Reliability maintenance advice generation method and electronic device

Abstract

The application discloses a reliability maintenance suggestion generation method and electronic equipment, and relates to the technical field of equipment maintenance, and comprises the following steps: acquiring first monitoring data and second monitoring data of a target device; determining a first probability distribution of a potential failure time according to the first monitoring data, and determining a second probability distribution of a time interval between the potential failure time and a functional failure time according to the second monitoring data; fusing the first probability distribution and the second probability distribution to obtain a third probability distribution of the functional failure time; determining a current reliability evaluation index of the target device according to the third probability distribution, and generating a reliability maintenance suggestion of the target device according to the current reliability evaluation index. The application accurately establishes a quantitative corresponding relationship between a real-time state of the device and a maintenance time, thereby effectively improving the efficiency and precision of reliability maintenance decision-making of the power equipment.

Description

Technical Field

[0001] This application relates to the field of equipment maintenance technology, and in particular to a method for generating reliability maintenance recommendations and an electronic device. Background Technology

[0002] In the field of reliability management and maintenance decision-making for power equipment, the RCM (Reliability Centered Maintenance) theory has been widely applied. RCM theory clearly defines the concepts of potential failure (P-point) and functional failure (F-point), as well as the PF curve describing the relationship between the two. It also introduces the failure rate curve to describe the change in equipment failure probability over time. The PF curve represents the physical process of equipment progressing from potential failure to functional failure, while the failure rate curve represents the statistical law governing the change in equipment failure probability over time. Both describe equipment failure behavior from different dimensions.

[0003] However, in practical engineering applications, the PF curve is mainly used qualitatively to describe the failure process or quantitatively to calculate the condition monitoring cycle, while the failure rate curve is mainly used qualitatively to describe the failure trend or quantitatively to calculate the periodic maintenance cycle. Neither can transform the physical process of equipment failure into an accurate probability model, nor can it transform the output of the probability model into a quantitative maintenance basis that can be used for decision-making. As a result, it is difficult to meet the actual needs of refined management of power equipment reliability. As a result, when making maintenance decisions on power equipment, maintenance personnel still find it difficult to establish an accurate quantitative correspondence between the real-time status of the equipment and the timing of maintenance. They can only rely on experience to make qualitative judgments or arrange maintenance according to a fixed cycle. Summary of the Invention

[0004] The main purpose of this application is to provide a method and electronic device for generating reliability maintenance recommendations, aiming to solve the technical problem of poor efficiency and accuracy in the maintenance management of power equipment under the current RCM theory.

[0005] To achieve the above objectives, this application proposes a method for generating reliability maintenance recommendations, the method comprising: Acquire first monitoring data and second monitoring data of the target device, wherein the first monitoring data is used to characterize the potential failure time of the target device, and the second monitoring data is used to characterize the time interval between the potential failure time and the functional failure time of the target device; A first probability distribution of the potential failure time is determined based on the first monitoring data, and a second probability distribution of the time interval is determined based on the second monitoring data; By combining the first probability distribution and the second probability distribution, a third probability distribution is obtained for the moment when the function fails. Based on the third probability distribution, the current reliability evaluation index of the target equipment is determined, and based on the current reliability evaluation index, a reliability maintenance suggestion for the target equipment is generated.

[0006] In one embodiment, the step of determining the first probability distribution of the potential failure time based on the first monitoring data includes: Identify potential failure events of the target device and the first occurrence time of the potential failure events from the first monitoring data, and use the first occurrence time as the potential failure time; A first time series is generated based on the potential failure time and the historical potential failure time corresponding to the target device, wherein the historical potential failure time is the occurrence time of at least one historical potential failure event under the device category to which the target device belongs; The first probability distribution is obtained by fitting the first time series with a Weibull distribution.

[0007] In one embodiment, the step of determining the second probability distribution of the time interval based on the second monitoring data includes: From the second monitoring data, identify potential failure events, functional failure events, the first occurrence time of the potential failure event and the second occurrence time of the functional failure event of the target device, and use the duration between the first occurrence time and the second occurrence time as the time interval. A second time series is generated based on the time interval and the historical time interval corresponding to the target device, wherein the historical time interval is the duration between at least one historical potential failure event and the corresponding historical functional failure event under the device category to which the target device belongs; The second probability distribution is obtained by fitting the second time series with a Weibull distribution.

[0008] In one embodiment, the step of fusing the first probability distribution and the second probability distribution to obtain the third probability distribution at the time of functional failure includes: Convolve the first probability density function corresponding to the first probability distribution and the second probability density function corresponding to the second probability distribution to obtain the third probability density function. The third probability density function is taken as the third probability distribution.

[0009] In one embodiment, the current reliability evaluation index includes the current failure rate and / or the current reliability, and the step of determining the current reliability evaluation index of the target device according to the third probability distribution includes: Extract the shape parameters of the third probability distribution and use the shape parameters as the current failure rate; and / or, Extract the shape parameter and scale parameter of the third probability distribution, and calculate the current reliability based on the shape parameter and the scale parameter, wherein the shape parameter is negatively correlated with the current reliability, and the scale parameter is positively correlated with the current reliability.

[0010] In one embodiment, the step of generating reliability maintenance recommendations for the target equipment based on the current reliability evaluation indicators includes: When the current reliability evaluation index includes the current failure rate, the shape parameter of the current failure rate is obtained, and the first reference suggestion corresponding to the target parameter interval in multiple preset parameter intervals is used as the reliability maintenance suggestion; When the current reliability evaluation index includes the current reliability, the target reliability of the target device is obtained, the reliability difference between the target reliability and the current reliability is determined, and the second reference suggestion corresponding to the target difference interval in multiple preset difference intervals is used as the reliability maintenance suggestion.

[0011] In one embodiment, before the steps of determining a first probability distribution of the potential failure time based on the first monitoring data and determining a second probability distribution of the time interval based on the second monitoring data, the method further includes: Acquire the status detection data of the target device, wherein the status detection data is used to characterize the failure cause of the target device; Based on the status detection data, identify the failure cause type of the target device; A probability distribution calculation strategy corresponding to the failure cause type is determined, and the steps of determining the first probability distribution of the potential failure time based on the first monitoring data and determining the second probability distribution of the time interval based on the second monitoring data are executed based on the probability distribution calculation strategy.

[0012] In one embodiment, the step of identifying the failure cause type of the target device based on the state detection data includes: Obtain the running time and load current of the target device; If the running time is less than a preset first time threshold and the status detection data meets the preset first defect index, the failure cause type is determined to be an explicit manufacturing defect. If the runtime is greater than or equal to a preset second duration threshold and the status detection data meets a preset second defect index, the failure cause type is determined to be a latent manufacturing defect. Specifically, if the relationship between the rate of change of the first characteristic parameter and the load current in the status detection data meets a preset nonlinear relationship, and / or the second characteristic parameter in the status detection data meets a preset aging deviation index, and / or the fluctuation of the third characteristic parameter in the status detection data meets a preset abnormal fluctuation index, the status detection data is determined to meet the preset second defect index. If the runtime is greater than or equal to a preset first duration threshold and less than a preset second duration threshold, or if the state detection data does not meet the preset first defect index and the preset second defect index, the failure cause type is determined to be runtime stress.

[0013] In one embodiment, the step of determining the probability distribution calculation strategy corresponding to the failure cause type includes: When the failure cause type is the explicit manufacturing defect, the probability distribution calculation strategy is determined to be a first calculation strategy, wherein the first calculation strategy is used to control the probability weight of the first probability distribution to be greater than the preset probability weight in a preset early stage. When the failure cause type is the latent manufacturing defect, the probability distribution calculation strategy is determined to be the second calculation strategy, wherein the second calculation strategy is used to control the probability weight of the first probability distribution in a preset early stage to be less than or equal to the preset probability weight. When the failure cause type is the operating stress, the probability distribution calculation strategy is determined to be a third calculation strategy, wherein the third calculation strategy is used to control the second probability distribution to dynamically change with the real-time operating parameters of the target device.

