Method and device for protecting mcu mode fusing fuse when power supply electronic switch fails

Abstract

The application provides a method and device for MCU mode fuse protection when power electronic switch fails, relates to the technical field of power electronic protection, and comprises the following steps: obtaining a target fusing temperature set by a user in advance through a dial switch, a real-time environment temperature of a temperature fuse and a real-time atmospheric environment temperature; performing self-diagnosis and judgment on the temperature fuse based on the real-time environment temperature of the temperature fuse and the real-time atmospheric environment temperature; dynamically controlling a heating module to heat the temperature fuse; when the real-time environment temperature of the temperature fuse is greater than the target fusing temperature, the temperature fuse is fused, so that the heating module is synchronously powered off to stop heating, and the circuit loop is cut off through the fusing of the temperature fuse. The application can completely solve the industry safety pain point that the main switch cannot be disconnected after failure.

Description

Technical Field

[0001] This invention relates to the field of power electronic protection technology, and more specifically, to a method and apparatus for MCU-based fuse protection when a power electronic switch fails. Background Technology

[0002] Power switches (MOSFETs, relays, etc.) in power electronic systems are core actuators controlling power transmission, and their reliability directly affects system safety. In practical applications, due to factors such as current surges, high-temperature aging, voltage spikes, and mechanical wear, switches may experience failures such as breakdown, shoot-through, and sticking. For example, the drain and source of a MOSFET may remain conductive after being turned off; the contacts of a relay may fail to separate after a disconnect command is issued. In such cases, even if the power supply controller detects serious faults such as overcurrent or overvoltage, it may be unable to execute the protection action of disconnecting the main circuit, leading to damage to the load or the power supply.

[0003] Currently, existing protection schemes are mainly divided into two categories: one is preventive protection, which can only reduce the probability of failure by selecting high-specification MOSFETs, parallel buffer circuits, and adding drive protection, but cannot eliminate failure; the other is post-failure remediation, which notifies the external circuit breaker to shut down after a fault is detected.

[0004] The drawback of existing technical solutions is that they fail to provide a highly reliable forced disconnection mechanism independent of the main switching device (MOSFET or relay). Furthermore, existing measures mainly protect the MOSFET from breakdown, but cannot completely prevent failures. When a problem occurs, the circuit still cannot be disconnected.

[0005] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0006] In view of this, the present invention provides a method and apparatus for MCU-based fuse protection when the power electronic switch fails, in order to solve the aforementioned problems.

[0007] To solve the above problems, the specific technical solution adopted by the present invention is as follows:

[0008] According to a first aspect of the present invention, a method for MCU-based fuse protection when a power electronic switch fails is provided, comprising the following steps:

[0009] S1. Obtain the target fusing temperature, the real-time ambient temperature of the thermal fuse, and the real-time atmospheric ambient temperature that the user has preset via the DIP switch.

[0010] S2. Based on the real-time ambient temperature and real-time atmospheric temperature of the thermal fuse, perform a self-diagnosis judgment on the thermal fuse. If it is determined to be normal, proceed to step S3; otherwise, send a warning message to the user to replace the thermal fuse and return to step S1. The self-diagnosis is performed by constructing a mapping database based on the temperature density analysis method of local sensitive hash function.

[0011] S3. After the main power controller detects that the main electronic switch is out of control, it sends a trigger signal for the main switch to be out of control to the microcontroller. After the microcontroller receives the trigger signal, it dynamically controls the heating module to heat the thermal fuse in combination with the ambient temperature of the thermal fuse.

[0012] S4. When the real-time ambient temperature of the thermal fuse is greater than the target melting temperature, the thermal fuse melts, so that the heating module is powered off and stops heating synchronously, and the circuit is cut off by the melting of the thermal fuse.

[0013] Preferably, the self-diagnostic judgment of the temperature fuse based on the real-time ambient temperature and the real-time atmospheric ambient temperature includes the following steps:

[0014] Historical ambient temperature data of thermal fuses under different atmospheric ambient temperatures were collected, and temperature-dense analysis method using local sensitive hash function was combined to construct a normal temperature range mapping database indexed by ambient temperature.

[0015] Based on the real-time ambient temperature, the upper and lower limits of the normal ambient temperature of the thermal fuse under the real-time ambient temperature are determined by mapping the database, thus obtaining the real-time temperature range.

[0016] The real-time temperature data of the thermal fuse is compared with the real-time temperature range. If the real-time temperature data exceeds the real-time temperature range, the thermal fuse is considered to be abnormal; otherwise, the thermal fuse is considered to be normal.

[0017] Preferably, the step of collecting historical ambient temperature data of thermal fuses under different atmospheric ambient temperatures, and constructing a normal temperature range mapping database indexed by ambient temperature using a temperature-dense analysis method based on locality-sensitive hash functions, includes the following steps:

[0018] Based on historical ambient temperature data of thermal fuses under different atmospheric ambient temperatures, a temperature feature plane is constructed with ambient temperature as the horizontal axis and historical ambient temperature as the vertical axis.

[0019] The temperature feature plane is divided into several temperature segments along the atmospheric temperature dimension, and the effective temperature sample set that meets the preset conditions is selected by calculating the density of historical ambient temperature sample points and the density threshold in each temperature segment.

[0020] Locality-sensitive hashing (LSH) is used to perform hash mapping on the atmospheric ambient temperature data in the effective temperature sample set, and multiple non-overlapping preliminary atmospheric ambient temperature ranges are generated based on the hash mapping results.

[0021] By using the inflection point analysis of the cumulative distribution curve, the peak point in the preliminary atmospheric temperature range is identified, and the preliminary atmospheric temperature range is adjusted based on the peak point to obtain the final atmospheric temperature range.

[0022] For the final atmospheric ambient temperature range, extract the historical ambient temperature values ​​corresponding to all valid temperature sample points within the final atmospheric ambient temperature range, and take the minimum value as the lower limit of the normal ambient temperature under the atmospheric ambient temperature range, and take the maximum value as the upper limit of the normal ambient temperature.

[0023] Using each atmospheric ambient temperature range as an index key, the upper and lower bounds of the corresponding normal ambient temperature are associated and stored to construct a normal temperature range mapping database indexed by atmospheric ambient temperature.

[0024] Preferably, the step of dividing the temperature feature plane into several temperature segments along the atmospheric ambient temperature dimension, and selecting an effective temperature sample set that meets preset conditions by calculating the density of historical temperature sample points and a density threshold within each temperature segment includes the following steps:

[0025] The temperature feature plane is divided into several continuous and non-overlapping temperature segments along the atmospheric ambient temperature dimension according to a preset fixed temperature step size. The total number of historical ambient temperature sample points in each temperature segment is counted and used as the density value of the temperature segment.

[0026] Calculate the average density value of all temperature segment intervals to obtain the average density, and use the product of the preset proportional coefficient and the average density as the density threshold.

