Big data-based electric drive system overheat protection monitoring method and system

Abstract

The present application relates to the technical field of state monitoring, in particular to an electric drive system overheat protection monitoring method and system based on big data, comprising the following steps: obtaining a power rising section and extracting a temperature curve, judging whether temperature rise lags behind a mark response delay, expanding a window to identify thermal fluctuation lag, extracting peaks and valleys to judge trend deviation, positioning nodes to track a thermal path, and identifying an interruption structure to generate an early warning mark. In the present application, by combining the coupling relationship between power change trend and temperature rise time sequence during the operation of the electric drive system, the accurate identification of temperature rise response delay is realized, by continuously tracing the sequence of thermal node temperature rise, the position distribution of the interruption structure in the thermal path is determined, by combining the linkage performance between thermal resistance interruption structures, the concurrent features of overheat abnormalities are identified, the response capability and early warning accuracy of the electric drive system in the early stage of overheat risk are improved, and the rapid positioning of potential thermal fault risk and intelligent division of risk level are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of state monitoring, and in particular to an electric drive system overheat protection monitoring method and system based on big data. BACKGROUND

[0002] The technical field of state monitoring involves real-time sensing, data collection, state judgment and early warning of various system operation states, covering power equipment monitoring, mechanical and electrical system health assessment, industrial automation equipment temperature rise detection, electronic device working state analysis and other aspects. It is an important supporting technology to ensure system safety, improve equipment stability and operating efficiency. Through sensor deployment, data collection terminals, data analysis tools and monitoring models, the running environment and internal parameters of the target equipment or system are monitored and recorded, abnormal states and potential risks are identified, and a systematic state management and maintenance system is formed. Among them, the traditional electric drive system overheat protection monitoring method refers to the means of monitoring and protection control of the heat generated by the electric drive system during operation. The technical matter it aims at is that the electric drive system is prone to overheat risk under high load, long time operation or environmental temperature rise, which leads to performance decline or damage. The traditional method measures the surface temperature of the key components of the electric drive system by presetting temperature thresholds and using thermal resistance or thermocouple sensors, and performs timing comparison and trigger control logic based on the collected temperature data to perform load reduction, power-off or alarm operation, to identify and protect the overheat state of the electric drive system.

[0003] The existing technology uses fixed temperature thresholds and single temperature measurement nodes for overheat monitoring, relies on immediate comparison of surface temperature and single sampling result to judge the running state, and is difficult to capture the temperature rise hysteresis effect after power fluctuation. When there is a time mismatch between power mutation and temperature response, potential overheat risks cannot be identified in time, especially in scenarios where load fluctuation is frequent or environmental temperature changes significantly. The traditional periodic sampling method lacks continuous analysis capability for temperature rise trend, leading to misjudgment or missed report of thermal fault in the early stage. The temperature rise caused by short-time high load does not exceed the threshold but has accumulated heat hazards to key devices, and cannot respond effectively in the early stage. SUMMARY

[0004] The purpose of the present application is to solve the shortcomings in the prior art, and to provide an electric drive system overheat protection monitoring method and system based on big data.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme, an electric drive system overheat protection monitoring method based on big data, comprising the following steps:

[0006] S1: obtaining an input power continuous sequence during the operation of the electric drive system, identifying a time period with a continuously rising direction of adjacent sampling points, extracting a temperature change curve of a thermal sensitive node before and after a section, and generating a power thermal response time sequence lag section label;

[0007] S2: using the time section in the power thermal response time sequence lag section label, expanding an observation window of the same length as the mutation section backward on the basis of the original monitoring period, identifying whether the temperature peak value change is completed within the expanded window node by node, and generating a thermal fluctuation lag behavior identification list;

[0008] S3: using the node temperature curve identified in the thermal fluctuation lag behavior identification list, extracting the time sequence of the wave crest and the wave trough, comparing the peak value amplitude and the interval of the continuous period, measuring the temperature peak value amplitude in the period section, judging the amplitude change direction between adjacent periods, and generating a thermal fluctuation evolution trend offset boundary set;

[0009] S4: according to the node position in the thermal fluctuation evolution trend offset boundary set, locating the front and rear nodes connected in the electric drive system in the thermal conduction structure, tracking the continuity according to the structure connection order between the nodes, and generating a structure chain internal thermal resistance breaking path node group.

[0010] As a further scheme of the application, the power thermal response time sequence lag section label includes a response delay section start and end time, a thermal sensitive node index of the power continuous rising section, and a time difference value of the power mutation start point and the temperature rise start point, the thermal fluctuation lag behavior identification list includes a lag node index, a temperature rise peak value appearance time and an expanded window time boundary, the thermal fluctuation evolution trend offset boundary set includes a trend offset node, a continuous period peak value growth trend and a period shortening feature, and the structure chain internal thermal resistance breaking path node group includes a thermal path interruption node pair, a connection order number and a time section without temperature rise response.

[0011] As a further scheme of the application, the acquisition step of the power thermal response time sequence lag section label is specifically:

[0012] S111: obtaining an input power continuous sequence during the operation of the electric drive system, comparing the power values between adjacent sampling points, identifying a continuous paragraph with continuously increasing power values, recording the start and end time corresponding to the paragraph, and generating a power continuous rising interval sequence;

[0013] S112: based on the power continuous rising interval sequence, calling the temperature sampling data of the thermal sensitive nodes before and after the section, judging whether the rising trend of the temperature curve in the section exists a lag response compared with the power mutation start point, comparing whether there is a time difference between the time interval between the section start point and the temperature mutation start time, taking the time difference as the response delay basis, obtaining a response lag matching determination result;

[0014] S113: Based on the response lag matching determination result, filter the power rise segment that meets the lag condition, extract the power rise amplitude, temperature rise start delay time and temperature change value of the thermal node for each segment, calculate the power thermal response lag intensity value, and mark it according to the segment order to obtain the power thermal response time lag segment mark.

[0015] As a further aspect of the present invention, the steps for obtaining the thermal fluctuation hysteresis behavior identification list are as follows:

[0016] S211: Based on the time segment marked in the power thermal response time lag segment, extend the observation time range to the same length as the abrupt change segment on the basis of the original monitoring period, collect the temperature sequence data of the corresponding thermal nodes in the extended segment, and obtain the temperature sequence set of the extended segment.

