fluctuation characteristics based on the temperature data of the power module to obtain temperature fluctuation characteristic data; The center-edge temperature transfer hysteresis is calculated based on the temperature fluctuation characteristic data to obtain the center-edge hysteresis coefficient data; Perform thermal balance processing according to the center edge hysteresis coefficient data and the power module temperature heat dissipation data to obtain the power module thermal balance data; The heat accumulation effect is calculated based on the thermal balance data of the power module and the temperature and heat dissipation data of the power module to obtain the center-edge heat accumulation effect data.

[0019] By extracting the temperature fluctuation characteristics in the present invention, short-term temperature anomalies, sudden temperature rises, and long-period temperature fluctuations can be identified, thereby improving the anomaly detection capability. There is a hysteresis effect when heat is transferred from the center to the edge of the power module. If the lag time is too long, heat will accumulate and reduce the heat dissipation efficiency. By analyzing the temperature fluctuation characteristics, the sensitivity of temperature anomaly detection is improved to ensure that the heat dissipation system can accurately respond to short-term and long-term temperature changes. This method optimizes the heat transfer process by calculating the center-edge temperature hysteresis coefficient, and improves the optimization capability of the heat dissipation path by calculating the temperature transfer hysteresis, ensuring that heat can be evenly diffused and preventing local heat accumulation. The present invention uses thermal balance processing to achieve the accuracy of the simulation of the center-edge heat accumulation effect, thereby providing accurate data support. According to the calculation of the heat accumulation effect, the heat distribution can be predicted and optimized to prevent local overheating and improve the long-term stability of the equipment.

[0020] Preferably, the center-edge temperature transfer hysteresis calculation is specifically as follows: Perform central peak-valley calculation and edge peak-valley calculation according to the temperature fluctuation characteristic data to obtain central peak-valley data and edge peak-valley data respectively; Acquire power heat dissipation data, wherein the power heat dissipation data includes power heat dissipation position data and power heat dissipation power data; According to the power supply heat dissipation data, the heat dissipation effect hysteresis time is calculated for the temperature position data corresponding to the power supply module temperature data to obtain the heat dissipation effect time coefficient data; According to the heat dissipation effect time coefficient data, the adjacent peak-valley transfer influence selection is performed on the central peak-valley data and the edge peak-valley data to obtain the adjacent peak-valley transfer data; The hysteresis coefficient is calculated based on the adjacent peak-to-valley transfer data to obtain the center-edge temperature transfer hysteresis data.

[0021] In the present invention, by calculating the peak and valley data of the center and the edge, the periodicity, amplitude and time delay of temperature fluctuations can be accurately monitored, and the accuracy of temperature anomaly detection can be improved. There is a time lag in the diffusion of heat from the center to the edge of the power module. By calculating the heat dissipation effect time coefficient, the heat dissipation effect can be more accurately reflected. Traditional heat dissipation management strategies do not consider the transfer relationship between temperature peaks and valleys, which can easily lead to overcooling or insufficient heat dissipation. Alternatively, traditional temperature control methods only focus on single-point temperature over-limit alarms or overall temperature mean analysis, ignoring the time transfer relationship between peaks and valleys. This method calculates the peak-to-valley time difference, transfer amplitude ratio, and similarity distribution of the center temperature and edge temperature, constructs a mathematical model, and accurately calculates the hysteresis coefficient, making temperature control more accurate and the heat dissipation system more intelligent. The present invention accurately depicts the temperature response delay at the center and edge of the power module, deeply reflects the heat transfer path, and improves the accuracy and dynamic adaptability of heat dissipation anomaly judgment.

[0022] Preferably, the present invention also provides an intelligent high-temperature power supply temperature monitoring system for executing the intelligent high-temperature power supply temperature monitoring method as described above, the intelligent high-temperature power supply temperature monitoring system comprising: A power module temperature data acquisition module is used to obtain the power module temperature data; The power module temperature analysis module is used to perform temperature heat dissipation processing and temperature change processing according to the power module temperature data, and obtain the power module temperature heat dissipation data and the power module temperature change data; A power module temperature heat dissipation abnormality judgment module is used to judge the power module temperature heat dissipation abnormality according to the power module temperature heat dissipation data to obtain the temperature heat dissipation abnormality data; A power module temperature change abnormality judgment module is used to judge the power module temperature change abnormality according to the power module temperature change data to obtain the temperature change abnormality data; The temperature anomaly curve construction module is used to construct the temperature anomaly curve according to the temperature heat dissipation anomaly data and the temperature change anomaly data to obtain the temperature anomaly curve data for auxiliary operation of high-temperature power supply temperature monitoring.

[0023] The beneficial effect of the present invention is that by combining the temperature data of the center position with the temperature data of the edge, a spatial temperature matrix is ​​constructed, which can accurately capture the local temperature hot spots, the heat dissipation path and the overall temperature distribution. By calculating the temperature heat dissipation rate of the local area and the cross-regional temperature heat dissipation rate, the heat dissipation efficiency of different parts can be analyzed, and the high temperature accumulation area can be accurately identified. Combined with thermal fluid simulation (CFD) to optimize the heat dissipation path of the air-cooled or liquid-cooled system, improve the heat dissipation uniformity, and avoid local overheating problems. By analyzing the thermal conductivity, thermal expansion coefficient, specific heat capacity and other parameters of the power module material, the heat dissipation anomaly caused by material aging or deformation can be identified. Combined with temperature and humidity compensation calculation, misjudgment caused by changes in ambient temperature or humidity can be avoided, and the accuracy of abnormal judgment can be improved. Through abnormal curve modeling, the long-term characteristics of temperature heat dissipation anomaly and temperature change anomaly can be comprehensively analyzed to realize the visualization of temperature anomaly trend. By comparing the standard temperature curve, it can be identified whether the current temperature curve deviates from the normal working range, and the accuracy of abnormal judgment can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting implementations made with reference to the following drawings: Figure 1 A flowchart showing a method for monitoring temperature of an intelligent high-temperature power supply according to an embodiment of the present invention is provided; Figure 2 A flowchart showing a method for collecting temperature data of a power module according to an embodiment of the present invention is shown; Figure 3 A flow chart showing the steps of a power module temperature analysis method according to an embodiment is shown; Figure 4 A flowchart showing a method for determining abnormal heat dissipation of a power module temperature according to an embodiment of the present invention is provided; Figure 5 A flow chart showing the steps of a method for constructing a temperature anomaly curve according to an embodiment is shown. DETAILED DESCRIPTION

