Intelligent testing method and device for frequency converter module

Abstract

The invention relates to the field of frequency converter testing. The intelligent testing method for the frequency converter module comprises the steps of determining a first qualified rate based on the state of basic working parameters of the frequency converter module; determining a second qualified rate based on the index characteristics of the frequency converter module; determining a third qualified rate based on the current voltage, the current current and the current power factor of the frequency converter module; determining a fourth qualified rate based on the current current curve, the current voltage curve and the current temperature curve of the frequency converter module; and determining a final test result of the frequency converter module according to the first qualified rate, the second qualified rate, the third qualified rate and the fourth qualified rate. According to the intelligent test method and device for the frequency converter module, the basic working parameters, the index characteristics, the real-time electrical parameters and the curve data of the frequency converter module are comprehensively considered, so that the finally given test result is more accurate and stable.

Description

Technical Field

[0001] This invention relates to the field of frequency converter testing technology, and specifically to an intelligent testing method and apparatus for frequency converter modules. Background Technology

[0002] In the field of industrial automation, the inverter module is the core drive unit of the compressor. Its product quality directly determines the operational stability and reliability of the terminal equipment. Therefore, accurate testing before leaving the factory is a key link to ensure product quality.

[0003] Currently, inverter module testing systems generally adopt a centralized architecture design, relying on a single sensor to collect basic data such as voltage and current, which is then centrally processed by a host computer. The test logic judges whether a value is qualified or not based solely on predefined hard thresholds. However, this traditional testing approach has many insurmountable drawbacks in practical applications:

[0004] First, the misjudgment rate remains high. Fixed thresholds cannot adapt to the subtle environmental fluctuations during the product manufacturing process, easily misjudging borderline qualified products as unqualified or missing products with minor defects. The actual misjudgment rate is usually as high as 5%-10%, which seriously affects production efficiency and product qualification rate.

[0005] Secondly, there is a lack of latent fault detection capabilities. Traditional solutions can only identify explicit faults such as voltage over-limit, and cannot detect potential defects such as early aging of components and poor soldering through multi-parameter correlation analysis. These latent problems will gradually worsen during the operation of the terminal equipment, leading to frequent failures in the later stages.

[0006] Therefore, developing an intelligent testing system for frequency converter modules that can improve testing accuracy and detect potential faults has become an urgent technical problem to be solved in the field of industrial automation testing. Summary of the Invention

[0007] This invention provides an intelligent testing method and device for inverter modules, which solves the defects of low testing accuracy and lack of fault detection capability in the prior art for inverter modules.

[0008] A smart testing method for frequency converter modules includes:

[0009] Obtain the basic operating parameters of the frequency converter module; determine the first pass rate based on the status of the basic operating parameters;

[0010] Obtain the index characteristics of the frequency converter module, and determine the second pass rate of the frequency converter module based on the index characteristics;

[0011] Obtain the current voltage, current current, and current power factor of the inverter module, and determine the third pass rate of the inverter module based on the current voltage, current current, and current power factor;

[0012] Obtain the current current curve, current voltage curve, and current temperature curve of the inverter module, and determine the fourth pass rate of the inverter module based on the current current curve, current voltage curve, and current temperature curve.

[0013] The final test results of the frequency converter module are determined based on the first pass rate, the second pass rate, the third pass rate, and the fourth pass rate.

[0014] According to the intelligent testing method for inverter modules described above, the basic operating parameters include: compressor operating frequency, fan speed, module bus voltage, and internal fault code status of the module status register.

[0015] The first pass rate is determined based on the status of the basic working parameters. include:

[0016] Determine whether the basic working parameters simultaneously meet the conditions. If they do, then the first pass rate is... The first pass rate is 1 if any condition is not met. =0;

[0017] The conditions include: whether the compressor operating frequency is stable at the test target frequency, whether the fan speed reaches the target speed, whether the module bus voltage is within the normal range, and whether the module status register displays an internal fault code.

[0018] According to the intelligent testing method for inverter modules described above, the index characteristics include: power deviation rate, voltage sampling deviation, current sampling deviation, power factor, module temperature, and time taken for the module to increase its frequency to 75Hz.

[0019] The determination of the second pass rate of the inverter module based on the aforementioned indicator characteristics includes:

[0020] Based on the aforementioned indicator characteristics, construct an indicator feature vector;

[0021] The indicator feature vector is input into the score prediction model, and based on the model, the initial total score of the indicator feature is predicted; the score prediction model is constructed using the GBDT model.

[0022] The initial total score is then converted into a pass / fail probability.

[0023] The calculated pass rate is compared with a preset threshold, and a second pass rate is determined based on the comparison result.

[0024] Based on the inverter module intelligent testing method described above, the second pass rate is determined according to the following formula. :

[0025] ᵀ

[0026]

[0027]

[0028]

[0029] in: , , For the score prediction model, is a constant. For learning rate, These correspond to the vectors of power deviation rate, voltage sampling deviation, current sampling deviation, power factor measured by electrical parameter instruments, module temperature, and time taken for the module to upclock to 75Hz, respectively. When the number of decision trees is t, the original predicted total score for feature vector X1 is... This represents the weight of vector X1 falling into the t-th decision tree. This represents the probability of passing.