[0027] The density value of each temperature segment interval is compared with the density threshold. If the density value of a certain temperature segment interval is greater than or equal to the density threshold, all historical ambient temperature sample points in that temperature segment interval are retained; otherwise, all historical ambient temperature sample points in that temperature segment interval are removed.

[0028] The historical environmental temperature sample points that are retained are summarized to form an effective temperature sample set.

[0029] Preferably, the step of performing hash mapping processing on the atmospheric ambient temperature data in the effective temperature sample set using a locality-sensitive hash function, and generating multiple non-overlapping continuous atmospheric ambient temperature intervals based on the hash mapping results, includes the following steps:

[0030] Extract the center temperature value of each temperature segment interval in the effective temperature sample set, and use the center temperature value as the characteristic temperature of that temperature segment interval to obtain a characteristic temperature sequence composed of all characteristic temperatures.

[0031] A family of locality-sensitive hash functions is used to perform hash calculations on each feature temperature in the feature temperature sequence to obtain the hash value corresponding to each temperature segment interval; temperature segment intervals with the same hash value and adjacent atmospheric ambient temperatures are merged into temporary temperature groups.

[0032] For each temporary temperature group, the minimum starting boundary of all temperature segment intervals it covers is taken as the left boundary, and the maximum ending boundary is taken as the right boundary, so as to generate multiple non-overlapping and continuous atmospheric ambient temperature intervals.

[0033] Preferably, the step of using a family of locality-sensitive hash functions to perform hash calculations on each feature temperature in the feature temperature sequence to obtain the hash value corresponding to each temperature segment interval; merging temperature segment intervals with the same hash value and adjacent atmospheric ambient temperatures into temporary temperature groups includes the following steps:

[0034] Initialize the configuration parameters of the locality-sensitive hash function family and construct a hash function family consisting of multiple independent locality-sensitive hash functions;

[0035] The feature temperature in each feature temperature sequence is hashed using all locality-sensitive hash functions in the hash function family, generating a hash value sequence corresponding to the number of hash functions, and the hash value sequence is used as the feature hash signature of the corresponding temperature segment interval.

[0036] Temperature segment intervals with completely identical feature hash signatures are divided into the same candidate group, resulting in multiple initial candidate groups;

[0037] The temperature segments within each initial candidate group are arranged in ascending order of atmospheric ambient temperature. Adjacent temperature segments with a temperature difference less than or equal to a preset temperature similarity threshold are merged to generate multiple temporary temperature groups.

[0038] Preferably, the method of identifying peak points in the preliminary atmospheric temperature range using cumulative distribution curve inflection point analysis, and adjusting the preliminary atmospheric temperature range based on the peak points to obtain the final atmospheric temperature range includes the following steps:

[0039] For each preliminary atmospheric temperature range, extract the historical ambient temperature values ​​of all valid temperature sample points, sort them in ascending order of value, calculate the cumulative probability corresponding to each temperature value, and plot the empirical cumulative distribution curve.

[0040] Based on the empirical cumulative distribution curve, candidate peak points are obtained by identifying the historical ambient temperature values ​​corresponding to local maxima.

[0041] If there is only one candidate peak point, the sample distribution within the preliminary atmospheric temperature range is determined to be unimodal, and the preliminary atmospheric temperature range is taken as the final atmospheric temperature range. If there is more than one candidate peak point, it is determined to be multimodal, and the point where the first derivative of the density estimation curve between adjacent candidate peak points turns from negative to positive is taken as the valley floor dividing point. The preliminary atmospheric temperature range is divided into multiple sub-ranges, and each sub-range is taken as a final atmospheric temperature range.

[0042] Preferably, the step of obtaining candidate peak points by identifying the historical ambient temperature values ​​corresponding to local maxima points based on empirical cumulative distribution curves includes the following steps:

[0043] The kernel density of the empirical cumulative distribution curve is estimated by using the Gaussian kernel function, and the probability density estimate at each historical ambient temperature value is calculated to generate a smoothed density estimation curve.

[0044] The first and second derivatives of the smoothed density estimation curve are calculated, and points where the first derivative is zero and the second derivative is negative are identified as local maxima.

[0045] The historical ambient temperature value corresponding to each local maximum point is used as a candidate peak point.

[0046] Preferably, the step of determining the upper and lower bounds of the normal ambient temperature for the thermal fuse under the real-time ambient temperature by mapping a database, based on the real-time atmospheric temperature, to obtain the real-time temperature range includes the following steps:

[0047] The real-time ambient temperature is compared with each ambient temperature range stored in the mapping database to determine the target range containing the real-time ambient temperature. If a target range containing the real-time temperature exists, the lower and upper bound values ​​of the normal ambient temperature associated with the target range are directly read.

[0048] If the real-time ambient temperature is less than the left boundary of the minimum interval in the mapping database or greater than the right boundary of the maximum interval, then the boundary interval closest to the real-time temperature is taken as the reference interval, and the lower and upper bounds of the real-time temperature are calculated using the reference interval and its adjacent intervals to obtain the real-time temperature interval.

[0049] If the real-time ambient temperature is located in the gap between two adjacent intervals in the mapping database, then linear interpolation is performed on the adjacent intervals on both sides of the gap, and the lower and upper bound interpolation values ​​are calculated respectively. The interpolation results are used as the lower and upper bound values ​​of the normal ambient temperature corresponding to the real-time temperature to obtain the real-time temperature interval.

[0050] According to a second aspect of the present invention, an apparatus for MCU-based fuse protection in the event of a power electronic switch failure is provided, the apparatus comprising:

[0051] The control and acquisition module is used to receive the runaway trigger signal from the main power controller, drive the heating module, and acquire the real-time ambient temperature.

[0052] The heating module uses ceramic heating elements, which generate heat when powered on, and also heats the thermal fuse.

[0053] A thermal fuse is connected in series in the main circuit. Under normal operating temperature, it can carry a rated current. When the sensed temperature reaches the preset target melting temperature, its internal alloy solder joints will irreversibly melt and break, thereby physically cutting off the main circuit.

[0054] Fuse connector, used as a port for electrical connection of the main circuit;

[0055] The microcontroller is used to execute the heating strategy and determine the temperature.

[0056] DIP switches are mechanical coded switches used by users to preset target melting temperatures and have a power-off memory function.

[0057] The signal control port is used to receive the runaway trigger signal from the power supply main controller;

[0058] The temperature sampling port is used to collect and upload the real-time ambient temperature of the thermal fuse.

[0059] The beneficial effects of this invention are as follows:

[0060] 1. This invention requires no short circuit, no large current impact or electric arc, has fast response, high reliability, simple structure, and low cost. It does not rely on the PWM or ADC capabilities of the power supply main control and is suitable for power supply, energy storage power supply and other systems. It can completely solve the industry safety pain point of not being able to disconnect after the main switch fails.

[0061] 2. This invention uses a thermal fuse as the final actuator. Its melting is an irreversible process based on the physical properties of the material. Once activated, the main circuit is permanently and reliably cut off, and there is no possibility of secondary recovery or failure. This provides an absolute safety isolation barrier for the power system and fundamentally eliminates accidents such as thermal runaway caused by switch malfunction.

