Impact prediction adaptive energy-saving method and device of frequency converter

Abstract

The application discloses an impact prediction self-adaptive energy-saving method and device of a frequency converter, and relates to the field of frequency converter energy saving.The method comprises the following steps: acquiring a current signal, judging whether a first impact event is over, acquiring an impact characteristic according to a current load type when it is judged that the first impact event is over, determining an impact intensity grade according to the impact characteristic, adding the impact characteristic to a preset historical event library, and predicting a predicted arrival time and a predicted intensity grade of a second impact event according to the preset historical event library; determining an adjustment percentage according to a predicted occurrence interval, and lowering an output voltage or a DC bus voltage by the adjustment percentage; and raising the output voltage or the DC bus voltage to a target voltage value corresponding to the predicted intensity grade at a preset advance time before the predicted arrival time. The technical scheme provided by the application can solve the problem of poor energy-saving effect of the frequency converter.

Description

Technical Field

[0001] This application relates to the field of frequency converter energy saving, specifically to a frequency converter impact prediction adaptive energy saving method and device. Background Technology

[0002] Variable frequency drives (VFDs) are core components of industrial drive systems, widely used in applications involving periodic impact loads such as stamping, crushing, and injection molding. In industrial production, motor drive systems are the primary energy consumers, making the energy-saving control of VFDs crucial for reducing electricity costs for businesses.

[0003] Existing frequency converters employ a responsive energy-saving strategy, which involves triggering a voltage reduction action only after detecting a light load. This approach suffers from response lag. When faced with periodic impact loads, this response lag makes it difficult for the frequency converter to effectively reduce and restore voltage within the brief impact interval. This manifests as delayed voltage recovery and insufficient torque to meet load demands when the impact occurs, resulting in poor energy-saving performance of the frequency converter.

[0004] Therefore, there is an urgent need for an adaptive energy-saving method and device for predicting the impact of frequency converters. Summary of the Invention

[0005] This application provides an adaptive energy-saving method and device for predicting impact in frequency converters, which solves the problem of poor energy-saving effect of frequency converters.

[0006] This application provides a first aspect of an adaptive energy-saving method for impact prediction in a frequency converter. The method includes: acquiring a current signal and determining whether a first impact event has ended based on the current signal; when the first impact event is determined to have ended, acquiring the impact characteristics of the first impact event based on the current load type, and determining the impact intensity level of the first impact event based on the impact characteristics, the impact characteristics including the impact occurrence time, current peak value, current rise slope, and impact duration; adding the impact characteristics to a preset historical event library, and predicting the predicted arrival time and predicted intensity level of a second impact event based on the updated preset historical event library, the second impact event being an adjacent event that occurs later than the first impact event and has not occurred at the current time; determining an adjustment percentage based on the predicted occurrence interval, and lowering the output voltage or DC bus voltage by the adjustment percentage to enter an energy-saving state, the predicted occurrence interval being the difference between the predicted arrival time and the current time, the adjustment percentage being positively correlated with the predicted occurrence interval; and, at a preset advance time before the predicted arrival time, raising the output voltage or DC bus voltage to a target voltage value corresponding to the predicted intensity level to provide torque matching the predicted intensity level when the second impact event occurs.

[0007] By adopting the above technical solution, the arrival time and intensity level of the next impact event are predicted, transforming the traditional passive mode of detecting light loads and then reducing voltage into an active prediction mode. This eliminates the response lag problem, enabling the frequency converter to safely achieve deep energy savings under periodic impact load conditions. The voltage reduction ratio is dynamically determined based on the predicted interval, automatically matching the voltage reduction depth with the interval length; the longer the interval, the deeper the voltage reduction, maximizing energy savings while ensuring safety. The voltage is boosted in advance to a target value matching the predicted intensity level, ensuring sufficient torque when the impact arrives, thus balancing deep energy savings with operational reliability.

[0008] Optionally, predicting the arrival time of the second impact event based on the updated preset historical event database includes: calculating the impact interval between two adjacent impact events based on the impact occurrence times of previous impact events in the preset historical event database, forming an impact interval sequence; calculating the statistical mean and variance of the impact interval sequence; using a recursive least squares algorithm with a forgetting factor, taking the statistical mean as the initial value of the recursive least squares algorithm state parameters and the variance as the initial covariance matrix, and recursively updating the impact interval sequence based on a constant mean state space model to obtain the current state parameter estimate and the current covariance matrix, and calculating the predicted impact interval of the second impact event based on the current state parameter estimate; adding the impact occurrence time to the predicted impact interval to obtain the predicted arrival time; calculating the confidence interval half-width of the predicted arrival time based on the current covariance matrix and the forgetting factor, and determining the confidence closed interval of the predicted arrival time as the range from the difference between the predicted arrival time and the confidence interval half-width to the sum of the predicted arrival time and the confidence interval half-width, the confidence closed interval being used to characterize the credible fluctuation range of the predicted arrival time.

[0009] By employing the above technical solution, a recursive least squares algorithm with a forgetting factor is used to progressively update the impact interval sequence. This allows the prediction model to adaptively track the slow drift of the impact interval, preventing the fixed-parameter strategy from failing over time and ensuring the long-term accuracy of the predicted arrival time. The output confidence interval for the predicted arrival time provides a quantitative and reliable fluctuation range for determining the depth of pressure reduction and the timing of early pressure increase, enabling energy-saving strategies to be safely implemented within the uncertainty boundary of the prediction.

[0010] Optionally, before determining the end of the first impact event, the method further includes: performing a windowed fast Fourier transform on the current signal to obtain the current spectrum; matching the current spectrum with multiple preset spectrum templates respectively, and calculating the corresponding Pearson correlation coefficients; identifying the load type represented by the preset spectrum template corresponding to the maximum value of the Pearson correlation coefficient as the current load type; the preset spectrum templates include the spectral characteristics of periodic impact loads.

[0011] By adopting the above technical solution, the current signal is subjected to windowed fast Fourier transform and matched with a preset spectrum template, and the current load type is automatically identified. This allows the frequency converter to adapt to the energy-saving needs of different load devices without manual configuration, thus improving the versatility and ease of deployment of the method.

