A power equipment fault early warning method and system based on multi-dimensional parameter perception

Abstract

The application provides a power equipment fault early warning method and system based on multi-dimensional parameter perception. A multi-dimensional state feature vector for representing a current operation state of power equipment is constructed through multi-dimensional operation parameters of the power equipment. A normal fluctuation interval of each operation parameter at a current time is generated based on the multi-dimensional state feature vector. A first deviation degree of actual multi-dimensional operation parameters at the current time relative to the normal fluctuation interval and a second deviation degree of a multi-dimensional feature distribution of the actual multi-dimensional operation parameters at the current time relative to a health state feature distribution of the power equipment are calculated to perform prediction analysis on the health state of the power equipment, so as to obtain a comprehensive health index of the power equipment. When the comprehensive health index is lower than a preset early warning threshold, a fault early warning signal is sent. The scheme of the application can effectively fuse multi-source parameter correlation and realize power equipment fault early warning with self-adaptive working condition.

Description

Technical Field

[0001] This application relates to the field of fault early warning technology, and more specifically, to a method and system for early warning of power equipment faults based on multi-dimensional parameter perception. Background Technology

[0002] Fault early warning is a predictive maintenance method based on data-driven and intelligent analysis. By integrating real-time monitoring of IoT sensors, historical operating patterns, and machine learning algorithms, it constructs a degradation trajectory model of equipment from health to failure.

[0003] As a core component of the power system, the accurate perception and health assessment of the operating status of power equipment are crucial for ensuring the safe and stable operation of the system. Currently, power equipment condition monitoring mainly relies on sensor data acquisition and threshold alarm mechanisms, using fixed upper and lower limits for operating parameters such as voltage, current, and temperature to identify anomalies. However, these methods typically employ a single-parameter independent discrimination analysis mode, neglecting the inherent coupling and correlation between different operating parameters. They also fail to fully consider the dynamic impact of external operating condition fluctuations on the reasonable range of operating parameter values, leading to normal operating condition fluctuations being easily misjudged as abnormal states, while early performance degradation signals characterized by multi-parameter coordinated deviations are difficult to effectively identify. Furthermore, existing methods generally lack the ability to assess the overall distribution of multi-dimensional state characteristics of equipment, failing to quantitatively compare the current operating state with the historical health state characteristics, making it difficult to capture the slow deterioration process of equipment transitioning from health to failure in a timely manner. Therefore, how to effectively integrate the correlation relationships of multi-source parameters and achieve adaptive early warning for power equipment faults has become a challenge for the industry. Summary of the Invention

[0004] This application provides a power equipment fault early warning method and system based on multi-dimensional parameter perception, which can effectively integrate the correlation of multi-source parameters and realize power equipment fault early warning with adaptive operating conditions.

[0005] In a first aspect, this application provides a method for early warning of power equipment faults based on multi-dimensional parameter sensing, comprising the following steps: Obtain multi-dimensional operating parameters of power equipment; The multidimensional operating parameters are preprocessed to extract the variation features of each parameter and construct a multidimensional state feature vector to characterize the current operating state of the power equipment. Based on the multidimensional state feature vector and the historical operating condition parameter sequence representing external operating conditions in the historical operating data of power equipment, the normal fluctuation range of each operating parameter at the current moment is generated. Calculate the first degree of deviation of the actual multidimensional operating parameters at the current moment from the normal fluctuation range, and the second degree of deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment from the health status feature distribution of the power equipment; The health status of the power equipment is predicted and analyzed based on the first deviation degree and the second deviation degree to obtain a comprehensive health index of the power equipment. When the comprehensive health index is lower than the preset warning threshold, a fault warning signal is issued.

[0006] In some embodiments, the multidimensional operating parameters include electrical parameters such as voltage, current, active power, reactive power, and power factor of the power equipment, as well as non-electrical parameters such as equipment surface temperature, ambient temperature, vibration amplitude, and partial discharge.

[0007] In some embodiments, preprocessing the multidimensional operating parameters, extracting the variation features of each parameter, and constructing a multidimensional state feature vector to characterize the current operating state of the power equipment specifically includes: The multidimensional operating parameters are preprocessed to obtain preprocessed multidimensional operating parameters; The statistical features and spectral features of each parameter are extracted from the preprocessed multidimensional operating parameters, and the statistical features and spectral features are fused to obtain the variation features of each parameter. The variation characteristics of each parameter are concatenated according to the preset feature dimensions to construct a multi-dimensional state feature vector that represents the current operating state of the power equipment.

[0008] In some embodiments, generating the normal fluctuation range of each operating parameter at the current moment based on the multidimensional state feature vector and the historical operating condition parameter sequence characterizing external operating conditions in the historical operating data of power equipment specifically includes: Extract a sequence of historical operating condition parameters from the historical operating data of power equipment that matches the current external operating conditions. The external operating conditions include equipment load rate, ambient temperature, and operating time. Statistical analysis is performed on each operating parameter corresponding to the historical operating condition parameter sequence to calculate the mean and standard deviation of each operating parameter under the same operating conditions. Based on the mean and standard deviation of each operating parameter, the upper and lower limits of normal fluctuation of each operating parameter under external operating conditions are determined according to the preset confidence level, so as to form the normal fluctuation range of each operating parameter at the current moment.

[0009] In some embodiments, calculating the first deviation of the actual multidimensional operating parameters at the current moment relative to the normal fluctuation range specifically includes: Each actual operating parameter in the current multidimensional operating parameters is compared with the corresponding normal fluctuation range to determine whether each actual operating parameter falls within the normal fluctuation range. For operating parameters that fall within the normal fluctuation range, their deviation is assigned a value of zero. For operating parameters that exceed the normal fluctuation range, calculate the ratio of the excess amount to the width of the normal fluctuation range, and use it as the individual deviation of the operating parameter. The individual deviations of all operating parameters are weighted and aggregated to obtain the first deviation of the actual multidimensional operating parameters at the current moment relative to the normal fluctuation range.

[0010] In some embodiments, calculating the second deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment from the health state feature distribution of the power equipment specifically includes: Obtain the historical multidimensional state feature vector of power equipment within a preset time period under healthy operating conditions, and construct the health state feature distribution; The health status feature distribution is modeled to obtain a health status feature distribution model; The multidimensional state feature vector of the actual multidimensional operating parameters at the current moment is input into the health state feature distribution model, and the probability density value of the current multidimensional state feature vector under the health state feature distribution model is calculated. The second degree of deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment from the health state feature distribution is determined based on the probability density value.

[0011] In some embodiments, predicting and analyzing the health status of power equipment based on the first degree of deviation and the second degree of deviation to obtain a comprehensive health index of the power equipment specifically includes: Determine the first weighting coefficient corresponding to the first degree of deviation and the second weighting coefficient corresponding to the second degree of deviation; The health status of power equipment is predicted based on the first degree of deviation and the first weighting coefficient, and the second degree of deviation and the second weighting coefficient, to obtain a comprehensive health index of the power equipment.