[0062] 3. The protection triggering mechanism of this invention is initiated by an independent power failure signal, and the protection action can be executed immediately regardless of whether the power supply is currently charging, discharging, or in a static state. The combined design of the heating module and the temperature fuse ensures high efficiency of heat conduction and guarantees a rapid response capability from fault identification to circuit disconnection.

[0063] 4. The device of the present invention has a simple structure, and the components are all mature industrial products with controllable costs. The power supply main controller only needs to output a simple on / off signal, which is especially suitable for power supply main control systems that do not have PWM output and ADC acquisition capabilities.

[0064] 5. This invention features a self-breaking point; the heating module automatically powers off after the fuse breaks, with no additional power consumption, no risk of overheating, and no reliance on external devices, providing a final safety barrier for the power supply. It can adapt to fuses with different melting temperatures via DIP switches, and can accommodate different customers and different specifications of temperature fuses without firmware modification. Furthermore, the DIP switches are mechanically encoded, retaining their state after power failure, preventing mismatches due to software malfunctions. The MCU independently completes heating drive, temperature sampling, and comparison judgment, reducing the performance requirements of the power supply controller, improving system compatibility, and even if the power supply controller subsequently malfunctions, the MCU can still complete the heating and melting process.

[0065] 6. This invention constructs an adaptive normal temperature range mapping database indexed by atmospheric ambient temperature by combining local sensitive hash functions, grid density filtering, and inflection point analysis of cumulative distribution curves. This enables highly reliable self-diagnosis of temperature fuses, significantly improving the robustness, adaptability, and accuracy of self-diagnosis. It avoids misjudgments caused by environmental changes, data sparsity, or multimodal distribution, providing intelligent pre-protection for reliable fuse blowing protection when power electronic switches malfunction. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0067] Figure 1 This is a flowchart of a method for MCU-based fuse protection when the power electronic switch fails, according to an embodiment of the present invention.

[0068] Figure 2 This is a flowchart of the self-diagnosis process in the MCU-based fuse protection method when the power electronic switch fails, according to an embodiment of the present invention.

[0069] Figure 3This is a flowchart of the database construction process in the method for protecting the fuse by MCU mode when the power electronic switch fails, according to an embodiment of the present invention.

[0070] Figure 4 This is a schematic diagram of a device for MCU-based fuse protection when the power electronic switch fails, according to an embodiment of the present invention.

[0071] Figure 5 This is a circuit connection diagram of a device for MCU-based fuse protection when the power electronic switch fails, according to an embodiment of the present invention.

[0072] In the picture:

[0073] 1. Control and acquisition module; 2. Heating module; 3. Temperature fuse; 4. Fuse interface; 5. Microcontroller; 6. DIP switch; 7. Signal control port; 8. Temperature sampling port; 9. Power supply controller; 10. Power supply positive port; 11. External devices. Detailed Implementation

[0074] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0075] According to embodiments of the present invention, a method and apparatus for MCU-based fuse protection when a power electronic switch fails are provided.

[0076] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, according to a first embodiment of the present invention, a method for MCU-based fuse protection when a power electronic switch fails is provided, comprising the following steps:

[0077] S1. Obtain the target fusing temperature, the real-time ambient temperature of the thermal fuse 3, and the real-time atmospheric ambient temperature that the user has preset via DIP switch 6.

[0078] It should be noted that the real-time ambient temperature of thermal fuse 3 needs to be collected by a temperature sensor. The real-time atmospheric ambient temperature can be obtained by acquiring weather information for the user's region.

[0079] S2. Based on the real-time ambient temperature and real-time atmospheric temperature of the thermal fuse 3, perform a self-diagnosis judgment on the thermal fuse. If it is determined to be normal, proceed to step S3; otherwise, send a warning message to the user to replace the thermal fuse 3 and return to step S1. The self-diagnosis is performed by constructing a mapping database based on the temperature density analysis method of local sensitive hash function.

[0080] like Figures 2-3 As shown, in a preferred embodiment, the self-diagnostic judgment of the temperature fuse 3 based on the real-time ambient temperature and the real-time atmospheric ambient temperature includes the following steps:

[0081] Historical ambient temperature data of thermal fuse 3 under different atmospheric ambient temperatures were collected, and temperature density analysis method using local sensitive hash function was combined to construct a normal temperature range mapping database indexed by ambient temperature.

[0082] In a preferred embodiment, the step of collecting historical ambient temperature data of the temperature fuse 3 under different atmospheric ambient temperatures, and constructing a normal temperature range mapping database indexed by ambient temperature using a temperature density analysis method based on locality-sensitive hash functions, includes the following steps:

[0083] Based on the historical ambient temperature data of thermal fuse 3 under different atmospheric ambient temperatures, a temperature feature plane is constructed with atmospheric ambient temperature as the horizontal axis and historical ambient temperature as the vertical axis.

[0084] Specifically, when constructing the temperature characteristic plane, it is necessary to collect multiple sets of data recorded by the thermal fuse 3 during its historical operation. Each set of data contains two key parameters: the ambient temperature, i.e., the air temperature of the environment where the thermal fuse 3 is located, which can be measured by an independent temperature sensor, and the ambient temperature of the thermal fuse 3 itself, i.e., the temperature measured close to the installation location of the thermal fuse 3. Then, with the ambient temperature as the horizontal axis (X-axis) and the fuse ambient temperature as the vertical axis (Y-axis), each set of historical data is plotted on a two-dimensional plane, resulting in a series of scattered points to form the temperature characteristic plane.

[0085] The temperature feature plane is divided into several temperature segments along the atmospheric temperature dimension, and the effective temperature sample set that meets the preset conditions is selected by calculating the density of historical ambient temperature sample points and the density threshold in each temperature segment.

[0086] In a preferred embodiment, the step of dividing the temperature feature plane into several temperature segment intervals along the atmospheric ambient temperature dimension, and filtering the effective temperature sample set that meets the preset conditions by calculating the density of historical temperature sample points and the density threshold within each temperature segment interval includes the following steps:

[0087] The temperature feature plane is divided into several continuous and non-overlapping temperature segments along the atmospheric ambient temperature dimension according to a preset fixed temperature step size. The total number of historical ambient temperature sample points in each temperature segment is counted and used as the density value of the temperature segment.

[0088] Specifically, the temperature feature plane is divided into several temperature segments of equal width along the atmospheric ambient temperature dimension (X-axis), for example, each segment is 1°C. For each temperature segment, the number of all historical ambient temperature sample points falling within that temperature segment is counted. This number is called the density value of that temperature segment.

[0089] Calculate the average density value of all temperature segment intervals to obtain the average density, and use the product of the preset proportional coefficient and the average density as the density threshold.

[0090] Specifically, the density threshold is obtained by calculating the average density across all temperature intervals and multiplying it by a preset scaling factor, such as 0.3-0.5. The preset scaling factor can be adjusted based on the quality of historical data: a smaller value, such as 0.3, can be used when the data noise is high, and a larger value, such as 0.5, can be used when the data quality is high.