[0017] S212: Based on the extended segment temperature sequence set, extract the time corresponding to the peak temperature value in the temperature time sequence of the thermal node, and compare it with the end time of the original monitoring cycle to determine whether there are nodes with peak temperature points distributed in the extended time period. Filter the nodes that meet the conditions and their corresponding time periods to obtain the peak delay node identification result.

[0018] S213: Based on the peak delay node identification results, extract the time difference between the original cycle termination time and the temperature peak occurrence time corresponding to the node, calculate the thermal fluctuation hysteresis index corresponding to the node, filter the nodes and time periods that meet the thermal peak delay, and establish a thermal fluctuation hysteresis behavior identification list.

[0019] As a further aspect of the present invention, the step of obtaining the thermal fluctuation evolution trend offset boundary set specifically includes:

[0020] S311: Based on the nodes identified in the thermal fluctuation hysteresis behavior identification list, extract the temperature change curves corresponding to the nodes, record the time positions corresponding to the peak and trough points within the continuous monitoring period, calculate the time span and temperature peak change within the period segment, and generate a periodic temperature feature sequence.

[0021] S312: Call the peak amplitude data and period interval data in the periodic temperature feature sequence, determine whether any three consecutive periods simultaneously satisfy the characteristic trend of increasing peak change amplitude and shrinking period interval, record the time range of the corresponding segment in the temperature curve for the segment that meets the conditions, and obtain the continuous offset trend identification result.

[0022] S313: Based on the continuous offset trend identification results, collect the temperature peak amplitude difference, local fluctuation intensity and peak position offset within each group of three periodic sequences, calculate the trend offset index of the three periodic segments, extract the range of segments where the offset trend meets the preset conditions, and obtain the thermal fluctuation evolution trend offset boundary set.

[0023] As a further aspect of the present invention, the step of obtaining the thermal blocking path node group within the structural chain specifically includes:

[0024] S411: Based on the location of the concentrated nodes of the offset boundary according to the thermal fluctuation evolution trend, locate the adjacent connected front and rear nodes one by one, extract the structural connection order between the nodes, and mark the connection status of each pair of nodes in the structural path to generate a structural connection order group.

[0025] S412: Based on the connection order of the node pairs in the structural connection order group, collect the corresponding temperature rise start time data, determine whether the subsequent node in the node pair shows a temperature rise phenomenon within the set monitoring window after the preceding node starts, mark the node pairs that do not show a temperature rise phenomenon, and generate a thermal blocking path node group within the structural chain.

[0026] As a further aspect of the present invention, the method further includes step S5:

[0027] S5: Based on the thermal blocking path node group in the structure chain, determine whether the overheating anomaly of the electric drive system is in the same time period. If the linkage triggering condition is met, classify the time period state as a high-level thermal warning state and generate a linkage thermal warning level identification result.

[0028] The results of the linkage thermal warning level identification include the high-level warning time period, linkage trigger node group, and abnormal status classification label.

[0029] As a further aspect of the present invention, the steps for obtaining the linkage thermal warning level identification result are specifically as follows:

[0030] S511: Based on the temperature change data of the nodes in the heat-blocking path node group within the structural chain, extract the time point at which the temperature rise start phenomenon first occurs within the monitoring period, call the node number and the continuous change segment of the temperature value in the corresponding time series, determine whether there is a stable upward behavior in the continuous temperature change trend, identify the correspondence between the node number and the time, and obtain the temperature rise start time series table.

[0031] S512: Based on the time distribution of nodes in the temperature rise start time sequence table, classify and identify the monitoring period to which they belong, and identify whether the start time of the nodes belongs to the same time period. The judgment basis is that the time period has a unique number in the preset time division. If the start time of all nodes belongs to the same numbered period, a linkage trigger state is formed, and a corresponding level label is assigned to the state, generating a linkage heat warning level label result.

[0032] The big data-based electric drive system overheat protection monitoring system is used to execute the above-mentioned big data-based electric drive system overheat protection monitoring method. The system includes:

[0033] The data monitoring module acquires the input power sequence data of the drive motor control unit in the electric drive system, compares the power values ​​at adjacent time points to construct a sequence of change direction, identifies the time period in which the power changes continuously, compares the time sequence of the temperature rise start time and the corresponding power change start time, and generates a power thermal response time lag segment mark.

[0034] The response identification module extends the observation window of the same length according to the power thermal response time lag section mark, retrieves the temperature curve of the thermistor in the extended section, identifies the peak position and time node by node, and obtains the thermal fluctuation lag behavior identification list.

[0035] The hysteresis extraction module calls the thermal fluctuation hysteresis behavior identification list to extract the time points of peaks and troughs within a continuous period, calculates the direction and interval of peak amplitude change in adjacent periods, and obtains the thermal fluctuation evolution trend offset boundary set.

[0036] The trend recognition module locates the connection node in the heat conduction structure of the electric drive system based on the thermal fluctuation evolution trend offset boundary set, collects the temperature rise start time of the preceding and following nodes, and determines whether the following node exhibits a temperature rise behavior within the response time window of the preceding node. If it does not exhibit a temperature rise behavior, it is determined that the thermal path is interrupted, and the thermal blocking path node group in the structural chain is obtained.

[0037] The abnormal early warning module calls the node information and temperature curve of the heat blocking path node group in the structure chain to analyze whether multiple nodes have abnormal temperature rise behavior in the same period. If the judgment result is true, the time period is marked as the linkage trigger state and the linkage heat warning level identification result is generated.

[0038] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0039] In this invention, by combining the coupling relationship between power change trends and temperature rise time series during the operation of the electric drive system, the precise identification of temperature rise response delay is achieved. By observing the dynamic process of temperature peaks at key nodes through an extended window, out-of-cycle temperature rise anomalies can be effectively captured. Based on the evolution trend of peak-valley time series, the increase in fluctuation amplitude and the compression of intervals are identified, enabling dynamic tracking of thermal anomaly trends. By continuously tracing the sequential relationship of temperature rise at hot nodes, the location distribution of thermal path interruption structures is clarified. Combined with the linkage performance between thermal blocking structures, the concurrent characteristics of overheating anomalies are identified, improving the response capability and early warning accuracy of the electric drive system in the early stages of overheating risk, and realizing rapid location of potential thermal fault risks and intelligent risk level classification. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the workflow of the present invention;

[0041] Figure 2 This is a flowchart illustrating the process of obtaining the power thermal response timing lag segment marker in this invention.