[0012] Secondly, this application provides a power equipment fault early warning system based on multi-dimensional parameter sensing, comprising: The acquisition module is used to acquire multi-dimensional operating parameters of power equipment; The processing module is used to preprocess the multidimensional operating parameters, extract the variation features of each parameter, and construct a multidimensional state feature vector to characterize the current operating state of the power equipment. The processing module is also used to generate the normal fluctuation range of each operating parameter at the current moment based on the multidimensional state feature vector and the historical operating condition parameter sequence characterizing the external operating conditions in the historical operating data of the power equipment. The processing module is also used to calculate the first degree of deviation of the actual multidimensional operating parameters at the current moment relative to the normal fluctuation range, and the second degree of deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment relative to the health status feature distribution of the power equipment. The execution module is used to predict and analyze the health status of the power equipment based on the first deviation degree and the second deviation degree to obtain a comprehensive health index of the power equipment. When the comprehensive health index is lower than a preset warning threshold, a fault warning signal is issued.

[0013] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described method for early warning of power equipment faults based on multi-dimensional parameter perception.

[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for early warning of power equipment faults based on multi-dimensional parameter perception.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The power equipment fault early warning method and system based on multi-dimensional parameter perception provided in this application firstly acquires multi-dimensional operating parameters of the power equipment, preprocesses these parameters, extracts the variation characteristics of each parameter, and constructs a multi-dimensional state feature vector to characterize the current operating state of the power equipment. This unifies and standardizes multi-source heterogeneous parameter data, mines the temporal variation patterns of single parameters, and transforms scattered and independent multi-source parameters into an integrated correlated feature vector, intuitively reflecting the intrinsic correlation between multiple parameters and achieving deep fusion expression of multi-source parameters to support quantitative judgment of differentiated operating conditions. Secondly, based on the multi-dimensional state feature vector and historical operating condition parameter sequences representing external operating conditions in the historical operating data of the power equipment, the normal fluctuation range of each operating parameter at the current moment is generated. Relying on the correspondence between historical operating conditions and equipment states, a normal range of parameters adapted to the current external environment is dynamically generated, abandoning fixed threshold standards and achieving automatic adjustment of the judgment benchmark according to operating conditions, thus realizing the underlying logic of adaptive early warning. Then, the first deviation of the current multidimensional operating parameters from the normal fluctuation range is calculated, as well as the second deviation of the multidimensional feature distribution of the current multidimensional operating parameters from the health status feature distribution of the power equipment. The degree of equipment anomaly is quantified from two dimensions: single-parameter operating condition boundary deviation and multi-parameter overall distribution matching degree. This approach considers both the fluctuation deviation of single parameters affected by operating conditions and fully utilizes the multi-source parameter correlation carried by the multidimensional feature vector, quantifying the degree of equipment deviation from its healthy state at multiple levels. Finally, the health status of the power equipment is predicted and analyzed based on the first and second deviations to obtain a comprehensive health index. When the comprehensive health index is lower than a preset warning threshold, a fault warning signal is issued. The single-operating condition adaptation deviation and the multi-parameter overall correlation distribution deviation are integrated to output a unified quantitative health index. The equipment operating risk is comprehensively judged based on the dual deviation dimensions. This approach achieves adaptive judgment based on dynamic operating condition ranges and fully mines the multi-source parameter coupled fault characteristics through multi-dimensional feature fusion, accurately triggering fault warnings. Using the scheme of this application, power equipment fault warnings can be effectively fused with multi-source parameter correlations and achieve adaptive warnings based on operating conditions. Attached Figure Description

[0016] Figure 1 This is an exemplary flowchart of a power equipment fault early warning method based on multi-dimensional parameter perception, according to some embodiments of this application; Figure 2 This is an exemplary flowchart illustrating the construction of a multidimensional state feature vector according to some embodiments of this application; Figure 3 This is a schematic diagram of the working scenario of a power equipment fault early warning method based on multi-dimensional parameter perception, according to some embodiments of this application; Figure 4This is a schematic diagram of the structure of a power equipment fault early warning system based on multi-dimensional parameter sensing, according to some embodiments of this application; Figure 5 This is a schematic diagram of the structure of a computer device that implements a power equipment fault early warning method based on multi-dimensional parameter perception, according to some embodiments of this application. Detailed Implementation

[0017] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] refer to Figure 1 The figure is an exemplary flowchart of a power equipment fault early warning method based on multi-dimensional parameter perception, according to some embodiments of this application. The power equipment fault early warning method based on multi-dimensional parameter perception mainly includes the following steps: In step 101, the multi-dimensional operating parameters of the power equipment are obtained.

[0019] It should be noted that the multidimensional operating parameters in this application refer to a set of multiple types of physical quantities that comprehensively characterize the operating status of power equipment. They reflect the real-time health level and dynamic change trend of power equipment under multiple dimensions such as electrical, thermal, mechanical and environmental factors. The multidimensional operating parameters include electrical quantity parameters such as voltage, current, active power, reactive power and power factor of power equipment, as well as non-electrical quantity parameters such as equipment surface temperature, ambient temperature, vibration amplitude and partial discharge.

[0020] In step 102, the multidimensional operating parameters are preprocessed to extract the variation features of each parameter and construct a multidimensional state feature vector to characterize the current operating state of the power equipment.

[0021] In some embodiments, reference Figure 2 The diagram is an exemplary flowchart of constructing a multidimensional state feature vector in some embodiments of this application. In this embodiment, the multidimensional operating parameters are preprocessed, and the change features of each parameter are extracted. The construction of a multidimensional state feature vector to characterize the current operating state of the power equipment can be achieved by the following steps: In step 1021, the multidimensional operating parameters are preprocessed to obtain preprocessed multidimensional operating parameters; In step 1022, statistical features and spectral features of each parameter are extracted from the preprocessed multidimensional operating parameters, and the statistical features and spectral features are fused to obtain the variation features of each parameter. In step 1023, the change features of each parameter are spliced ​​together according to the preset feature dimensions to construct a multi-dimensional state feature vector that characterizes the current operating state of the power equipment.

[0022] It should be noted that the preprocessed multidimensional operating parameters in this application are a set of operating parameter data that has undergone outlier removal, missing value completion, and dimensional normalization for feature extraction. The variation characteristics of each parameter are a set of quantitative indicators used to characterize the dynamic changes of each original operating parameter in both the time-domain statistical characteristics and the frequency-domain energy distribution. The multidimensional state feature vector is a numerical vector composed of all operating parameter variation characteristics concatenated in a fixed order, used to characterize the overall operating state of the power equipment at the current moment.