[0012] Optionally, after adjusting the output voltage or DC bus voltage to the target voltage value corresponding to the predicted intensity level, the method further includes: calculating the reduction in input active power within a preset statistical period, and using the ratio between the reduction and a preset reference power as the energy saving rate; acquiring the actual torque of the second impact event through a sensorless torque observer, calculating the difference between the preset rated torque and the actual torque to obtain the torque margin; acquiring the actual speed of the second impact event, calculating the difference between the actual speed and the preset commanded speed to obtain the speed drop amplitude; and reducing the adjustment percentage by a first preset adjustment amount or increasing the preset advance time by a second preset adjustment amount when any of the following conditions are met: the energy saving rate is lower than a preset energy saving rate threshold; the torque margin is lower than a preset torque margin threshold; or the speed drop amplitude is greater than a preset drop amplitude threshold.

[0013] By adopting the above technical solution, the energy saving rate, torque margin, and speed drop amplitude are calculated in real time. When any indicator is abnormal, the voltage reduction ratio or preset advance time is automatically adjusted. This allows the energy-saving strategy to automatically correct itself when operating conditions change, avoiding energy-saving degradation or safety hazards caused by fixed parameters. The torque margin directly quantifies the torque safety redundancy during the impact period, and the speed drop amplitude provides intuitive feedback on the impact response effect from a motion control perspective. Multi-dimensional evaluation ensures the reliability of the adjustment strategy.

[0014] Optionally, after obtaining the torque margin, the method further includes: obtaining the current peak value of a preset number of impact events from a preset historical event library, and calculating the average value of the current peak value of the preset number of impact events; calculating the ratio of the average value to a preset reference value as the load health status index; when the load health status index is greater than the preset health index threshold, increasing the preset torque margin threshold by a preset ratio, and reducing the adjustment percentage by a third preset adjustment amount.

[0015] By adopting the above technical solution, the load health status index is calculated, and the long-term mechanical degradation trend of the equipment is incorporated into the energy-saving decision. When the health status deterioration is detected, the torque margin threshold is actively increased and the voltage reduction ratio is reduced, so that the system can automatically balance energy saving and safety throughout the entire life cycle of the equipment, and avoid stalling or damage caused by continuing to use the original aggressive energy-saving strategy due to equipment aging.

[0016] Optionally, the predicted intensity level of the second impact event is predicted based on the updated preset historical event database. Specifically, this includes: extracting the peak current, current rise slope, and impact duration of each impact event from the preset historical event database to form a multi-dimensional feature vector; inputting the multi-dimensional feature vector into a pre-trained autoregressive integral moving average model to obtain the predicted peak current of the second impact event; and determining the predicted intensity level of the second impact event based on the preset peak range into which the predicted peak current falls, with each preset peak range corresponding to a predicted intensity level.

[0017] By adopting the above technical solution and using the autoregressive integral moving average model to model the multidimensional characteristics of each impact event, the trend of impact intensity changes can be captured more accurately, so that the target voltage value of the advance voltage rise is precisely matched with the actual impact demand, avoiding energy waste caused by overcompensation or insufficient torque caused by undercompensation.

[0018] Optionally, determining the impact intensity level of the first impact event based on impact characteristics specifically includes: matching the current peak value with multiple preset peak value intervals to determine the preset peak value interval into which the current peak value falls, and determining a first score based on the preset peak value interval and score correspondence; matching the current rise slope with multiple preset slope intervals to determine the preset slope interval into which the current rise slope falls, and determining a second score based on the preset slope interval and score correspondence; matching the impact duration with multiple preset duration intervals to determine the preset duration interval into which the impact duration falls, and determining a third score based on the preset duration interval and score correspondence; weighting and summing the first score, second score, and third score according to preset weighting coefficients to obtain a comprehensive impact score; and determining the impact intensity level based on the preset level interval into which the comprehensive impact score falls, with each preset level interval corresponding to one impact intensity level.

[0019] By adopting the above technical solution, the impact characteristics of the three dimensions of current peak value, current rise slope and impact duration are weighted and comprehensively evaluated, so that the intensity level classification comprehensively reflects the severity of the impact event in terms of peak value, impact intensity and duration, providing a more accurate historical calibration data basis for subsequent prediction of the intensity level of the second impact event.

[0020] In a second aspect, this application provides an impact prediction adaptive energy-saving device for a frequency converter, the device including an acquisition unit and a processing unit;

[0021] The acquisition unit is used to acquire the current signal and determine whether the first impact event has ended based on the current signal.

[0022] The processing unit is configured to, upon determining the end of the first impact event, acquire the impact characteristics of the first impact event based on the current load type, determine the impact intensity level of the first impact event based on the impact characteristics, including the impact occurrence time, peak current, current rise slope, and impact duration; add the impact characteristics to a preset historical event database, and predict the predicted arrival time and predicted intensity level of a second impact event based on the updated preset historical event database, wherein the second impact event is an adjacent event that occurs later than the first impact event and has not yet occurred at the current time; determine an adjustment percentage based on the predicted occurrence interval, and lower the output voltage or DC bus voltage by the adjustment percentage to enter an energy-saving state, wherein the predicted occurrence interval is the difference between the predicted arrival time and the current time, and the adjustment percentage is positively correlated with the predicted occurrence interval; and increase the output voltage or DC bus voltage to a target voltage value corresponding to the predicted intensity level at a preset advance time before the predicted arrival time, so as to provide torque matching the predicted intensity level when the second impact event occurs.

[0023] In a third aspect, this application provides an electronic device including a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the first aspect or any possible implementation of the first aspect.

[0024] In a fourth aspect, this application provides a computer-readable storage medium storing a computer program, which is executed by a processor as described in the first aspect or any possible implementation thereof.

[0025] In summary, one or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. By predicting the arrival time and intensity of the next impact event, the system transforms the traditional passive mode of detecting light loads and then reducing voltage into an active prediction mode, eliminating response lag and enabling the inverter to safely achieve deep energy savings under periodic impact load conditions. The voltage reduction ratio is dynamically determined based on the predicted interval, automatically matching the voltage reduction depth to the interval length; the longer the interval, the deeper the voltage reduction, maximizing energy savings while ensuring safety. The voltage is boosted in advance to a target value matching the predicted intensity level, ensuring sufficient torque when the impact arrives, balancing deep energy savings with operational reliability.