[0030] According to the intelligent testing method for inverter modules described above, determining the third pass rate of the inverter module based on the current voltage, current current, and current power factor includes:

[0031] Based on the current voltage, current, and power factor, determine the power index of the inverter module. Health Index ,temperature ;

[0032] According to the power qualification rate Health Index Temperature index Determine the third pass rate of the inverter module. .

[0033] According to the intelligent testing method for inverter modules described above, the power index Determined according to the following formula:

[0034]

[0035]

[0036]

[0037]

[0038] in, In the case of a single-phase system, In the case of a three-phase system, , Indicates the current voltage; Indicates the current; Indicates the current power factor; Indicates the current power; Theoretical power, This is the power deviation value; For adaptive tolerance;

[0039] The health index Determined according to the following formula:

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] in, For historical qualified datasets, Indicates the number of qualified samples. These are the voltage, current, and power factor of the i-th qualified sample, respectively. For parameter vector X2 and historical qualified dataset Mahalanobis distance between them; This is the parameter vector of the module; These are the module's power supply voltage vector, power supply current vector, and actual power factor vector, respectively. This is the mean vector of parameters for the healthy cluster. These are the average voltage, average current, and average power factor, respectively. The health assessment threshold for Mahalanobis distance;

[0047] The temperature index Determined according to the following formula:

[0048]

[0049]

[0050]

[0051]

[0052] in, To predict temperature, For ambient temperature, The measured power value is given by the electrical parameter meter, and t is the running time. The power temperature rise coefficient, The time-temperature rise coefficient, For actual measured temperature, It is designed for adaptive temperature tolerance.

[0053] According to the intelligent testing method for inverter modules described above, determining the fourth pass rate of the inverter module based on the current curve, voltage curve, and current temperature curve of the current test includes:

[0054] Obtain the test sequence; the test sequence consists of the instantaneous voltage, instantaneous current, and instantaneous temperature values ​​that change with time t during this test;

[0055] The test sequence is compared one by one with the typical fault curves of each type in the pre-set fault mode library to determine the fault distance between the test sequence and each type of fault sequence.

[0056] The fourth pass rate is determined based on the fault distance.

[0057] According to the intelligent testing method for inverter modules described above, determining the final test result of the inverter module based on the first pass rate, the second pass rate, the third pass rate, and the fourth pass rate includes:

[0058] The final test result of the frequency converter module is determined according to the following formula. :

[0059]

[0060]

[0061] in, The weighted score for the inverter module. The highest pass rate, The second pass rate, The third pass rate, It is the fourth qualified;

[0062] , These are the weights corresponding to each pass rate.

[0063] Based on the intelligent testing method for inverter modules described above, the score prediction model is continuously updated according to the final result corresponding to the module to be reviewed.

[0064] An intelligent testing system for frequency converter modules includes:

[0065] Electrical parameter meter, used by users to measure the voltage, current, power and power factor of frequency converter modules;

[0066] Voltage detection board, used to detect the DC bus voltage inside the inverter module;

[0067] The variable frequency drive is configured to receive test commands from the edge intelligent processing layer, drive the load simulation unit to run according to preset parameters, realize frequency upsampling control from 0Hz to 75Hz, and provide feedback on the frequency upsampling execution status.

[0068] The compressor is used in conjunction with the frequency converter to complete the full-condition test of startup, frequency increase, and stable operation, so as to output the fan speed;

[0069] The lower-level machine is used to receive detection data via the RS485 communication interface;

[0070] The host computer includes an acquisition unit and a determination unit; wherein, the acquisition unit is used to acquire the basic operating parameters of the frequency converter module;

[0071] The determining unit is used to determine a first pass rate based on the status of the basic working parameters;

[0072] The acquisition unit is also used to acquire the index characteristics of the frequency converter module;

[0073] The determining unit is further configured to determine a second pass rate of the inverter module based on the index characteristics;

[0074] The acquisition unit is also used to acquire the current voltage, current current and current power factor of the inverter module;

[0075] The determining unit is also used to determine the third pass rate of the inverter module based on the current voltage, current current and current power factor;

[0076] The acquisition unit is also used to acquire the current current curve, current voltage curve, and current temperature curve of the inverter module;

[0077] The determining unit is also used to determine the fourth pass rate of the inverter module based on the current current curve, the current voltage curve, and the current temperature curve.

[0078] The determining unit is further configured to determine the final test result of the frequency converter module based on the first pass rate, the second pass rate, the third pass rate, and the fourth pass rate.