[0023] The multidimensional operating parameters are preprocessed to obtain preprocessed multidimensional operating parameters. Specifically, the following steps are taken: First, a statistical outlier detection method based on the Laida criterion is used to scan each parameter sequence in the acquired multidimensional operating parameters, including voltage, current, active power, reactive power, power factor, equipment surface temperature, ambient temperature, vibration amplitude, and partial discharge. Extreme data points exceeding three standard deviations of the parameter sequence's average value are identified as outliers and removed. Then, for missing positions in the parameter sequences after outlier removal, linear interpolation is used to calculate and fill in the missing position based on the data values ​​of two consecutive valid time points adjacent to the missing position and the time interval between the two time points. Finally, a minimum-maximum normalization method is used to map all valid data of each type of operating parameter to a closed interval between zero and one, eliminating numerical scale differences between parameters of different dimensions such as voltage, current, temperature, and vibration amplitude. After the above processing, standardized, continuous, and non-abnormally disturbed preprocessed multidimensional operating parameters are obtained.

[0024] Statistical features and spectral features of each parameter are extracted from the preprocessed multidimensional operating parameters. These features are then fused to obtain the variation characteristics of each parameter. Specifically, during statistical feature extraction, for each operating parameter's data sequence within a preset time window, time-domain statistical analysis is used to calculate the mean, standard deviation, maximum value, minimum value, peak-to-peak value, and skewness of the sequence. These six statistical measures are combined to form the statistical feature set of the parameter. During spectral feature extraction, for the same operating parameter's data sequence within the same time window, a discrete Fourier transform is first used to convert the time-domain sequence into a frequency-domain sequence. Then, the amplitude corresponding to the frequency component with the largest amplitude is extracted as the dominant frequency amplitude, and the amplitude corresponding to the frequency component with the second largest amplitude is extracted as the secondary dominant frequency amplitude. The energy proportion of different frequency bands in the entire frequency-domain sequence is then calculated. The dominant frequency amplitude, secondary dominant frequency amplitude, and the energy proportion of each frequency band are combined to form the spectral feature set of the parameter. Subsequently, for each operating parameter, its statistical feature set and spectral feature set are concatenated in a predetermined order, with time-domain statistics first and frequency-domain features last, to form a comprehensive feature vector for that parameter. This comprehensive feature vector contains both the fluctuation amplitude information of the parameter in the time domain and the energy distribution information in the frequency domain, which is the variation characteristic of each parameter.

[0025] The variation characteristics of various parameters are concatenated according to preset feature dimensions to construct a multi-dimensional state feature vector characterizing the current operating state of power equipment. Specifically, a unified dimensional arrangement order is pre-defined based on the actual needs of power equipment condition monitoring. This order is as follows: voltage variation characteristics, current variation characteristics, active power variation characteristics, reactive power variation characteristics, power factor variation characteristics, equipment surface temperature variation characteristics, ambient temperature variation characteristics, and vibration amplitude variation characteristics are arranged sequentially; finally, partial discharge variation characteristics are arranged. Following this preset order, each component of the variation characteristics of each operating parameter is extracted sequentially and concatenated into a complete one-dimensional vector. The total dimension of this vector is equal to the sum of the number of variation characteristic components of all operating parameters. Each element in this one-dimensional vector has a clear physical meaning, corresponding to a specific time-domain statistic or a specific frequency-domain feature of a particular operating parameter. Finally, this concatenated complete one-dimensional vector is used as the multi-dimensional state feature vector for power equipment health status analysis and fault early warning judgment.

[0026] In step 103, the normal fluctuation range of each operating parameter at the current moment is generated based on the multidimensional state feature vector and the historical operating condition parameter sequence representing the external operating conditions in the historical operating data of the power equipment.

[0027] In some embodiments, generating the normal fluctuation range of each operating parameter at the current moment based on the multidimensional state feature vector and the historical operating condition parameter sequence characterizing external operating conditions in the historical operating data of power equipment can be achieved by the following steps: Extract historical operating condition parameter sequences from historical operating data of power equipment that match the current external operating conditions; Statistical analysis is performed on each operating parameter corresponding to the historical operating condition parameter sequence to calculate the mean and standard deviation of each operating parameter under the same operating conditions. Based on the mean and standard deviation of each operating parameter, the upper and lower limits of normal fluctuation of each operating parameter under external operating conditions are determined according to the preset confidence level, so as to form the normal fluctuation range of each operating parameter at the current moment.

[0028] It should be noted that the historical operating condition parameter sequence in this application is a data set composed of operating condition parameters such as load rate, ambient temperature, and operating time arranged in chronological order, used to characterize the trajectory of changes in external operating conditions during the historical operation of power equipment. The mean and standard deviation are two statistical quantities used to quantify the location and dispersion of each operating parameter under specific external operating conditions. The mean reflects the central tendency of each operating parameter, while the standard deviation reflects the fluctuation range of each operating parameter around the mean. The normal fluctuation range is a closed numerical interval consisting of the upper and lower limits of normal fluctuation, used to define the reasonable range of variation of each operating parameter under current external operating conditions.

[0029] The process involves extracting a sequence of historical operating parameters from historical power equipment operating data to match the current external operating conditions. Specifically, this historical operating data is pre-stored in the historical database of the power equipment condition monitoring system, recording multi-dimensional operating parameters and corresponding external operating parameters at various points in the equipment's past operation. First, the current external operating conditions are obtained, specifically including three dimensions: equipment load rate, ambient temperature, and runtime. Then, a load rate matching tolerance range is set centered on the current load rate value, an ambient temperature matching tolerance range is set centered on the current ambient temperature value, and a runtime matching tolerance range is set centered on the current runtime value. Each historical record in the database is then evaluated. If the load rate, ambient temperature, and runtime of a given historical record all fall within the load rate matching tolerance range, the record is considered to match the current external operating conditions. Extract the multidimensional operating parameters corresponding to each record from all the historical records that are determined to be matched, and arrange the extracted multidimensional operating parameters in chronological order to form a historical operating condition parameter sequence that matches the external operating conditions at the current moment. This historical operating condition parameter sequence contains the specific values ​​of the multidimensional operating parameters at each moment in the historical operation of the power equipment under similar external operating conditions.

[0030] Statistical analysis is performed on the corresponding operating parameters in the historical operating condition parameter sequence to calculate the mean and standard deviation of each operating parameter under the same operating conditions. Specifically, the historical operating condition parameter sequence extracted in the first step is decomposed according to the type of operating parameter to obtain the historical sequences for voltage, current, active power, reactive power, power factor, equipment surface temperature, ambient temperature, vibration amplitude, and partial discharge. For each type of historical sequence, the arithmetic mean is calculated by summing all values ​​in the sequence and dividing by the total number of data points to obtain the mean of that type of operating parameter under the same operating conditions. Simultaneously, for the same type of historical sequence, the standard deviation is calculated by first calculating the square of the difference between each value in the sequence and the mean, summing all squares and dividing by the total number of data points to obtain the variance, and then taking the square root of the variance to obtain the standard deviation of that type of operating parameter under the same operating conditions. The above calculation process for the mean and standard deviation is performed for each type of operating parameter to finally obtain the mean and standard deviation of each type of operating parameter under the current external operating conditions.