[0026] 2. The current signal is subjected to windowed fast Fourier transform and matched with a preset spectrum template to automatically identify the current load type. This allows the frequency converter to adapt to the energy-saving needs of different load devices without manual configuration, improving the versatility and ease of deployment of the method.

[0027] 3. Real-time calculation of energy saving rate, torque margin, and speed drop magnitude; automatic adjustment of voltage reduction ratio or preset advance time when any indicator is abnormal, enabling the energy-saving strategy to automatically correct deviations when operating conditions change, avoiding energy-saving degradation or safety hazards caused by fixed parameters. Torque margin directly quantifies the torque safety redundancy during the impact period, and speed drop magnitude provides intuitive feedback on the impact response effect from a motion control perspective. Multi-dimensional evaluation ensures the reliability of the adjustment strategy. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating an adaptive energy-saving method for impact prediction of a frequency converter provided in an embodiment of this application.

[0029] Figure 2 This is a schematic diagram of the structure of an adaptive energy-saving device for impact prediction of a frequency converter provided in an embodiment of this application.

[0030] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0031] Explanation of reference numerals in the attached drawings: 201, acquisition unit; 202, processing unit; 300, electronic device; 301, processor; 302, communication bus; 303, user interface; 304, network interface; 305, memory. Detailed Implementation

[0032] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0033] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0034] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0035] Existing frequency converters employ a responsive energy-saving strategy, which involves triggering a voltage reduction action only after detecting a light load. This approach suffers from response lag. When faced with periodic impact loads, this lag makes it difficult for the frequency converter to effectively reduce and recover voltage within the brief impact interval. This manifests as delayed voltage recovery and insufficient torque to meet load demands upon the arrival of an impact, resulting in poor energy-saving performance. Therefore, this embodiment provides an impact prediction adaptive energy-saving method and device for frequency converters.

[0036] The impact prediction and adaptive energy-saving method for frequency converters provided in this application can be referenced. Figure 1 , Figure 1 This is a flowchart illustrating an adaptive energy-saving method for impact prediction in a frequency converter, provided in an embodiment of this application, and applied to a frequency converter. The method includes steps S101 to S105.

[0037] S101. Obtain the current signal and determine whether the first impact event has ended based on the current signal.

[0038] In the above steps, the frequency converter continuously collects the output current signal through its built-in current sensor during the operation of driving the periodic impact load. The occurrence and end of the impact event will produce obvious changes in the current signal: when the impact occurs, the current rapidly rises to a high level, and when the impact ends, the current quickly drops back to the no-load or light-load level. The frequency converter monitors the collected current signal in real time. When it detects that the current amplitude has rapidly dropped from a high level and stabilized at a low level, it determines that an impact event has ended. Taking a frequency converter driving a stamping press as an example, the peak current of the stamping press can reach 150% to 200% of the rated current during the stamping stroke. After the stamping is completed, the current drops back to 20% to 30% of the rated current within tens of milliseconds. The frequency converter can set the end judgment threshold to 35% of the rated current. When the current amplitude is detected to be lower than this threshold and the duration reaches the preset confirmation time, the impact event is determined to have ended. The confirmation time setting is used to eliminate false judgments caused by instantaneous current fluctuations. For the stamping press operation, the confirmation time can be set to 10 to 20 ms.

[0039] S102. When it is determined that the first impact event has ended, the impact characteristics of the first impact event are obtained according to the current load type, and the impact intensity level of the first impact event is determined according to the impact characteristics. The impact characteristics include the time of impact occurrence, peak current, current rise slope and impact duration.

[0040] In the above steps, when the first impact event is determined to have ended, the frequency converter extracts the impact characteristics of that event from the current signal. These characteristics include the moment of impact occurrence (the point at which the impact current begins to rise); the peak current (the maximum current value reached during the impact); the current rise slope (the rate at which the current rises from a baseline level to the peak); and the impact duration (the length of time the current remains at a high level). Depending on the current load type, the extraction parameters for these impact characteristics may vary. For example, a crusher's peak impact current may reach 180% to 250% of the rated current, with a steeper rise slope and a longer impact duration than a press. Therefore, the judgment threshold in the extraction parameters needs to be adjusted accordingly to avoid misidentifying normal load fluctuations as impact events. After extracting the impact characteristics, the frequency converter maps these characteristics to an impact intensity level to characterize the severity of the impact. A larger peak current, a steeper rise slope, and a longer duration correspond to a higher impact intensity level. The mapping process can be completed using a multi-dimensional weighted scoring method. In the specific implementation, the peak current can be divided into several intervals, each interval corresponding to a different score. For example, 1 point corresponds to 150% of the rated current, 2 points to 150% to 200%, and 3 points to 200% or more. The rising slope and duration adopt the same interval scoring mechanism. The specific values ​​can be set by the staff according to actual needs. Finally, the weighted sum is used to obtain the comprehensive score and determine the level.

[0041] In one possible implementation, before determining the end of the first impact event, the method further includes: performing a windowed fast Fourier transform on the current signal to obtain the current spectrum; matching the current spectrum with multiple preset spectrum templates respectively, and calculating the corresponding Pearson correlation coefficients; identifying the load type represented by the preset spectrum template corresponding to the maximum value of the Pearson correlation coefficient as the current load type; the preset spectrum template includes the spectral characteristics of periodic impact loads.

[0042] Specifically, before the system initially determines the end of the impact event, the frequency converter uses the acquired current signal to identify the load type. A windowed Fast Fourier Transform (FFT) is performed on the current signal, using either a Hanning or Hamming window to suppress spectral leakage and obtain the true current spectrum distribution. The frequency converter internally stores multiple preset spectrum templates for typical loads, covering common periodic impact load types such as presses, crushers, and injection molding machines. Each load type's spectrum template contains the characteristic spectral distribution of that load during steady-state operation. For example, the low-frequency band of the press spectrum shows a clear peak corresponding to the pressing frequency, while the crusher spectrum has abundant high-frequency harmonic components. The real-time calculated current spectrum is matched with each preset spectrum template, and the Pearson correlation coefficient is calculated as a similarity measure. The Pearson correlation coefficient ranges from -1 to 1; a value closer to 1 indicates a higher degree of matching, and vice versa. The load type represented by the template with the highest correlation coefficient is selected as the current load type. For example, if the correlation coefficient with the stamping press template is 0.92, the correlation coefficient with the crusher template is 0.65, and the correlation coefficient with the injection molding machine template is 0.48, then the current load is identified as a stamping press. After identification, the frequency converter can extract parameters based on the corresponding impact characteristics according to the load type, without manual intervention.