[0079] The present invention provides an intelligent testing method and apparatus for frequency converter modules. The method provided in this application, by comprehensively considering the basic operating parameters, index characteristics, real-time electrical parameters and curve data of the frequency converter module, makes the final test results more accurate and stable. Attached Figure Description

[0080] Figure 1 This is one of the flowcharts of the intelligent testing method for frequency converter modules provided by the present invention;

[0081] Figure 2 The second flowchart is for the intelligent testing method of frequency converter module provided by the present invention. Detailed Implementation

[0082] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0083] Traditional testing systems are typically built on a centralized processing architecture. Their typical operating mode involves using a single sensor (e.g., an electrical parameter meter) to collect key electrical parameters (such as voltage and current) of the module under test, and uploading all data to a host computer for centralized processing and judgment. The system's testing logic relies entirely on preset fixed thresholds (e.g., voltage range, frequency standards). However, the use of fixed thresholds fails to adapt to normal variations or environmental fluctuations between different products, resulting in many products bordering on pass being falsely judged as fail, or some defective products being missed. The overall false judgment rate is typically as high as 5%-10%. Moreover, traditional solutions can only identify obvious, overt faults exceeding hard thresholds (such as voltage overruns), and cannot detect potential "soft faults" such as early performance degradation or poor soldering by analyzing the correlation between multiple characteristic parameters. Furthermore, the system lacks the ability to learn from historical data and self-optimize. When new products are introduced or production processes change, manual experience must be relied upon to readjust test parameters and thresholds, a cumbersome and slow process. Therefore, this application employs a multimodal intelligent diagnostic mechanism, integrating basic rule verification, adaptive learning models, parameter correlation cross-validation, and sequence fault mode matching to comprehensively analyze the inverter module from different dimensions, thereby improving test accuracy and reliability.

[0084] Figure 1 The flowchart of the intelligent testing method for frequency converter modules provided by this invention is as follows: Figure 1 As shown, the method includes the following steps:

[0085] Step 101: Obtain the basic operating parameters of the frequency converter module; determine the first pass rate based on the status of the basic operating parameters.

[0086] Specifically, firstly, the system needs to collect the basic operating parameters of the frequency converter module. Then, the system checks whether these basic operating parameters simultaneously meet preset conditions. If all conditions are met, the first pass rate is 1; otherwise, it is 0. This application verifies whether the basic operating status of the frequency converter module is normal through its basic operating parameters.

[0087] Step 102: Obtain the index characteristics of the frequency converter module, and determine the second pass rate of the frequency converter module based on the index characteristics.

[0088] Specifically, the system collects the performance characteristics of the frequency converter module and uses these characteristics to calculate the second pass rate through a decision tree-based model (or other adaptive learning model). This model considers the relative importance and correlation between the indicators, thus providing more accurate evaluation results.

[0089] Step 103: Obtain the current voltage, current current, and current power factor of the inverter module, and determine the third pass rate of the inverter module based on the current voltage, current current, and current power factor.

[0090] Specifically, the system measures the current voltage, current, and power factor of the inverter module in real time, and calculates the power index, health index, and temperature index based on these real-time electrical data, thereby determining the third pass rate. This allows for a comprehensive assessment of the health status of electrical parameters to provide the pass rate.

[0091] Step 104: Obtain the current current curve, current voltage curve, and current temperature curve of the inverter module. Based on the current current curve, current voltage curve, and current temperature curve, determine the fourth pass rate of the inverter module.

[0092] Specifically, the system records the current, voltage, and temperature curves of the inverter module during the test. By comparing these curves with typical fault curves in a pre-set fault mode library, the system calculates the fault distance between the test sequence and each type of fault sequence to determine the fourth pass rate. The fourth pass rate helps to identify potential sequential fault modes.

[0093] Step 105: Determine whether the frequency converter module is qualified based on the first pass rate, the second pass rate, the third pass rate, and the fourth pass rate.

[0094] Specifically, the system comprehensively considers the first, second, third, and fourth pass rates to give the final pass / fail judgment of the inverter module, ensuring the comprehensiveness and accuracy of the evaluation.

[0095] The method provided in this application, by comprehensively considering the basic operating parameters, index characteristics, real-time electrical parameters and curve data of the frequency converter module, makes the final test results more accurate and stable.

[0096] Furthermore, the basic operating parameters provided in this application include: compressor operating frequency, fan speed, module bus voltage, and internal fault code status of the module status register.

[0097] Specifically, the compressor is the main load of the inverter module, and its operating frequency directly reflects the module's output control capability. If the frequency is unstable or deviates from the target value, it will lead to decreased compressor operating efficiency, increased noise, and even mechanical failure. This application, by monitoring whether the compressor's operating frequency is stable within a preset range, can quantitatively evaluate the inverter module's frequency tracking accuracy and dynamic response capability, ensuring that it meets the drive requirements of the terminal equipment. For example, in an air conditioning system, excessive compressor frequency fluctuations may lead to unstable cooling / heating effects; therefore, this parameter is a key indicator of the first pass rate.

[0098] Fans typically work in conjunction with compressors (such as in the outdoor unit of an air conditioner), and their speed directly affects heat dissipation efficiency and airflow performance. Abnormal fan speed may indirectly trigger compressor overheat protection or reduce system energy efficiency. Meeting the fan speed requirements verifies the inverter module's ability to coordinate control across multiple loads.