[0031] Based on the mean and standard deviation of each operating parameter, and using a preset confidence level, the upper and lower limits of normal fluctuation for each operating parameter under external operating conditions are determined to constitute the normal fluctuation range of each operating parameter at the current moment. Specifically, the implementation is as follows: First, a preset confidence level is obtained, which is a probability threshold pre-set according to the reliability requirements of power equipment condition monitoring. Using the mean of a certain operating parameter calculated in the second step as the center, the mean is multiplied by the standard deviation and the confidence coefficient corresponding to the confidence level, and then summed to obtain the upper limit of normal fluctuation for that operating parameter under external operating conditions; specifically, this is the mean plus the product of the confidence coefficient and the standard deviation. The lower limit of normal fluctuation for that operating parameter under external operating conditions is obtained by subtracting the product of the confidence coefficient and the standard deviation from the mean. This calculation method of adding and subtracting the above products from the mean ensures that the numerical range between the upper and lower limits of normal fluctuation probabilistically covers most of the possible values ​​of that operating parameter under the same external operating conditions according to the stated confidence level. For each type of operating parameter, the upper and lower limits of normal fluctuation are calculated using the same method, based on their respective means, standard deviations, and confidence coefficients. The calculated lower and upper limits of normal fluctuation for each type of operating parameter are then used as the two endpoints of the normal fluctuation interval for that parameter. The endpoint with the smaller value is designated as the left endpoint, and the endpoint with the larger value as the right endpoint. These left and right endpoints together form a closed interval of normal fluctuation. Finally, the normal fluctuation intervals corresponding to each type of operating parameter are combined to form the normal fluctuation interval for each operating parameter at the current moment, used to determine the degree of deviation of the actual multidimensional operating parameters.

[0032] In step 104, the first deviation of the actual multidimensional operating parameters at the current moment from the normal fluctuation range is calculated, and the second deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment from the health status feature distribution of the power equipment is calculated.

[0033] In some embodiments, calculating the first deviation of the actual multidimensional operating parameters at the current moment relative to the normal fluctuation range can be achieved by the following steps: Each actual operating parameter in the current multidimensional operating parameters is compared with the corresponding normal fluctuation range to determine whether each actual operating parameter falls within the normal fluctuation range. For operating parameters that fall within the normal fluctuation range, their deviation is assigned a value of zero. For operating parameters that exceed the normal fluctuation range, calculate the ratio of the excess amount to the width of the normal fluctuation range, and use it as the individual deviation of the operating parameter. The individual deviations of all operating parameters are weighted and aggregated to obtain the first deviation of the actual multidimensional operating parameters at the current moment relative to the normal fluctuation range.

[0034] It should be noted that the normal fluctuation range in this application is a closed numerical interval consisting of the upper and lower limits of normal fluctuation, used to define the reasonable range of variation of each operating parameter under current external operating conditions. The individual deviation is a dimensionless value used to quantify the degree to which a single operating parameter exceeds its normal fluctuation range. The first deviation is a weighted aggregate value characterizing the overall deviation level of all actual operating parameters relative to their normal fluctuation ranges at the current moment.

[0035] Each actual operating parameter in the current multidimensional operating parameters is compared with its corresponding normal fluctuation range to determine whether it falls within the normal fluctuation range. Specifically, the following steps are taken: First, the actual multidimensional operating parameters collected at the current moment are obtained. These parameters include the specific values ​​of nine categories of operating parameters: voltage, current, active power, reactive power, power factor, equipment surface temperature, ambient temperature, vibration amplitude, and partial discharge. Simultaneously, the normal fluctuation ranges for each operating parameter, constructed in the previous steps, are obtained. These ranges provide corresponding upper and lower limits for each category of operating parameters. Then, starting with the first category of operating parameters, the actual value of that category at the current moment is extracted, along with its corresponding lower and upper normal fluctuation limits. The actual value is then compared with both the lower and upper normal fluctuation limits to determine whether it is greater than or equal to the lower normal fluctuation limit and less than or equal to the upper normal fluctuation limit. If both of the above conditions are met, the actual operating parameter is determined to fall within the corresponding normal fluctuation range; if the actual value is less than the lower limit of normal fluctuation or greater than the upper limit of normal fluctuation, the actual operating parameter is determined not to fall within the corresponding normal fluctuation range. Following the above comparison method, the same numerical comparison and fall-in judgment operation is performed on the actual value of each type of operating parameter and its corresponding normal fluctuation range in sequence, until all types of operating parameters have completed the fall-in judgment.

[0036] For operating parameters falling within the normal fluctuation range, their deviation is set to zero. Specifically, following the judgment result for each type of operating parameter in the first step, for those parameters determined to fall within the normal fluctuation range, their corresponding deviation value is directly set to zero. This zero value indicates that the actual value of this type of operating parameter at the current moment is within the normal fluctuation range, requiring no deviation calculation; it has no deviation relative to the normal fluctuation range. This operation of setting the deviation to zero is performed on all operating parameters determined to fall within the normal fluctuation range, ensuring that each type of operating parameter within the normal fluctuation range obtains a definite deviation value of zero.

[0037] For operating parameters that exceed the normal fluctuation range, the ratio of the excess amount to the width of the normal fluctuation range is calculated as the individual deviation of the operating parameter. Specifically, the following steps are taken: First, identify the type of operating parameter that is determined not to fall within the normal fluctuation range. Then, determine the specific direction of the excess. Compare the actual value with the lower limit of the normal fluctuation range. If the actual value is less than the lower limit, it is considered a downward excess. In this case, subtract the actual value from the lower limit to obtain the downward excess. Next, compare the actual value with the upper limit of the normal fluctuation range. If the actual value is greater than the upper limit, it is considered an upward excess. In this case, subtract the upper limit from the upper limit to obtain the upward excess. For both downward and upward excess cases, the excess is taken as a positive value. Then, calculate the width of the normal fluctuation range corresponding to this type of operating parameter. Specifically, subtract the lower limit from the upper limit of the normal fluctuation range for this type of operating parameter; the difference is the width of the normal fluctuation range for that operating parameter. Finally, the calculated excess amount is divided by the width of the normal fluctuation range to obtain the ratio of the excess amount to the width of the normal fluctuation range. This ratio is the individual deviation of the operating parameter. This individual deviation reflects the relative magnitude of the actual value of the operating parameter deviating from its normal fluctuation range; the larger the deviation, the more severe the abnormality of the operating parameter. For each type of operating parameter determined to be outside the normal fluctuation range, the above operations of excess amount calculation, normal fluctuation range width calculation, and ratio calculation are performed to ensure that each type of operating parameter outside the normal fluctuation range obtains its corresponding individual deviation.