[0043] In one possible implementation, determining the impact intensity level of the first impact event based on impact characteristics specifically includes: matching the current peak value with multiple preset peak value intervals to determine the preset peak value interval into which the current peak value falls, and determining a first score based on the preset peak value interval and score correspondence; matching the current rise slope with multiple preset slope intervals to determine the preset slope interval into which the current rise slope falls, and determining a second score based on the preset slope interval and score correspondence; matching the impact duration with multiple preset duration intervals to determine the preset duration interval into which the impact duration falls, and determining a third score based on the preset duration interval and score correspondence; weighting and summing the first score, second score, and third score according to preset weighting coefficients to obtain a comprehensive impact score; and determining the impact intensity level based on the preset level interval into which the comprehensive impact score falls, with each preset level interval corresponding to one impact intensity level.

[0044] Specifically, when determining the impact intensity level of the first impact event, the frequency converter adopts a multi-dimensional feature weighted comprehensive scoring method, rather than relying on a simple threshold judgment based on a single indicator. The peak current is matched with multiple preset peak ranges to determine the range it falls into, and a first score is obtained based on the preset peak range and score correspondence. For example, below 150% of the rated current corresponds to 1 point, 150% to 200% corresponds to 2 points, 200% to 250% corresponds to 3 points, and above 250% corresponds to 4 points. The current rise rate is matched with multiple preset slope ranges to determine the range it falls into, obtaining a second score. For example, a rise rate per millisecond below 5% of the rated current corresponds to 1 point, 5% to 10% corresponds to 2 points, and so on. The impact duration is matched with multiple preset duration ranges to determine the range it falls into, obtaining a third score. For example, below 50ms corresponds to 1 point, 50ms to 100ms corresponds to 2 points, and above 100ms corresponds to 3 points. These three dimensions reflect the magnitude, severity, and duration of the impact, respectively. The comprehensive impact score is obtained by weighting and summing the three scores according to preset weighting coefficients. The weighting coefficients can be configured based on the importance of each dimension in the actual application scenario. Taking a stamping press as an example, the peak current and the rise slope better reflect the instantaneous state of the die impact; the peak current weight can be set to 0.4, the rise slope weight to 0.4, and the duration weight to 0.2. Taking a crusher as an example, the impact duration better reflects the material crushing process; the duration weight can be increased to 0.4, the peak current weight to 0.3, and the rise slope weight to 0.3. After the comprehensive impact score is calculated, the final impact intensity level is determined based on the preset level range it falls into; each preset level range corresponds to one intensity level. This multi-dimensional comprehensive scoring method provides a more accurate historical calibration data basis for predicting the intensity level of the next impact, and compared to the judgment method that relies solely on a single peak threshold, it can more comprehensively reflect the true severity of the impact event.

[0045] S103. Add the impact features to the preset historical event database, and predict the predicted arrival time and predicted intensity level of the second impact event based on the updated preset historical event database. The second impact event is an adjacent event that occurs later than the first impact event and has not occurred at the current time.

[0046] In the above steps, the frequency converter adds the extracted impact features to a preset historical event database. This database is a continuously maintained collection of data during inverter operation, storing impact feature records for each impact event, including the occurrence time, peak current, current rise rate, and duration of each impact. After each impact event, its feature data is appended to the database. To control storage space and computational complexity, the database can be maintained using a sliding window mechanism, retaining only records of the most recent impact events. The oldest record exceeding the window size is automatically removed. The window size can be configured according to the load type; for example, a press can have a window size of 50 events, while a crusher can have a window size of 30 events. Based on the updated preset historical event database, the frequency converter performs dual predictions for the next impact event. Regarding arrival time prediction, the frequency converter extracts the occurrence times of previous impact events from the database, calculates the time intervals between adjacent impact events to form an impact interval sequence, and extrapolates the possible arrival time of the next impact by statistically analyzing the patterns in this sequence. In terms of intensity level prediction, the inverter extracts the peak current, current rise slope, and impact duration of previous impact events from its database to form a multi-dimensional feature vector input prediction model, thereby obtaining the predicted intensity level of the next impact. Through this dual prediction, the inverter can know when the next impact will occur and how strong it will be after each impact, providing a basis for subsequent energy-saving actions.

[0047] In one possible implementation, predicting the arrival time of the second impact event based on an updated preset historical event database specifically includes: calculating the impact interval between two adjacent impact events based on the impact occurrence times of previous impact events in the preset historical event database, forming an impact interval sequence; calculating the statistical mean and variance of the impact interval sequence; using a recursive least squares algorithm with a forgetting factor, taking the statistical mean as the initial value of the recursive least squares algorithm state parameters and the variance as the initial covariance matrix, and recursively updating the impact interval sequence based on a constant mean state space model to obtain the current state parameter estimate and the current covariance matrix, and calculating the predicted impact interval of the second impact event based on the current state parameter estimate; adding the impact occurrence time to the predicted impact interval to obtain the predicted arrival time; calculating the confidence interval half-width of the predicted arrival time based on the current covariance matrix and the forgetting factor, and determining the confidence closed interval of the predicted arrival time as the range from the difference between the predicted arrival time and the confidence interval half-width to the sum of the predicted arrival time and the confidence interval half-width, the confidence closed interval being used to characterize the credible fluctuation range of the predicted arrival time.