[0099] Bus voltage is the voltage of the DC link within the inverter module, directly affecting the module's output capability and stability. Stable bus voltage is a prerequisite for the normal operation of the inverter module; excessively high or low bus voltages can lead to module damage or performance degradation. By monitoring whether the bus voltage is within the normal range, the inverter module's power management capability and anti-interference capability can be assessed. Abnormal bus voltage indicates a problem with the module's power management, which may affect its long-term stability and reliability. Therefore, bus voltage is a crucial parameter for determining the initial pass rate.

[0100] Internal fault code status is a direct reflection of the inverter module's self-diagnostic function, indicating whether the module has hardware faults or software errors. By reading the internal fault code status, problems in the module can be quickly located, preventing the fault from escalating or affecting the operation of terminal equipment. If the module displays an internal fault code, it indicates a hardware or software problem, which may prevent it from working properly or pose a safety hazard. Therefore, the internal fault code status is a parameter that must be considered when determining the initial pass rate.

[0101] Traditional single-parameter testing (such as measuring only bus voltage) may miss latent faults (such as frequency fluctuations but normal voltage). This application achieves full-chain verification from power supply stability to output control, load coordination, and fault diagnosis by integrating four sets of parameters, which significantly improves test coverage and enables a more comprehensive evaluation of the inverter module's operating status, thereby reducing the false alarm rate.

[0102] The following section explains how to determine the first pass rate based on the status of the basic working parameters described above. The implementation schemes are introduced, including the following:

[0103] Determine whether the basic working parameters simultaneously meet the conditions. If they do, then the first pass rate is... The first pass rate is 1 if any condition is not met. The condition is 0; the conditions include: whether the compressor operating frequency is stable at the test target frequency, whether the fan speed reaches the target speed, whether the module bus voltage is within the normal range, and whether the module status register displays an internal fault code.

[0104] Specifically, this application tests whether the actual operating frequency of the compressor is stable within the target value of 75Hz ± δ, where δ is the allowable deviation range, which needs to be specified according to the equipment's accuracy requirements. If the frequency fluctuation exceeds the allowable range, it is deemed unqualified. It also tests whether the actual fan speed reaches the target value of 800RPM ± 20RPM; if the speed exceeds this range, it is deemed unqualified. The module bus voltage is also tested. Check if the voltage is within the normal operating range (single-phase input: 280V≤V_bus≤380V, three-phase input: 500V≤V_bus≤700V), or check if the PFC (Power Factor Correction) function is normal. If the voltage exceeds the limit or the PFC function is abnormal, it is considered unqualified. Internal fault code verification is performed by reading the module status register through a communication interface (such as CAN, RS485, etc.) to confirm the presence of non-zero internal fault codes (such as overcurrent, overvoltage, overheating, communication errors, etc.). If any fault code is detected, it is considered unqualified.

[0105] This application employs a veto system, meaning that if any test item fails to meet the requirements, the subsequent testing process is immediately terminated, and the final result is output. The final result is as follows:

[0106]

[0107] Traditional testing methods can only identify explicit faults (such as voltage over-limit) and cannot detect potential defects such as early performance degradation and poor soldering. This application, by analyzing the stability and consistency of operating frequency, fan speed, and bus voltage, helps to identify potential performance degradation problems that may occur during long-term operation of the module. Simultaneously, the internal fault code status can directly reflect whether there are hardware or software faults in the module, thereby enhancing the detection capability of latent faults.

[0108] The method provided in this application constructs a comprehensive pass rate evaluation system covering power supply, control, load, and fault by coordinating the detection of four sets of parameters: compressor operating frequency, fan speed, module bus voltage, and internal fault code status, effectively improving the accuracy and reliability of testing.

[0109] Furthermore, the index features provided in this application include: power deviation rate. Voltage sampling deviation Current sampling deviation Power factor Module temperature Time required for module to upclock to 75Hz .

[0110] Specifically, power deviation rate It is used to measure the deviation ratio between actual output power and expected power, reflecting the stability of power control. The larger the deviation, the more likely there is an anomaly. Its calculation formula is as follows:

[0111]

[0112] in, This is the measured power value from the electrical parameter meter. This represents the expected power value.

[0113] Voltage sampling deviation This measurement measures the deviation between the internal sampled voltage of an IPM (Intelligent Power Module) and the measured power supply voltage by an electrical parameter analyzer, reflecting the accuracy of the voltage sampling circuit. Voltage is fundamental to equipment operation; abnormal deviations can affect equipment performance. The calculation formula is as follows:

[0114]

[0115] in, Measure the voltage of the module. For measuring voltage with an electrical parameter instrument.

[0116] Current sampling deviation This is used to measure the deviation ratio between the internal sampling current of the IPM and the actual power supply current measured by the electrical parameter instrument. It reflects the accuracy of the current sampling circuit and its fault diagnosis capability. An abnormal current may indicate that the equipment has problems such as short circuit or overload. The calculation formula is as follows:

[0117]

[0118] in, Measure the current of the module. For measuring current using electrical parameters.

[0119] Power factor It reflects the efficiency of electricity utilization. If it is too low, it will cause energy waste. It is also an important indicator for judging whether the equipment is operating normally.