[0038] The weighted aggregation of individual deviations of all operating parameters yields the first deviation of the actual multidimensional operating parameter at the current moment relative to the normal fluctuation range. Specifically, this is achieved by first obtaining a pre-set weight coefficient for each category of operating parameters. This weight coefficient is a pre-assigned value based on the importance of each category of operating parameter to the operating state of the power equipment, and the sum of the weight coefficients for all categories of operating parameters equals one. For each category of operating parameters whose deviation values ​​have been obtained—namely, the operating parameters that fall within the normal fluctuation range and are assigned a value of zero in the second step, and the operating parameters that exceed the normal fluctuation range and whose individual deviation values ​​are calculated in the third step—the deviation value of each category of operating parameters is multiplied by the corresponding weight coefficient to obtain the weighted deviation of that category of operating parameters. Then, the weighted deviations of all categories of operating parameters are summed, that is, the weighted deviations of voltage, current, active power, reactive power, power factor, equipment surface temperature, ambient temperature, vibration amplitude, and partial discharge are all added together to obtain a total value. This total value is the first deviation of the actual multidimensional operating parameters at the current moment from the normal fluctuation range.

[0039] In some embodiments, calculating the second deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment relative to the health state feature distribution of the power equipment can be achieved by the following steps: Obtain the historical multidimensional state feature vector of power equipment within a preset time period under healthy operating conditions, and construct the health state feature distribution; The health status feature distribution is modeled to obtain a health status feature distribution model; The multidimensional state feature vector of the actual multidimensional operating parameters at the current moment is input into the health state feature distribution model, and the probability density value of the current multidimensional state feature vector under the health state feature distribution model is calculated. The second degree of deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment from the health state feature distribution is determined based on the probability density value.

[0040] It should be noted that the health state feature distribution in this application is a probability distribution formed by a large number of historical state feature vectors in a high-dimensional space, used to characterize the value patterns of multidimensional state feature vectors at various times during the healthy operation of power equipment. The health state feature distribution model is a probability density function expression used to describe the mathematical form of the health state feature distribution of power equipment, composed of distribution type parameters and distribution feature parameters. The probability density value is a non-negative value used to quantify the probability of the current multidimensional state feature vector appearing in the health state feature distribution model. The second deviation degree is a comprehensive value used to characterize the overall difference level between the multidimensional feature distribution of the actual multidimensional operating parameters at the current time and the overall health state feature distribution.

[0041] To obtain historical multidimensional state feature vectors of power equipment within a preset time period under healthy operating conditions, and construct the health state feature distribution, the specific implementation method is as follows: First, determine the criteria for judging the healthy operating state of the power equipment. This criterion is set based on the factory test qualification report and the load commissioning records during the initial commissioning period. The operating phase within the first three months after the equipment is put into operation without any abnormal alarms is marked as a healthy operating state. Then, retrieve the multidimensional state feature vectors for each sampling moment within the time period corresponding to the above-mentioned healthy operating state from the historical database of the power equipment condition monitoring system. Each element in the multidimensional state feature vector corresponds to the time-domain statistical characteristics or frequency-domain feature components of a type of operating parameter. Arrange all the retrieved historical multidimensional state feature vectors in chronological order according to their respective sampling moments to form a historical multidimensional state feature vector sequence. Statistical analysis is performed on the distribution of all vectors in the feature space of the historical multidimensional state feature vector sequence. The value range and frequency distribution of all historical vectors in each feature dimension are statistically analyzed. At the same time, the joint distribution of values ​​between different feature dimensions is statistically analyzed. The distribution of all feature dimensions and the joint distribution between feature dimensions are jointly constructed into the health state feature distribution.

[0042] The health state feature distribution is modeled to obtain a health state feature distribution model. Specifically, based on the health state feature distribution constructed in the first step, a probability density estimation method is used to mathematically model the distribution. Specifically, a kernel density estimation method is employed. Each historical vector in the historical multidimensional state feature vector sequence obtained in the first step is used as the kernel center. A smooth kernel function that conforms to the basic characteristics of a probability density function is selected. This kernel function contributes probability density to the surrounding space with each kernel center as its center of symmetry, and the influence range of the kernel function is controlled by a preset bandwidth parameter. The probability density contributions generated by all kernel centers are superimposed and averaged in an equal-weighted manner. That is, the probability density value at each position in the feature space is equal to the arithmetic mean of the values ​​of all kernel functions at that position, thus forming a continuous, smooth probability density function form that can be completely driven by historical data. This probability density function form is used as the health state feature distribution model. This health state feature distribution model can provide a corresponding probability density value for any position in the feature space.

[0043] The multidimensional state feature vector of the current multidimensional operating parameters is input into the health state feature distribution model. The probability density value of the current multidimensional state feature vector under the health state feature distribution model is calculated. Specifically, the multidimensional state feature vector corresponding to the current multidimensional operating parameters is obtained. This multidimensional state feature vector is constructed after preprocessing and feature extraction of the current multidimensional operating parameters in the previous steps. This multidimensional state feature vector is then substituted as an input variable into the health state feature distribution model established in the second step, that is, the specific value of each feature dimension of the multidimensional state feature vector is substituted into the corresponding variable position of the probability density function. The kernel density estimate at the input variable position is calculated in the health state feature distribution model. This kernel density estimate is equal to the arithmetic mean of the values ​​of all kernel functions at that input variable position. The calculated kernel density estimate is used as the probability density value of the current multidimensional state feature vector under the health state feature distribution model. The magnitude of this probability density value directly reflects the degree of matching between the current multidimensional state feature vector of the power equipment and the feature distribution under healthy operating conditions.

[0044] The second deviation degree of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment relative to the health state feature distribution is determined based on the probability density value. Specifically, the maximum probability density value that the health state feature distribution model can provide is first determined. This maximum probability density value appears in the central region of the health state feature distribution, i.e., where the historical multidimensional state feature vector distribution is most densely distributed. The current probability density value obtained in the third step is then compared to the maximum probability density value, i.e., the current probability density value is divided by the maximum probability density value, resulting in a dimensionless ratio between zero and one. This ratio reflects the relative density level of the current feature vector within the health state feature distribution. Then, this ratio is subtracted from the value one, i.e., the complement of the aforementioned relative density level is used as the second deviation degree. The value of this second deviation degree also ranges between zero and one. When the current multidimensional state feature vector is located in the central dense region of the health state feature distribution, the probability density value is close to the maximum value, the relative density level is close to one, and the second deviation degree is close to zero. When the current multidimensional state feature vector is far from the central dense region of the health state feature distribution, the probability density value decreases rapidly, the relative density level approaches zero, and the second deviation degree approaches one.