[0042] Figure 3 This is a flowchart illustrating the process of obtaining the thermal fluctuation hysteresis behavior identification list in this invention.

[0043] Figure 4 This is a flowchart illustrating the process of obtaining the boundary set of thermal fluctuation evolution trend offset in this invention.

[0044] Figure 5 This is a flowchart illustrating the process of obtaining the thermal blocking path node group within the structural chain in this invention.

[0045] Figure 6 This is a flowchart illustrating the process of obtaining the linkage thermal early warning level identification result in this invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0047] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0048] Please see Figure 1 This invention provides a technical solution: a big data-based method for monitoring and protecting the overheating of an electric drive system, comprising the following steps:

[0049] S1: Obtain the continuous sequence of input power during the operation of the electric drive system, identify the time period of continuous increase in the direction of adjacent sampling points, record the start and end times, extract the temperature change curve of the thermal node before and after the segment, determine whether the temperature rise start time lags behind the power change start point, if the lag condition is met, mark it as a response delay segment, and generate a power thermal response time lag segment mark.

[0050] S2: Using the time segment marked in the power thermal response time lag section, the observation window of the same length as the abrupt change segment is extended backward on the basis of the original monitoring period. The temperature sequence data of key thermal nodes in the window are retrieved, and the temperature peak change is identified node by node to see if it is completed within the extended window. If there are nodes whose temperature rise peak is delayed outside the original period, the node and the time period in which it is located are marked as thermal fluctuation lag events, and a thermal fluctuation lag behavior identification list is generated.

[0051] S3: Using the temperature curves of the nodes identified in the thermal fluctuation hysteresis behavior identification list, extract the time series of peaks and troughs, compare the peak amplitude and interval of continuous cycles, measure the peak amplitude of temperature within the cycle segment and determine the direction of amplitude change between adjacent cycles. If three consecutive cycles show an increasing trend of fluctuation amplitude and the cycle interval is shortening, mark the trend offset boundary segment on the corresponding node and generate the thermal fluctuation evolution trend offset boundary set.

[0052] S4: Based on the location of the boundary concentration node according to the thermal fluctuation evolution trend, locate the connected front and rear nodes one by one in the heat conduction structure of the electric drive system, collect the time sequence of temperature rise start of adjacent nodes, and continuously track according to the structural connection order between nodes. If the rear node does not show temperature rise in the monitoring window after the front node heats up, the node pair is recorded as a thermal path interruption structure, and a thermal blocking path node group in the structural chain is generated.

[0053] S5: Based on the thermal blocking path node group within the structural chain, determine whether the overheating anomaly of the electric drive system occurs in the same time period. If the linkage triggering condition is met, classify the time period status as a high-level thermal warning state and generate a linkage thermal warning level identifier result.

[0054] The power thermal response time lag segment markers include the start and end times of the response delay segment, the thermal node index of the continuous power rise segment, and the time difference between the power mutation start point and the temperature rise start point. The thermal fluctuation lag behavior identification list includes the delayed node index, the temperature rise peak occurrence time, and the extended window time boundary. The thermal fluctuation evolution trend offset boundary set includes trend offset nodes, continuous period peak growth trend, and period shortening characteristics. The thermal blockage path node group within the structural chain includes thermal path interruption node pairs, connection sequence number, and time segments without temperature rise response. The linkage thermal warning level identification results include the high-level warning time period, linkage trigger node group, and abnormal state classification label.

[0055] Please see Figure 2 The specific steps for obtaining the power thermal response time lag segment markers are as follows:

[0056] S111: Acquire the continuous sequence of input power during the operation of the electric drive system, compare the power values ​​between adjacent sampling points, identify continuous segments with continuously increasing power values, record the start and end times of the segments, and generate a sequence of continuously increasing power intervals.

[0057] By collecting input power data point by point within the sampling period, with a sampling frequency of 5 times per second, a total of 150 sample points were collected within a continuous running time of 30 seconds, forming an array. The power value differences between adjacent sample points are compared sequentially, and a judgment threshold is set. ,when When the condition is met for 5 consecutive sample periods, the starting point of the sequence is recorded as the starting point of the continuous power increase segment. until the difference is less than And the continuous decline indicates the termination point. Based on actual operational examples, if a certain sampling segment contains the following data segments:

[0058] If the start time of the power segment is identified as the 1st second and the end time as the 6th second, the segment can be identified as a continuously increasing power segment. This identification requires a sliding window-style segment-by-segment judgment of the entire sequence to extract the power increasing segments that meet the continuous increasing condition and generate a continuously increasing power interval sequence.

[0059] S112: Based on the power continuous rise interval sequence, call the temperature sampling data of the thermal nodes before and after the segment, determine whether the upward trend of the temperature curve in the segment has a lag response compared with the power change start point, and obtain the response lag matching judgment result by comparing whether there is a time difference between the time interval between the segment start point and the temperature change start time.

[0060] Retrieve temperature data recorded by the temperature sensors at each thermistor node within the corresponding time interval. Assuming a temperature sampling frequency of twice per second, extract the corresponding temperature rise trend data for each segment within the aforementioned interval to construct a temperature change curve for each segment. The time interval between the inflection point of the temperature curve and the start time of the power mutation point is synchronously compared on the time axis to determine the time interval between the two. ,in, This is the point where the temperature curve shows a significant increase. As the starting point of the power surge, the judgment threshold interval is set as follows: ,like If the value falls within this range, a response lag matching relationship is determined to exist. For example, if the power surge occurs at the 10th second and the significant temperature rise at the thermistor occurs at the 13th second, then... If the hysteresis interval is satisfied, the response is considered valid. The matching relationship between the power range and the thermal response point is recorded to form a power-temperature response matching range array, such as... This type of judgment process requires traversing the power rise segment and the temperature curve of each node to obtain the effective matching segment and obtain the response lag matching judgment result.