[0014] Furthermore, to achieve the above objectives, this application also proposes a reliability maintenance suggestion generation device, which includes: The data acquisition module is used to acquire first monitoring data and second monitoring data of the target device, wherein the first monitoring data is used to characterize the potential failure time of the target device, and the second monitoring data is used to characterize the time interval between the potential failure time and the functional failure time of the target device. The probability calculation module is used to determine a first probability distribution of the potential failure time based on the first monitoring data, and to determine a second probability distribution of the time interval based on the second monitoring data; A probability fusion module is used to fuse the first probability distribution and the second probability distribution to obtain a third probability distribution at the moment of functional failure; The suggestion generation module is used to determine the current reliability evaluation index of the target device based on the third probability distribution, and generate reliability maintenance suggestions for the target device based on the current reliability evaluation index.

[0015] In addition, to achieve the above objectives, this application also proposes an electronic device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the reliability maintenance recommendation generation method as described above.

[0016] One or more technical solutions proposed in this application have at least the following technical effects: This application obtains first monitoring data characterizing the potential failure time of a target device and second monitoring data characterizing the time interval between the potential failure time and the functional failure time. It then determines the first probability distribution of the potential failure time and the second probability distribution of the time interval, respectively. The first and second probability distributions are then fused to obtain the third probability distribution of the functional failure time. Based on this, the current reliability evaluation index of the target device is determined, and reliability maintenance recommendations are generated. This establishes a complete mapping chain from the physical process of device failure to the probability quantification model and then to the maintenance decision output. Specifically, the potential failure time and the time interval correspond to two stages with clear physical meaning in the PF curve: the P point and the PF interval, respectively. By treating these two as independent random variables and establishing their respective probability distributions, the physical process described by the PF curve can be quantified into a computable probability input. Next, by fusing the first and second probability distributions to obtain the third probability distribution of the functional failure time, the temporal evolution characteristics of the physical process are fully mapped to the failure probability distribution, thus establishing an intrinsic relationship between the PF curve and the failure probability. Finally, based on the third probability distribution, the current reliability evaluation index is determined, and reliability maintenance recommendations are generated accordingly. This allows the output of the probability model to be directly transformed into specific and quantifiable maintenance decision-making criteria. This allows maintenance personnel to establish a quantitative correlation between the real-time status of the equipment and the timing of maintenance based on the monitoring data of the target equipment itself, without relying on experience for qualitative judgment or fixed-cycle maintenance. This effectively improves the efficiency and accuracy of reliability maintenance decisions for power equipment. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the first embodiment of the reliability maintenance suggestion generation method of this application; Figure 2 A flowchart illustrating Embodiment 2 of the reliability maintenance suggestion generation method for this application; Figure 3 A simplified flowchart illustrating the reliability maintenance suggestion generation method provided in Embodiment 2 of this application; Figure 4 This is a schematic diagram of the module structure of the reliability maintenance suggestion generation device according to an embodiment of this application; Figure 5 This is a schematic diagram of the device structure of the hardware operating environment involved in the reliability maintenance suggestion generation method in this application embodiment.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0022] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of this application embodiment is as follows: First monitoring data and second monitoring data of the target device are obtained, wherein the first monitoring data is used to characterize the potential failure time of the target device, and the second monitoring data is used to characterize the time interval between the potential failure time and the functional failure time of the target device; a first probability distribution of the potential failure time is determined based on the first monitoring data, and a second probability distribution of the time interval is determined based on the second monitoring data; the first probability distribution and the second probability distribution are fused to obtain a third probability distribution of the functional failure time; the current reliability evaluation index of the target device is determined based on the third probability distribution, and a reliability maintenance recommendation for the target device is generated based on the current reliability evaluation index.

[0024] In practical engineering applications, the PF curve is mainly used qualitatively to describe the failure process or quantitatively to calculate the condition monitoring cycle, while the failure rate curve is mainly used qualitatively to describe the failure trend or quantitatively to calculate the periodic maintenance cycle. Neither can transform the physical process of equipment failure into an accurate probability model, nor can it transform the output of the probability model into a quantitative maintenance basis that can be used for decision-making. As a result, it is difficult to meet the actual needs of refined management of power equipment reliability. As a result, when making maintenance decisions for power equipment, maintenance personnel still find it difficult to establish an accurate quantitative correspondence between the real-time status of the equipment and the timing of maintenance. They can only rely on experience to make qualitative judgments or arrange maintenance according to a fixed cycle.

[0025] This application provides a solution that establishes a complete mapping chain from the physical process of equipment failure to a probabilistic quantitative model and then to the output of maintenance decisions. Specifically, the potential failure time and time interval correspond to two stages with clear physical meaning in the Power Process (PF) curve: the P point and the PF interval. These are treated as independent random variables, and probability distributions are established for each. This allows the physical process described by the PF curve to be quantified into a computable probabilistic input. Then, by fusing the first and second probability distributions, a third probability distribution for the functional failure time is obtained, fully mapping the temporal evolution characteristics of the physical process to the failure probability distribution, thus establishing an intrinsic relationship between the PF curve and the failure probability. Finally, based on the third probability distribution, current reliability evaluation indicators are determined, and reliability maintenance recommendations are generated accordingly. This allows the output of the probabilistic model to be directly transformed into specific and quantifiable maintenance decision-making criteria. Therefore, maintenance personnel no longer need to rely on experience-based qualitative judgments or fixed-cycle maintenance. Instead, they can accurately establish a quantitative correspondence between the real-time status of the equipment and the timing of maintenance based on the monitoring data of the target equipment itself, effectively improving the efficiency and accuracy of reliability maintenance decisions for power equipment.

[0026] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions. The following description uses an electronic device as an example to illustrate this embodiment and the subsequent embodiments.

[0027] Based on this, embodiments of this application provide a method for generating reliability maintenance recommendations, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the reliability maintenance suggestion generation method of this application.

[0028] In this embodiment, the reliability maintenance suggestion generation method includes steps S10 to S40: Step S10: Obtain first monitoring data and second monitoring data of the target device, wherein the first monitoring data is used to characterize the potential failure time of the target device, and the second monitoring data is used to characterize the time interval between the potential failure time and the functional failure time of the target device. It should be noted that the target equipment is electrical equipment that requires reliability assessment and maintenance decisions. In this embodiment, the target equipment is the object of the reliability maintenance suggestion generation method. By acquiring and analyzing relevant monitoring data of the target equipment, maintenance suggestions for the target equipment are ultimately generated. The target equipment can be various main power equipment in a substation, such as transformers, circuit breakers, disconnect switches, and reactors.

[0029] The first monitoring data is used to characterize the potential failure moment of the target equipment. This data reflects the time point when the target equipment transitions from normal operation to a potential failure state, i.e., the moment of point P. In actual engineering, the first monitoring data can be state parameters of the target equipment such as vibration amplitude, temperature, and partial discharge. By monitoring the real-time values ​​of these parameters and comparing them with preset thresholds, the potential failure moment can be determined when the parameter exceeds the preset threshold.

[0030] The second monitoring data is used to characterize the time interval between the potential failure point and the functional failure point of the target equipment, i.e., the monitoring data used to determine the duration of the PF interval. In actual engineering, the second monitoring data can be a continuous monitoring value of a state parameter of the same type as the first monitoring data. For example, the duration of the PF interval can be determined by monitoring the time it takes for the vibration amplitude to go from exceeding the P-point threshold to exceeding the F-point threshold. Alternatively, it can be other state parameters that can reflect the degradation process of the equipment, and the duration required for the target equipment to develop from a potential failure state to a functional failure state can be determined by monitoring the trend of this parameter.