[0025] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0026] In addition, the drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0027] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0028] See also Figures 1 to 5 The present application provides a method for monitoring temperature of an intelligent high-temperature power supply, comprising the following steps: Step S1: Acquire power module temperature data; In one embodiment, temperature sensors (such as thermocouples, PT100, NTC thermistors, etc.) are arranged at multiple key parts of the power module (such as power devices, heat sinks, PCB boards), and temperature data is collected through A / D converters. In extreme high temperature environments (such as above 175°C), fiber Bragg grating (FBG) temperature sensors are used to collect temperature data to improve measurement accuracy and anti-interference capabilities. For power modules that are not suitable for wiring, wireless temperature sensors (such as Bluetooth, Zigbee, LoRa, etc.) are used to transmit temperature data to the main control unit.

[0029] In one embodiment, the temperature sensor is arranged at a key position of the power module, for example, directly mounted on the top of the power module (MOSFET / IGBT) package. Installed on the heat sink, the heat sink is located in the area where the heat flow is most concentrated, such as the center and edge of the heat sink. Installed in the PCB thermal sensitive area, the area with thick copper on the power PCB board (such as high power loop). Using NTC thermistor or K-type thermocouple, the measurement range can reach , accuracy The sensor is connected to a 24-bit ADC (such as ADS1115) for high-precision data acquisition. 10Hz (10 acquisitions per second) to ensure temperature response speed. Use I²C or SPI bus to transmit to a microcontroller (such as STM32). If the environment is not suitable for wiring, LoRa / Zigbee wireless temperature sensors can be used to transmit data.

[0030] Step S2: performing temperature heat dissipation processing and temperature change processing according to the temperature data of the power module to obtain the temperature heat dissipation data of the power module and the temperature change data of the power module; In one embodiment, based on finite element analysis (FEA) or heat transfer model (such as Fourier heat conduction equation), the theoretical heat dissipation performance under different working conditions is calculated and compared with the measured data. The fan speed is controlled by PWM (pulse width modulation) to increase or decrease the heat dissipation capacity. The coolant flow rate is monitored and the flow rate is adjusted according to the temperature change to improve the heat dissipation efficiency. When the temperature is too high, the output power is reduced by adjusting the PWM control signal of the power module to prevent thermal runaway.

[0031] In one embodiment, the real-time heat dissipation efficiency is calculated, and the thermal resistance is equal to the module temperature minus the ambient temperature divided by the power consumption. If the thermal resistance is greater than 1.5 K / W (standard value), it indicates poor heat dissipation. The fan is controlled by PWM to adjust the speed. When T < 60°C, the fan speed is maintained at 40%. When 60°C ≤ T < 80°C, the fan speed increases linearly to 100%. When T ≥ 80°C, the fan runs at full speed and an alarm is sounded. The coolant flow rate is monitored to ensure that the flow rate is ≥ 1.5 L / min. The system is normal. When the flow rate is < 1.5 L / min, the coolant pump is triggered to increase the pressure. When the flow rate is < 0.5 L / min, an alarm is sounded and the power output is reduced.

[0032] Step S3: judging abnormality of heat dissipation of the power module temperature according to the heat dissipation data of the power module temperature, and obtaining abnormal heat dissipation data of the power module temperature; In one embodiment, based on the temperature heat dissipation curve under normal operation, a threshold is set, and when the actual heat dissipation speed is lower than the preset standard (such as the cooling rate <5°C / min), the heat dissipation is determined to be abnormal. The ratio of T = Tmodule-Tenvironment) to power consumption P (Rth = T / P), to determine whether the heat dissipation performance has decreased. When the thermal resistance exceeds the design value (such as Rth>1.5K / W), it indicates poor heat dissipation. Monitor the fan speed. If the speed drops by more than 30% or stops running, it is determined that the air cooling system is abnormal. Monitor the coolant temperature and flow rate. If the flow rate drops or the temperature continues to rise, it is determined that the liquid cooling system is faulty.

[0033] Step S4: judging abnormal temperature change of the power module according to the temperature change data of the power module, and obtaining abnormal temperature change data; In one embodiment, the temperature rise rate (dT / dt) per unit time is calculated. If the temperature rise rate exceeds the safety value (e.g. >10°C / min), the temperature is judged to be abnormal. The sliding window method (e.g. the temperature change range in the past 5 minutes >20°C) is used to determine whether there is abnormal temperature fluctuation. If the temperature exceeds a certain threshold (e.g. T>150°C), an alarm is triggered or the frequency is automatically reduced. If the temperature exceeds the limit (e.g. T>175°C), an emergency shutdown is triggered.

[0034] Step S5: constructing a temperature anomaly curve according to the temperature heat dissipation abnormality data and the temperature change abnormality data to obtain temperature anomaly curve data for performing auxiliary work of high-temperature power supply temperature monitoring.

[0035] In one embodiment, based on the collected temperature anomaly data, a temperature anomaly curve is constructed using algorithms such as polynomial regression, neural network, or support vector regression (SVR). Combined with historical fault data, a machine learning model (such as LSTM, CNN) is used to identify potential abnormal patterns. Predict future temperature anomalies and take countermeasures in advance. When the temperature anomaly curve exceeds the set range, the system automatically sends an alarm message (such as SMS, APP notification). Linked temperature protection mechanism, such as adjusting the heat dissipation strategy or triggering an emergency power outage.

[0036] In one embodiment, polynomial regression is used to fit the temperature anomaly curve by the least square method: ,in is the temperature anomaly curve data, For temperature heat dissipation abnormal data, is the abnormal temperature change data, is the acceleration data of abnormal temperature change data, is the cubic term of abnormal temperature change data, a high-order nonlinear effect, representing the dramatic fluctuation of the temperature change rate. is the time parameter item. If the temperature is too high, the linear temperature rise rate increases, indicating that the heat dissipation capacity decreases. It is necessary to increase the fan speed to improve the heat dissipation capacity, or increase the coolant flow to improve the heat exchange efficiency. If the temperature is too high, heat accumulation will accelerate. Due to aging of heat dissipation materials or environmental influences, it is necessary to replace the heat sink with a high thermal conductivity (such as copper instead of aluminum), or reduce the ambient temperature (such as increasing air flow). If the temperature is too high, the temperature curve fluctuates violently, which is a dynamic heat dissipation anomaly (such as fan fluctuation). It is necessary to stabilize the fan power supply to ensure a constant speed, or use a coolant with higher heat capacity to reduce local temperature changes.