[0045] In step 105, the health status of the power equipment is predicted and analyzed based on the first deviation degree and the second deviation degree to obtain a comprehensive health index of the power equipment. When the comprehensive health index is lower than a preset warning threshold, a fault warning signal is issued.

[0046] In some embodiments, predicting and analyzing the health status of power equipment based on the first degree of deviation and the second degree of deviation to obtain a comprehensive health index of the power equipment can be achieved through the following steps: Determine the first weighting coefficient corresponding to the first degree of deviation and the second weighting coefficient corresponding to the second degree of deviation; The health status of power equipment is predicted based on the first degree of deviation and the first weighting coefficient, and the second degree of deviation and the second weighting coefficient, to obtain a comprehensive health index of the power equipment.

[0047] It should be noted that the first weighting coefficient in this application is a dimensionless value used to quantify the proportion of influence of the first deviation in the calculation of the comprehensive health index of power equipment. The second weighting coefficient is a dimensionless value used to quantify the proportion of influence of the second deviation in the calculation of the comprehensive health index of power equipment, and the sum of the first and second weighting coefficients is equal to one. The comprehensive health index is a normalized value used to characterize the overall health level of power equipment at a given moment, and its value ranges from zero to one hundred. The higher the value, the better the health status of the equipment.

[0048] The determination of the first weighting coefficient corresponding to the first deviation degree and the second weighting coefficient corresponding to the second deviation degree is specifically implemented as follows: First, the current operating condition information of the power equipment is obtained. This operating condition information includes the current load rate value, the total accumulated operating time, and the current ambient temperature value. The load rate value is compared with a preset load rate threshold. If the load rate is less than or equal to 60% of the rated load, the current condition is determined to be light load; if the load rate is greater than 60% of the rated load but less than or equal to 90% of the rated load, the current condition is determined to be medium load; and if the load rate is greater than 90% of the rated load, the current condition is determined to be heavy load. Simultaneously, the total accumulated operating time is compared with a preset aging stage division threshold. When the total operating time is less than or equal to one-third of the design life, it is determined to be the initial operation period; when the total operating time is greater than one-third of the design life but less than or equal to two-thirds of the design life, it is determined to be the stable operation period; and when the total operating time is greater than two-thirds of the design life, it is determined to be the end of the operation period. Then, the weighting coefficients are determined based on the combination of the operating condition information and the aging stage information. When the current operating condition is heavy load or nearing the end of its operation, the equipment's operating parameters are prone to significant instantaneous fluctuations due to load shocks and component aging. In this case, the first deviation level can more sensitively reflect parameter exceeding limits; therefore, the first weighting coefficient is set to the first value, and the second weighting coefficient is set to the second value, with the first weighting greater than the second weighting. When the current operating condition is light load or in the initial stage of commissioning, the equipment operates smoothly overall, and operating parameters rarely exceed the normal fluctuation range. The overall deviation of the multidimensional characteristic distribution can reflect potential performance degradation trends earlier. In this case, the first weighting coefficient is set to the third value, and the second weighting coefficient is set to the fourth value, with the second weighting greater than the first weighting. When the current operating condition is medium load and in a stable operating period, the first weighting coefficient is set to the fifth value, and the second weighting coefficient is set to the sixth value, with the first and second weighting coefficients being equal in value. The first and fourth values ​​are the same, the second and third values ​​are the same, and the fifth and sixth values ​​are both between the first and third values. All weight coefficients are chosen to ensure that the sum of the first and second weight coefficients equals one. Finally, the first weight coefficient determined in the above process is used as the influence weight of the first degree of deviation in subsequent calculations, and the second weight coefficient is used as the influence weight of the second degree of deviation in subsequent calculations.

[0049] The health status of power equipment is predicted based on the first deviation degree and the first weighting coefficient, and the second deviation degree and the second weighting coefficient, to obtain a comprehensive health index for the power equipment. Specifically, the method is as follows: First, obtain the first and second weighting coefficients determined in the first step, and simultaneously obtain the first and second deviation degree values ​​calculated in the preceding steps. Multiply the first deviation degree value by the first weighting coefficient to obtain a first weighted deviation value. Multiply the second deviation degree value by the second weighting coefficient to obtain a second weighted deviation value. Then, add the first and second weighted deviation values ​​to obtain a comprehensive deviation value, which comprehensively reflects the overall deviation level of the first and second deviation degrees under their respective weights. Subsequently, the comprehensive deviation value is converted into a health index. Specifically, the conversion method is to subtract the product of the comprehensive deviation value and a preset scaling factor from 100, where the preset scaling factor is set so that when the comprehensive deviation value is in the range of zero to one, it can be mapped to a health index range of zero to one hundred. When both the first and second deviation degrees are zero, the comprehensive deviation value is zero, and the comprehensive health index equals one hundred. As the first or second deviation degree increases, the comprehensive deviation value increases accordingly, and the comprehensive health index decreases accordingly. A comprehensive health index closer to 100 indicates a smaller deviation between the operating state and the health state of the electrical equipment, and a higher level of equipment health. Conversely, a comprehensive health index closer to zero indicates a greater deviation between the operating state and the health state of the electrical equipment, and a lower level of equipment health.

[0050] In some embodiments, issuing a fault warning signal when the comprehensive health index is lower than a preset warning threshold can be achieved through the following steps: Multiple preset warning threshold levels; The comprehensive health index is compared with each warning threshold level to determine the warning level of the comprehensive health index. Based on the aforementioned warning level, a corresponding fault warning signal is generated and pushed to the operation and maintenance terminal.

[0051] It should be noted that the multiple warning threshold levels in this application are several pre-defined health index critical values ​​used to gradient-divide the numerical range of the comprehensive health index. Each critical value corresponds to a severity level of fault risk. The warning level is a level identifier used to characterize the severity of fault risk corresponding to the current health status of the power equipment; this level identifier is uniquely determined by the threshold range into which the comprehensive health index falls. The fault warning signal is an alarm message used to convey the current fault risk level and the urgency of the warning to maintenance personnel. The maintenance terminal is a terminal device used to receive and display the fault warning signal to prompt maintenance personnel to take timely and appropriate measures.