[0061] S113: Based on the response lag matching determination results, select the power rise segments that meet the lag conditions, and extract the power rise amplitude, temperature rise start-up delay time, and temperature change value of the thermistor for each segment using the following formula:

[0062] ;

[0063] Calculate the power thermal response hysteresis intensity value and mark it according to the segment sequence to obtain the power thermal response time hysteresis segment mark;

[0064] in, Representing the The power thermal response hysteresis intensity value of each power rise segment Representing the The increase in power of the segment Representing the Temperature rise start-up delay rate factor, Representing the Duan Di Temperature change value of each thermal node Representing the Duan Di The thermal power response coupling factor corresponding to each node This represents the number of thermal nodes;

[0065] Formula calculation logic: By increasing the total power of a certain power rise segment... Temperature rise start-up delay coefficient corresponding to this segment Multiplication reflects the rapidity of power change over time, while also incorporating the temperature change value of the thermistor. With the corresponding response coupling factor The sum of the products of these two values ​​is used to quantify the magnitude and extent of the thermal response of each node to this power range. The two results are added together and their absolute values ​​are taken, expressed as a function of the number of nodes. Normalized division is performed to obtain the response intensity value under a unified standard. The formula couples power, time delay and multi-point temperature rise data in multiple dimensions to construct a numerical expression with the ability to describe the response intensity of the target object with directionality, amplitude and action, which is convenient for objective quantification and ranking of the thermal response effect of different power change segments.

[0066] The power thermal response hysteresis intensity value is used to measure the degree of response of a certain power rise segment to the temperature change of each thermistor node. It is the product of the power rise amplitude, temperature rise delay time, temperature change of each node and response coupling factor. The larger the value, the stronger the temperature response of the power change to multiple nodes, reflecting its higher thermal influence capability in the electric drive system.

[0067] Filter the power rise range that meets the response lag range and set the response judgment index. Used to quantify the strength of the power band's response to temperature rise, and to calculate the response amplitude, delay time, and temperature change for each matched band;

[0068] Parameter explanation and calculation process:

[0069] : No. Power increase per segment, in kW;

[0070] : No. The corresponding temperature rise start-up delay coefficient is for the segment Normalized adjustment value, in s ;

[0071] : No. The corresponding segment Temperature change values ​​of individual thermal nodes, in °C;

[0072] : No. Section 1 The coupling factor corresponding to the node temperature change takes values ​​in the range [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 1, 1, 1, 2, 1, 1, 2, 1, 2, 3, 1, 2, 3, 1, 2, 3, 4, 5 ...6, 7, 8, 9, 1 1], used to quantify the degree of thermal response;

[0073] Set in the second power response range Corresponding temperature rise delay time Normalize and adjust its settings There are 3 thermistor points, and their temperature changes are respectively , , The corresponding coupling factors are respectively , , The response intensity of this segment is calculated as follows:

[0074] ;

[0075] The results show that in the second power increase segment, the power change and temperature hysteresis response strength is 4.13, and the higher the value, the more significant the response.

[0076]

[0077] As shown in Table 1, the response intensity values ​​corresponding to different power response segments are different, and the power segments can be sorted and labeled according to these values.

[0078] Please see Figure 3 The specific steps for obtaining the thermal fluctuation hysteresis behavior identification list are as follows:

[0079] S211: Based on the time segment marked in the power thermal response time lag segment, the observation time range is extended backward by the same length as the abrupt change segment on the basis of the original monitoring period. Temperature sequence data of the corresponding thermal nodes in the extended segment are collected to obtain the temperature sequence set of the extended segment.

[0080] The time interval to be selected is an extension of the original monitoring period. Assume the original monitoring period is 60 seconds, the extended period is 30 seconds, and the total extended time interval is 90 seconds. Temperature data is sampled from each thermal node within the extended time interval at a frequency of twice per second. Assume there are 5 thermal nodes, denoted as nodes A1 to A5. A temperature sequence is constructed for each node. ,in Up to 5, This results in five sets of temperature extension sequences, each 60 units in length. The data structure in each set is normalized, removing non-monotonic jump points caused by noise, and retaining the original sampled data points and change points directly related to system load fluctuations. In practice, the difference between two consecutive sampled points in the sampled sequence can be processed using moving difference calculus. Zero-intersection analysis is performed on the difference sequence to confirm the location of the temperature rise initiation point. In the actual operation of the electric drive system, if the extended sequence contains the following node A3 sampling temperature: {45.0℃, 45.6℃, 46.3℃, 46.8℃, 47.1℃}, and its rate of change continues to increase, it indicates that there is an effective thermal excitation reaction process at this node, and the extended segment temperature sequence set is obtained.

[0081] S212: Based on the extended segment temperature sequence set, extract the peak temperature value corresponding to the time in the temperature time series of the thermal node, and compare it with the end time of the original monitoring cycle to determine whether there are nodes with peak temperature points distributed in the extended time period. Filter the nodes that meet the conditions and their corresponding time periods to obtain the peak delay node identification results.

[0082] For each sequence, extract the peak temperature value within the extended time period, and use a traversal algorithm to find the maximum value in each sequence. Record the corresponding time points And compare it with the original cycle termination time. Compare and calculate the time difference This time difference must meet the set peak value judgment interval [3s, 10s] to be considered a valid hysteresis response peak. For example, if the temperature of node A2 reaches its maximum value of 48.9℃ at the 68th second, while the original cycle end time is 60 seconds, then... If the value meets the criteria for hysteresis response, this point is recorded as the hysteresis peak point of node A2. During the screening process, false peaks at the boundaries of the extended interval and non-thermal rise points caused by reverse abrupt changes must also be excluded. After screening, the valid set of nodes and time difference sequences that meet the criteria are retained. The peak delay node identification results are obtained.

[0083] S213: Based on the peak delay node identification results, extract the time difference between the original cycle termination time and the temperature peak occurrence time corresponding to the node, using the formula:

[0084] ;

[0085] Calculate the thermal fluctuation hysteresis index corresponding to the node, filter the nodes and time periods that meet the thermal peak delay, and establish a thermal fluctuation hysteresis behavior identification list;

[0086] in, Representing the The thermal fluctuation lag index corresponding to each node Indicates the first Temperature fluctuation range of each node within the extended section Indicates the first Peak temperature rise delay at nodes Indicates the first The node at the th The rate of temperature change at each sampling point Indicates the number of sampling points;

[0087] Formula calculation logic: By extracting the first... Maximum temperature variation of the node during the extended period and the corresponding average heating rate Calculate its theoretical thermal response value This value represents the maximum temperature change behavior of the node under ideal conditions at the current heating rate; it represents the temperature difference corresponding to each sampling of the node within the extended period. Perform weighted calculations to obtain the average response value of the measured changes. Then, the measured average value is subtracted from the theoretical value and the absolute value is taken to obtain the response deviation between the node and the ideal model within the extended period. The larger the value, the stronger the hysteresis of its thermal fluctuation response. This index takes into account the heating amplitude, rate and distribution process, and is suitable for node-level response delay screening of thermal behavior.