[0037] Preferably, step S1 specifically includes: Step S11: controlling a temperature sensor preset at the center of the power module to collect the temperature of the power module to obtain the center temperature data of the power module; In one embodiment, a K-type thermocouple (measuring range -40°C~175°C, accuracy ±1°C) or an NTC thermistor (suitable for high temperature environment) is selected. For extremely high temperature environments (such as >175°C), a fiber Bragg grating (FBG) sensor is used. Paste the thermocouple in the center of the power device (such as MOSFET, IGBT), and fix it with high-temperature resistant silicone or welding to ensure optimal heat transfer. If a PCB embedded temperature sensor is used, a sensor pad is reserved at the central power path copper layer of the PCB, and data is collected through an I²C or SPI interface. The temperature signal is read through a 24-bit ADC (such as ADS1115), sampling 10 times per second (10Hz) to ensure sufficient response speed. If the sensor uses wireless transmission, the ZigBee / LoRa protocol is used to collect temperature data and send it to the control unit.

[0038] Step S12: controlling a temperature sensor preset at an edge of the power module to collect the temperature of the power module to obtain edge temperature data of the power module; In one embodiment, a K-type thermocouple, NTC thermistor or FBG fiber optic temperature sensor is used in the same position as the center. A sensor is installed at each of the four corners of the power module to monitor the temperature in an evenly distributed manner. The temperature sensor is installed by bolting, thermally conductive adhesive bonding or directly embedded in the PCB to ensure stable sensing. Samples are taken 10 times per second (10Hz). Send to the control unit via I²C / SPI or LoRa / ZigBee.

[0039] Step S13: constructing a spatial matrix for the power module center temperature data and the power module edge temperature data to obtain the power module temperature data.

[0040] In one embodiment, the matrix form is set: ,in, is the power module temperature data, is the temperature data of the first edge sensor, is the temperature data of the second edge sensor, is the temperature data of the third edge sensor, is the temperature data of the fourth edge sensor, wherein the first edge sensor, the second edge sensor, the third edge sensor and the fourth edge sensor are sensors installed at different edge positions of the power module, It is the temperature data of the center sensor.

[0041] Preferably, step S2 specifically includes: Step S21: performing temperature heat dissipation rate processing according to the temperature data of the power module to obtain temperature heat dissipation rate data; In one embodiment, the temperature heat dissipation rate is calculated using the central difference method, that is, 1 second is used as the sampling interval, and the temperature data of the current sampling is subtracted from the temperature data of the previous sampling interval and divided by the sampling interval to obtain the temperature heat dissipation rate data.

[0042] Step S22: constructing a spatial gradient map of the temperature heat dissipation rate data to obtain the temperature heat dissipation data of the power module; In one embodiment, the gradient is calculated using the finite difference method, that is, along the X direction: , and along the Y direction: ,in is the distance between the temperature data of the first edge sensor and the temperature data of the second edge sensor, is the distance between the temperature data of the second edge sensor and the temperature data of the fourth edge sensor.

[0043] Step S23: performing a center-edge heat accumulation effect simulation according to the power module temperature heat dissipation data and the power module temperature data to obtain the center-edge heat accumulation effect data; In one embodiment, the temperature difference between the center and the edge is calculated : .like , indicating that heat accumulates in the center and the heat dissipation is uneven; if , indicating that the heat distribution is uniform and there is no obvious heat accumulation effect. Calculate the heat flow per unit area : , is the thermal conductivity, which is obtained by querying the database based on the material properties. The temperature gradient is the temperature heat dissipation data of the power module. Finite element analysis (FEA) is used to simulate the heat flow and generate isotherm diagrams to intuitively display the heat accumulation area to calculate the temperature distribution under different environmental conditions (wind speed, coolant flow rate).

[0044] Step S24: Calculate the constant temperature change based on the center-edge heat accumulation effect data to obtain the power module temperature change data.

[0045] In one embodiment, the power module transfers heat to the environment mainly in the form of convection, satisfying the temperature change law under Newton cooling conditions: ,in is the current temperature of the power module (instantaneous temperature), that is, the module temperature observed at a certain moment, is the convection heat transfer coefficient (W / m²·K). is the ambient temperature, It is approximated by differential form in the actual system. Set the steady-state temperature formula: , is the heat generation power per unit surface area of ​​the module (unit: W / m²), which can be obtained by converting the input power and the area of ​​the heating area. , it is considered to have reached a steady state.

[0046] Preferably, step S3 specifically includes: Step S31: judging the abnormality of heat dissipation of material properties according to the heat dissipation data of the power module temperature, and obtaining abnormal heat dissipation data of material properties; In one embodiment, the thermal resistance of the material is calculated : ,in, is the power module temperature, is the ambient temperature, is the module power consumption, is the thickness of the material. Calculate the thermal conductivity of the module material (such as aluminum, copper, ceramic substrate): ,in, is the thermal conductivity (W / m·K), such as 385W / m·K for copper and 205W / m·K for aluminum. is the heat dissipation area, is the temperature gradient. Set the base thermal conductivity. If the calculated thermal resistance If it is more than 10% higher than the design value, the material thermal resistance is considered abnormal (due to material aging, oxidation, and surface contamination).