[0031] The potential failure moment refers to the point in time when a target device transitions from a normal operating state to a potential failure state, i.e., the moment corresponding to point P. In RCM theory, a potential failure refers to a situation where, although the device has not yet lost its intended function, a detectable abnormal state has emerged, indicating that the device is progressing towards functional failure. In practical engineering, the potential failure moment can be determined by setting a threshold for the first monitoring data; that is, when the state parameter represented by the first monitoring data first exceeds the threshold, that moment is identified as the potential failure moment.

[0032] The functional failure moment refers to the point in time when the target equipment loses its specified function, i.e., the time corresponding to point F. In RCM theory, functional failure means that the equipment cannot perform the function required by its design, such as a transformer failing to transform voltage normally or a circuit breaker failing to open and close normally. In practical engineering, the functional failure moment can be determined by setting a functional failure threshold for the state parameters. That is, when the state parameters exceed the functional failure threshold, the target equipment is considered to have lost its specified function, and this moment is determined as the functional failure moment.

[0033] Step S20: Determine a first probability distribution of the potential failure time based on the first monitoring data, and determine a second probability distribution of the time interval based on the second monitoring data; It should be noted that the first probability distribution is the probability distribution followed by the potential failure time of the target device. This probability distribution is obtained by statistically fitting the historical potential failure time data of the target device. The first probability distribution can quantitatively describe the probability of a potential failure event occurring at any point in time during the entire life cycle of the target device, that is, it can answer the question "What is the probability that the target device will experience a potential failure at a certain point in time?" In this embodiment, the first probability distribution is a key intermediate result determined based on the first monitoring data.

[0034] The second probability distribution is the probability distribution followed by the time interval between the potential failure time and the functional failure time of the target device. This probability distribution is obtained by statistically fitting the historical PF interval data of the target device. The second probability distribution can quantitatively describe the uncertainty of the time required for the target device to continue deteriorating to functional failure after a potential failure, that is, it can answer the question "How long will it take for the target device to reach the functional failure state after a potential failure?" In this embodiment, the second probability distribution is another key intermediate result determined based on the second monitoring data.

[0035] Step S30: Combine the first probability distribution and the second probability distribution to obtain the third probability distribution at the moment of functional failure; It should be noted that the third probability distribution refers to the probability distribution followed by the moment of functional failure of the target device. This probability distribution is obtained by fusing the first and second probability distributions. In this embodiment, the moment of functional failure is physically equal to the sum of the potential failure moment and the PF interval. These two components are probabilistically independent random variables. Therefore, the probability distribution of the moment of functional failure can be obtained, but is not limited to, by performing a convolution operation on the first and second probability distributions. The third probability distribution can quantitatively describe the probability of the target device functional failure event occurring at any point in time.

[0036] Step S40: Determine the current reliability evaluation index of the target equipment according to the third probability distribution, and generate a reliability maintenance suggestion for the target equipment based on the current reliability evaluation index.

[0037] It should be noted that the current reliability evaluation index refers to a quantitative parameter used to characterize the reliability level of the target equipment at the current moment, which may include the current failure rate and / or the current reliability. The current failure rate reflects the instantaneous rate at which the target equipment fails at the current moment, while the current reliability reflects the probability that the target equipment can continue to perform its specified functions at the current moment. In this embodiment, the current reliability evaluation index is an intermediate output result determined based on a third probability distribution, which provides a quantitative decision-making basis for subsequently generating reliability maintenance recommendations.

[0038] Reliability maintenance recommendations are quantitative suggestions generated based on the current reliability evaluation indicators of the target equipment, used to guide maintenance personnel in making maintenance decisions for the target equipment. These recommendations reflect the current reliability status of the target equipment, providing maintenance personnel with clear quantitative basis for judging whether the equipment is currently reliable and what level of maintenance measures should be taken. In this embodiment, the reliability maintenance recommendation is the final output result, enabling maintenance personnel to make accurate and quantitative maintenance decisions based on the target equipment's own monitoring data, without relying on experience-based qualitative judgments or fixed-cycle maintenance.

[0039] Understandably, because the physical process of equipment failure from potential failure to functional failure, as represented by the PF curve, lacks an inherent mathematical correlation with the statistical law of equipment failure probability changing over time, as represented by the failure rate curve, the physical process described by the PF curve cannot be converted into a probabilistic input for failure rate calculation, nor can the output of the probabilistic model be converted into a quantitative maintenance basis for decision-making. Maintenance personnel find it difficult to establish an accurate quantitative correspondence between the real-time status of the equipment and the timing of maintenance, and can only rely on experience to make qualitative judgments or arrange maintenance according to fixed cycles, resulting in poor efficiency and accuracy in the maintenance and management of power equipment.

[0040] Therefore, this embodiment acquires first monitoring data to characterize the potential failure time of the target device and second monitoring data to characterize the time interval between the potential failure time and the functional failure time of the target device. Based on the first and second monitoring data, a first probability distribution of point P and a second probability distribution of the PF interval time are determined. Since the functional failure time is physically equal to the sum of the potential failure time and the PF interval time, and the two components are probabilistically independent random variables, the third probability distribution of the functional failure time can be obtained by fusing the first and second probability distributions. Based on this, the current reliability evaluation index of the target device is determined according to the third probability distribution, and then a reliability maintenance suggestion for the target device is generated based on the current reliability evaluation index.

[0041] Therefore, based on the probability distributions of two random variables with clear physical meanings, P point and PF interval time, a third probability distribution of the functional failure time is obtained by fusion. This realizes a complete mathematical mapping from the PF physical process to the probability quantification model and then to the maintenance decision output. This allows the time evolution characteristics described by the PF curve to be transformed into a precise basis for failure rate quantification calculation. Furthermore, the output of the probability model is transformed into quantitative maintenance suggestions that can be directly used by operation and maintenance personnel, thereby effectively improving the efficiency and accuracy of reliability maintenance decisions for power equipment.

[0042] For example, a transformer in a substation is used as the target device, and the vibration amplitude of the transformer's tank wall is used as the first and second monitoring data. Maintenance personnel continuously collect tank wall vibration amplitude data during transformer operation. The moment when the vibration amplitude first exceeds a first threshold (e.g., 60 μm) is designated as the potential failure moment, and the moment when the vibration amplitude reaches a second threshold (e.g., 100 μm) is designated as the functional failure moment. Based on this, vibration amplitude data of the transformer itself and multiple historical transformers of the same voltage level and model are obtained during operation. The potential failure moment and functional failure moment of each transformer are extracted, and the power factor interval (PF) between the potential failure moment and the functional failure moment is calculated for each transformer. Using these historical data as samples, statistical fitting is performed on the distribution of potential failure moments and the distribution of PF intervals to obtain a first probability distribution and a second probability distribution. The first and second probability distributions are then fused probabilistically (e.g., through convolution) to obtain a third probability distribution, which reflects the probability distribution of the timing of functional failure of the transformer during future operation. The failure rate and / or reliability of the transformer at the current moment are calculated based on the third probability distribution. Finally, the calculated failure rate and / or reliability are used as the basis for generating maintenance recommendations, and a reliability maintenance recommendation for the transformer is output.