[0048] Specifically, the frequency converter extracts the impact times of past impact events from a pre-set historical event database, calculates the time interval between each two adjacent impact events, and forms an impact interval sequence. The statistical mean and variance of this sequence are calculated as the initial input for subsequent recursive algorithms. A recursive least squares algorithm with a forgetting factor is used, with the statistical mean as the initial value of the state parameters and the variance as the initial covariance matrix. The impact interval sequence is recursively updated based on a constant-mean state-space model. The forgetting factor determines the rate at which the weights of historical data decay during the recursive process. A typical forgetting factor value is between 0.9 and 0.99, for example, 0.95. The closer the forgetting factor is to 1, the more historical data is retained, the less sensitive the model is to random fluctuations, but the slower the response to changes in operating conditions; the smaller the forgetting factor, the more sensitive the model is, but the more susceptible it is to noise interference. After each recursive update, the current state parameter estimate and the current covariance matrix are obtained. Based on the current state parameter estimate, the predicted impact interval for the next impact is calculated. The predicted arrival time is obtained by adding the time of the most recent shock event to the predicted shock interval. Simultaneously, the half-width of the confidence interval is calculated based on the current covariance matrix and the forgetting factor. A closed confidence interval is then determined with the predicted arrival time as the center and the half-width of the confidence interval as the radius. For example, if the predicted arrival time is 1.02 s after the current time, and the half-width of the confidence interval is 0.05 s, then the arrival time of the next shock is expected to be between 0.97 s and 1.07 s. This interval provides a quantitative measure of uncertainty for deciding the depth of depressurization and setting the timing of early pressurization. A wider interval indicates greater prediction uncertainty, allowing for a more conservative depressurization strategy.

[0049] In one possible implementation, the predicted intensity level of the second impact event is predicted based on an updated preset historical event database. Specifically, this includes: extracting the peak current, current rise slope, and impact duration of each impact event from the preset historical event database to form a multidimensional feature vector; inputting the multidimensional feature vector into a pre-trained autoregressive integral moving average model to obtain the predicted peak current of the second impact event; and determining the predicted intensity level of the second impact event based on the preset peak range into which the predicted peak current falls, with each preset peak range corresponding to a predicted intensity level.

[0050] Specifically, the frequency converter extracts the peak current, current rise slope, and impact duration of each impact event from a pre-set historical event database, forming a multi-dimensional feature vector. This vector comprehensively describes the intensity of each impact event. This feature vector is then input into a pre-trained autoregressive integral moving average model. Model training is completed before the frequency converter leaves the factory or during the initial commissioning phase, using impact data sequences of the load under typical operating conditions to fit the model parameters. After training, the model parameters are stored in the frequency converter, and only forward inference calculations are performed during online operation, resulting in low computational load. By analyzing the time series patterns of the peak current, the model captures the periodic fluctuations and slow changing trends of the impact intensity, outputting the predicted peak current for the next impact. Based on the pre-set peak range into which the predicted current peak falls, the corresponding predicted intensity level is determined. The division of peak ranges can be set according to the specific load type. For example, for a stamping press with a rated current of 100A, ranges below 150A can be set as range one, corresponding to level one; ranges from 150A to 200A as range two, corresponding to level two; and ranges above 200A as range three, corresponding to level three. The larger the predicted peak value, the higher the corresponding target voltage value, and the more sufficient the torque provided after the voltage is boosted in advance, thus achieving precise matching for impacts of different intensities.

[0051] S104. Determine the adjustment percentage based on the predicted occurrence interval, and lower the output voltage or DC bus voltage by the adjustment percentage to enter the energy-saving state. The predicted occurrence interval is the difference between the predicted arrival time and the current time. The adjustment percentage is positively correlated with the predicted occurrence interval.

[0052] In the above steps, after predicting the arrival time of the next impact, the frequency converter calculates the difference between that arrival time and the current time to obtain the predicted occurrence interval, which is the estimated time between the end of the current impact and the arrival of the next impact. Based on the length of the predicted occurrence interval, the frequency converter dynamically determines the voltage reduction ratio for this energy-saving phase. The longer the predicted occurrence interval, the larger the voltage reduction ratio; the shorter the predicted occurrence interval, the smaller the voltage reduction ratio. Taking a frequency converter driving a stamping press as an example, if the stamping cycle is stable at 60 times per minute and the impact interval is approximately 1 second, the voltage reduction ratio can be set to 20% to 30% of the rated voltage. If, due to process adjustments, the cycle speed increases to 80 times per minute and the impact interval shortens to 0.75 seconds, the voltage reduction ratio automatically decreases to 10% to 15% to ensure safe voltage recovery within a shorter time. This strategy of adaptively adjusting the voltage reduction depth based on the interval length balances energy saving and operational safety. The voltage reduction target can be the DC bus voltage or the output voltage. Inverters that can directly control the DC bus voltage reduce energy consumption at the source and reduce inverter switching losses; inverters that can only adjust the output voltage achieve a similar effect by reducing the modulation depth. The choice of which method to use depends on the inverter's hardware architecture.

[0053] S105. Before the predicted arrival time, the output voltage or DC bus voltage is increased to the target voltage value corresponding to the predicted intensity level, so as to provide torque matching the predicted intensity level when the second impact event occurs.

[0054] In the above steps, before the predicted arrival time, the inverter initiates a voltage boost operation a preset lead time, raising the voltage from the low voltage value in energy-saving mode to the target voltage value. The preset lead time setting needs to comprehensively consider the voltage rise rate and system response margin; a typical setting range is 5% to 15% of the impact interval. For example, when the impact interval is 1 second, the lead time can be set to 50ms to 150ms. The target voltage value corresponds to the predicted intensity level; the higher the predicted intensity level, the larger the target voltage value. Taking three intensity levels as an example, the target voltage value corresponding to level one can be 90% of the rated voltage, level two is 95%, and level three is 100%. This graded matching mechanism ensures that the inverter does not waste energy when dealing with smaller impacts and provides sufficient torque when dealing with larger impacts. The purpose of the advance voltage boost is to ensure that the voltage has recovered to a sufficient level when the next impact actually arrives, ensuring sufficient torque and avoiding speed drops or stalling due to untimely voltage recovery. After the actual impact arrives, the inverter continuously monitors the current signal. When it detects a current drop and meets the termination criteria, it re-enters the next cycle.