[0047] Step S32: Correct the heat dissipation effect of ambient temperature and humidity according to the material property heat dissipation abnormality data and the power module temperature heat dissipation data to obtain ambient temperature and humidity heat dissipation correction data; In one embodiment, the air convection heat dissipation coefficient ( ): ,in is the convective heat transfer constant measured experimentally, is the ambient temperature, is the surface temperature of the power module, is the temperature correction coefficient, which is used to adjust the temperature dependence of the heat transfer coefficient, taking into account the effect of temperature on air flow characteristics (such as changes in Reynolds number and Prandtl number). The empirical value is , depending on the cooling conditions. Natural convection: , forced convection: , is the wind speed. Calculate the effect of humidity on thermal conductivity. When the humidity is high (>70%), if the radiator surface is wet, the thermal conductivity will decrease by 5%~15%. If the humidity H>80%, the thermal resistance An increase of 10% results in a decrease in heat dissipation capacity. High humidity affects the thermal conductivity of the heat sink: ,in, Because high humidity affects the thermal conductivity of the radiator, The thermal conductivity of the heat dissipation material under normal conditions (such as aluminum ), is the humidity impact factor (range 5% -15%, depending on humidity). Under normal circumstances: If humidity H=80%: A 10% drop in thermal conductivity means a reduction in heat dissipation capacity.

[0048] Step S33: Perform heat transfer medium interaction processing according to the ambient temperature and humidity heat dissipation correction data to obtain temperature heat dissipation abnormality data.

[0049] In one embodiment, the heat capacity of the coolant is calculated as: ,in, is the mass of coolant (kg), is the specific heat capacity of the coolant (J / kg·K), is the temperature difference between the coolant inlet and outlet (°C). Monitor the coolant flow rate. If the flow rate is <1.5L / min, the coolant flow rate is too low. If the temperature difference is <3°C, the coolant heat exchange efficiency is reduced. Calculate the air cooling system efficiency: , is the heat dissipation power, is the convective heat transfer coefficient, The heat dissipation area refers to the heat dissipation surface area of ​​the power module in contact with the air, including the surface area of ​​the heat sink and heat pipe. is the surface temperature of the power module, is the ambient air temperature. If the wind speed decreases, If any of the above conditions is triggered (including: too low flow rate, small liquid cooling temperature difference, reduced air cooling power), the system will generate temperature and heat dissipation abnormality data and mark it according to the abnormality type: liquid cooling heat dissipation abnormality; air cooling heat dissipation efficiency abnormality; multi-mode heat exchange coordination failure abnormality.

[0050] Preferably, step S4 is specifically: Step S41: judging abnormal temperature change of material properties according to the temperature change data of the power module, and obtaining abnormal temperature change data of the material; In one embodiment, the central difference method is used to calculate the temperature change rate per unit time. Calculate the thermal expansion rate of the material ,in, is the length change of the material due to temperature change (m), is the initial length of the material (m), is the temperature change data. If the temperature change rate |dT / dt|>10°C / s and the thermal expansion rate of the material exceeds the standard range by 10%, the material is judged to be abnormal. To judge that the material temperature change is abnormal, set the material stress calculation: ,in, is the internal thermal stress of the material (Pa), is the elastic modulus of the material (such as copper Pa). If the stress is >108 Pa (material stress limit), the material will crack or be damaged, which is considered to be an abnormal temperature change of the material.

[0051] Step S42: Perform temperature, humidity and heat dissipation correction according to the abnormal temperature change data of the material to obtain the abnormal temperature change data.

[0052] In one embodiment, the system obtains environmental humidity data H (unit: %), which is sampled regularly (for example, updated every 10 seconds) by a humidity sensor deployed in the power module housing or cooling channel, and is used to sense the humidity level of the air or cooling medium in real time. Based on the humidity value, the system sets the humidity influence factor , which is used to reflect the degree to which humidity weakens the thermal conductivity of materials. The specific settings are as follows: , , The system then calculates the base thermal conductivity (For example: aluminum material =205 W / m\cdotpK) to calculate the corrected thermal conductivity: ,in is the corrected thermal conductivity (W / m·K), is the thermal conductivity of the material under standard dry conditions, is the humidity influence factor (5%-15%, depending on the humidity ).

[0053] Preferably, step S5 is specifically: Step S51: constructing an abnormal curve according to the temperature heat dissipation abnormal data and the temperature change abnormal data to obtain abnormal curve data; In one embodiment, the temperature heat dissipation abnormal data and temperature change abnormal data in the last 30 minutes are selected, the time is aligned, and a smooth abnormal curve is generated using cubic spline interpolation: ,in is the temperature anomaly curve data, For temperature heat dissipation abnormal data, is the abnormal temperature change data, is the acceleration data of abnormal temperature change data, is the cubic term of abnormal temperature change data, a high-order nonlinear effect, representing the dramatic fluctuation of the temperature change rate. It is the time parameter item.

[0054] Step S52: Acquire standard power module temperature data; In one embodiment, normal operating data with similar ambient temperature and humidity and the same load power are selected, and a moving window method (10-minute window) is used to obtain short-term standard temperature data.

[0055] Step S53: constructing a temperature curve according to the standard power module temperature data to obtain temperature curve data; In one embodiment, the standard temperature curve is calculated: , is the smoothing factor, take 0.3, for Standard power module temperature data at the moment, for Standard power module temperature data at the moment.

[0056] Step S54: extracting similarities and differences based on the abnormal curve data and the temperature curve data to obtain the temperature abnormal curve data to perform auxiliary work for high-temperature power supply temperature monitoring.

[0057] In one embodiment, the error between the abnormal curve and the standard curve is calculated, the mean square error is calculated, the data segment with an error less than or equal to a threshold is marked as convergent feature data, the data segment with an error greater than the threshold is marked as significant abnormal feature data, and the two are combined to obtain temperature anomaly curve data.

[0058] In one embodiment, further, time-frequency quantization analysis is performed based on the abnormal curve data and the temperature curve data to obtain time-frequency quantum data; a fusion network is constructed based on the time-frequency quantum data to obtain temperature curve fusion network data; entropy flow dynamics evolution is performed based on the temperature curve fusion network data to obtain temperature curve entropy flow data; phase change detection is performed on the temperature curve entropy flow data to obtain temperature anomaly curve data, so as to perform auxiliary operations for high-temperature power supply temperature monitoring.

[0059] Map the temperature signal to the quantum state through wavelet transform: ;in is the quantum state corresponding to the temperature signal, is the normalization factor, is the wavelet component index (indicating different frequency levels), is the base of natural logarithms, is an imaginary unit, For the The frequency of the wavelet component, is the time variable, is the energy ground state corresponding to the abnormal curve data / temperature curve data. Assume the abnormal event as the observation operator and calculate its collapse probability distribution: ,in The quantum state is in the ground state The projection probability on the temperature signal is extracted by wavelet transform and mapped to the quantum state. The probability distribution of the quantum state reflects the contribution of the temperature signal in different frequency and energy modes. The probability of abnormal events is calculated and high-probability abnormal points are detected. Combined with wavelet transform, the time scale of abnormal occurrence can be accurately located.