[0120] Module temperature ( The measured values ​​of key points inside the IPM module (such as IGBT junction temperature and heat sink temperature) reflect the thermal management performance and device reliability. Excessive operating temperature of the equipment can affect the lifespan of components and even cause failure.

[0121] Time required for module to upclock to 75Hz ( The frequency rise time is the time required for the compressor or motor to accelerate from 0Hz to 75Hz. The rise time reflects the module's response speed and performance. An excessively long rise time may indicate that the module is aging or has a fault.

[0122] The following describes how to determine the second pass rate of the inverter module based on the aforementioned indicator characteristics, specifically including the following scheme:

[0123] Based on the indicator characteristics, an indicator feature vector is constructed; the indicator feature vector is input into a score prediction model, and based on the model, an initial total score for the indicator characteristics is predicted; the score prediction model is constructed using a GBDT model; the initial total score is converted into a pass probability; the calculated pass probability is compared with a preset threshold, and a second pass rate is determined based on the comparison result.

[0124] Specifically, this application determines the second pass rate according to the following formula. :

[0125] ᵀ

[0126]

[0127]

[0128]

[0129] in: , , For the score prediction model, is a constant. The learning rate is 0 < η ≤ 1, which controls the contribution strength of each tree and prevents overfitting. These are vectors corresponding to the power deviation rate, voltage sampling deviation, current sampling deviation, power factor measured by electrical parameter instruments, module temperature, and time taken for the module to upclock to 75Hz, respectively. When the number of decision trees is t, the original predicted total score for feature vector X1 is... This represents the weight of vector X1 falling into the t-th decision tree. This represents the probability of passing.

[0130] Specifically, firstly, six characteristic data points of the frequency converter module are collected. Then, set the core parameters of the GBDT model: initial model constants. Learning rate The number of decision trees, t (in this embodiment, t=6), is used to divide the feature vectors. Given 6 decision trees, calculate the contribution value of each tree in turn. Then follow the formula Obtain the original predicted total score, and convert the original total score to the predicted total score. Substitute into the Sigmoid function to calculate the probability of passing. Finally, the probability is compared with the preset threshold, and the output is... ,like If it is less than 0.3, then it is determined that... If it is 0; If it is greater than 0.8, then it is determined that... If it is 1, If the value is greater than or equal to 0.3 and less than or equal to 0.8, then it is determined that... It is 0.5.

[0131] Traditional methods rely solely on threshold classification for a single feature (such as power deviation rate), easily overlooking complex relationships between multiple features. This application, however, uses the GBDT model to learn feature patterns from a massive number of qualified / unqualified samples, capturing implicit correlations between features such as power deviation and upsampling time, thus reducing the false positive rate. Furthermore, the model, through multi-tree collaborative analysis of the interaction relationships of six features, can identify soft faults that traditional methods cannot detect, such as poor soldering and early device degradation, allowing for early detection of potential module risks and reducing post-shipment failure rates. In addition, this application can output three levels of results: unqualified, pending, and qualified. Pending samples can enter a manual review process, ensuring testing efficiency while avoiding missed critical defects, thus avoiding the shortcomings of traditional solutions that only output qualified / unqualified results without distinguishing edge samples.

[0132] Furthermore, the following section introduces a scheme for determining the third pass rate of the inverter module based on the current voltage, current current, and current power factor, specifically including:

[0133] Based on the current voltage, current, and power factor, determine the power index of the inverter module. Health Index ,temperature According to the power qualification rate Health Index Temperature index Determine the third pass rate of the inverter module. .

[0134] This application comprehensively evaluates the qualification status of the frequency converter module through three key indices (power index, health index, and temperature index), and finally outputs a third qualification rate. .

[0135] Among them, power index Based on the electrical power calculation formula, the accuracy of the power data and the stability of the module output are verified by comparing the deviation between the measured power and the theoretically calculated power. The power is determined according to the following formula:

[0136]

[0137]

[0138]

[0139]

[0140] in, In the case of a single-phase system, In the case of a three-phase system, , Indicates the current voltage; Indicates the current; Indicates the current power factor; Indicates the current power; Theoretical power, This is the power deviation value; For adaptive tolerance;

[0141] Health Index A cluster of healthy parameters is constructed based on historical qualified samples to verify whether the voltage-current-power factor combination of the current module is within a healthy range. It is determined according to the following formula:

[0142]

[0143]

[0144] X2

[0145]

[0146]

[0147]

[0148] in, For historical qualified datasets, Indicates the number of qualified samples. These are the voltage, current, and power factor of the i-th qualified sample, respectively. For parameter vector X2 and historical qualified dataset Mahalanobis distance between them; This is the parameter vector of the module; These are the module's power supply voltage vector, power supply current vector, and actual power factor vector, respectively. This is the mean vector of parameters for the healthy cluster. These are the average voltage, average current, and average power factor, respectively. The health assessment threshold for Mahalanobis distance;

[0149] Temperature Index Based on the physical relationship between power and temperature (higher power and longer operating time result in higher temperature), this method verifies whether the current temperature matches the power and operating time, avoiding hidden heat dissipation or power consumption defects. The calculation formula is as follows:

[0150]

[0151]

[0152]

[0153]

[0154] in, To predict temperature, For ambient temperature, The measured power value is given by the electrical parameter meter, and t is the running time. The power temperature rise coefficient, The time-temperature rise coefficient, For actual measured temperature, It is designed for adaptive temperature tolerance.