[0043] This embodiment provides a method for generating reliability maintenance recommendations. It establishes a complete mapping chain from the physical process of equipment failure to a probabilistic quantification model and then to the output of maintenance decisions. Specifically, the potential failure time and time interval correspond to two stages with clear physical meaning in the Power Process (PF) curve: the P point and the PF interval. These are treated as independent random variables to establish probability distributions, allowing the physical process described by the PF curve to be quantified into a computable probabilistic input. Then, by fusing the first and second probability distributions, a third probability distribution for the functional failure time is obtained, fully mapping the temporal evolution characteristics of the physical process to the failure probability distribution, thus establishing an intrinsic correlation between the PF curve and the failure probability. Finally, based on the third probability distribution, the current reliability evaluation index is determined, and reliability maintenance recommendations are generated accordingly. This allows the output of the probabilistic model to be directly transformed into specific and quantifiable maintenance decision-making criteria. Therefore, maintenance personnel no longer need to rely on experience-based qualitative judgments or fixed-cycle maintenance. Instead, they can accurately establish a quantitative correspondence between the real-time status of the equipment and the timing of maintenance based on the monitoring data of the target equipment itself, effectively improving the efficiency and accuracy of reliability maintenance decisions for power equipment.

[0044] In one feasible implementation, the step of determining the first probability distribution of the potential failure time based on the first monitoring data in step S20 may include steps S21 to S23: Step S21: Identify potential failure events of the target device and the first occurrence time of the potential failure events from the first monitoring data, and take the first occurrence time as the potential failure time; It should be noted that a potential failure event refers to the specific event that occurs when the target equipment first enters a potential failure state from a normal operating state. Within the RCM theoretical framework, a potential failure state refers to a detectable abnormal state in which the equipment, although not yet having lost its intended function, has exhibited abnormalities indicating that the equipment is progressing towards functional failure. The determination of a potential failure event relies on initial monitoring data. For example, by setting a threshold for the initial monitoring data, when the state parameter represented by the initial monitoring data first exceeds this threshold, a potential failure event is considered to have occurred in the target equipment. The potential failure event is the starting point of the entire probabilistic modeling process in this technical solution; it marks the transition of the target equipment from a healthy state to a degraded state, and the subsequent PF interval is calculated from the moment this event occurs.

[0045] The first occurrence time refers to the specific point in time when a potential failure event occurs, that is, the moment when the state parameter represented by the first monitoring data first exceeds a preset threshold. The first occurrence time can be determined by continuously monitoring the state parameters of the target device, recording the trajectory of these parameters over time, and recording the current time point as the first occurrence time when the parameter first reaches or exceeds the preset threshold. The first occurrence time is numerically equal to the potential failure time, and its purpose is to provide crucial time data input for subsequently constructing the first time series and fitting the first probability distribution.

[0046] Step S22: Generate a first time series based on the potential failure time and the historical potential failure time corresponding to the target device, wherein the historical potential failure time is the occurrence time of at least one historical potential failure event under the device category to which the target device belongs; It should be noted that historical potential failure times refer to the times of potential failure events recorded by the target equipment itself during its past operation, as well as the times of potential failure events recorded by other similar equipment within the same equipment category as the target equipment during its past operation. The core function of historical potential failure times is to serve as statistical samples to fit the probability distribution of the potential failure times of the target equipment. Since the potential failure time of a single piece of equipment is a random variable, its probability distribution cannot be directly known. Only by collecting a large amount of historical data from similar equipment can the probability distribution form and parameters of this random variable be obtained through statistical fitting methods. Sources of historical potential failure times can include the equipment operation and maintenance database of power companies, historical data recorded by online monitoring systems, etc.

[0047] The first time series refers to a data set formed by arranging multiple potential failure moments in chronological order. Specifically, this data set includes the currently identified potential failure moment of the target device, as well as the corresponding historical potential failure moments. The first time series can be constructed by combining the current potential failure moment identified from the first monitoring data with multiple historical potential failure moments obtained from historical data to form a one-dimensional time data vector. The first time series serves as the input data for Weibull distribution fitting. By statistically fitting the first time series, the first probability distribution can be obtained, which is the probability distribution function followed by the potential failure moments of the target device.

[0048] Historical potential failure events refer to potential failure events that occurred during the past operation of other similar equipment within the same equipment category as the target equipment, as well as potential failure events that occurred during the past operation of the target equipment itself. Each historical potential failure event corresponds to a historical potential failure time. Data on historical potential failure events typically originates from the historical operation and maintenance records of power companies or historical data from online monitoring systems. In this embodiment, instead of directly obtaining the numerical value of the historical potential failure time, the corresponding time information is extracted from the identification and recording of historical potential failure events, thereby forming a data sample of historical potential failure times. This provides a sufficient statistical sample size for subsequent Weibull distribution fitting.

[0049] Step S23: Fit the first time series using a Weibull distribution to obtain the first probability distribution.

[0050] Understandably, by identifying potential failure events and their first occurrence times from the initial monitoring data, and using this first occurrence time as the potential failure time, and further generating a time series containing multiple sample points based on this potential failure time and the historical potential failure times of historical potential failure events within the target equipment category, a first probability distribution is obtained by fitting this time series to a Weibull distribution. Thus, through explicit and reproducible data processing steps, the probability distribution followed by the potential failure time of the target equipment is accurately obtained, providing a reliable quantitative input basis for subsequent fusion operations and maintenance suggestion generation.

[0051] For example, a transformer in a substation is used as the target equipment, and the vibration amplitude of the transformer's tank wall is used as the first monitoring data. Maintenance personnel continuously collect tank wall vibration amplitude data during transformer operation. When the vibration amplitude first exceeds a preset threshold (e.g., 60 μm), a potential failure event is determined, and the current time is recorded as the first occurrence time of this potential failure event, which is also taken as the potential failure time of the transformer. Based on this, vibration amplitude data of multiple historical transformers of the same voltage level and model in the power grid of the region where the transformer is located are obtained from the power company's equipment maintenance database. The times when potential failure events occurred for each historical transformer are extracted and taken as historical potential failure times. The currently identified potential failure times are combined with the multiple historical potential failure times and arranged chronologically to form a time series containing multiple sample points. Finally, a Weibull distribution is used to statistically fit this time series. The shape and scale parameters of the Weibull distribution are determined using methods such as maximum likelihood estimation, thereby obtaining the first probability distribution followed by the potential failure time of the transformer.

[0052] In one feasible implementation, the step of determining the second probability distribution of the time interval based on the second monitoring data in step S20 may include steps S24-S26: Step S24: Identify potential failure events, functional failure events, the first occurrence time of the potential failure event, and the second occurrence time of the functional failure event from the second monitoring data, and use the duration between the first occurrence time and the second occurrence time as the time interval; It should be noted that a functional failure event refers to a specific event that occurs when the target equipment loses its specified function. Within the RCM theoretical framework, a functional failure state refers to the equipment's inability to perform its designed functions, such as a transformer failing to transform voltage normally, a circuit breaker failing to open or close normally, or a disconnector failing to switch on or off normally. The determination of a functional failure event relies on second monitoring data. A functional failure threshold is set for the state parameters represented by the second monitoring data; that is, when the state parameter continuously deteriorates and reaches or exceeds the functional failure threshold, a functional failure event is considered to have occurred in the target equipment. The functional failure event is the endpoint of the entire probabilistic modeling process in this technical solution. It signifies that the target equipment has completed the full degradation process from potential failure to functional failure. The difference between the time of this event and the time of the potential failure event is the PF interval time.

[0053] The second occurrence time refers to the specific point in time when the functional failure event occurs, that is, the moment when the state parameter represented by the second monitoring data reaches or exceeds the functional failure threshold. The second occurrence time can be determined by continuously monitoring the state parameter of the target device, recording the trajectory of this parameter over time, and recording the current time point as the second occurrence time when the parameter first reaches or exceeds the preset functional failure threshold. The second occurrence time and the first occurrence time together constitute the two time endpoints for calculating the PF interval duration; that is, the duration obtained by subtracting the first occurrence time from the second occurrence time is the time elapsed from the potential failure to the functional failure of the target device.