[0060] Construct a 3rd-order tensor: time × temperature gradient × anomaly level, and use Tucker decomposition to extract core features: , It is a data structure containing three dimensions: time (T) × temperature gradient (G) × abnormal level (L). For (modal The product) represents the Dimensional matrix multiplication of tensors, taking values ​​of 1, 2, 3, is the core tensor, a compressed low-dimensional representation containing the main feature information. is the time modal feature matrix (i.e., time-frequency quantum data), which describes the characteristics of the temperature signal in the time dimension. is the temperature gradient modal characteristic matrix, describing the pattern of temperature change rate, is the abnormal level modal feature matrix, reflecting the characteristic distribution of the abnormal level, , It is a spatiotemporal correlation map tensor. After removing the abnormal level data, it only contains the core features of time and temperature gradients.

[0061] , is the total entropy of the system, the total amount of disorder or information entropy of the system, which is used to measure the changing trend of the system state. is the time variable of system evolution, describing the change of entropy over time, is the heat transferred to the system and is used to calculate the thermal entropy contribution, The temperature corresponding to the heat transfer affects the change of thermal entropy. Status in the system The probability of occurrence is used to calculate information entropy. is the anomaly correlation tensor, which represents the interaction strength between anomaly energies, is the abnormal event energy, which expresses the energy of a certain abnormal state. is the energy of the abnormal event, the energy of another abnormal state, and interaction, is the thermodynamic entropy term, describing the entropy change due to heat transfer, It is the information entropy term, which describes the uncertainty of the system state. The parameters in the entropy term are directly or indirectly calculated from the temperature curve fusion network data.

[0062] , is the order parameter used to detect the change of entropy with temperature and energy Variation of phase change sensitivity, is the system entropy, the total amount of disorder or information entropy of the system, which is used to measure the change of system state during phase transition. is the temperature of the system, one of the main variables controlling phase change, It is the total energy of the system or the energy of an abnormal event, which together with temperature affects the change in entropy. , is the phase change threshold. Phase change detection , calculate entropy Temperature and energy The second-order partial derivative of reflects whether the system is in a phase transition state. Sudden changes indicate that a phase transition has occurred in the system. Set the phase transition threshold ,when Exceeded time , the system enters an abnormal temperature state, thereby realizing anomaly detection based on convergent feature analysis of quantum computing, which is used to detect the key points of temperature anomaly and determine whether the cooling system is stable.

[0063] Preferably, the temperature heat dissipation rate processing is specifically as follows: Performing heat dissipation rate processing of the original area temperature according to the temperature data of the power module to obtain first temperature heat dissipation rate data; In one embodiment, the original area refers to the temperature change rate of a single component (such as a MOSFET, IGBT, or a certain area of ​​a heat sink). The central difference method is used for calculation, that is, the temperature difference of the original area is divided by the sampling time. The spatial variation of the heat dissipation rate is calculated, that is, the temperature gradient of the original area is calculated, and the temperature gradient along the X direction and the temperature gradient along the Y direction are calculated respectively to obtain the first temperature heat dissipation rate data.

[0064] Perform cross-region temperature heat dissipation rate processing according to the power module temperature data to obtain cross-region temperature heat dissipation rate data; In one embodiment, the temperature change behavior between multiple spatially adjacent devices (such as multiple MOSFETs, IGBT chips and heat dissipation structures) is analyzed to quantify the difference in heat dissipation efficiency across components, thereby identifying risk points such as uneven heat dissipation and heat accumulation. Based on a high-precision temperature sensor network, the system collects temperature time series data of each adjacent area and calculates the difference in temperature change rate: ,in For the The temperature profile of adjacent device regions, For the The temperature profile of adjacent device regions, is the temperature change rate calculated by forward difference or sliding average, The difference in temperature rise rate between adjacent areas. The larger the difference, the uneven heat diffusion and the potential existence of thermal barriers. The system builds a cross-region temperature gradient model and calculates the average temperature difference in the spatial direction: ,in For the , The physical distance between devices (unit: m) is used to estimate the directional temperature change. The system introduces Fourier's law of thermal conductivity to calculate the cross-region heat flux density: , For the With The cross-region heat flux density between devices (unit: W / m²), is the thermal conductivity of the corresponding material (unit: W / m·K), which is determined by the structural material properties. The system integrates the heat flux density calculation results between all adjacent regions into the cross-region temperature heat dissipation rate data.

[0065] The heat flow simulation of the heat dissipation medium is performed according to the cross-region temperature heat dissipation rate data to obtain the second temperature heat dissipation rate data.

[0066] In one embodiment, the medium flow rate and heat transfer capacity are calculated: , is the convective heat transfer value, is the convection heat transfer coefficient (W / m²·K), is the heat dissipation surface area (m²), is the surface temperature (°C), is the cooling medium or ambient temperature (°C). The system and specific heat capacity , calculate the temperature heat dissipation rate per unit mass (i.e. the second temperature heat dissipation rate data): , which reflects the heat dissipation capacity of the cooling medium under unit mass and unit specific heat conditions, and is an important indicator for evaluating heat flow efficiency. To further determine the flow state of the cooling medium, the system calculates the Reynolds number (Re) as follows: ,in is the coolant density (kg / m³), is the coolant flow rate (m / s), is the pipe diameter (m), which is calculated from the inner diameter of the pipe through which the coolant or air flows. is the dynamic viscosity (Pa·s). If Re<2300, it is judged as laminar flow with low heat dissipation efficiency. If Re>4000, it is judged as turbulent flow with high heat dissipation efficiency. If Re is between the two, it is a transition interval, and its trend can be dynamically evaluated. Calculate the heat transfer efficiency, and divide the heat dissipation capacity actually calculated by the current flow rate, medium properties and temperature difference by the ideal maximum heat exchange capacity calculated based on the optimal working condition of the cooling system design (which can be preset or obtained by simulation) to obtain the heat dissipation efficiency.