[0091] The density value of each temperature segment interval is compared with the density threshold. If the density value of a certain temperature segment interval is greater than or equal to the density threshold, all historical ambient temperature sample points in that temperature segment interval are retained; otherwise, all historical ambient temperature sample points in that temperature segment interval are removed.

[0092] The historical environmental temperature sample points that are retained are summarized to form an effective temperature sample set.

[0093] Locality-sensitive hashing (LSH) is used to perform hash mapping on the atmospheric ambient temperature data in the effective temperature sample set, and multiple non-overlapping preliminary atmospheric ambient temperature ranges are generated based on the hash mapping results.

[0094] In a preferred embodiment, the step of performing hash mapping processing on the atmospheric ambient temperature data in the effective temperature sample set using a locality-sensitive hash function, and generating multiple non-overlapping continuous atmospheric ambient temperature intervals based on the hash mapping results, includes the following steps:

[0095] Extract the center temperature value of each temperature segment interval in the effective temperature sample set, and use the center temperature value as the characteristic temperature of that temperature segment interval to obtain a characteristic temperature sequence composed of all characteristic temperatures.

[0096] Specifically, to calculate the center temperature value, it is necessary to determine the starting and ending boundaries of each temperature segment interval, and then calculate the average value based on the starting and ending boundaries to obtain the center temperature value.

[0097] A family of locality-sensitive hash functions is used to perform hash calculations on each feature temperature in the feature temperature sequence to obtain the hash value corresponding to each temperature segment interval; temperature segment intervals with the same hash value and adjacent atmospheric ambient temperatures are merged into temporary temperature groups.

[0098] In a preferred embodiment, the step of using a family of locality-sensitive hash functions to perform hash calculations on each feature temperature in the feature temperature sequence to obtain a hash value corresponding to each temperature segment interval; merging temperature segment intervals with the same hash value and adjacent atmospheric ambient temperatures into temporary temperature groups includes the following steps:

[0099] Initialize the configuration parameters of the locality-sensitive hash function family and construct a hash function family consisting of multiple independent locality-sensitive hash functions;

[0100] Specifically, a set of configurable Locality Sensitive Hash (LSH) functions is established to map the characteristic temperatures of temperature segment intervals to discrete hash buckets, so that values ​​that are close on the temperature axis have a higher probability of falling into the same bucket.

[0101] The configuration parameters include the number of hash functions, the number of hash buckets, and the temperature similarity threshold. This refers to the number of independent hash functions contained in a family of hash functions. Each feature temperature will be processed through... Each function is calculated separately to obtain a length of The hash signature. The number of hash buckets controls the probability of hash collisions. Fewer buckets result in a higher probability of collisions, as different temperatures are more likely to fall into the same bucket; more buckets result in a lower probability of collisions, and stricter similarity requirements. The temperature similarity threshold is used to determine whether two temperature ranges should be considered numerically similar.

[0102] Specifically, based on the above parameters, construct An independent locality-sensitive hash function. For a one-dimensional feature temperature. (Unit: °C), the commonly used and effective LSH function form is:

[0103] ;

[0104] In the formula, Denotes the i-th constructed locality-sensitive hash function. Represents the random scaling factor, typically derived from the standard normal distribution. The value is obtained by sampling and rounding to an integer, or by directly using a floating-point number. This allows different functions to have different projection directions. Represents a random offset, which can be derived from a uniform distribution. The samples were obtained from sampling to avoid periodic deviations in binning caused by fixed boundaries. This represents the bucket width parameter, which controls how much variation in the input value will cause the hash value to change. It is usually set to be related to the desired temperature similarity threshold, for example Sure, This indicates the threshold for determining temperature similarity. A larger threshold indicates a lower threshold. This makes it easier for similar values ​​to fall into the same bucket. This indicates rounding down to the nearest integer. This indicates taking the modulo of the number of buckets, ensuring the output is within the range of buckets. Within the range.

[0105] The feature temperature in each feature temperature sequence is hashed using all locality-sensitive hash functions in the hash function family, generating a hash value sequence corresponding to the number of hash functions, and the hash value sequence is used as the feature hash signature of the corresponding temperature segment interval.

[0106] Specifically, for each feature temperature value in the feature temperature sequence (i.e., the center temperature of each temperature segment interval), each locality-sensitive hash function in the hash function family is called in sequence for calculation. Each function outputs an integer hash value. The results of all L hash functions are combined in a fixed order into an L-dimensional vector or encoded string, which serves as the feature hash signature of that temperature segment interval. This signature can guarantee with a high probability that two feature temperatures that are numerically similar will generate the same signature, thus providing a basis for quickly identifying similar intervals in the future.

[0107] Temperature segment intervals with completely identical feature hash signatures are divided into the same candidate group, resulting in multiple initial candidate groups;

[0108] Specifically, the feature hash signature of each temperature segment interval is used as a unique identifier. All intervals are grouped according to the signature. Intervals with the same signature are grouped into the same candidate group, while intervals with different signatures are grouped into different groups, thus forming multiple initial candidate groups.

[0109] The temperature segments within each initial candidate group are arranged in ascending order of atmospheric ambient temperature. Adjacent temperature segments with a temperature difference less than or equal to a preset temperature similarity threshold are merged to generate multiple temporary temperature groups.

[0110] Specifically, for each initial candidate group, all temperature segment intervals within that initial candidate group are extracted and sorted in ascending order according to the starting boundary or center temperature of each interval to obtain an ordered interval sequence. Then, starting from the first interval, the difference between two adjacent intervals on the atmospheric ambient temperature dimension is checked sequentially. If the difference is less than or equal to a preset temperature similarity judgment threshold, the two intervals are merged into a temporary temperature group, and the process continues to check whether this group can be merged with the next interval. If the difference is greater than the threshold, the merging of the current group ends, and a new temporary temperature group is started with the current interval as the starting point. Finally, each initial candidate group is split into one or more temporary temperature groups. The temperature segment intervals within each temporary temperature group not only have the same hash signature, but are also continuous and close to each other on the atmospheric temperature axis, ensuring the physical similarity and continuity of the merged intervals.

[0111] For each temporary temperature group, the minimum starting boundary of all temperature segment intervals it covers is taken as the left boundary, and the maximum ending boundary is taken as the right boundary, so as to generate multiple non-overlapping and continuous atmospheric ambient temperature intervals.

[0112] Specifically, for each generated temporary temperature group, by traversing all the original temperature segment intervals contained in the temporary temperature group, the smallest starting boundary value is found from these intervals as the left boundary of the group, and the largest ending boundary value is found as the right boundary of the group, thus forming a continuous and closed atmospheric temperature interval. Since different temporary temperature groups do not overlap, the generated atmospheric temperature intervals do not overlap with each other and can completely cover the atmospheric temperature range corresponding to all valid temperature sample points.