[0055] In one possible implementation, after adjusting the output voltage or DC bus voltage to the target voltage value corresponding to the predicted intensity level, the method further includes: calculating the reduction in input active power within a preset statistical period, and using the ratio between the reduction and a preset reference power as the energy saving rate; acquiring the actual torque of the second impact event through a sensorless torque observer, calculating the difference between the preset rated torque and the actual torque to obtain the torque margin; acquiring the actual rotational speed of the second impact event, calculating the difference between the actual rotational speed and the preset commanded rotational speed to obtain the speed drop amplitude; and reducing the adjustment percentage by a first preset adjustment amount or increasing the preset advance time by a second preset adjustment amount when any of the following conditions are met: the energy saving rate is lower than a preset energy saving rate threshold; the torque margin is lower than a preset torque margin threshold; or the speed drop amplitude is greater than a preset drop amplitude threshold.

[0056] Specifically, after each pre-voltage boost, subsequent actual occurrence, and completion of the next impact, the inverter evaluates the effectiveness of this energy-saving cycle. The evaluation is based on three indicators. Regarding energy saving rate, the reduction in input active power within a preset statistical period is calculated, and the ratio of this reduction to a preset baseline power is used as the energy saving rate. The preset baseline power is the input active power under the same operating conditions without the energy-saving strategy enabled, and can be calibrated at the initial stage of equipment commissioning. For example, if the baseline power is 10kW, and the average input power during the statistical period after enabling energy saving is 7.5kW, the reduction is 2.5kW, then the energy saving rate is 25%. The preset energy saving rate threshold can be set according to the application scenario, such as 20%. Regarding torque margin, the actual torque value during the impact period is obtained through a sensorless torque observer. The sensorless torque observer estimates the torque in real time based on the motor's voltage, current, and speed signals using a motor mathematical model. Using the rated torque as a benchmark, the difference between the rated torque and the actual torque is calculated as the torque margin. For example, if the rated torque is 100 N·m and the maximum actual torque observed during the impact period is 85 N·m, then the torque margin is 15 N·m. The preset torque margin threshold can be set to 10% to 15% of the rated torque. Regarding the speed drop, the actual speed value during the impact period is obtained, and the difference between the actual speed and the commanded speed is calculated. The commanded speed is the target speed value set in the inverter control system. For example, if the commanded speed is 1500 rpm and the actual speed drops to a minimum of 1450 rpm during the impact period, then the speed drop is 50 rpm. The preset drop threshold can be set according to process requirements, such as 30 rpm. When the energy saving rate is lower than the preset energy saving rate threshold, it indicates that the voltage reduction is too conservative, and the voltage reduction ratio is reduced by the first preset adjustment amount, for example, from the current 20% to 18%. When the torque margin is lower than the preset torque margin threshold, it indicates that the voltage reduction is too aggressive, and the voltage reduction ratio is reduced by the first preset adjustment amount. When the engine speed drops more than the preset drop threshold, it indicates that the boost timing is too late. The preset advance time will be increased by a second preset adjustment amount, for example, from the current 100ms to 120ms. The three evaluations can be triggered independently or simultaneously, with each adjustment amount independent.

[0057] In one possible implementation, after obtaining the torque margin, the method further includes: obtaining the current peak value of a preset number of impact events from a preset historical event library, and calculating the average value of the current peak value of the preset number of impact events; calculating the ratio of the average value to a preset reference value as a load health status index; when the load health status index is greater than a preset health index threshold, increasing the preset torque margin threshold by a preset percentage, and reducing the adjustment percentage by a third preset adjustment amount.

[0058] Specifically, during long-term operation, the mechanical components of the load may experience gradual changes in impact characteristics due to wear or aging. For example, after the die of a stamping press wears down, the peak impact current at the moment of stamping may gradually increase; after the bearing of a crusher wears down, the impact duration may gradually lengthen. The frequency converter selects the most recent preset number of impact events from a preset historical event library, such as the last 50 events, extracts the peak current of each event, and calculates the average value. This average value is compared with a preset baseline value to obtain the load health status index. The preset baseline value is the baseline value calibrated under normal operating conditions during the initial operation of the equipment, such as the average peak current of the first 100 impacts before the equipment was put into operation. The load health status index can be expressed as the ratio of the average peak current of the most recent impacts to the baseline value. For example, if the baseline value is 150% of the rated current and the average value of the last 50 impacts is 165%, then the health status index is 1.1. When the health status index exceeds the preset health index threshold, such as 1.15, it indicates that the mechanical condition of the equipment is deteriorating, and the risk of insufficient torque under the same voltage reduction depth increases. At this point, the safety threshold of the torque margin is increased, for example, from 10% to 15% of the rated torque, and the voltage reduction ratio is reduced by the third preset adjustment amount, for example, from 20% to 15%. This makes the energy-saving strategy automatically more conservative, prioritizing the safe operation of the equipment and avoiding stalling or damage caused by continuing the original aggressive energy-saving strategy due to equipment aging.

[0059] This application also provides an impact prediction adaptive energy-saving device for frequency converters, referring to... Figure 2 The device is a frequency converter, which includes an acquisition unit 201 and a processing unit 202.

[0060] The acquisition unit 201 is used to acquire the current signal and determine whether the first impact event has ended based on the current signal.

[0061] Processing unit 202 is configured to, upon determining the end of the first impact event, acquire the impact characteristics of the first impact event based on the current load type, determine the impact intensity level of the first impact event based on the impact characteristics, including the impact occurrence time, peak current, current rise slope, and impact duration; add the impact characteristics to a preset historical event library, and predict the predicted arrival time and predicted intensity level of a second impact event based on the updated preset historical event library, wherein the second impact event is an adjacent event that occurs later than the first impact event and has not occurred at the current time; determine an adjustment percentage based on the predicted occurrence interval, and lower the output voltage or DC bus voltage by the adjustment percentage to enter an energy-saving state, wherein the predicted occurrence interval is the difference between the predicted arrival time and the current time, and the adjustment percentage is positively correlated with the predicted occurrence interval; and increase the output voltage or DC bus voltage to a target voltage value corresponding to the predicted intensity level at a preset advance time before the predicted arrival time, so as to provide torque matching the predicted intensity level when the second impact event occurs.