[0088] The thermal fluctuation hysteresis index is used to measure the degree of hysteresis response of a thermal node to temperature changes after a power disturbance. It is constructed by the difference between the theoretical temperature rise and the sampling mean, reflecting the response gap between the ideal model and the actual behavior. The larger the value, the less timely the node's temperature rise behavior and response to the power disturbance are, and the more significant the hysteresis phenomenon exists.

[0089] Extract the temperature difference between the peak occurrence time and the original cycle termination time at each node. And combined with the rate of temperature change of the nodes in the extended section With peak temperature increment Joint quantification is performed to form a thermal fluctuation hysteresis index. This metric is used to measure whether a single-node temperature fluctuation has significant response characteristics;

[0090] Parameter description and logical explanation:

[0091] : indicates the first The thermal fluctuation hysteresis index corresponding to each node;

[0092] : for the first The temperature fluctuation range of a node during the expansion cycle is taken as the difference between the maximum value and the initial value;

[0093] : for the first The average rate of temperature change of a node within the extended segment, measured in changes per second;

[0094] : for the first Node number Temperature change value at the time of the next sampling;

[0095] Parameter assignment and calculation:

[0096] Let the initial temperature of node A4 within the extended segment be 45.3℃ and the maximum temperature be 48.6℃, then we have If the temperature rise is 3.3℃, ​​and the extension period is 30 seconds, then the rate of change is ℃. ℃ / s, Calculation Assuming the sampling frequency is 2Hz, the total number of samples is... The temperature changes at some sampling points are as follows: ℃, ℃, ℃;

[0097] Take the above three examples for calculation:

[0098] ;

[0099] ;

[0100] ;

[0101] The results show that the thermal fluctuation hysteresis index of node A4 is 0.973, which is significantly higher than that of A1 and A2. This indicates that the temperature rise process of this node in the extended section has a significant delay behavior compared with the theoretical thermal response. It will be given priority in the discrimination sequence in the subsequent thermal fluctuation identification.

[0102]

[0103] As shown in Table 2, the thermal fluctuation hysteresis index of each node is obtained by combining the sampled temperature difference and the heating rate. The intensity of each node’s participation in the delay behavior can be judged based on the index value.

[0104] The advantage of the formula is that it constructs a theoretical thermal response model by multiplying the maximum temperature difference and the rate, and introduces the average response of multiple measured points as a comparison benchmark, which effectively characterizes the degree of lag in the response of each node during the temperature rise process, and achieves accurate screening and calibration of significant lag in local thermal behavior.

[0105] Please see Figure 4 The specific steps for obtaining the boundary set of thermal fluctuation evolution trend offset are as follows:

[0106] S311: Based on the nodes identified in the thermal fluctuation hysteresis behavior identification list, extract the temperature change curves corresponding to the nodes, record the time positions corresponding to the peak and trough points within the continuous monitoring period, calculate the time span and temperature peak change within the period segment, and generate a periodic temperature feature sequence.

[0107] For each node, temperature change curves are extracted within the marked thermal hysteresis zone. The temperature sampling sequence for the corresponding node is recorded, and interpolation is performed twice per second to construct continuous temperature change data for that node. Peak and trough values ​​are located and identified within three consecutive monitoring periods. Local maxima and minima are identified by judging the gradient change trend in each temperature sequence. The sampling time in each period segment is used as the peak and trough marker to form a temperature fluctuation location sequence. Then set the time interval between peaks to the periodic time. The change in fluctuation amplitude is calculated based on the time difference between adjacent periods. The temperature change within each period is extracted, and a three-period segment sequence is constructed. The temperature difference, fluctuation period, and position interval within each segment are processed in a data structured manner. For example, the peak values ​​of node P3 in the periods from q=1 to q=3 are 45.3℃, 47.0℃, and 48.6℃, respectively, and the peak occurrence times are 12s, 24s, and 35s, respectively. Then the period intervals are 12s and 11s, generating a periodic temperature feature sequence.

[0108] S312: Call the peak amplitude data and period interval data in the periodic temperature feature sequence, determine whether the characteristic trend of increasing peak amplitude and shrinking period interval is met simultaneously in any three consecutive periods, record the time range of the corresponding segment in the temperature curve for the segment that meets the conditions, and obtain the continuous offset trend identification result.

[0109] The amplitude changes for each period With periodic time variation value The analysis process determines whether the amplitude increases periodically and whether the period length decreases periodically. Calculations are performed for each period segment. , Set a judgment threshold , To satisfy the judgment interval of increasing and decreasing trends, if three consecutive segments satisfy... , ,at the same time , If all conditions for trend determination are met, then the periodic segment is recorded as a trend offset segment of continuous hot behavior, and the corresponding time period index is recorded. The peak amplitude and time interval change values ​​are used as the basis for subsequent offset calculation to obtain the continuous offset trend recognition results.

[0110] S313: Based on the continuous offset trend identification results, collect the temperature peak amplitude difference, local fluctuation intensity, and peak position offset within each group of three periodic sequences, using the following formula:

[0111] ;

[0112] Calculate the trend offset index of the three periodic segments, extract the range of segments where the offset trend meets the preset conditions, and obtain the thermal fluctuation evolution trend offset boundary set;

[0113] in, Indicates the first The trend deviation index of each node within three period segments. Indicates the first The node The magnitude of temperature peak variation within a periodic segment. This represents the ratio of changes in the period interval corresponding to the period. Indicates the first The node The local fluctuation intensity of the segment Indicates the first The node The time position of the main wave peak within a segment period Indicates the time position of the main wave peak in the previous cycle;

[0114] Formula calculation logic: The calculation process integrates the temperature peak difference of the nodes within three periods. Degree of periodic contraction Local fluctuation intensity and changes in the position of the main wave peak Four key parameters, through analysis of each cycle and Weighted summation yields the offset measure for that period. Averaging across three periods provides an overall trend assessment index. This formula uses square root transformation to reflect the nonlinear contribution of period compression to the offset intensity, and enhances the proportion of the main wave peak's forward shift in the index using the reciprocal of the time difference. It can effectively characterize whether the temperature fluctuation at this node shows a trend of continuous increase and earlier peak, and provide quantitative support for the system to identify trend behavior;

[0115] The trend deviation index is used to measure whether the temperature fluctuation behavior of a node shows an increasing and forward trend over multiple consecutive cycles. It is calculated by combining factors such as the increase in temperature difference, the degree of cycle compression, the intensity of local disturbance, and the time shift of the peak. The larger the value, the more obviously the thermal behavior of the node deviates from the original rhythm and has a continuously changing trend.