[0067] Preferably, the simulation of the center-edge heat collection effect is specifically as follows: Extract fluctuation characteristics based on the temperature data of the power module to obtain temperature fluctuation characteristic data; In one embodiment, the temperature fluctuation characteristics are extracted: ,in, is the temperature fluctuation characteristic data, is the sampling quantity data, is the sampling order term, For the The temperature value measured is the average temperature in the window.

[0068] The center-edge temperature transfer hysteresis is calculated based on the temperature fluctuation characteristic data to obtain the center-edge hysteresis coefficient data; In one embodiment, the system sets the edge temperature response of the power module to be approximately described as a time-delay expression of the center temperature, that is: ,in is the edge temperature, is the center temperature, is the sampling time, is the temperature transfer lag time (seconds). To preliminarily estimate the lag time, the heat diffusion theory formula is used: ,in is the estimated value of the basic lag time derived from the physical model, is the physical distance from the center to the edge (m), is the thermal diffusion coefficient of the material (m² / s). To enhance the adaptability of the model under complex working conditions, the system further introduces statistical correlation analysis between the center-edge temperature curves. Define the center temperature sequence and edge temperature series The mutual correlation coefficient for: , is the center temperature of the power module, is the average center temperature of the power module, is the edge temperature of the power module, is the average temperature at the edge of the power module, is the temperature fluctuation characteristic data of the center temperature of the power module, is the temperature fluctuation characteristic data of the edge temperature of the power module, , Approximately equal to The reciprocal of , , is the empirical correction coefficient (1-2), indicating that when the center and edge temperature changes are highly synchronized (i.e. →1), then →0; if the temperature changes of the two are unrelated ( approaches 0), the system determines that there is a significant lag in thermal response.

[0069] Perform thermal balance processing according to the center edge hysteresis coefficient data and the power module temperature heat dissipation data to obtain the power module thermal balance data; In one embodiment, the heat conduction equation is used: ,in, is the input heat (W) (i.e. calculated by power), is the heat dissipation (W) (calculated by temperature sensing through the sensor), is the heat capacity (J / K). For an accurate estimate , further introduce the heat transfer rate estimation model (such as one-dimensional steady-state heat conduction): ,in, is the heat transfer rate, is the thermal conductivity of the material (W / m·K), which is approximately regarded as , is the center temperature of the power module, is the heat transfer area (m²), is the edge temperature of the power module, is the heat transfer path (m). The system calculates the input and output heat difference at the current moment based on the above model and makes a relative error judgment: , It is the thermal balance judgment threshold data. If the heat income and expenditure are basically consistent (the error is less than 5%), the system will judge it as a thermal balance state, record the current state and enter the temperature stability trend modeling module.

[0070] The heat accumulation effect is calculated based on the thermal balance data of the power module and the temperature and heat dissipation data of the power module to obtain the center-edge heat accumulation effect data.

[0071] In one embodiment, in order to identify the heat concentration phenomenon (heat accumulation effect) inside the power module, the system uses a multi-dimensional method to calculate the heat accumulation effect based on the heat balance data and temperature heat dissipation data, and outputs the center-edge heat accumulation effect data as the key basis for abnormal judgment and temperature control strategy optimization. The system calculates the center temperature measurement point of the power module in real time With edge temperature measurement point The temperature difference between : ,like , then it is determined that there is a heat accumulation trend, where The temperature difference threshold (e.g. 5°C) is set according to the system material, environment, and load conditions. It supports static setting or dynamic adjustment based on historical data training. The heat flux intensity between each temperature measurement point is calculated using Fourier's Law: ,in is the heat flux intensity, is the thermal conductivity of the material (W / m·K), is the temperature gradient vector, which is obtained by dividing the temperature change in the center-edge direction by the distance difference. The difference in heat flux intensity between the center and the edge is further calculated: ,in is the center-edge heat flux intensity difference, is the heat flux density in the central area, is the heat flux density in the edge area, if , it means that the heat conduction efficiency in the center area is much lower than that in the edge area, and there is heat accumulation, which can be judged as a significant heat accumulation phenomenon. Can be set based on system heat load capacity.

[0072] In high-performance application scenarios, the system calls the heat conduction finite element analysis module to build a temperature field simulation model based on the current temperature distribution, material properties and boundary conditions to simulate the temperature evolution trend in the short term in the future. Taking the current moment as the starting point, the system predicts the temperature field changes in the next 5 minutes. If the simulation results show that the temperature in the central area has increased by more than 10% relative to the current value, it is determined that there is an enhanced heat accumulation trend, which can trigger the heat dissipation strategy adjustment and safety warning mechanism in advance. The center-edge heat accumulation effect data may include parameters such as center-edge temperature difference, heat flow difference, predicted temperature rise percentage, etc., and supports input into the abnormal curve construction module and temperature control strategy optimization module in the form of graphics, matrices or structured data, which is used for intelligent scheduling of air cooling / liquid cooling systems, adjustment of heat dissipation path design or output of user warning information.

[0073] Preferably, the center-edge temperature transfer hysteresis calculation is specifically as follows: Perform central peak-valley calculation and edge peak-valley calculation according to the temperature fluctuation characteristic data to obtain central peak-valley data and edge peak-valley data respectively; In one embodiment, the sliding window method is used to extract the temperature data of the last 10 minutes, the window size is set to 60 seconds, the first-order derivative is used to detect the extreme point, and the temperature change rate is calculated. If the temperature change rate changes its sign before and after the peak point (positive to negative), it is identified as a peak point. If the temperature change rate changes its sign before and after the valley point (negative to positive), it is identified as a valley point.

[0074] Acquire power heat dissipation data, wherein the power heat dissipation data includes power heat dissipation position data and power heat dissipation power data; In one embodiment, the position of a radiator of an air cooling system or a liquid cooling system is obtained, and the heat dissipation power of each heat dissipation point is recorded.