[0062] In one possible implementation, the processing unit 202 is used to calculate the impact interval between two adjacent impact events based on the impact occurrence times of previous impact events in a preset historical event database, forming an impact interval sequence; calculate the statistical mean and variance of the impact interval sequence; use a recursive least squares algorithm with a forgetting factor, taking the statistical mean as the initial value of the recursive least squares algorithm state parameter and the variance as the initial covariance matrix, and recursively update the impact interval sequence based on a constant mean state space model to obtain the current state parameter estimate and the current covariance matrix, and calculate the predicted impact interval of the second impact event based on the current state parameter estimate; add the impact occurrence time to the predicted impact interval to obtain the predicted arrival time; calculate the confidence interval half-width of the predicted arrival time based on the current covariance matrix and the forgetting factor, and determine the confidence closed interval of the predicted arrival time as the range from the difference between the predicted arrival time and the confidence interval half-width to the sum of the predicted arrival time and the confidence interval half-width, which is used to characterize the reliable fluctuation range of the predicted arrival time.

[0063] In one possible implementation, the processing unit 202 is used to perform a windowed fast Fourier transform on the current signal to obtain the current spectrum; match the current spectrum with multiple preset spectrum templates respectively, and calculate the corresponding Pearson correlation coefficients; identify the load type represented by the preset spectrum template corresponding to the maximum value of the Pearson correlation coefficient as the current load type; the preset spectrum templates include the spectral characteristics of periodic impact loads.

[0064] In one possible implementation, the processing unit 202 is used to calculate the reduction in input active power within a preset statistical period, and use the ratio between the reduction and a preset reference power as the energy saving rate; when any of the following conditions are met, the adjustment percentage is reduced by a first preset adjustment amount, or the preset advance time is increased by a second preset adjustment amount: the energy saving rate is lower than a preset energy saving rate threshold; the torque margin is lower than a preset torque margin threshold; the speed drop is greater than a preset drop magnitude threshold; the acquisition unit 201 is used to acquire the actual torque of the second impact event through a sensorless torque observer, calculate the difference between the preset rated torque and the actual torque to obtain the torque margin; acquire the actual speed of the second impact event, calculate the difference between the actual speed and the preset commanded speed to obtain the speed drop magnitude.

[0065] In one possible implementation, the acquisition unit 201 is used to acquire the current peak value of a preset number of impact events from a preset historical event library and calculate the average value of the current peak value of the preset number of impact events; the processing unit 202 is used to calculate the ratio of the average value to a preset reference value as the load health status index; when the load health status index is greater than the preset health index threshold, the preset torque margin threshold is increased by a preset ratio and the adjustment percentage is reduced by a third preset adjustment amount.

[0066] In one possible implementation, the acquisition unit 201 is used to extract the current peak value, current rise slope and impact duration of each impact event from a preset historical event database to form a multi-dimensional feature vector; the processing unit 202 is used to input the multi-dimensional feature vector into a pre-trained autoregressive integral moving average model to obtain the predicted current peak value of the second impact event; and to determine the predicted intensity level of the second impact event based on the preset peak range into which the predicted current peak value falls, with each preset peak range corresponding to a predicted intensity level.

[0067] In one possible implementation, the processing unit 202 is configured to match the current peak value with multiple preset peak value intervals to determine the preset peak value interval into which the current peak value falls, and determine a first score according to the preset correspondence between peak value intervals and scores; match the current rise slope with multiple preset slope intervals to determine the preset slope interval into which the current rise slope falls, and determine a second score according to the preset correspondence between slope intervals and scores; match the impact duration with multiple preset duration intervals to determine the preset duration interval into which the impact duration falls, and determine a third score according to the preset duration interval and score correspondence; perform a weighted summation of the first score, second score, and third score according to preset weighting coefficients to obtain a comprehensive impact score; and determine the impact intensity level according to the preset level interval into which the comprehensive impact score falls, with each preset level interval corresponding to an impact intensity level.

[0068] It should be noted that the above embodiments of the apparatus are only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0069] This application also provides an electronic device. (See reference...) Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device 300 may include: at least one processor 301, at least one communication bus 302, at least one user interface 303, a network interface 304, and a memory 305.

[0070] The communication bus 302 is used to enable communication between these components.

[0071] The user interface 303 may include a display screen and a camera. Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0072] The network interface 304 may include standard wired interfaces and wireless interfaces (such as Wi-Fi interfaces).

[0073] The processor 301 may include one or more processing cores. The processor 301 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 305, and by calling data stored in memory 305. Optionally, the processor 301 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 301 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 301 and may be implemented as a separate chip.

[0074] The memory 305 may include random access memory (RAM) or read-only memory. Optionally, the memory 305 may include a non-transitory computer-readable storage medium. The memory 305 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. The memory 305 may also be at least one storage device located remotely from the aforementioned processor 301. (Refer to...) Figure 3 The memory 305, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for an inverter's impulse prediction adaptive energy-saving method.

[0075] exist Figure 3In the illustrated electronic device 300, the user interface 303 is mainly used to provide an input interface for the user and to acquire user input data; while the processor 301 can be used to call an application program of an inverter's impulse prediction adaptive energy-saving method stored in the memory 305. When executed by one or more processors 301, the electronic device 300 performs one or more methods as described in the above embodiments. It should be noted that, for the foregoing method embodiments, for the sake of simplicity, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0076] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0077] This application also provides a computer-readable storage medium storing instructions. When executed by one or more processors, these instructions cause an electronic device to perform one or more of the methods described in the above embodiments.

[0078] In the various embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between apparatuses or units may be electrical or other forms.

[0079] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0080] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0081] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0082] The above description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure.

[0083] This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are to be considered exemplary only, and the scope of this disclosure is defined by the claims.

Claims

1. An adaptive energy-saving method for impact prediction in frequency converters, characterized in that, The method includes: Acquire a current signal and determine whether the first impact event has ended based on the current signal; When the first impact event is determined to have ended, the impact characteristics of the first impact event are obtained according to the current load type, and the impact intensity level of the first impact event is determined according to the impact characteristics. The impact characteristics include the impact occurrence time, current peak value, current rise slope, and impact duration. The impact features are added to a preset historical event database, and the predicted arrival time and predicted intensity level of the second impact event are predicted based on the updated preset historical event database. The second impact event is an adjacent event that occurs later than the first impact event and has not occurred at the current time. The adjustment percentage is determined based on the predicted occurrence interval, and the output voltage or DC bus voltage is reduced by the adjustment percentage to enter the energy-saving state. The predicted occurrence interval is the difference between the predicted arrival time and the current time, and the adjustment percentage is positively correlated with the predicted occurrence interval. At a preset lead time before the predicted arrival time, the output voltage or the DC bus voltage is increased to a target voltage value corresponding to the predicted intensity level, so as to provide torque matching the predicted intensity level when the second impact event occurs.