[0113] By using the inflection point analysis of the cumulative distribution curve, the peak point in the preliminary atmospheric temperature range is identified, and the preliminary atmospheric temperature range is adjusted based on the peak point to obtain the final atmospheric temperature range.

[0114] In a preferred embodiment, the step of identifying peak points in the preliminary atmospheric temperature range using cumulative distribution curve inflection point analysis, and adjusting the preliminary atmospheric temperature range based on these peak points to obtain the final atmospheric temperature range includes the following steps:

[0115] For each preliminary atmospheric temperature range, extract the historical ambient temperature values ​​of all valid temperature sample points, sort them in ascending order of value, calculate the cumulative probability corresponding to each temperature value, and plot the empirical cumulative distribution curve.

[0116] Specifically, for each obtained preliminary atmospheric ambient temperature range, i.e. a continuous atmospheric temperature range, such as [20°C, 25°C), the historical ambient temperature values ​​corresponding to all valid temperature sample points within the preliminary atmospheric ambient temperature range are extracted, and these temperature values ​​are arranged in ascending order from smallest to largest to obtain an ordered sequence.

[0117] For the The cumulative probability of each sorted temperature value is as follows: Calculate, and then use each temperature value and its corresponding cumulative probability as a point pair. Plotting these points on a coordinate system yields the Empirical Cumulative Distribution Curve (ECDF). This curve is a step function that monotonically rises from zero to one, and its shape reflects the distribution of ambient temperature values.

[0118] Based on the empirical cumulative distribution curve, candidate peak points are obtained by identifying the historical ambient temperature values ​​corresponding to local maxima.

[0119] In a preferred embodiment, the step of obtaining candidate peak points by identifying the historical ambient temperature values ​​corresponding to local maxima points based on the empirical cumulative distribution curve includes the following steps:

[0120] The kernel density of the empirical cumulative distribution curve is estimated by using the Gaussian kernel function, and the probability density estimate at each historical ambient temperature value is calculated to generate a smoothed density estimation curve.

[0121] Specifically, after obtaining an ordered sequence of all historical ambient temperature values ​​within the initial atmospheric temperature range, in order to extract peak points from the empirical cumulative distribution curve, it is necessary to first convert it into a continuously differentiable probability density curve. Specifically, the Gaussian kernel density estimation method is used: for the interval... Historical ambient temperature samples Its density estimation function is defined as:

[0122] ;

[0123] In the formula, Represents the standard Gaussian kernel function. Indicates bandwidth. Typically based on the sample standard deviation and sample size Determined using Silverman's rule of thumb: .

[0124] Specifically, by uniformly selecting a series of discrete points (e.g., 200 points) between the minimum and maximum historical ambient temperature values, the kernel density estimate is calculated for each point, resulting in a set of density values. Connecting these points with a smooth curve yields the smoothed density estimation curve. The ordinate of this curve represents the probability density at different historical ambient temperature values, with its peak (local maxima) corresponding to the temperature region where the sample is most concentrated.

[0125] The first and second derivatives of the smoothed density estimation curve are calculated, and points where the first derivative is zero and the second derivative is negative are identified as local maxima.

[0126] Specifically, after obtaining the smoothed density estimation curve, in order to accurately find the positions of all the peaks on the curve, it is necessary to perform numerical calculations of the first and second derivatives of the curve.

[0127] The first derivative reflects the trend of the density estimation curve at various points: when the first derivative is positive, the curve is in an upward segment; when it is negative, the curve is in a downward segment; when it is zero, the curve may be at a peak, trough, or plateau. To further distinguish peaks, the second derivative needs to be calculated: if the first derivative is zero and the second derivative is negative at a certain position, it indicates that the curve at that point changes from upward to downward, and the curve near that point has a downward convex shape. In actual numerical implementation, since the derivative at discrete sampling points cannot be exactly zero, the sign change method is usually used: traversing the sampling points on the curve, if the first derivative at a certain point changes from positive to negative, that is, the previous derivative was positive and the current derivative is close to zero or directly negative, then that point is identified as a local maximum point.

[0128] The historical ambient temperature value corresponding to each local maximum point is used as a candidate peak point.

[0129] If there is only one candidate peak point, the sample distribution within the preliminary atmospheric temperature range is determined to be unimodal, and the preliminary atmospheric temperature range is taken as the final atmospheric temperature range. If there is more than one candidate peak point, it is determined to be multimodal, and the point where the first derivative of the density estimation curve between adjacent candidate peak points turns from negative to positive is taken as the valley floor dividing point. The preliminary atmospheric temperature range is divided into multiple sub-ranges, and each sub-range is taken as a final atmospheric temperature range.

[0130] It should be noted that after obtaining the number of candidate peak points, the processing method for the current preliminary atmospheric temperature range needs to be determined based on this number. If there is only one candidate peak point, it means that all historical ambient temperature values ​​within this range are clustered around a single center, i.e., the sample distribution has a single-peak shape. In this case, there is no need to split the range; the preliminary atmospheric temperature range can be directly used as the final atmospheric temperature range. If there are two or more candidate peak points, it means that there are multiple clustering centers in the sample distribution within this range. For example, within the same atmospheric temperature range, fuses may exhibit two or more typical ambient temperatures due to differences in heat dissipation conditions or aging states, i.e., a multi-peak shape. In this case, the multi-peak shape needs to be split into multiple single-peak sub-ranges.

[0131] Specifically, the splitting method involves finding the valley on the density estimation curve between every two adjacent candidate peak points. Mathematically, the valley corresponds to the point where the first derivative of the curve changes from negative to positive, where the density value is minimum, making it the optimal location to split the two peaks. Using the historical ambient temperature value corresponding to this valley as the splitting threshold, the initial interval is divided into two sub-intervals at this threshold. This splitting operation is repeated for each pair of adjacent peak points, ultimately splitting the original interval into multiple consecutive sub-intervals, each containing exactly one peak point, thus ensuring a unimodal sample distribution in each sub-interval. Each sub-interval serves as a final atmospheric ambient temperature interval for subsequent steps to extract the upper and lower bounds of normal ambient temperature. If a sub-interval contains too few sample points, it can be merged into an adjacent sub-interval with higher confidence, or marked as a low-confidence interval.

[0132] For the final atmospheric ambient temperature range, extract the historical ambient temperature values ​​corresponding to all valid temperature sample points within the final atmospheric ambient temperature range, and take the minimum value as the lower limit of the normal ambient temperature under the atmospheric ambient temperature range, and take the maximum value as the upper limit of the normal ambient temperature.

[0133] Specifically, for each final atmospheric temperature range obtained after adjustment, all historical ambient temperature sample points belonging to that range are first extracted from the effective temperature sample set; these are the temperature measurements of the thermal fuse 3 itself. These sample points have been purified by previous density screening and unimodal splitting, representing the typical ambient temperature for normal fuse operation under that atmospheric temperature condition. These historical ambient temperature values ​​are then scanned to find the minimum and maximum values. The minimum value is taken as the lower bound of the normal ambient temperature within that range, and the maximum value is taken as the upper bound. The lower and upper bounds represent the normal fluctuation range of the fuse's ambient temperature within that atmospheric temperature range.