[0054] Step S25: Generate a second time series based on the time interval and the historical time interval corresponding to the target device, wherein the historical time interval is the duration between at least one historical potential failure event and the corresponding historical functional failure event under the device category to which the target device belongs; It should be noted that historical time intervals refer to the duration recorded by the target device itself during its past operation, from the potential failure moment to the functional failure moment, as well as the duration recorded by other similar devices within the same device category during their past operation. Each historical time interval corresponds to a complete historical failure process, that is, the time span between a historical potential failure event and its corresponding historical functional failure event. The core function of historical time intervals is to serve as statistical samples to fit the probability distribution followed by the PF interval duration of the target device. Because the PF interval duration of a single device is a random variable, its probability distribution cannot be directly known. Only by collecting a large amount of historical PF interval data from similar devices can the probability distribution form and parameters of this random variable be obtained through statistical fitting methods.

[0055] The second time series refers to a data set formed by arranging multiple PF interval durations in chronological order. This data set specifically includes the PF interval duration currently identified by the target device, as well as the corresponding historical time intervals for the target device. The second time series can be constructed by combining the current PF interval duration calculated from the second monitoring data with multiple historical time intervals obtained from the historical data to form a one-dimensional duration data vector. The second time series serves as the input data for Weibull distribution fitting. By statistically fitting the second time series, a second probability distribution can be obtained, which is the probability distribution function followed by the PF interval durations of the target device.

[0056] Historical functional failure events refer to functional failure events that occurred during the past operation of other similar equipment within the same equipment category as the target equipment, as well as functional failure events that occurred during the past operation of the target equipment itself. Each historical functional failure event corresponds to a specific historical functional failure time, and there is a one-to-one correspondence between this historical functional failure event and a historical potential failure event. That is, in a complete failure process, a potential failure event necessarily corresponds to a functional failure event, and the two together constitute a complete historical failure sample. The data for historical functional failure events may, but is not limited to, come from the historical operation and maintenance records of power companies or historical data from online monitoring systems.

[0057] Step S26: Fit the second time series using a Weibull distribution to obtain the second probability distribution.

[0058] Understandably, by identifying potential failure events and corresponding functional failure events of the target device from the second monitoring data, and determining the first and second occurrence times of these two events respectively, the duration between the first and second occurrence times is used as the PF interval. Furthermore, based on this PF interval and the historical time intervals in historical failure processes within the target device's category, a second time series containing multiple sample points is generated. Finally, this second time series is fitted using a Weibull distribution to obtain a second probability distribution. Thus, through explicit and reproducible data processing steps, the probability distribution followed by the target device's PF interval is accurately obtained, providing a reliable PF interval quantification input basis for subsequent fusion operations.

[0059] For example, a transformer in a substation is used as the target device, and the vibration amplitude of the transformer's tank wall is used as the second monitoring data. Maintenance personnel continuously collect tank wall vibration amplitude data during transformer operation. When the vibration amplitude first exceeds a first threshold (e.g., 60 μm), a potential failure event is determined, and the current time is recorded as the first occurrence time. When the vibration amplitude continues to increase and reaches a second threshold (e.g., 100 μm), a functional failure event is determined, and the current time is recorded as the second occurrence time. The duration between the second and first occurrence times is used as the transformer's power PF interval time. Based on this, complete failure data of multiple historical transformers of the same voltage level and model in the power grid of the region where the transformer is located are obtained from the power company's equipment maintenance database. The historical PF interval time between the potential failure event and the corresponding functional failure event of each historical transformer is extracted. The currently calculated PF interval time is combined with the multiple historical PF interval times to form a second time series containing multiple sample points. Finally, the second time series is statistically fitted using the Weibull distribution, and the shape and scale parameters of the Weibull distribution are determined by methods such as maximum likelihood estimation, thereby obtaining the second probability distribution that the transformer PF interval time follows.

[0060] In one feasible implementation, step S30 may include steps S31-S32: Step S31: Convolve the first probability density function corresponding to the first probability distribution and the second probability density function corresponding to the second probability distribution to obtain the third probability density function; Step S32: Use the third probability density function as the third probability distribution.

[0061] Understandably, in practical engineering applications, the time of functional failure is physically equal to the sum of the potential failure time and the power failure (PF) interval. That is, the target device must first experience a potential failure, then endure a PF interval, before finally experiencing functional failure. When two random variables are independent, the probability distribution of their sum is not simply added or multiplied.

[0062] To this end, this embodiment further convolves the first probability density function corresponding to the first probability distribution and the second probability density function corresponding to the second probability distribution to obtain a third probability density function, and uses this third probability density function as the third probability distribution. Thus, by introducing convolution operation, the physical process of "the functional failure time is equal to the sum of the potential failure time and the PF interval time" is accurately mapped to the probability operation process of "the third probability density function is equal to the convolution of the first probability density function and the second probability density function". This avoids the problem of being unable to implement in engineering due to unclear fusion method, and ensures that the third probability distribution on which the current reliability evaluation index and reliability maintenance recommendations are subsequently determined can be calculated accurately and repeatably.

[0063] For example, based on the aforementioned scheme having determined the first probability distribution and the second probability distribution respectively, a fusion step is further performed, that is, the first probability density function corresponding to the first probability distribution is extracted from the computer memory, and the second probability density function corresponding to the second probability distribution is extracted from the computer memory. Since the functional failure time is physically equal to the sum of the potential failure time and the PF interval time, and the potential failure time and the PF interval time are two physically independent random variables, the first probability density function and the second probability density function are convolved.

[0064] Specifically, for each possible functional failure time t, the product of the first probability density function at a certain time τ and the second probability density function at time t-τ is calculated, and the product is integrated and summed over all possible τ values ​​to obtain the value of the third probability density function at time t. The convolution calculation results corresponding to all possible functional failure times t are collected to obtain the third probability density function. This third probability density function is stored as the third probability distribution for subsequent use in determining the current reliability evaluation index.

[0065] In one feasible implementation, the current reliability evaluation index includes the current failure rate and / or the current reliability. The step S40, which involves determining the current reliability evaluation index of the target device based on the third probability distribution, may include at least one of steps S41 and S42: Step S41: Extract the shape parameters of the third probability distribution and use the shape parameters as the current failure rate; It should be noted that the shape parameter, denoted as β, is a statistical parameter in the Weibull distribution used to control the geometric shape of the probability density function curve of the Weibull distribution. In this embodiment, the shape parameter is extracted from the third probability distribution and is used to characterize the trend of the failure rate of the target equipment over time. When the shape parameter takes different values, the failure rate curve exhibits different shapes; that is, when the shape parameter is greater than a certain value, the failure rate increases over time; when the shape parameter is equal to that value, the failure rate remains constant; and when the shape parameter is less than that value, the failure rate decreases over time. This shape parameter is the direct basis for determining the current failure rate in this embodiment. By using the shape parameter as a quantitative representation of the current failure rate, the failure rate characteristics of the target equipment at the current moment can be quantitatively described, thus providing one of the core input parameters for generating reliability maintenance recommendations.

[0066] The current failure rate refers to the instantaneous rate at which the target device fails at the current moment; that is, the conditional probability density of the target device failing in the next unit of time given that it has already been operating to the current moment. In this embodiment, the current failure rate is determined by extracting the shape parameter of the third probability distribution and directly using this shape parameter as the current failure rate. Based on the characteristics of the Weibull distribution, the shape parameter itself determines the shape and trend of the failure rate curve. That is, when the shape parameter is greater than a certain value, the failure rate increases with time; when the shape parameter is equal to the specific value, the failure rate is constant; and when the shape parameter is less than the specific value, the failure rate decreases with time. Therefore, using the shape parameter as a quantitative representation of the current failure rate has a clear theoretical basis. The current failure rate reflects the degree of failure risk of the target device in its current state and is one of the two core decision bases for generating reliability maintenance recommendations.