[0075] According to the power supply heat dissipation data, the heat dissipation effect hysteresis time is calculated for the temperature position data corresponding to the power supply module temperature data to obtain the heat dissipation effect time coefficient data; In one embodiment, in order to achieve modeling and compensation for the temperature response hysteresis effect of different temperature measurement points of the power module, the system calculates the heat dissipation response time using the heat transfer time delay model based on the heat diffusion theory. Specifically, based on the spatial distribution relationship of the temperature measurement point relative to the heat source (i.e., the heat dissipation position), the following is defined: , is the distance from the heat dissipation point to the temperature measurement point (m), is the thermal diffusion coefficient (m² / s). By calculating the , a time-delay response model between temperature changes and actual heat dissipation behavior can be established. This model is not only used to determine whether the temperature responses of the center and the edge are synchronized, but can also be used as a dynamic correction factor to adjust the timing of the temperature curve in subsequent peak-valley analysis and hysteresis coefficient judgment to improve the accuracy of hysteresis judgment. In addition, to ensure calculation accuracy, the system supports real-time updating of dynamic parameters of the heat diffusion path based on the thermal sensor network. When the system detects a sudden change in high load or a large change in ambient temperature and humidity, it automatically recalculates , thereby adapting to the impact of changes in the state of the heat dissipation medium on the heat transfer rate and achieving accurate time domain temperature prediction and abnormal warning.

[0076] According to the heat dissipation effect time coefficient data, the adjacent peak-valley transfer influence selection is performed on the central peak-valley data and the edge peak-valley data to obtain the adjacent peak-valley transfer data; In one embodiment, in order to identify whether there is a temperature peak-valley transfer effect between the center area and the edge area of ​​the power module, the system selects the neighboring peak-valley transfer influence of the center peak-valley data and the edge peak-valley data based on the principles of timing comparison and neighbor matching. Based on the peak points and valley points (i.e., local maximum and minimum values) extracted from the temperature data at the center position, the system searches for corresponding peak-valley points in the edge area temperature data within a ±5 second time window centered on the peak-valley time point. To ensure accurate comparison, the center and edge temperature data have been processed synchronously through sampling, and interpolation is introduced for time alignment when necessary. The system calculates the center-edge peak-valley time difference: ,in is the central peak-valley data, is the edge peak and valley data, seconds, it is considered that there is an effective thermal response transfer effect between the central peak valley and the edge peak valley, and it is recorded as effective adjacent peak valley transfer data. When a central peak valley point corresponds to multiple candidate edge peak valley points, the system selects the pair with the smallest time difference for matching and constructs the peak valley transfer mapping relationship between the center and the edge.

[0077] The hysteresis coefficient is calculated based on the adjacent peak-to-valley transfer data to obtain the center-edge temperature transfer hysteresis data.

[0078] In one embodiment, in order to further determine the heat conduction state between the center area and the edge area of ​​the power module, the system calculates the relative hysteresis ratio of the center and edge temperature responses based on the adjacent peak-to-valley transfer data obtained above. , used to characterize the thermal diffusion efficiency and heat retention. Calculate the relative hysteresis ratio between the center temperature and the edge temperature , is the lag of the center temperature change relative to the edge temperature, is the length of a single temperature fluctuation cycle in the center temperature curve of the power module, that is, the time required to drop from a peak value to a valley value, or to rise from a valley value to the next peak value, where Specifically, it represents the average time difference between the central peak and valley and the edge peak and valley (i.e., thermal response delay), which is calculated through multiple matching adjacent peak and valley pairs: , is the number of effective peak-valley pairs, is the ordinal index of the peak-valley pair, For the The timestamp of the central peak and valley point indicates the time point (in seconds) of the first identified peak or valley value in the temperature curve of the central area of ​​the power module. For the The timestamp of the edge peak and valley point indicates the temperature curve of the edge area of ​​the power module. The time point of the peak or valley value (unit: second) obtained by matching the time window, adjacent to the central peak and valley points. The relative rate of the central heat response rhythm and the heat diffusion to the edge. The larger the value, the more serious the lag is and the more difficult it is for the central heat to diffuse. When >1.5, it is considered that there is obvious heat retention in the center and the edge area responds slowly due to local blocking of the heat conduction structure, deviation of the cooling system, and decrease in the thermal diffusion coefficient of the material. At this time, the system will mark the value as the center-edge temperature transfer lag data; record abnormal events, and start the temperature control strategy adaptive adjustment module, such as increasing the edge fan speed, adjusting the liquid cooling flow distribution, etc.; provide input to the temperature anomaly curve modeling module to optimize the response timing of the temperature anomaly judgment model; issue a heat dissipation warning to the user terminal, prompting to check the local heat dissipation system or material status. The threshold of 1.5 can be dynamically configured based on experimental data, equipment model or operating environment. The system supports historical data training to optimize this value, and dynamically adjusts the judgment criteria according to the material properties of different power modules (such as copper, aluminum, ceramic) and layout structures to ensure the applicability and robustness of the model.

[0079] Calculate the relative hysteresis ratio of the center temperature to the edge temperature , is the lag of the center temperature change relative to the edge temperature, is the time interval from the peak value to the valley value or from the valley value to the peak value of the center temperature of the power module. >1.5, indicating that the center heat is seriously retained and there is a problem of uneven heat dissipation.

[0080] Preferably, the present invention also provides an intelligent high-temperature power supply temperature monitoring system for executing the intelligent high-temperature power supply temperature monitoring method as described above, the intelligent high-temperature power supply temperature monitoring system comprising: A power module temperature data acquisition module is used to obtain the power module temperature data; The power module temperature analysis module is used to perform temperature heat dissipation processing and temperature change processing according to the power module temperature data, and obtain the power module temperature heat dissipation data and the power module temperature change data; A power module temperature heat dissipation abnormality judgment module is used to judge the power module temperature heat dissipation abnormality according to the power module temperature heat dissipation data to obtain the temperature heat dissipation abnormality data; A power module temperature change abnormality judgment module is used to judge the power module temperature change abnormality according to the power module temperature change data to obtain the temperature change abnormality data; The temperature anomaly curve construction module is used to construct the temperature anomaly curve according to the temperature heat dissipation anomaly data and the temperature change anomaly data to obtain the temperature anomaly curve data for auxiliary operation of high-temperature power supply temperature monitoring.

[0081] Therefore, from any point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the attached application documents rather than the above description, and it is intended that all changes falling within the meaning and scope of equivalent elements of the application documents are included in the present invention.