[0134] Using each atmospheric ambient temperature range as an index key, the upper and lower bounds of the corresponding normal ambient temperature are associated and stored to construct a normal temperature range mapping database indexed by atmospheric ambient temperature.

[0135] It's important to note that when constructing the mapping database, each final atmospheric temperature range is used as an index key, and the corresponding lower and upper bounds of the normal ambient temperature range are used as association values, stored in the mapping database as key-value pairs. The database can be implemented using a hash table (dictionary), an in-memory database, or a relational database table. For example, the table structure could be: Atmospheric Temperature Range Start Value, Atmospheric Temperature Range End Value, Normal Lower Bound, Normal Upper Bound. During storage, it's crucial to ensure that each range is non-overlapping and covers all valid temperature ranges. Furthermore, to improve the efficiency of subsequent real-time queries, the range keys can be sorted and a range index can be built, such as using a segment tree or a binary search table. The final mapped database will serve as the core basis for self-diagnosis. When it's necessary to query the normal temperature range based on the real-time atmospheric ambient temperature, simply searching the mapping database for the atmospheric temperature range containing that real-time temperature allows for quick retrieval of the corresponding upper and lower bounds for comparison.

[0136] Based on the real-time atmospheric ambient temperature, the upper and lower limits of the normal ambient temperature of the thermal fuse 3 are determined by mapping the database, thus obtaining the real-time temperature range.

[0137] In a preferred embodiment, the step of determining the upper and lower bounds of the normal ambient temperature of the thermal fuse 3 based on the real-time atmospheric ambient temperature and obtaining the real-time temperature range through a mapping database includes the following steps:

[0138] The real-time ambient temperature is compared with each ambient temperature range stored in the mapping database to determine the target range containing the real-time ambient temperature. If a target range containing the real-time temperature exists, the lower and upper bound values ​​of the normal ambient temperature associated with the target range are directly read.

[0139] It should be noted that after obtaining the real-time atmospheric ambient temperature, the system searches the existing mapping database for atmospheric ambient temperature ranges containing that temperature. Each range in the mapping database is defined by a start boundary and an end boundary, for example... Furthermore, all intervals are non-overlapping and continuously cover most of the temperature range. By sequential scanning or binary search, a solution is found that satisfies... The interval, i.e., the target interval If found, the corresponding stored lower and upper bounds of the normal ambient temperature range are directly read as the normal temperature range under the current real-time atmospheric temperature for subsequent comparison. This scenario corresponds to a real-time temperature that is within the range fully covered by historical data.

[0140] If the real-time ambient temperature is less than the left boundary of the minimum interval in the mapping database or greater than the right boundary of the maximum interval, then the boundary interval closest to the real-time temperature is taken as the reference interval, and the lower and upper bounds of the real-time temperature are calculated using the reference interval and its adjacent intervals to obtain the real-time temperature interval.

[0141] It should be noted that if the real-time ambient temperature is less than the left boundary of the minimum interval in the mapping database, or greater than the right boundary of the maximum interval, an interval containing that real-time temperature cannot be directly found. In this case, the boundary interval closest to that real-time temperature is taken as the baseline interval; for example, if the temperature is too low, the minimum temperature interval is taken; if the temperature is too high, the maximum temperature interval is taken. To estimate the upper and lower bounds of the normal ambient temperature under this extreme temperature, a linear extrapolation method can be used: the slope of the upper and lower bounds changing with the ambient temperature is calculated using the baseline interval and one adjacent interval (if any), and then the extrapolated lower and upper bounds are obtained by extending outwards according to the slope based on the distance between the real-time temperature and the boundary of the baseline interval. If there is only a baseline interval and no adjacent intervals, the upper and lower bound values ​​of the baseline interval are directly used. The extrapolation result is marked as low confidence, and the comparison tolerance can be appropriately relaxed in subsequent self-diagnosis.

[0142] If the real-time ambient temperature is located in the gap between two adjacent intervals in the mapping database, then linear interpolation is performed on the adjacent intervals on both sides of the gap, and the lower and upper bound interpolation values ​​are calculated respectively. The interpolation results are used as the lower and upper bound values ​​of the normal ambient temperature corresponding to the real-time temperature to obtain the real-time temperature interval.

[0143] It should be noted that if the real-time ambient temperature falls within the gap between two adjacent intervals in the mapping database (i.e., greater than the end boundary of the previous interval and less than the start boundary of the next interval), it indicates that there are gaps in the historical data for that temperature range. In this case, the lower and upper bounds of the normal ambient temperature are obtained for the two adjacent intervals on the left and right sides of the gap. Then, using the ambient temperature as the x-axis, linear interpolation is performed on the lower and upper bounds. Specifically, the relative position of the real-time temperature between the boundaries of the two intervals is calculated, and the corresponding lower and upper bound interpolation values ​​are calculated proportionally. The interpolation results are used as the lower and upper bounds of the normal ambient temperature for that real-time ambient temperature, thus obtaining the real-time temperature interval.

[0144] The real-time temperature data of thermal fuse 3 is compared with the real-time temperature range. If the real-time temperature data exceeds the real-time temperature range, it is determined that thermal fuse 3 is abnormal; otherwise, it is considered that thermal fuse 3 is normal.

[0145] After detecting that the main control electronic switch is out of control, the main power controller 9 sends a trigger signal for the main switch to be out of control to the microcontroller 5. After receiving the trigger signal, the microcontroller 5 dynamically controls the heating module 2 to heat the thermal fuse 3 in combination with the ambient temperature of the thermal fuse 3.

[0146] When the real-time ambient temperature of the thermal fuse 3 is greater than the target melting temperature, the thermal fuse 3 melts, so that the heating module 2 is powered off and stops heating, and the circuit is cut off by the melting of the thermal fuse 3.

[0147] like Figures 4-5 As shown, according to a second embodiment of the present invention, a device for MCU-based fuse protection when the power electronic switch fails is provided. The device includes:

[0148] The control and acquisition module 1 is used to receive the runaway trigger signal from the main power controller 9, drive the heating module 2, and acquire the real-time ambient temperature;

[0149] Heating module 2 is used to generate heat when powered on using ceramic heating elements and to heat the temperature fuse 3.

[0150] The thermal fuse 3 is connected in series in the main circuit. Under normal operating temperature, it can carry a rated current. When the sensed temperature reaches the preset target melting temperature, its internal alloy solder joints will irreversibly melt, thus physically cutting off the main circuit. The heating module 2 and the thermal fuse 3 are integrated into a single package, ensuring efficient heat conduction and shortening the melting response time through a tight fit. The power supply circuit of the heating module 2 relies on the thermal fuse 3 for conduction. When the thermal fuse 3 melts, it simultaneously cuts off the heating power supply, automatically stopping the heating function and avoiding the risk of continuous power consumption or overheating.