[0155] When the temperature testing module is a scale block, the power temperature rise coefficient α is 1.2℃ / KW, and the time temperature rise coefficient β is 1.1℃ / KW. s: seconds.

[0156] Finally, the third pass rate was calculated based on a weighted average. :

[0157]

[0158] The method provided in this application, on the one hand, quantifies module performance through three dimensions: power stability, electrical health status, and thermal management efficiency, avoiding misjudgment based on a single indicator; on the other hand, it outputs a continuous pass rate through exponential weighting or multiplicative fusion, which can capture hidden faults that are difficult to identify by traditional methods, reduce the false judgment rate, and improve the accuracy of detection.

[0159] Furthermore, the following describes how to determine the fourth pass rate of the inverter module based on the current curve, voltage curve, and current temperature curve of the current test, specifically including the following scheme:

[0160] First, a test sequence is obtained; the test sequence consists of the instantaneous voltage, instantaneous current, and instantaneous temperature values ​​that change with time t during this test; second, the test sequence is compared one by one with various typical fault curves in a pre-set fault mode library to determine the fault distance between the test sequence and each type of fault sequence; finally, the fourth pass rate is determined based on the fault distance.

[0161] Specifically, during the testing process, this application collects three key parameters in real time that change with time t: instantaneous voltage value Reflects the dynamic fluctuations of the module's input / output voltage. Instantaneous current value. Capture transient changes in current (such as inrush current and harmonic current). Instantaneous temperature value. Monitor temperature changes in key components of the monitoring module (such as the IGBT substrate and heat sink). Combine the above parameters into a three-dimensional test sequence along the time axis. Typical failure modes of the frequency converter module are extracted through historical fault data, simulation, or accelerated aging tests in the laboratory. The formula for calculating the fourth pass rate is as follows:

[0162]

[0163]

[0164]

[0165]

[0166]

[0167]

[0168]

[0169] in, Let be the instantaneous values ​​of voltage, current, and temperature at time t, respectively. The test sequence at time t; This represents the optimal matching relationship between the i-th data point in the test sequence and the j-th data point in the fault sequence. N represents the dynamic sequence of the k-th typical fault in the fault mode library; N represents the test sequence. The total number of data points, M is the k-th fault sequence. The total number of data points; For the i-th data point in the test sequence, For the j-th data point in the fault sequence, Euclidean distance. To test the similarity between the sequence and the k-th failure mode.

[0170] The fault mode library provided in this application Store dynamic curves corresponding to K typical faults (such as voltage fluctuation curves corresponding to oscillation startup and temperature rise curves corresponding to excessively rapid temperature rise), and use the Dynamic Time Warping (DTW) algorithm to achieve curve comparison, thereby solving the problem of inconsistent duration / rhythm between test sequences and fault curves. This represents the Euclidean distance between the i-th point of the test sequence and the j-th point of the fault curve; For optimal regularization, the smaller the DTW distance, the higher the matching degree between the test sequence and the fault curve. The maximum normalized distance is the preset similarity. The closer it is to 1, the higher the degree of matching between the test sequence and the k-th type of fault.

[0171] The method provided in this application, which compares dynamic test sequences with a fault mode library, enables a quantitative evaluation of the dynamic performance of inverter modules. It can identify transient shocks (such as starting overcurrent and voltage drops caused by load changes), provide early warning of potential faults, avoid the shortcomings of traditional methods that only focus on steady-state parameters (such as rated voltage and current), and also avoid misjudgment of parameters at a single time point, thus further improving the accuracy of the test.

[0172] Furthermore, this application determines the final qualification of the frequency converter module according to the following formula:

[0173]

[0174]

[0175] in, The weighted score for the inverter module. The highest pass rate, The second pass rate, The third pass rate, It is the fourth qualified;

[0176] , These are the weights corresponding to each pass rate.

[0177] As a basic parameter weight, if it is not qualified, it means that the module cannot start / work normally, and it is a prerequisite for qualification. As a real-time parameter weight, abnormal real-time parameters (such as power deviation or sudden increase in module temperature) may lead to operational failure. As prediction weights for the model, they generally have a certain error and are only used for edge case determination to avoid affecting the overall situation; As the detection of latent faults is crucial to product reliability, it has the highest priority. Therefore, in this embodiment of the invention, the weight allocation is as follows: =0.3、 =0.2、 =0.15、 =0.35.

[0178] This application achieves the following balance through the above weight allocation:

[0179] Accuracy: Prioritize addressing the issue of missed detection of hidden faults in inverter testing, while reducing misjudgments of basic functions;

[0180] Engineering practicality: Combining real-time monitoring and model prediction, it adapts to the needs of complex working conditions;

[0181] Robustness: While ensuring the reliability of core functions, optimize test accuracy and model applicability.