[0082] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An intelligent high-temperature power supply temperature monitoring method, characterized in that: The following steps are involved: Step S1: Control the temperature sensor preset at the center position of the power module to collect the temperature of the power module to obtain the center temperature data of the power module; control the temperature sensor preset at the edge position of the power module to collect the temperature of the power module to obtain the edge temperature data of the power module; construct a spatial matrix for the center temperature data of the power module and the edge temperature data of the power module to obtain the power module temperature data; Step S2: Process the temperature heat dissipation rate according to the power module temperature data and construct a spatial gradient map to obtain the power module temperature heat dissipation data; perform center-edge temperature transfer hysteresis calculation according to the power module temperature data to obtain center-edge hysteresis coefficient data; perform heat accumulation effect calculation according to the center-edge hysteresis coefficient data and the power module temperature heat dissipation data to obtain center-edge heat accumulation effect data; perform temperature constant change calculation according to the center-edge heat accumulation effect data to obtain power module temperature change data; Step S3: judging abnormality of heat dissipation of the power module temperature according to the heat dissipation data of the power module temperature, and obtaining abnormal heat dissipation data of the power module temperature; Step S4: judging abnormal temperature change of the power module according to the temperature change data of the power module, and obtaining abnormal temperature change data; Step S5: constructing a temperature anomaly curve according to the temperature heat dissipation anomaly data and the temperature change anomaly data to obtain temperature anomaly curve data for performing auxiliary work of high-temperature power supply temperature monitoring.

2. The method according to claim 1, characterized in that Step S3 is specifically as follows: According to the power module temperature heat dissipation data, the material property heat dissipation abnormality is judged to obtain the material property heat dissipation abnormality data; Correct the heat dissipation effect of ambient temperature and humidity based on the abnormal heat dissipation data of material properties and the heat dissipation data of the power module temperature to obtain the corrected heat dissipation data of ambient temperature and humidity; According to the ambient temperature and humidity heat dissipation correction data, the heat transfer medium interaction processing is performed to obtain the temperature heat dissipation abnormality data.

3. The method according to claim 1, characterized in that Step S4 is specifically as follows: According to the temperature change data of the power module, the abnormal temperature change of the material property is judged to obtain the abnormal temperature change data of the material; Temperature and humidity heat dissipation correction is performed according to the abnormal temperature change data of the material to obtain the abnormal temperature change data.

4. The method according to claim 1, characterized in that Step S5 is specifically as follows: An abnormal curve is constructed according to the temperature heat dissipation abnormal data and the temperature change abnormal data to obtain abnormal curve data; Get standard power module temperature data; Construct a temperature curve according to the temperature data of the standard power module to obtain temperature curve data; The difference and similarity features are extracted based on the abnormal curve data and the temperature curve data to obtain the temperature abnormal curve data for auxiliary work of high-temperature power supply temperature monitoring.

5. The method according to claim 1, characterized in that The specific temperature heat dissipation rate processing is: Performing heat dissipation rate processing of the original area temperature according to the temperature data of the power module to obtain first temperature heat dissipation rate data; Perform cross-region temperature heat dissipation rate processing according to the power module temperature data to obtain cross-region temperature heat dissipation rate data; The heat flow simulation of the heat dissipation medium is performed according to the cross-region temperature heat dissipation rate data to obtain the second temperature heat dissipation rate data.

6. The method according to claim 1, characterized in that The simulation of the center-edge heat accumulation effect is as follows: Extract the fluctuation characteristics according to the temperature data of the power module to obtain the temperature fluctuation characteristic data; The center-edge temperature transfer hysteresis is calculated based on the temperature fluctuation characteristic data to obtain the center-edge hysteresis coefficient data; Perform thermal balance processing according to the center edge hysteresis coefficient data and the power module temperature heat dissipation data to obtain the power module thermal balance data; The heat accumulation effect is calculated based on the thermal balance data of the power module and the temperature and heat dissipation data of the power module to obtain the center-edge heat accumulation effect data.

7. The method according to claim 6, characterized in that The calculation of center-edge temperature transfer hysteresis is as follows: Perform central peak-valley calculation and edge peak-valley calculation according to the temperature fluctuation characteristic data to obtain central peak-valley data and edge peak-valley data respectively; Acquire power heat dissipation data, wherein the power heat dissipation data includes power heat dissipation position data and power heat dissipation power data; According to the power supply heat dissipation data, the heat dissipation effect hysteresis time is calculated for the temperature position data corresponding to the power supply module temperature data to obtain the heat dissipation effect time coefficient data; According to the heat dissipation effect time coefficient data, the adjacent peak-valley transfer influence selection is performed on the central peak-valley data and the edge peak-valley data to obtain the adjacent peak-valley transfer data; The hysteresis coefficient is calculated based on the adjacent peak-to-valley transfer data to obtain the center-edge temperature transfer hysteresis data.

8. An intelligent high-temperature power supply temperature monitoring system, characterized in that: Used to execute the intelligent high-temperature power supply temperature monitoring method as claimed in claim 1, the intelligent high-temperature power supply temperature monitoring system comprises: The power module temperature data acquisition module is used to control the temperature sensor preset at the center of the power module to collect the power module temperature and obtain the power module center temperature data; control the temperature sensor preset at the edge of the power module to collect the power module temperature and obtain the power module edge temperature data; construct a spatial matrix for the power module center temperature data and the power module edge temperature data to obtain the power module temperature data; The power module temperature analysis module is used to process the temperature heat dissipation rate and construct a spatial gradient map according to the power module temperature data to obtain the power module temperature heat dissipation data; perform center-edge temperature transfer hysteresis calculation according to the power module temperature data to obtain center-edge hysteresis coefficient data; perform heat accumulation effect calculation according to the center-edge hysteresis coefficient data and the power module temperature heat dissipation data to obtain center-edge heat accumulation effect data; perform temperature constant change calculation according to the center-edge heat accumulation effect data to obtain the power module temperature change data; A power module temperature heat dissipation abnormality judgment module is used to judge the power module temperature heat dissipation abnormality according to the power module temperature heat dissipation data to obtain the temperature heat dissipation abnormality data; A power module temperature change abnormality judgment module is used to judge the power module temperature change abnormality according to the power module temperature change data to obtain the temperature change abnormality data; The temperature anomaly curve construction module is used to construct the temperature anomaly curve according to the temperature heat dissipation anomaly data and the temperature change anomaly data to obtain the temperature anomaly curve data for auxiliary operation of high-temperature power supply temperature monitoring.

Images