[0151] In addition, the thermal fuse 3 is a one-time, non-resettable, temperature-sensitive fuse element that physically and irreversibly cuts off the main circuit when the temperature reaches the target melting threshold.

[0152] Fuse interface 4 is used as a port for electrical connection of the main circuit;

[0153] Microcontroller 5, or MCU, is used to execute heating strategies and temperature judgments. It is equipped with a self-diagnostic module to perform self-diagnostic judgments on thermal fuse 3 based on the real-time ambient temperature and the real-time atmospheric temperature.

[0154] The DIP switch 6 is a mechanical coded switch used by the user to preset the target melting temperature and has a power-off memory function.

[0155] Signal control port 7 is used to receive the runaway trigger signal from the power supply main controller 9;

[0156] Temperature sampling port 8 is used to collect and upload the real-time ambient temperature of temperature fuse 3. The collected real-time ambient temperature is transmitted to control and acquisition module 1.

[0157] The control and acquisition module 1 and the power supply main controller 9 are connected through the signal control port 7.

[0158] It should be noted that the MCU-based fuse protection device, which trips when the power electronic switch fails, is also connected to the main power controller 9, the positive power port 10, and the positive port of the external device 11 (such as a load), thus establishing a circuit connection. The circuit connection is divided into two parts: the main circuit and the signal control.

[0159] 1. Main circuit: Thermal fuse 3 is connected in series in the main circuit, with one end connected to the positive terminal of the power supply and the other end connected to the positive terminal of the external device (load, etc.);

[0160] 2. Signal control: The signal control port is connected to the main power controller 9 and receives the power controller 9's failure trigger signal.

[0161] Among them, the power supply main controller 9 is a third-party device, and its methods for detecting the failure of the main control electronic switch vary. Here, we only take the BMS control charging and discharging circuit as an example: when the power supply main controller BMS controls the electronic switch of its charging circuit to be open, and a continuous charging current can still be detected in the loop, then the main control electronic switch circuit is determined to be failed; when the power supply main controller BMS controls the electronic switch of its discharging circuit to be open, and a continuous discharging current can still be detected in the loop, then the main control electronic switch circuit is determined to be failed.

[0162] Furthermore, the main switch malfunctions in this invention include: MOSFET drain-source breakdown short circuit, relay contact sticking, and relay coil jamming. Regardless of whether the power supply is charging, discharging, or stationary, once a main switch malfunction is detected, a heating fuse will be triggered, achieving full-scenario protection. The entire protection process does not generate short circuits, large current surges, or electric arcs, and does not damage the power supply, wiring harness, or other components. After the thermal fuse blows, the main circuit is permanently disconnected and cannot be automatically restored, achieving final-level safety isolation.

[0163] Specifically, taking a 48V / 100A DC power supply system as an example, the specific implementation method is as follows:

[0164] 1. Thermal fuse: set to 100A, operating temperature 102℃;

[0165] 2. Heating module: Ceramic heating element, 60W;

[0166] 3. Control: After receiving the trigger signal, the MCU heats at a fixed power. At the same time, the temperature sensor detects that the temperature fuse 3 melts and automatically cuts off the power when the temperature reaches the preset threshold.

[0167] 4. Response time: ≤10 seconds;

[0168] 5. Action result: The main circuit is permanently disconnected, and heating module 2 is simultaneously powered off.

[0169] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for MCU-based fuse protection when a power electronic switch fails, characterized in that, include: S1. Obtain the target fusing temperature, the real-time ambient temperature of the thermal fuse, and the real-time atmospheric ambient temperature that the user has preset via the DIP switch. S2. Based on the real-time ambient temperature and real-time atmospheric temperature of the thermal fuse, perform a self-diagnosis judgment on the thermal fuse. If it is determined to be normal, proceed to step S3; otherwise, send a warning message to the user to replace the thermal fuse and return to step S1. The self-diagnosis is performed by constructing a mapping database based on the temperature density analysis method of local sensitive hash function. S3. After the microcontroller receives the trigger signal of the out-of-control situation, it dynamically controls the heating module to heat the thermal fuse based on the ambient temperature of the thermal fuse. S4. When the real-time ambient temperature of the thermal fuse is greater than the target melting temperature, the thermal fuse will melt.

2. The method for MCU-based fuse protection when the power electronic switch fails, as described in claim 1, is characterized in that... The self-diagnostic judgment of the temperature fuse based on the real-time ambient temperature and real-time atmospheric temperature includes the following steps: Historical ambient temperature data of thermal fuses under different atmospheric ambient temperatures were collected, and temperature-dense analysis method using local sensitive hash function was combined to construct a normal temperature range mapping database indexed by ambient temperature. Based on the real-time ambient temperature, the upper and lower limits of the normal ambient temperature of the thermal fuse under the real-time ambient temperature are determined by mapping the database, thus obtaining the real-time temperature range. The real-time temperature data of the thermal fuse is compared with the real-time temperature range. If the real-time temperature data exceeds the real-time temperature range, the thermal fuse is considered to be abnormal; otherwise, the thermal fuse is considered to be normal.

3. The method for MCU-based fuse protection when the power electronic switch fails, as described in claim 2, is characterized in that... The process of collecting historical ambient temperature data of thermal fuses under different atmospheric ambient temperatures, and constructing a normal temperature range mapping database indexed by ambient temperature using a temperature-dense analysis method based on locality-sensitive hash functions, includes the following steps: Based on historical ambient temperature data of thermal fuses under different atmospheric ambient temperatures, a temperature feature plane is constructed with ambient temperature as the horizontal axis and historical ambient temperature as the vertical axis. The temperature feature plane is divided into several temperature segments along the atmospheric temperature dimension, and the effective temperature sample set that meets the preset conditions is selected by calculating the density of historical ambient temperature sample points and the density threshold in each temperature segment. Locality-sensitive hashing (LSH) is used to perform hash mapping on the atmospheric ambient temperature data in the effective temperature sample set, and multiple non-overlapping preliminary atmospheric ambient temperature ranges are generated based on the hash mapping results. By using the inflection point analysis of the cumulative distribution curve, the peak point in the preliminary atmospheric temperature range is identified, and the preliminary atmospheric temperature range is adjusted based on the peak point to obtain the final atmospheric temperature range. For the final atmospheric ambient temperature range, extract the historical ambient temperature values ​​corresponding to all valid temperature sample points within the final atmospheric ambient temperature range, and take the minimum value as the lower limit of the normal ambient temperature under the atmospheric ambient temperature range, and take the maximum value as the upper limit of the normal ambient temperature. Using each atmospheric ambient temperature range as an index key, the upper and lower bounds of the corresponding normal ambient temperature are associated and stored to construct a normal temperature range mapping database indexed by atmospheric ambient temperature.