[0052] Multiple warning threshold levels are preset. The specific implementation is as follows: First, based on the basic requirements for the safe operation of power equipment and the needs of equipment lifecycle health management, three warning threshold levels are preset in descending order of the comprehensive health index. The first warning threshold level is the attention warning level, corresponding to the first preset threshold. When the comprehensive health index falls below the first preset threshold, it indicates that the equipment has slightly deviated from its healthy state, requiring attention from maintenance personnel. The second warning threshold level is the fault warning level, corresponding to the second preset threshold, which is lower than the first. When the comprehensive health index falls below the second preset threshold, it indicates that the equipment has shown significant deterioration in its operating condition, posing a high risk of failure, requiring timely maintenance personnel to arrange repairs. The third warning threshold level is the emergency warning level, corresponding to the third preset threshold, which is lower than the second preset threshold. When the comprehensive health index falls below the third preset threshold, it indicates that the equipment is in a severely abnormal operating state, with a potential failure at any time, requiring immediate shutdown measures from maintenance personnel. The preset thresholds for the three warning threshold levels are all specific values ​​between zero and one hundred, and are arranged in a strict order: the third preset threshold is less than the second preset threshold, and the second preset threshold is less than the first preset threshold. These three warning threshold levels, along with their corresponding health index thresholds, are stored together to form a warning threshold level system. This system serves as the benchmark for subsequent warning determination. Ultimately, the first, second, and third preset thresholds are used as multiple warning threshold levels.

[0053] The comprehensive health index is compared with each warning threshold level to determine the warning level of the comprehensive health index. Specifically, the following steps are taken: First, obtain the specific value of the comprehensive health index calculated in the preceding steps for the current moment, and simultaneously obtain the specific values ​​of the first, second, and third preset thresholds set in the first step. Then, compare the comprehensive health index sequentially with the thresholds corresponding to each warning threshold level, from highest to lowest. First, compare the comprehensive health index with the first preset threshold. If the comprehensive health index is greater than or equal to the first preset threshold, it is determined that the comprehensive health index has not triggered any warning level at the current moment, the power equipment is in normal operation, and no warning signal is required. If the comprehensive health index is less than the first preset threshold, continue comparing the comprehensive health index with the second preset threshold. When the comprehensive health index is greater than or equal to the second preset threshold and less than the first preset threshold, it is determined that the comprehensive health index is within the value range corresponding to the attention warning level, and the attention warning level is determined as the current warning level. When the comprehensive health index is less than the second preset threshold, continue comparing the comprehensive health index with the third preset threshold. If the comprehensive health index is greater than or equal to the third preset threshold and less than the second preset threshold, the comprehensive health index is determined to be within the numerical range corresponding to the fault warning level, and the fault warning level is determined as the current warning level. If the comprehensive health index is less than the third preset threshold, the comprehensive health index is determined to be within the numerical range corresponding to the emergency warning level, and the emergency warning level is determined as the current warning level. Finally, the warning level determined through the above comparison process is taken as the current warning level of the comprehensive health index, which directly reflects the severity of the current equipment failure risk.

[0054] Based on the aforementioned warning levels, corresponding fault warning signals are generated and pushed to the operation and maintenance terminal. Specifically: when no warning level is triggered, no warning signal is generated, and normal operation monitoring continues. When the current warning level is determined to be a "pay attention" warning level, a Level 1 fault warning signal is generated. This Level 1 signal includes the specific value of the current comprehensive health index, the "pay attention" warning level identifier that triggered the warning, and a suggestion for operation and maintenance personnel to strengthen their observation of the operational status. When the current warning level is determined to be a "fault" warning level, a Level 2 fault warning signal is generated. This Level 2 signal includes the specific value of the current comprehensive health index, the fault warning level identifier that triggered the warning, and a suggestion for operation and maintenance personnel to arrange equipment maintenance as soon as possible. It also includes information on the operational parameter category with the largest deviation for operation and maintenance personnel's reference. When the current warning level is determined to be an "emergency" warning level, a Level 3 fault warning signal is generated. This Level 3 signal includes the specific value of the current comprehensive health index, the "emergency" warning level identifier that triggered the warning, and a suggestion for operation and maintenance personnel to immediately stop equipment operation and conduct a comprehensive investigation. It also includes information on the main contributing factors to the decline in the comprehensive health index. After the aforementioned fault warning signals at each level are generated, they are pushed to the designated maintenance terminal via real-time data transmission through the communication network between the power equipment condition monitoring system and the maintenance terminal. Upon receiving the fault warning signal, the maintenance terminal displays the warning level and content to the maintenance personnel through a pop-up interface, and simultaneously triggers different urgency-based audio-visual alerts in ascending order of warning level. Ultimately, the fault warning signals pushed to the maintenance terminal serve as the direct basis for communicating the equipment health risk status to maintenance personnel and guiding maintenance decisions.

[0055] In some embodiments, reference Figure 3 The figure is a schematic diagram of the working scenario of the power equipment fault early warning method based on multi-dimensional parameter perception in some embodiments of this application. In the figure, the power equipment is used to collect multi-dimensional operating parameters such as output voltage, temperature, and load and transmit them to the intermediate data processing equipment. The data processing equipment sequentially completes multi-dimensional parameter preprocessing, parameter change feature extraction and multi-dimensional state feature vector construction. Combined with historical operating condition parameter sequences, it generates a normal fluctuation range of parameters adapted to the real-time operating conditions. It calculates the first deviation degree corresponding to the parameter range and the second deviation degree corresponding to the multi-dimensional feature distribution matching degree. The two types of deviation indicators are fused to calculate the comprehensive health index of the equipment. After comparing the health index with the preset early warning threshold, if it is determined that the equipment has a fault risk, it sends an instruction to the signal early warning device on the right. The signal early warning device outputs an alarm prompt signal, which fully presents the entire process of power equipment fault early warning that is condition-adaptive and integrates multi-source parameter correlation features.

[0056] Furthermore, in another aspect of this application, in some embodiments, this application provides a power equipment fault early warning system based on multi-dimensional parameter perception, referring to... Figure 4 The figure is a schematic diagram of the structure of a power equipment fault early warning system based on multi-dimensional parameter perception according to some embodiments of this application. The power equipment fault early warning system 400 based on multi-dimensional parameter perception includes: an acquisition module 401, a processing module 402, and an execution module 403, which are described below: The acquisition module 401 in this application is mainly used to acquire multi-dimensional operating parameters of power equipment; Processing module 402, in this application, is used to preprocess the multidimensional operating parameters, extract the change features of each parameter, and construct a multidimensional state feature vector to characterize the current operating state of the power equipment; It should be noted that the processing module 402 in this application is also used to generate the normal fluctuation range of each operating parameter at the current moment based on the multidimensional state feature vector and the historical operating condition parameter sequence characterizing the external operating conditions in the historical operating data of the power equipment. In addition, it should be noted that the processing module 402 in this application is also used to calculate the first degree of deviation of the actual multidimensional operating parameters at the current moment relative to the normal fluctuation range, and the second degree of deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment relative to the health status feature distribution of the power equipment. The execution module 403 in this application is mainly used to predict and analyze the health status of the power equipment based on the first deviation degree and the second deviation degree to obtain the comprehensive health index of the power equipment. When the comprehensive health index is lower than the preset warning threshold, a fault warning signal is issued.