[0067] Step S42: Extract the shape parameter and scale parameter of the third probability distribution, and calculate the current reliability based on the shape parameter and the scale parameter, wherein the shape parameter is negatively correlated with the current reliability, and the scale parameter is positively correlated with the current reliability.

[0068] It should be noted that the scaling parameter, denoted as η, is another key statistical parameter in the Weibull distribution, used to characterize the scaling scale of the Weibull distribution on the time axis, i.e., the time order of the target device failure. In this embodiment, the scaling parameter and shape parameter are extracted together from the third probability distribution, and are used together to calculate the current reliability. The scaling parameter reflects the characteristic lifetime of the target device; that is, the larger the scaling parameter, the later the overall failure time distribution of the target device, i.e., the longer the device's lifetime; the smaller the scaling parameter, the earlier the failure time distribution, i.e., the shorter the device's lifetime. The scaling parameter is one of the two necessary input parameters for calculating the current reliability in this embodiment.

[0069] Current reliability refers to the probability that a target device can continue to perform its specified functions at the current moment; that is, the probability that the target device has not experienced functional failure since its commissioning. In this context, current reliability is calculated by extracting the shape and scale parameters of the third probability distribution. Based on the properties of the Weibull distribution, the reliability function is jointly determined by the shape and scale parameters. Specifically, the shape parameter and reliability are negatively correlated (the larger the shape parameter, the lower the reliability); conversely, the scale parameter and reliability are positively correlated (the larger the scale parameter, the higher the reliability). Current reliability reflects the reliability level of the target device's remaining lifespan at the current moment and is one of the two core decision bases for generating reliability maintenance recommendations. Together with the current failure rate, it describes the reliability status of the target device from different dimensions.

[0070] Understandably, in practical engineering applications, maintenance personnel usually need to refer to two different dimensions of reliability quantification parameters when making maintenance decisions: the instantaneous rate at which the equipment fails (i.e., failure rate) and the probability that the equipment can continue to perform its specified functions (i.e., reliability). Both reflect the reliability status of the equipment from different perspectives.

[0071] To this end, this embodiment further extracts the shape parameter of the third probability distribution directly as the current failure rate, and / or extracts the shape parameter and scale parameter of the third probability distribution and calculates the current reliability based on the mathematical relationship between the two. This allows maintenance personnel to accurately calculate the current failure rate and current reliability from the third probability distribution. These two parameters can jointly reflect the reliability status of the equipment from different dimensions, providing a comprehensive and accurate quantitative input basis for generating subsequent reliability maintenance recommendations.

[0072] For example, based on the third probability distribution already obtained in the aforementioned scheme, the system extracts the shape parameter β from the third probability distribution and directly uses this shape parameter β as the current failure rate. Simultaneously (or alternatively), the system extracts the shape parameter β and the scale parameter η from the third probability distribution. According to the expression for reliability R(t) in the Weibull distribution, R(t) = e^{-(t / η)^β}, the extracted shape parameter β and scale parameter η are substituted into this expression, and combined with the value of the current time t, the reliability R(t) of the target device at the current time is calculated and used as the current reliability. During the calculation process, according to the mathematical properties of the Weibull distribution reliability function, the shape parameter β is negatively correlated with the current reliability; that is, when β increases, e^{-(t / η)^β} decreases, and the current reliability decreases. Conversely, the scale parameter η is positively correlated with the current reliability; that is, when η increases, e^{-(t / η)^β} increases, and the current reliability increases. The system will output the calculated current failure rate and / or current reliability as the current reliability evaluation index for subsequent generation of reliability maintenance recommendations.

[0073] In one feasible implementation, the step of generating reliability maintenance recommendations for the target equipment based on the current reliability evaluation index in step S40 may include steps S43-S44: Step S43: If the current reliability evaluation index includes the current failure rate, obtain the shape parameter of the current failure rate, and use the first reference suggestion corresponding to the target parameter interval in multiple preset parameter intervals as the reliability maintenance suggestion; It should be noted that the first reference suggestion refers to the candidate maintenance suggestion generated based on the preset parameter range into which the shape parameter of the current failure rate falls when the current reliability evaluation index of the target equipment includes the current failure rate.

[0074] Specifically, the first reference suggestion can be generated as follows: the range of shape parameter values ​​is pre-divided into multiple preset parameter intervals, and each preset parameter interval is pre-associated with a maintenance suggestion. When the system obtains the shape parameter of the current failure rate, it determines which preset parameter interval the shape parameter falls into, and uses the maintenance suggestion corresponding to that interval as the first reference suggestion. Subsequently, this first reference suggestion is output as the final reliability maintenance suggestion. The first reference suggestion is selected from multiple candidate suggestions, and its specific content depends on the interval into which the shape parameter falls. Different intervals correspond to different maintenance suggestion types to reflect the differences in the failure rate trend of the target equipment under different shape parameter values.

[0075] Step S44: If the current reliability evaluation index includes the current reliability, obtain the target reliability of the target device, determine the reliability difference between the target reliability and the current reliability, and use the second reference suggestion corresponding to the target difference interval in a plurality of preset difference intervals as the reliability maintenance suggestion.

[0076] It should be noted that the second reference suggestion refers to the candidate maintenance suggestion generated based on the preset difference range that the reliability difference between the target reliability and the current reliability falls into when the current reliability evaluation index of the target equipment includes the current reliability.

[0077] Specifically, the second reference suggestion can be generated as follows: the range of reliability difference values ​​is pre-divided into multiple preset difference intervals, and each preset difference interval is pre-associated with a maintenance suggestion. When the computer system obtains the target reliability and the current reliability and calculates the reliability difference between the two, it determines which preset difference interval the reliability difference falls into, and uses the maintenance suggestion corresponding to that interval as the second reference suggestion. Subsequently, this second reference suggestion is output as the final reliability maintenance suggestion. The second reference suggestion is selected from multiple candidate suggestions, and its specific content depends on the interval into which the reliability difference falls. Different intervals correspond to different maintenance suggestion types to reflect the degree to which the current reliability of the target equipment meets the target reliability under different reliability differences.

[0078] Understandably, in actual operation and maintenance scenarios, even if operation and maintenance personnel obtain the precise value of the current failure rate, they still need to judge for themselves whether "this value means that immediate repair is needed" and "what level of maintenance measures should be taken". This results in a decision gap between quantitative parameters and maintenance decisions.

[0079] Therefore, this embodiment further obtains the shape parameter of the current failure rate when the current reliability evaluation index includes the current failure rate, and uses the first reference suggestion corresponding to the target parameter interval in multiple preset parameter intervals as the reliability maintenance suggestion; when the current reliability evaluation index includes the current reliability, the target reliability of the target device is obtained, the reliability difference between the target reliability and the current reliability is determined, and the second reference suggestion corresponding to the target difference interval in multiple preset difference intervals is used as the reliability maintenance suggestion.

[0080] Accordingly, in the present embodiment, through the correspondence between the preset parameter intervals and the first reference suggestions as well as the correspondence between the preset difference intervals and the second reference suggestions, the shape parameter of the current failure rate and the reliability difference are respectively mapped to specific maintenance suggestions, thereby avoiding the problem that operation and maintenance personnel still need to manually interpret quantitative parameters and judge maintenance strategies by themselves, enabling the two quantitative parameters of the current failure rate and current reliability of the target equipment to be automatically and accurately converted into directly executable reliability maintenance suggestions, and effectively realizing a full-link quantitative closed loop from monitoring data to maintenance decisions.