[0116] Parameter definition and logical interpretation:

[0117] : indicates the first The trend deviation index of a node within a three-segment period;

[0118] : indicates the first Node number The magnitude of the peak temperature change within the period;

[0119] : indicates the first Node number The change in the periodic interval, in seconds;

[0120] : indicates the first Node number The intensity of local fluctuations in a periodic segment is quantified as the standard deviation of the previous period. The percentage increase, for example, 10% is represented as 0.1;

[0121] , : These are the time positions corresponding to the main wave peaks of the current cycle and the previous cycle, respectively, in seconds;

[0122] Parameter assignment and example derivation:

[0123] Let the peak temperature difference within the three periods be at node r=1. , , The periodic intervals are respectively , , The local fluctuation intensities are respectively , , The peak time positions are respectively , , The formula is as follows:

[0124] Item 1: ;

[0125] Item 2: ;

[0126] Item 3: ;

[0127] Substitute into the formula to calculate:

[0128] ;

[0129] The results show that the temperature peak of node 1 gradually increases within the three-period segment, the period interval gradually shortens, the local fluctuations are enhanced, and the position of the main wave peak continues to move forward, constituting a significant shift behavior. Its index value of 1.7877 is much higher than the benchmark judgment value of 1.5, confirming it as a key node of thermal trend shift.

[0130]

[0131] As shown in Table 3, the calculated trend deviation index of node 1 in the three-period segment is 1.7877;

[0132] The advantage of the formula is that it measures the increasing trend of thermodynamics by combining the peak amplitude with the periodic contraction synergy term, and measures the degree of deviation of dynamic thermal change by superimposing the fluctuation intensity with the peak forward shift amplitude. The combination of the two forms a unified trend judgment index, which supports the rapid identification of high-risk thermal behavior nodes with forward shift of electric drive system response in continuous periodic data.

[0133] Please see Figure 5 The specific steps for obtaining the thermal blocking path node group within the structural chain are as follows:

[0134] S411: Based on the location of the boundary concentration node according to the thermal fluctuation evolution trend, locate the adjacent connected front and rear nodes one by one, extract the structural connection order between nodes, and mark the connection status of each pair of nodes in the structural path to generate a structural connection order group.

[0135] The spatial distribution of nodes within the offset boundary is clearly defined. The node coordinate list and topological connection matrix in the heat conduction structure diagram are retrieved. Based on the node numbers and connections, the connections are identified one by one, starting from the thermal wave front region. For each node pair in the structure diagram, the connection relationship between the preceding and following nodes is extracted according to their numbering order. In the heat conduction module, the node numbers are... If at that time, and connect and connect and A connection will generate the corresponding node pair. Extracting such node pairs requires combining the node numbering order with the physical adjacency judgment in the actual connection path, excluding discontinuous node pairs caused by cross-region connections. Then, the connection status of each identified node pair is marked, and it needs to be determined whether the connection is a direct thermal path connection on the heat conduction path. The determination method is whether there is a synchronous temperature rise trend along the path within the last three cycles (e.g., 30 seconds). This involves calling the two nodes on the connection path... The temperature data recorded at each time point is used to determine the synchronization response threshold between adjacent nodes, which is set to a temperature difference of no more than 1.2℃. If, within three time points, the temperature of the subsequent node is not lower than the temperature of the preceding node by more than 1.2℃, it is considered to have a direct connection relationship with heat conduction. exist The time points were 35.2℃, 36.1℃, and 37.0℃, respectively. If the temperatures are 34.3℃, 35.0℃, and 36.1℃, then the node pair is a valid thermal path connection and is marked as a valid connection. After the sequential extraction and connection status marking of the node pairs are completed, the recorded connection pairs and validity identifiers are integrated to form a structural connection sequence group.

[0136] S412: Based on the connection order of the node pairs in the structural connection order group, collect the corresponding temperature rise start time data, determine whether the temperature rise phenomenon occurs in the set monitoring window after the start of the preceding node in the node pair, mark the node pairs that do not show temperature rise, and generate the thermal blocking path node group in the structural chain.

[0137] Extract the temperature rise start-up time of the preceding and following nodes after the onset of thermal fluctuations from the node pair. Call the temperature monitoring sequence of the nodes in the electric drive system, and identify the time point when the temperature curve first continuously exceeds the temperature rise judgment threshold as the start-up time. The temperature rise judgment threshold is set to 1.5℃ above the base stable temperature. If the base temperature of the preceding node is 33.0℃, then the temperature rise start-up threshold is set to 34.5℃. When its temperature first continuously exceeds 34.5℃ for at least 5 seconds within the monitoring period, this moment is recorded as the temperature rise start-up time. After extracting the temperature rise start-up time of the preceding and following nodes, determine whether the following node shows temperature rise behavior within the monitoring window after the start-up time of the preceding node. The monitoring window is set to 30 seconds. If the start-up time of the preceding node is... If the time is less than 1 second, it is necessary to determine whether the next node is within 1 second. to If the temperature rise threshold is exceeded for the first time within a few seconds, the base temperature of the node is set to 32.5℃ and the temperature rise threshold is set to 34.0℃. If the temperature requirement is not met within the range of 180-210 seconds, the node pair is judged to be a connection path without effective thermal response, and is marked. The node pairs marked as having no thermal response are extracted, and their node numbers are combined with the structural path numbers to form a thermal blocking path node group within the structural chain.

[0138] Please see Figure 6 The specific steps for obtaining the linkage heat warning level indicator result are as follows:

[0139] S511: Based on the temperature change data of nodes in the thermal blocking path node group within the structural chain, extract the time point at which the temperature rise start phenomenon first occurs within the monitoring period, call the node number and the continuous change segment of the temperature value in the corresponding time series, determine whether there is a stable upward behavior in the continuous temperature change trend, identify the correspondence between node number and time, and obtain the temperature rise start time series table.