2. The method according to claim 1, characterized in that, The step of predicting the arrival time of the second impact event based on the updated preset historical event database specifically includes: Based on the impact times of each impact event in the preset historical event database, the impact interval between two adjacent impact events is calculated to form an impact interval sequence. Calculate the statistical mean and variance of the impact interval sequence; The recursive least squares algorithm with a forgetting factor is adopted. The statistical mean is used as the initial value of the state parameter of the recursive least squares algorithm, and the variance is used as the initial covariance matrix. The impact interval sequence is updated successively based on the constant mean state space model to obtain the current state parameter estimate and the current covariance matrix. The predicted impact interval of the second impact event is calculated based on the current state parameter estimate. The predicted arrival time is obtained by adding the time of the impact occurrence to the predicted impact interval. Based on the current covariance matrix and the forgetting factor, the confidence interval half-width of the predicted arrival time is calculated, and the confidence closed interval of the predicted arrival time is determined as the sum of the predicted arrival time and the confidence interval half-width, from the difference between the predicted arrival time and the confidence interval half-width. The confidence closed interval is used to characterize the credible fluctuation range of the predicted arrival time.

3. The method according to claim 1, characterized in that, Before determining the end of the first impact event, the method further includes: The current signal is subjected to a windowed Fast Fourier Transform to obtain the current spectrum; The current spectrum is matched with multiple preset spectrum templates, and the corresponding Pearson correlation coefficient is calculated. The load type represented by the preset spectrum template corresponding to the maximum value of the Pearson correlation coefficient is identified as the current load type. The preset spectrum template includes the spectral characteristics of periodic impact loads.

4. The method according to claim 1, characterized in that, After adjusting the output voltage or the DC bus voltage to a target voltage value corresponding to the predicted intensity level, the method further includes: Calculate the reduction in input active power within a preset statistical period, and use the ratio between the reduction and a preset baseline power as the energy saving rate; The actual torque of the second impact event is obtained by acquiring it through a sensorless torque observer, and the difference between the preset rated torque and the actual torque is calculated to obtain the torque margin. Obtain the actual rotational speed of the second impact event, calculate the difference between the actual rotational speed and the preset commanded rotational speed, and obtain the rotational speed drop amplitude; When any of the following conditions are met, the adjustment percentage will be reduced by a first preset adjustment amount, or the preset advance time will be increased by a second preset adjustment amount: The energy saving rate is lower than the preset energy saving rate threshold; The torque margin is lower than a preset torque margin threshold; The speed drop is greater than a preset drop threshold.

5. The method according to claim 4, characterized in that, After obtaining the torque margin, the method further includes: Obtain the peak current of a preset number of impact events from the preset historical event database, and calculate the average value of the peak current of the preset number of impact events; The ratio of the average value to the preset benchmark value is calculated as the load health status index; When the load health status index is greater than the preset health index threshold, the preset torque margin threshold is increased by a preset ratio, and the adjustment percentage is reduced by a third preset adjustment amount.

6. The method according to claim 1, characterized in that, The step of predicting the intensity level of the second impact event based on the updated preset historical event database specifically includes: Extract the peak current, current rise slope, and impact duration of each impact event from the preset historical event database to form a multi-dimensional feature vector; The multidimensional feature vector is input into a pre-trained autoregressive integral moving average model to obtain the predicted peak current of the second impact event. Based on the preset peak range into which the predicted current peak falls, the predicted intensity level of the second impact event is determined, with each preset peak range corresponding to a predicted intensity level.

7. The method according to claim 1, characterized in that, Determining the impact intensity level of the first impact event based on the impact characteristics specifically includes: The current peak value is matched with multiple preset peak ranges to determine the preset peak range in which the current peak value falls, and the first score is determined according to the preset peak range and score correspondence relationship. The current rise slope is matched with multiple preset slope intervals to determine the preset slope interval into which the current rise slope falls, and a second score is determined according to the preset slope interval and score correspondence. The impact duration is matched with multiple preset duration intervals to determine the preset duration interval into which the impact duration falls, and a third score is determined based on the preset duration interval and score correspondence. The first score, the second score, and the third score are weighted and summed according to preset weighting coefficients to obtain the comprehensive impact score. The impact intensity level is determined based on the preset level range into which the comprehensive impact score falls, with each preset level range corresponding to an impact intensity level.

8. An impulse prediction adaptive energy-saving device for a frequency converter, characterized in that, The device includes an acquisition unit (201) and a processing unit (202): The acquisition unit (201) is used to acquire a current signal and determine whether the first impact event has ended based on the current signal. The processing unit (202) is used to obtain the impact characteristics of the first impact event according to the current load type when it is determined that the first impact event has ended, and to determine the impact intensity level of the first impact event according to the impact characteristics. The impact characteristics include the impact occurrence time, current peak value, current rise slope and impact duration. The processing unit (202) is further configured to add the impact features to a preset historical event database, and predict the predicted arrival time and predicted intensity level of the second impact event based on the updated preset historical event database, wherein the second impact event is an adjacent event that occurs later than the first impact event and has not occurred at the current time; The processing unit (202) is further configured to determine the adjustment percentage based on the predicted occurrence interval, and reduce the output voltage or DC bus voltage by the adjustment percentage to enter the energy-saving state. The predicted occurrence interval is the difference between the predicted arrival time and the current time, and the adjustment percentage is positively correlated with the predicted occurrence interval. The processing unit (202) is further configured to increase the output voltage or the DC bus voltage to a target voltage value corresponding to the predicted intensity level by a preset advance time before the predicted arrival time, so as to provide torque matching the predicted intensity level when the second impact event occurs.

9. An electronic device, characterized in that, The device includes a processor (301), a memory (305), a user interface (303), and a network interface (304). The memory (305) is used to store instructions. The user interface (303) and the network interface (304) are used to communicate with other devices. The processor (301) is used to execute the instructions stored in the memory (305) to cause the electronic device (300) to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7 above.

Images