[0057] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described power equipment fault early warning method based on multi-dimensional parameter perception.

[0058] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device implementing a power equipment fault early warning method based on multi-dimensional parameter perception, according to some embodiments of this application. The power equipment fault early warning method based on multi-dimensional parameter perception in the above embodiments can... Figure 5 The computer device shown is used to implement this, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0059] Processor 501 can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0060] The communication bus 502 can be used to transmit information between the aforementioned components.

[0061] Memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory 503 may exist independently and be connected to processor 501 via communication bus 502. Memory 503 may also be integrated with processor 501.

[0062] The memory 503 stores program code for executing the scheme of this application, and its execution is controlled by the processor 501. The processor 501 executes the program code stored in the memory 503. The program code may include one or more software modules. The method used in the above embodiments can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0063] Communication interface 504 uses any transceiver-like device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0064] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0065] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0066] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for early warning of power equipment faults based on multi-dimensional parameter perception.

[0067] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0068] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for early warning of power equipment faults based on multi-dimensional parameter sensing, characterized in that, Includes the following steps: Obtain multi-dimensional operating parameters of power equipment; The multidimensional operating parameters are preprocessed to extract the variation features of each parameter and construct a multidimensional state feature vector to characterize the current operating state of the power equipment. Based on the multidimensional state feature vector and the historical operating condition parameter sequence representing external operating conditions in the historical operating data of power equipment, the normal fluctuation range of each operating parameter at the current moment is generated. Calculate the first degree of deviation of the actual multidimensional operating parameters at the current moment from the normal fluctuation range, and the second degree of deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment from the health status feature distribution of the power equipment; The health status of the power equipment is predicted and analyzed based on the first deviation degree and the second deviation degree to obtain a comprehensive health index of the power equipment. When the comprehensive health index is lower than the preset warning threshold, a fault warning signal is issued.

2. The method as described in claim 1, characterized in that, The multidimensional operating parameters include electrical parameters such as voltage, current, active power, reactive power, and power factor of the power equipment, as well as non-electrical parameters such as equipment surface temperature, ambient temperature, vibration amplitude, and partial discharge.

3. The method as described in claim 1, characterized in that, Preprocessing the multidimensional operating parameters, extracting the variation features of each parameter, and constructing a multidimensional state feature vector to characterize the current operating state of the power equipment specifically includes: The multidimensional operating parameters are preprocessed to obtain preprocessed multidimensional operating parameters; The statistical features and spectral features of each parameter are extracted from the preprocessed multidimensional operating parameters, and the statistical features and spectral features are fused to obtain the variation features of each parameter. The variation characteristics of each parameter are concatenated according to the preset feature dimensions to construct a multi-dimensional state feature vector that represents the current operating state of the power equipment.

4. The method as described in claim 1, characterized in that, The normal fluctuation range of each operating parameter at the current moment is generated based on the multidimensional state feature vector and the historical operating condition parameter sequence representing external operating conditions in the historical operating data of power equipment. Specifically, this includes: Extract a sequence of historical operating condition parameters from the historical operating data of power equipment that matches the current external operating conditions. The external operating conditions include equipment load rate, ambient temperature, and operating time. Statistical analysis is performed on each operating parameter corresponding to the historical operating condition parameter sequence to calculate the mean and standard deviation of each operating parameter under the same operating conditions. Based on the mean and standard deviation of each operating parameter, the upper and lower limits of normal fluctuation of each operating parameter under external operating conditions are determined according to the preset confidence level, so as to form the normal fluctuation range of each operating parameter at the current moment.

5. The method as described in claim 1, characterized in that, Calculating the first deviation of the actual multidimensional operating parameters at the current moment from the normal fluctuation range specifically includes: Each actual operating parameter in the current multidimensional operating parameters is compared with the corresponding normal fluctuation range to determine whether each actual operating parameter falls within the normal fluctuation range. For operating parameters that fall within the normal fluctuation range, their deviation is assigned a value of zero. For operating parameters that exceed the normal fluctuation range, calculate the ratio of the excess amount to the width of the normal fluctuation range, and use it as the individual deviation of the operating parameter. The individual deviations of all operating parameters are weighted and aggregated to obtain the first deviation of the actual multidimensional operating parameters at the current moment relative to the normal fluctuation range.

6. The method as described in claim 1, characterized in that, The second deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment from the health status feature distribution of the power equipment specifically includes: Obtain the historical multidimensional state feature vector of power equipment within a preset time period under healthy operating conditions, and construct the health state feature distribution; The health status feature distribution is modeled to obtain a health status feature distribution model; The multidimensional state feature vector of the actual multidimensional operating parameters at the current moment is input into the health state feature distribution model, and the probability density value of the current multidimensional state feature vector under the health state feature distribution model is calculated. The second degree of deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment from the health state feature distribution is determined based on the probability density value.

7. The method as described in claim 1, characterized in that, Based on the first and second deviation degrees, the health status of power equipment is predicted and analyzed to obtain a comprehensive health index for the power equipment, which specifically includes: Determine the first weighting coefficient corresponding to the first degree of deviation and the second weighting coefficient corresponding to the second degree of deviation; The health status of power equipment is predicted based on the first degree of deviation and the first weighting coefficient, and the second degree of deviation and the second weighting coefficient, to obtain a comprehensive health index of the power equipment.

8. A power equipment fault early warning system based on multi-dimensional parameter sensing, characterized in that, include: The acquisition module is used to acquire multi-dimensional operating parameters of power equipment; The processing module is used to preprocess the multidimensional operating parameters, extract the variation features of each parameter, and construct a multidimensional state feature vector to characterize the current operating state of the power equipment. The processing module is also used to generate the normal fluctuation range of each operating parameter at the current moment based on the multidimensional state feature vector and the historical operating condition parameter sequence characterizing the external operating conditions in the historical operating data of the power equipment. The processing module is also used to calculate the first degree of deviation of the actual multidimensional operating parameters at the current moment relative to the normal fluctuation range, and the second degree of deviation of the multidimensional feature distribution of the actual multidimensional operating parameters at the current moment relative to the health status feature distribution of the power equipment. The execution module is used to predict and analyze the health status of the power equipment based on the first deviation degree and the second deviation degree to obtain a comprehensive health index of the power equipment. When the comprehensive health index is lower than a preset warning threshold, a fault warning signal is issued.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the power equipment fault early warning method based on multi-dimensional parameter perception 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 the processor, it implements the power equipment fault early warning method based on multi-dimensional parameter perception as described in any one of claims 1 to 7.

Images