[0140] Temperature sequences recorded by each node within a continuous monitoring period need to be collected. These sequences should be generated by the sampling module at fixed time intervals, arranged in chronological order, and bound to unique node identification information. Continuous change segments are extracted from the temperature sequences to determine if there is a continuous increase in numerical change. Segments that satisfy the continuous upward trend and have a number of consecutive data points not less than the judgment benchmark set by the electric drive system are selected as candidate upward trend segments. The start time of this segment is recorded as the temperature rise start time. Short-term sudden increases in data need to be eliminated during operation. In the temperature record set at node number N005, if the temperatures of the 10th to 14th sampling points are 31.2, 31.6, 32.0, 32.3, and 32.7, it can be determined as a continuous upward trend. If there is a drop in value in the middle, the data segment needs to be excluded. Based on this judgment method, the temperature rise start time of all nodes is extracted, and a one-to-one mapping relationship between node number and its temperature rise start time is established to obtain the temperature rise start time sequence table.

[0141] S512: Based on the time distribution of nodes in the temperature rise start time sequence table, classify and identify the monitoring period to which they belong, and identify whether the start time of the nodes belongs to the same time period. The judgment basis is that the time period has a unique number in the preset time division. If the start time of all nodes belongs to the same numbered period, a linkage trigger state is formed, and the corresponding level label is assigned to the state, generating a linkage heat warning level label result.

[0142] The node start-up time distribution needs to be categorized. Based on the monitoring cycle set by the electric drive system, multiple continuous time periods are constructed, each assigned a unique number. The electric drive system is set to divide the entire running time into multiple consecutive numbered segments starting from 0 seconds according to fixed time periods, and a correspondence between time and number is established. The start-up time of each node in the temperature rise start-up time sequence table is determined to belong to a segment. Nodes whose start-up time falls into the same segment number are grouped together and identified by the same number. The number of different segment numbers in the node group is counted. If the start-up time of a node corresponds to only one segment number, it is marked as constituting a linkage state. Based on this, combined with the number of nodes, the proportion of thermal blocking paths, and regional distribution characteristics, the corresponding level identifier for the group state is assigned according to the warning level set internally by the electric drive system. If the linkage trigger state is met, a thermal warning level code of number L3 is assigned, and the linkage thermal warning level identifier result is output.

[0143] The big data-based electric drive system overheat protection monitoring system is used to execute the above-mentioned big data-based electric drive system overheat protection monitoring method. The system includes:

[0144] The data monitoring module acquires the input power sequence data of the drive motor control unit in the electric drive system, compares the power values ​​at adjacent time points to construct a sequence of change direction, identifies the time period in which the power changes continuously, compares the time sequence of the temperature rise start time and the corresponding power change start time, and generates a power thermal response time lag segment mark.

[0145] The response identification module extends the observation window of the same length according to the power thermal response time lag section mark, retrieves the temperature curve of the thermistor in the extended section, identifies the peak position and time node by node, and obtains the thermal fluctuation lag behavior identification list.

[0146] The hysteresis extraction module calls the thermal fluctuation hysteresis behavior identification list to extract the time points of peaks and troughs within a continuous cycle, calculates the direction and interval of peak amplitude change in adjacent cycles, and obtains the thermal fluctuation evolution trend offset boundary set.

[0147] The trend recognition module locates the connection node in the heat conduction structure of the electric drive system based on the boundary set offset by the thermal fluctuation evolution trend. It collects the temperature rise start time of the nodes before and after the node and determines whether the node after the node shows a temperature rise behavior within the response time window of the node before the node. If it does not show a temperature rise behavior, it is determined that the thermal path is interrupted, and the thermal blocking path node group in the structural chain is obtained.

[0148] The abnormal early warning module calls the node information and temperature curve of the heat blocking path node group in the structural chain to analyze whether multiple nodes have abnormal temperature rise behavior in the same period. If the judgment result is true, the time period is marked as the linkage trigger state and the linkage heat warning level identification result is generated.

[0149] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for monitoring overheat protection of electric drive systems based on big data, characterized in that, Includes the following steps: S1: Obtain the continuous sequence of input power during the operation of the electric drive system, identify the time period when the direction of adjacent sampling points rises continuously, extract the temperature change curve of the thermal node before and after the segment, and generate power thermal response time lag segment markers. S2: Using the time segment marked in the power thermal response time lag segment, the observation window of the same length as the abrupt change segment is extended backward on the basis of the original monitoring period. The temperature peak change is identified node by node to determine whether it is completed within the extended window, and a thermal fluctuation lag behavior identification list is generated. S3: Using the node temperature curves identified in the thermal fluctuation hysteresis behavior identification list, extract the time series of peaks and troughs, compare the peak amplitude and interval of continuous cycles, determine the peak amplitude of temperature within the cycle segment and determine the direction of amplitude change between adjacent cycles, and generate a thermal fluctuation evolution trend offset boundary set. S4: Based on the location of the concentrated nodes at the offset boundary of the thermal fluctuation evolution trend, locate the connected front and rear nodes one by one in the thermal conduction structure of the electric drive system, and continuously track the structural connection sequence between the nodes to generate a group of thermal blocking path nodes within the structural chain.

2. The method for monitoring overheat protection of an electric drive system based on big data according to claim 1, characterized in that, The power thermal response timing lag segment marker includes the start and end times of the response delay segment, the thermal node index of the continuous power rise segment, and the time difference between the power mutation start point and the temperature rise start point. The thermal fluctuation lag behavior identification list includes the delayed node index, the temperature rise peak occurrence time, and the extended window time boundary. The thermal fluctuation evolution trend offset boundary set includes trend offset nodes, continuous period peak growth trend, and period shortening characteristics. The thermal blocking path node group within the structural chain includes thermal path interruption node pairs, connection sequence number, and time segments without temperature rise response.

3. The method for monitoring overheat protection of an electric drive system based on big data according to claim 1, characterized in that, The specific steps for obtaining the power thermal response time lag segment marker are as follows: S111: Acquire the continuous sequence of input power during the operation of the electric drive system, compare the power values ​​between adjacent sampling points, identify continuous segments with continuously increasing power values, record the start and end times of the segments, and generate a sequence of continuously increasing power intervals. S112: Based on the power continuous rise interval sequence, call the temperature sampling data of the thermal nodes before and after the segment, determine whether the upward trend of the temperature curve in the segment has a lag response compared with the power change start point, and obtain the response lag matching judgment result by comparing whether there is a time difference between the time interval between the segment start point and the temperature change start time. S113: Based on the response lag matching determination result, filter the power rise segment that meets the lag condition, extract the power rise amplitude, temperature rise start delay time and temperature change value of the thermal node for each segment, calculate the power thermal response lag intensity value, and mark it according to the segment order to obtain the power thermal response time lag segment mark.