[0182] The method provided in this application, through multi-dimensional weighted fusion (covering basic functions, model judgment, parameter correlation, and fault matching), takes into account the judgment results of each stage, while also... As a strong constraint, it avoids missing modules that are good in some areas but poor overall, and also reduces the misjudgment of modules that are poor in some areas but good overall. Moreover, through the hierarchical decision rules (qualified / unqualified / pending review), the results of samples that are clearly qualified / unqualified are directly output (accounting for about 80%), and only marginal samples are subject to manual review (accounting for about 20%), so that the testing efficiency is effectively improved while ensuring the rigor of the judgment.

[0183] like Figure 2 As shown below, the overall flow of the testing method provided in this application is described in detail:

[0184] Step 1: Adaptive Test Initialization

[0185] By scanning the product barcode, the product type is accurately identified. Based on the identified product model, customized test parameters and thresholds are loaded from the database to ensure the relevance and effectiveness of the test.

[0186] Step 2: Collaborative Data Acquisition and Preprocessing

[0187] The lower-level computer is responsible for real-time data acquisition and performs basic safety checks, such as detecting excessive voltage, to ensure system safety. The upper-level computer processes data from multiple sensors to determine four pass rates.

[0188] Step 3: Multimodal Intelligent Diagnosis

[0189] A comprehensive assessment from multiple perspectives is conducted to determine the final judgment.

[0190] Step 4: Feedback Learning and System Evolution

[0191] Detailed records are kept of the judgments of manually reviewed cases, providing a basis for system optimization. New judgments are used to fine-tune the model through online learning, continuously improving its accuracy. Newly identified fault characteristics are promptly updated to the fault mode library, enriching the system's fault diagnosis knowledge base.

[0192] This invention also provides an intelligent testing system for frequency converter modules, comprising:

[0193] Electrical parameter meter, used by users to measure the voltage, current, power and power factor of frequency converter modules;

[0194] Voltage detection board, used to detect the DC bus voltage inside the inverter module;

[0195] The variable frequency drive is configured to receive test commands from the edge intelligent processing layer, drive the load simulation unit to run according to preset parameters, realize frequency upsampling control from 0Hz to 75Hz, and provide feedback on the frequency upsampling execution status.

[0196] The compressor is used in conjunction with the frequency converter to complete the full-condition test of startup, frequency increase, and stable operation, so as to output the fan speed;

[0197] The lower-level machine is used to receive detection data via the RS485 communication interface;

[0198] The host computer includes an acquisition unit and a determination unit; wherein, the acquisition unit is used to acquire the basic operating parameters of the frequency converter module;

[0199] The determining unit is used to determine a first pass rate based on the status of the basic working parameters;

[0200] The acquisition unit is also used to acquire the index characteristics of the frequency converter module;

[0201] The determining unit is further configured to determine a second pass rate of the inverter module based on the index characteristics;

[0202] The acquisition unit is also used to acquire the current voltage, current current and current power factor of the inverter module;

[0203] The determining unit is also used to determine the third pass rate of the inverter module based on the current voltage, current current and current power factor;

[0204] The acquisition unit is also used to acquire the current current curve, current voltage curve, and current temperature curve of the inverter module;

[0205] The determining unit is also used to determine the fourth pass rate of the inverter module based on the current current curve, the current voltage curve, and the current temperature curve.

[0206] The determining unit is further configured to determine the final test result of the frequency converter module based on the first pass rate, the second pass rate, the third pass rate, and the fourth pass rate.

[0207] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart testing method for a frequency converter module, characterized in that, include: Obtain the basic operating parameters of the frequency converter module; determine the first pass rate based on the status of the basic operating parameters; Obtain the index characteristics of the frequency converter module, and determine the second pass rate of the frequency converter module based on the index characteristics; Obtain the current voltage, current current, and current power factor of the inverter module, and determine the third pass rate of the inverter module based on the current voltage, current current, and current power factor; Obtain the current current curve, current voltage curve, and current temperature curve of the inverter module, and determine the fourth pass rate of the inverter module based on the current current curve, current voltage curve, and current temperature curve. The final test results of the frequency converter module are determined based on the first pass rate, the second pass rate, the third pass rate, and the fourth pass rate.

2. The intelligent testing method for frequency converter modules according to claim 1, characterized in that, The basic operating parameters include: compressor operating frequency, fan speed, module bus voltage, and internal fault code status of the module status register; The first pass rate is determined based on the status of the basic working parameters. include: Determine whether the basic working parameters simultaneously meet the conditions. If they do, then the first pass rate is... The first pass rate is 1 if any condition is not met. =0; The conditions include: whether the compressor operating frequency is stable at the test target frequency, whether the fan speed reaches the target speed, whether the module bus voltage is within the normal range, and whether the module status register displays an internal fault code.