4. The method for MCU-based fuse protection when the power electronic switch fails, as described in claim 3, is characterized in that... The process of dividing the temperature feature plane into several temperature segments along the atmospheric temperature dimension, and then selecting an effective temperature sample set that meets preset conditions by calculating the density of historical temperature sample points and a density threshold within each temperature segment, includes the following steps: The temperature feature plane is divided into several continuous and non-overlapping temperature segments along the atmospheric ambient temperature dimension according to a preset fixed temperature step size. The total number of historical ambient temperature sample points in each temperature segment is counted and used as the density value of the temperature segment. Calculate the average density value of all temperature segment intervals to obtain the average density, and use the product of the preset proportional coefficient and the average density as the density threshold. The density value of each temperature segment interval is compared with the density threshold. If the density value of a certain temperature segment interval is greater than or equal to the density threshold, all historical ambient temperature sample points in that temperature segment interval are retained; otherwise, all historical ambient temperature sample points in that temperature segment interval are removed. The historical environmental temperature sample points that are retained are summarized to form an effective temperature sample set.

5. The method for MCU-based fuse protection when the power electronic switch fails, as described in claim 3, is characterized in that... The process of using a locality-sensitive hash function to perform hash mapping on the atmospheric ambient temperature data in the effective temperature sample set, and generating multiple non-overlapping continuous atmospheric ambient temperature intervals based on the hash mapping results, includes the following steps: Extract the center temperature value of each temperature segment interval in the effective temperature sample set, and use the center temperature value as the characteristic temperature of that temperature segment interval to obtain a characteristic temperature sequence composed of all characteristic temperatures. A family of locality-sensitive hash functions is used to perform hash calculations on each feature temperature in the feature temperature sequence to obtain the hash value corresponding to each temperature segment interval; temperature segment intervals with the same hash value and adjacent atmospheric ambient temperatures are merged into temporary temperature groups. For each temporary temperature group, the minimum starting boundary of all temperature segment intervals it covers is taken as the left boundary, and the maximum ending boundary is taken as the right boundary, so as to generate multiple non-overlapping and continuous atmospheric ambient temperature intervals.

6. The method for MCU-based fuse protection when the power electronic switch fails, as described in claim 5, is characterized in that... The step of using a family of locality-sensitive hash functions to perform hash calculations on each feature temperature in the feature temperature sequence to obtain the hash value corresponding to each temperature segment interval; merging temperature segment intervals with the same hash value and adjacent atmospheric ambient temperatures into temporary temperature groups includes the following steps: Initialize the configuration parameters of the locality-sensitive hash function family and construct a hash function family consisting of multiple independent locality-sensitive hash functions; The feature temperature in each feature temperature sequence is hashed using all locality-sensitive hash functions in the hash function family, generating a hash value sequence corresponding to the number of hash functions, and the hash value sequence is used as the feature hash signature of the corresponding temperature segment interval. Temperature segment intervals with completely identical feature hash signatures are divided into the same candidate group, resulting in multiple initial candidate groups; The temperature segments within each initial candidate group are arranged in ascending order of atmospheric ambient temperature. Adjacent temperature segments with a temperature difference less than or equal to a preset temperature similarity threshold are merged to generate multiple temporary temperature groups.

7. The method for MCU-based fuse protection when the power electronic switch fails, as described in claim 5, is characterized in that... The method of using cumulative distribution curve inflection point analysis to identify peak points in the preliminary atmospheric temperature range, and adjusting the preliminary atmospheric temperature range based on these peak points to obtain the final atmospheric temperature range includes the following steps: For each preliminary atmospheric temperature range, extract the historical ambient temperature values ​​of all valid temperature sample points, sort them in ascending order of value, calculate the cumulative probability corresponding to each temperature value, and plot the empirical cumulative distribution curve. Based on the empirical cumulative distribution curve, candidate peak points are obtained by identifying the historical ambient temperature values ​​corresponding to local maxima. If there is only one candidate peak point, the sample distribution within the preliminary atmospheric temperature range is determined to be unimodal, and the preliminary atmospheric temperature range is taken as the final atmospheric temperature range. If there is more than one candidate peak point, it is determined to be multimodal, and the point where the first derivative of the density estimation curve between adjacent candidate peak points turns from negative to positive is taken as the valley floor dividing point. The preliminary atmospheric temperature range is divided into multiple sub-ranges, and each sub-range is taken as a final atmospheric temperature range.

8. The method for MCU-based fuse protection when the power electronic switch fails, as described in claim 7, is characterized in that... The process of obtaining candidate peak points by identifying historical environmental temperature values ​​corresponding to local maxima based on empirical cumulative distribution curves includes the following steps: The kernel density of the empirical cumulative distribution curve is estimated by using the Gaussian kernel function, and the probability density estimate at each historical ambient temperature value is calculated to generate a smoothed density estimation curve. The first and second derivatives of the smoothed density estimation curve are calculated, and points where the first derivative is zero and the second derivative is negative are identified as local maxima. The historical ambient temperature value corresponding to each local maximum point is used as a candidate peak point.

9. The method for MCU-based fuse protection when the power electronic switch fails, as described in claim 2, is characterized in that... The process of determining the upper and lower bounds of the normal ambient temperature for the thermal fuse based on real-time atmospheric temperature, through a mapping database, to obtain the real-time temperature range includes the following steps: The real-time ambient temperature is compared with each ambient temperature range stored in the mapping database to determine the target range containing the real-time ambient temperature. If a target range containing the real-time temperature exists, the lower and upper bound values ​​of the normal ambient temperature associated with the target range are directly read. If the real-time ambient temperature is less than the left boundary of the minimum interval in the mapping database or greater than the right boundary of the maximum interval, then the boundary interval closest to the real-time temperature is taken as the reference interval, and the lower and upper bounds of the real-time temperature are calculated using the reference interval and its adjacent intervals to obtain the real-time temperature interval. If the real-time ambient temperature is located in the gap between two adjacent intervals in the mapping database, then linear interpolation is performed on the adjacent intervals on both sides of the gap, and the lower and upper bound interpolation values ​​are calculated respectively. The interpolation results are used as the lower and upper bound values ​​of the normal ambient temperature corresponding to the real-time temperature to obtain the real-time temperature interval.

10. A device for MCU-based fuse protection in the event of a power electronic switch failure, used to implement the MCU-based fuse protection method for the event of a power electronic switch failure as described in any one of claims 1-9, characterized in that, The device includes: The control and acquisition module is used to receive the runaway trigger signal from the main power controller, drive the heating module, and acquire the real-time ambient temperature. The heating module uses ceramic heating elements, which generate heat when powered on, and also heats the thermal fuse. A thermal fuse is connected in series in the main circuit. Under normal operating temperature, it can carry a rated current. When the sensed temperature reaches the preset target melting temperature, its internal alloy solder joints will irreversibly melt and break, thereby physically cutting off the main circuit. Fuse connector, used as a port for electrical connection of the main circuit; The microcontroller is used to execute the heating strategy and determine the temperature. DIP switches are mechanical coded switches used by users to preset target melting temperatures and have a power-off memory function. The signal control port is used to receive the runaway trigger signal from the power supply main controller; The temperature sampling port is used to collect and upload the real-time ambient temperature of the thermal fuse.

Images