[0081] Exemplarily, on the basis that the foregoing solution has determined the current reliability evaluation index, the system judges the specific content included in the current reliability evaluation index: When the current reliability evaluation index includes the current failure rate, the computer system acquires the shape parameter β of the current failure rate from the current reliability evaluation index, compares the shape parameter β with a plurality of preset parameter intervals pre-stored in the system one by one, determines the target parameter interval into which the shape parameter β falls, then reads the first reference suggestion corresponding to the target parameter interval from the system, and outputs the first reference suggestion as a reliability maintenance suggestion. For example, when β>2.5, it indicates that the failure rate of the equipment increases sharply with time, and the corresponding first reference suggestion at this time is to reduce the β value through technical improvement; when 1<β≤2.5, it indicates that the failure rate of the equipment increases with time but at a relatively slow growth rate, and the equipment shows a certain aging trend, and the corresponding first reference suggestion at this time is to strengthen monitoring and conduct reliability risk assessment; when β≤1, it indicates that the failure rate of the equipment decreases with time, and the corresponding first reference suggestion at this time indicates that there is no time-related failure mode; When the current reliability evaluation index includes the current reliability, the computer system acquires the target reliability R_target pre-input by the operation and maintenance personnel, acquires the current reliability R_current from the current reliability evaluation index, calculates the reliability difference RD=R_target-R_current, compares the reliability difference RD with a plurality of preset difference intervals pre-stored in the computer system one by one, determines the target difference interval into which the reliability difference RD falls, then reads the second reference suggestion corresponding to the target difference interval from the computer system, and outputs the second reference suggestion as a reliability maintenance suggestion. For example, when RD≤0, it indicates that the overall reliability of the equipment meets the standard, and the corresponding second reference suggestion at this time indicates that the equipment is in good condition; when 0<RD≤0.05, it indicates that the overall reliability of the equipment is close to the target value, and the corresponding second reference suggestion at this time indicates that the equipment is in a fair condition; when 0.05<RD≤0.10, it indicates that the overall reliability does not meet the standard, and the corresponding second reference suggestion at this time indicates that the equipment has a reliability risk; when RD>0.10, it indicates that the overall reliability is seriously below the standard, and the corresponding second reference suggestion at this time indicates that the equipment has a serious risk; When the current reliability evaluation index includes both the current failure rate and the current reliability, the system independently generates the first reference suggestion and the second reference suggestion according to the two paths mentioned above, and outputs both as reliability maintenance suggestions.

[0082] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 Before step S20, the reliability maintenance suggestion generation method may further include steps S01 to S03: Step S01: Obtain the status detection data of the target device, wherein the status detection data is used to characterize the failure cause of the target device; It should be noted that condition monitoring data refers to monitoring data used to characterize the causes of failure in target equipment. The difference between this condition monitoring data and the first and second monitoring data lies in the fact that the first and second monitoring data are used to determine the potential failure time and power factor (PF) interval, focusing on when failure occurs, while condition monitoring data is used to identify the type of failure cause, focusing on why failure occurs. In practical engineering, condition monitoring data can be various monitoring data generated during the operation of the target equipment that reflect changes in the internal physical and chemical state of the equipment. For example, for oil-immersed transformers, condition monitoring data can be data on the concentration of dissolved gases in the oil (including acetylene concentration, hydrogen concentration, total hydrocarbon content, etc.) and furfural content, which are obtained through periodic oil sample testing or online monitoring devices. For other types of power equipment, condition monitoring data can be the corresponding equipment condition characteristic parameters.

[0083] Step S02: Based on the status detection data, identify the failure cause type of the target device; It should be noted that the failure cause type refers to the category to which the cause of the target equipment failure belongs. In this embodiment, the failure cause types can be divided into two main categories: manufacturing defects and operational stress. Manufacturing defects refer to inherent problems left over from insufficient quality control during the design, production, assembly, or installation of the target equipment, such as material defects, insufficient processing precision, improper assembly, and installation errors. These defects usually become apparent within a certain period after the equipment is put into operation and may manifest in different forms, either overt or covert. Operational stress refers to the damage accumulated by the target equipment during operation due to external operating conditions (such as load size, ambient temperature, electric field strength, etc.), such as thermal aging caused by overload operation and insulation degradation caused by voltage fluctuations. These failure causes are related to the actual operating conditions and operating history of the equipment.

[0084] Step S03: Determine the probability distribution calculation strategy corresponding to the failure cause type, and execute the steps of determining the first probability distribution of the potential failure time based on the first monitoring data and determining the second probability distribution of the time interval based on the second monitoring data based on the probability distribution calculation strategy.

[0085] It should be noted that the probability distribution calculation strategy refers to the pre-determined selection of specific logic or models for calculating the probability distribution for different failure cause types. The core function of the probability distribution calculation strategy is that different failure cause types lead to different distribution characteristics of failures over time. Specifically, manufacturing defects often result in a higher failure probability in the early stages of equipment operation, while operational stress causes the probability factor (PF) interval to change more dynamically with real-time operating conditions. Therefore, it is necessary to select an appropriate calculation strategy based on the failure cause type to ensure that the calculated probability distribution accurately reflects the actual failure behavior of the target equipment under that failure cause type.

[0086] Specifically, when the failure cause is an explicit manufacturing defect, the probability distribution calculation strategy can be determined as the first calculation strategy, which is used to control the first probability distribution to have a higher probability weight in a preset early stage; when the failure cause is a latent manufacturing defect, the probability distribution calculation strategy can be determined as the second calculation strategy, which is used to control the first probability distribution to have a relatively lower probability weight in a preset early stage; when the failure cause is operating stress, the probability distribution calculation strategy can be determined as the third calculation strategy, which is used to control the second probability distribution to dynamically change with the real-time operating parameters of the target equipment.

[0087] It is understandable that if all failure events are assumed to follow the same probability distribution calculation strategy, without considering the impact of the failure cause type on the probability distribution characteristics, the calculated first and second probability distributions will not match the actual failure behavior, thus affecting the accuracy of the final reliability maintenance recommendations.

[0088] Therefore, this embodiment further introduces a failure cause identification and strategy determination mechanism before performing the step of determining the first probability distribution and the second probability distribution. That is, it acquires state detection data to characterize the failure cause of the target device, then identifies the failure cause type of the target device based on the state detection data, finally determines the probability distribution calculation strategy corresponding to the failure cause type, and performs the step of determining the first probability distribution and the second probability distribution based on the strategy.

[0089] Therefore, this embodiment uses failure cause type identification as a prerequisite for probability distribution calculation, allowing the calculation methods for the first and second probability distributions to adapt to different calculation strategies based on the different failure cause types. Manufacturing defects affect the first probability distribution (i.e., the distribution of potential failure times), while operating stress affects the second probability distribution (i.e., the distribution of PF interval times). This avoids the problem of probability distribution calculation results not matching the actual failure behavior of the equipment due to ignoring differences in failure cause types. The resulting reliability maintenance recommendations more accurately reflect the actual failure patterns of the target equipment under its specific failure cause type, further improving the relevance and effectiveness of the maintenance recommendations.

[0090] For example, taking a transformer in a substation as the target device, before executing the steps of determining a first probability distribution based on first monitoring data and a second probability distribution based on second monitoring data, the system first acquires the transformer's condition monitoring data, such as obtaining dissolved gas concentration data in transformer oil through oil sample analysis. Based on the acquired condition monitoring data, the computer system identifies the failure cause type of the transformer. For example, when the transformer has been in operation for a short time (e.g., less than 72 hours) and the acetylene or hydrogen concentration in the oil shows an abnormal increase, the failure cause type can be identified as a manufacturing defect; when the transformer has been in operation for a long time and the condition monitoring data shows a steady-state change trend highly correlated with the load current, the failure cause type can be identified as operating stress.