4. The method for monitoring overheat protection of an electric drive system based on big data according to claim 3, characterized in that, The specific steps for obtaining the thermal fluctuation hysteresis behavior identification list are as follows: S211: Based on the time segment marked in the power thermal response time lag segment, extend the observation time range to the same length as the abrupt change segment on the basis of the original monitoring period, collect the temperature sequence data of the corresponding thermal nodes in the extended segment, and obtain the temperature sequence set of the extended segment. S212: Based on the extended segment temperature sequence set, extract the time corresponding to the peak temperature value in the temperature time sequence of the thermal node, and compare it with the end time of the original monitoring cycle to determine whether there are nodes with peak temperature points distributed in the extended time period. Filter the nodes that meet the conditions and their corresponding time periods to obtain the peak delay node identification result. S213: Based on the peak delay node identification results, extract the time difference between the original cycle termination time and the temperature peak occurrence time corresponding to the node, calculate the thermal fluctuation hysteresis index corresponding to the node, filter the nodes and time periods that meet the thermal peak delay, and establish a thermal fluctuation hysteresis behavior identification list.

5. The method for monitoring overheat protection of an electric drive system based on big data according to claim 4, characterized in that, The specific steps for obtaining the boundary set of the thermal fluctuation evolution trend offset are as follows: S311: Based on the nodes identified in the thermal fluctuation hysteresis behavior identification list, extract the temperature change curves corresponding to the nodes, record the time positions corresponding to the peak and trough points within the continuous monitoring period, calculate the time span and temperature peak change within the period segment, and generate a periodic temperature feature sequence. S312: Call the peak amplitude data and period interval data in the periodic temperature feature sequence, determine whether any three consecutive periods simultaneously satisfy the characteristic trend of increasing peak change amplitude and shrinking period interval, record the time range of the corresponding segment in the temperature curve for the segment that meets the conditions, and obtain the continuous offset trend identification result. S313: Based on the continuous offset trend identification results, collect the temperature peak amplitude difference, local fluctuation intensity and peak position offset within each group of three periodic sequences, calculate the trend offset index of the three periodic segments, extract the range of segments where the offset trend meets the preset conditions, and obtain the thermal fluctuation evolution trend offset boundary set.

6. The method for monitoring overheat protection of an electric drive system based on big data according to claim 5, characterized in that, The specific steps for obtaining the thermal blocking path node group within the structural chain are as follows: S411: Based on the location of the concentrated nodes of the offset boundary according to the thermal fluctuation evolution trend, locate the adjacent connected front and rear nodes one by one, extract the structural connection order between the nodes, and mark the connection status of each pair of nodes in the structural path to generate a structural connection order group. S412: Based on the structural connection sequence group, collect the corresponding temperature rise start time data, determine whether the subsequent node in the node pair experiences a temperature rise phenomenon within the set monitoring window after the preceding node starts, mark the node pair that does not show a temperature rise phenomenon, and generate a thermal blocking path node group within the structural chain.

7. The method for monitoring overheat protection of an electric drive system based on big data according to claim 1, characterized in that, The method further includes step S5: S5: Based on the thermal blocking path node group in the structure chain, determine whether the overheating anomaly of the electric drive system is in the same time period. If the linkage triggering condition is met, classify the time period state as a high-level thermal warning state and generate a linkage thermal warning level identification result. The results of the linkage thermal warning level identification include the high-level warning time period, linkage trigger node group, and abnormal status classification label.

8. The method for monitoring overheat protection of an electric drive system based on big data according to claim 7, characterized in that, The specific steps for obtaining the linkage thermal warning level identifier result are as follows: S511: Based on the temperature change data of the nodes in the heat-blocking path node group within the structural chain, extract the time point at which the temperature rise start phenomenon first occurs within the monitoring period, call the node number and the continuous change segment of the temperature value in the corresponding time series, determine whether there is a stable upward behavior in the continuous temperature change trend, identify the correspondence between the node number and the time, and obtain the temperature rise start time series table. S512: Based on the time distribution of nodes in the temperature rise start time sequence table, classify and identify the monitoring period to which they belong, and identify whether the start time of the nodes belongs to the same time period. The judgment basis is that the time period has a unique number in the preset time division. If the start time of all nodes belongs to the same numbered period, a linkage trigger state is formed, and a corresponding level label is assigned to the state, generating a linkage heat warning level label result.

9. A big data-based overheat protection monitoring system for electric drive systems, characterized in that: The system is used to implement the big data-based overheat protection monitoring method for electric drive systems according to any one of claims 1-8, and the system includes: The data monitoring module acquires the input power sequence data of the drive motor control unit in the electric drive system, compares the power values ​​at adjacent time points to construct a sequence of change direction, identifies the time period in which the power changes continuously, compares the time sequence of the temperature rise start time and the corresponding power change start time, and generates a power thermal response time lag segment mark. The response identification module extends the observation window of the same length according to the power thermal response time lag section mark, retrieves the temperature curve of the thermistor in the extended section, identifies the peak position and time node by node, and obtains the thermal fluctuation lag behavior identification list. The hysteresis extraction module calls the thermal fluctuation hysteresis behavior identification list to extract the time points of peaks and troughs within a continuous period, calculates the direction and interval of peak amplitude change in adjacent periods, and obtains the thermal fluctuation evolution trend offset boundary set. The trend recognition module locates the connection node in the heat conduction structure of the electric drive system based on the thermal fluctuation evolution trend offset boundary set, collects the temperature rise start time of the preceding and following nodes, and determines whether the following node exhibits a temperature rise behavior within the response time window of the preceding node. If it does not exhibit a temperature rise behavior, it is determined that the thermal path is interrupted, and the thermal blocking path node group in the structural chain is obtained. The abnormal early warning module calls the node information and temperature curve of the heat blocking path node group in the structure chain to analyze whether multiple nodes have abnormal temperature rise behavior in the same period. If the judgment result is true, the time period is marked as the linkage trigger state and the linkage heat warning level identification result is generated.

Images