3. The intelligent testing method for frequency converter modules according to claim 1, characterized in that, The key performance indicators include: power deviation rate, voltage sampling deviation, current sampling deviation, power factor, module temperature, and time taken for the module to upclock to 75Hz. The determination of the second pass rate of the inverter module based on the aforementioned indicator characteristics includes: Based on the aforementioned indicator characteristics, construct an indicator feature vector; The indicator feature vector is input into the score prediction model, and based on the model, the initial total score of the indicator feature is predicted; the score prediction model is constructed using the GBDT model. The initial total score is then converted into a pass / fail probability. The calculated pass rate is compared with a preset threshold, and a second pass rate is determined based on the comparison result.

4. The intelligent testing method for frequency converter modules according to claim 3, characterized in that, The second pass rate is determined according to the following formula. : ᵀ in: , , For the score prediction model, is a constant. This is the learning rate, typically 0.

1. These correspond to the vectors of power deviation rate, voltage sampling deviation, current sampling deviation, power factor measured by electrical parameter instruments, module temperature, and time taken for the module to upclock to 75Hz, respectively. When the number of decision trees is t, the original predicted total score for feature vector X1 is... This represents the weight of vector X1 falling into the t-th decision tree. This represents the probability of passing.

5. The intelligent testing method for frequency converter modules according to claim 1, characterized in that, The determination of the third pass rate of the inverter module based on the current voltage, current current, and current power factor includes: Based on the current voltage, current, and power factor, determine the power index of the inverter module. Health Index ,temperature ; According to the power qualification rate Health Index Temperature index Determine the third pass rate of the inverter module. .

6. The intelligent testing method for frequency converter modules according to claim 5, characterized in that, The power index Determined according to the following formula: in, In the case of a single-phase system, In the case of a three-phase system, , Indicates the current voltage; Indicates the current; Indicates the current power factor; Indicates the current power; Theoretical power, This is the power deviation value; For adaptive tolerance; The health index Determined according to the following formula: in, For historical qualified datasets, Indicates the number of qualified samples. These are the voltage, current, and power factor of the i-th qualified sample, respectively. For parameter vector X2 and historical qualified dataset Mahalanobis distance between them; This is the parameter vector of the module; These are the module's power supply voltage vector, power supply current vector, and actual power factor vector, respectively. This is the mean vector of parameters for the healthy cluster. These are the average voltage, average current, and average power factor, respectively. The health assessment threshold is defined by Mahalanobis distance. The temperature index Determined according to the following formula: in, To predict temperature, For ambient temperature, The measured power value is given by the electrical parameter meter, and t is the running time. The power temperature rise coefficient, The time-temperature rise coefficient, For actual measured temperature, It is designed for adaptive temperature tolerance.

7. The intelligent testing method for frequency converter modules according to claim 1, characterized in that, The determination of the fourth pass rate of the inverter module based on the current test current curve, voltage curve, and current temperature curve includes: Obtain the test sequence; the test sequence consists of the instantaneous voltage, instantaneous current, and instantaneous temperature values ​​that change with time t during this test; The test sequence is compared one by one with the typical fault curves of each type in the pre-set fault mode library to determine the fault distance between the test sequence and each type of fault sequence. The fourth pass rate is determined based on the fault distance.

8. The intelligent testing method for frequency converter modules according to claim 3, characterized in that, The determination of the final test result of the inverter module based on the first pass rate, the second pass rate, the third pass rate, and the fourth pass rate includes: The final test result of the frequency converter module is determined according to the following formula. : in, The weighted score for the inverter module. The highest pass rate, The second pass rate, The third pass rate, It is the fourth qualified; , These are the weights corresponding to each pass rate.

9. The intelligent testing method for frequency converter modules according to claim 8, characterized in that, The score prediction model is continuously updated based on the final result corresponding to the one to be reviewed.

10. An intelligent testing system for frequency converter modules, characterized in that, include: Electrical parameter meter, used by users to measure the voltage, current, power and power factor of frequency converter modules; Voltage detection board, used to detect the DC bus voltage inside the inverter module; The variable frequency drive is configured to receive test commands from the edge intelligent processing layer, drive the load simulation unit to run according to preset parameters, realize frequency upsampling control from 0Hz to 75Hz, and provide feedback on the frequency upsampling execution status. The compressor is used in conjunction with the frequency converter to complete the full-condition test of startup, frequency increase, and stable operation, so as to output the fan speed; The lower-level machine is used to receive detection data via the RS485 communication interface; The host computer includes an acquisition unit and a determination unit; wherein, the acquisition unit is used to acquire the basic operating parameters of the frequency converter module; The determining unit is used to determine a first pass rate based on the status of the basic working parameters; The acquisition unit is also used to acquire the index characteristics of the frequency converter module; The determining unit is further configured to determine a second pass rate of the inverter module based on the index characteristics; The acquisition unit is also used to acquire the current voltage, current current and current power factor of the inverter module; The determining unit is also used to determine the third pass rate of the inverter module based on the current voltage, current current and current power factor; The acquisition unit is also used to acquire the current current curve, current voltage curve, and current temperature curve of the inverter module; The determining unit is also used to determine the fourth pass rate of the inverter module based on the current current curve, the current voltage curve, and the current temperature curve. The determining unit is further configured to determine the final test result of the frequency converter module based on the first pass rate, the second pass rate, the third pass rate, and the fourth pass rate.

Images