[0091] After identifying the failure cause type, the system determines the probability distribution calculation strategy corresponding to that failure cause type. For example, when the failure cause type is manufacturing defect, a strategy is determined to be used for subsequent calculation of the first probability distribution and the second probability distribution; when the failure cause type is operating stress, another strategy is determined to be used for subsequent calculation of the first probability distribution and the second probability distribution.

[0092] After determining the probability distribution calculation strategy, the system executes the steps of determining the first probability distribution of potential failure time based on the first monitoring data and the second probability distribution of PF interval time based on the second monitoring data, so that the subsequent probability distribution calculation method can match the failure cause type of the target device.

[0093] In this embodiment, by identifying the failure cause type as a prerequisite for probability distribution calculation, and determining the corresponding probability distribution calculation strategy according to different failure cause types, the subsequently calculated first and second probability distributions can adopt different calculation rules according to different failure cause types. This avoids the problem that the probability distribution calculation results do not match the actual failure behavior of the equipment due to ignoring the differences in failure cause types. As a result, the final reliability maintenance recommendations can more accurately reflect the actual failure pattern of the target equipment under its specific failure cause type, further improving the pertinence and effectiveness of the maintenance recommendations.

[0094] In one feasible implementation, step S02 may include steps S021 to S024: Step S021: Obtain the running time and load current of the target device; It should be noted that runtime refers to the cumulative operating time of the target device from the moment of commissioning to the present moment. In this embodiment, different failure causes often exhibit different manifestations at different stages of device operation: explicit manufacturing defects can usually be identified through status detection data within a very short time after the device is commissioned (i.e., when the runtime is short), while implicit manufacturing defects can only be identified through specific data features after the device has been running for a longer period of time (i.e., when the runtime reaches a certain threshold). Therefore, runtime is the initial judgment condition that all three judgment branches in this embodiment depend on. It can be obtained by reading the device's commissioning record and calculating the time difference with the current moment, or it can be directly read from the device's operation log.

[0095] Load current refers to the magnitude of the current carried by the target equipment during operation, and its value reflects the current load level of the equipment. In this embodiment, load current is an important auxiliary judgment parameter for identifying latent manufacturing defects. That is, when the equipment is in a state of latent manufacturing defects, there is a specific relationship between the rate of change of characteristic parameters and load current, which is specifically a non-linear relationship. Load current can be obtained by directly reading the real-time measurement value of the current transformer on the equipment, or it can be obtained from the operating data of the power monitoring system.

[0096] Step S022: If the running time is less than a preset first duration threshold and the status detection data meets the preset first defect index, the failure cause type is determined to be an explicit manufacturing defect. It should be noted that the preset first defect index refers to a pre-set judgment condition for explicit manufacturing defects, used to determine whether the condition detection data meets the characteristics of explicit manufacturing defects. In this embodiment, the specific content of the preset first defect index may be that the detection value of a certain feature parameter in the condition detection data exceeds a certain preset threshold, or it may be that a combination of multiple feature parameters meets a certain preset logical condition. The specific value of the preset first defect index can be preset based on the type of target equipment, historical fault data, and engineering experience.

[0097] Explicit manufacturing defects refer to manufacturing defects that can be clearly identified through condition monitoring data within a very short time after the target equipment is put into operation. In contrast to latent manufacturing defects, explicit manufacturing defects are characterized by their early manifestation. Because explicit manufacturing defects are often severe or located in critical positions, they cause obvious anomalies in condition monitoring data shortly after the equipment is put into operation, thus allowing them to be identified within a very short operating time. In this embodiment, the identification criteria for explicit manufacturing defects are that the operating time is less than a preset first time threshold and the condition monitoring data meets a preset first defect index. After an explicit manufacturing defect is identified, a first calculation strategy is adopted in subsequent steps to control the calculation method of the first probability distribution, that is, assigning a higher probability weight in the preset early stage.

[0098] Step S023: If the running time is greater than or equal to a preset second duration threshold and the status detection data meets a preset second defect index, the failure cause type is determined to be a latent manufacturing defect. Specifically, if the relationship between the rate of change of the first characteristic parameter and the load current in the status detection data meets a preset nonlinear relationship, and / or the second characteristic parameter in the status detection data meets a preset aging deviation index, and / or the fluctuation of the third characteristic parameter in the status detection data meets a preset abnormal fluctuation index, the status detection data is determined to meet the preset second defect index. It should be noted that the preset second defect index refers to a pre-defined judgment condition for latent manufacturing defects, used to determine whether the condition monitoring data meets the characteristics of latent manufacturing defects. Unlike the preset first defect index, the preset second defect index is not a single judgment condition, but can be composed of three parallel judgment methods: the relationship between the rate of change of the first characteristic parameter in the condition monitoring data and the load current conforms to a preset nonlinear relationship; the second characteristic parameter in the condition monitoring data conforms to a preset aging deviation index; and the fluctuation of the third characteristic parameter in the condition monitoring data conforms to a preset abnormal fluctuation index. If any one or all of the three judgment methods are satisfied, the condition monitoring data can be considered to meet the preset second defect index.

[0099] Latent manufacturing defects refer to manufacturing defects that do not immediately manifest as abnormalities in target equipment after commissioning, but gradually emerge after a relatively long period of operation. In contrast to explicit manufacturing defects, latent manufacturing defects are characterized by their delayed onset and concealment. Because latent manufacturing defects are usually minor or located in hidden areas, they do not cause obvious anomalies in the status monitoring data during the initial stages of equipment commissioning. However, the impact of the defect gradually becomes apparent as operating time accumulates or specific operating conditions (such as increased load or temperature rise) occur. In this embodiment, the identification criteria for latent manufacturing defects are an operating time greater than or equal to a preset second time threshold and status monitoring data conforming to a preset second defect index. After a latent manufacturing defect is identified, a second calculation strategy is adopted in subsequent steps to control the calculation method of the first probability distribution, that is, assigning a relatively low probability weight (i.e., lower than the probability weight corresponding to explicit manufacturing defects) in the preset early stage.

[0100] Here, the first characteristic parameter refers to a specific characteristic parameter in the condition detection data used in the first determination method of latent manufacturing defects. In this embodiment, the specific selection of the first characteristic parameter depends on the type of target equipment and the type of condition detection data available. For oil-immersed transformers, the first characteristic parameter can be the concentration value of a characteristic gas; for other types of power equipment, the first characteristic parameter can be the corresponding equipment condition characteristic quantity.

[0101] The preset nonlinear relationship refers to a preset judgment condition that establishes a nonlinear mathematical relationship between the rate of change of the first characteristic parameter and the load current. In the identification logic of latent manufacturing defects, when equipment has latent manufacturing defects (such as poor contact, loose bolts, etc.), the rate of change of the characteristic parameter and the load current will exhibit a nonlinear relationship; while when the equipment is operating normally, the rate of change of the characteristic parameter and the load current usually exhibit an approximately linear relationship. In this embodiment, the preset nonlinear relationship is one of the three parallel judgment methods for identifying latent manufacturing defects. When the relationship between the rate of change of the first characteristic parameter and the load current conforms to this preset nonlinear relationship, the judgment condition of the preset second defect index may be met, and the failure cause type may be determined as a latent manufacturing defect.

[0102] The second characteristic parameter refers to a specific characteristic parameter in the condition detection data used in the second determination method of latent manufacturing defects. In this embodiment, the specific selection of the second characteristic parameter also depends on the type of target equipment. For oil-immersed transformers, the second characteristic parameter may be the content of a specific component; for other types of power equipment, the second characteristic parameter may be the corresponding equipment condition characteristic quantity.

[0103] The preset aging deviation index refers to a preset judgment condition used to determine whether the aging behavior of equipment deviates from the normal aging pattern. Under normal operating conditions, multiple characteristic parameters usually follow a consistent aging trend; however, when the equipment has latent manufacturing defects, the aging states reflected by different characteristic parameters will show a deviation. In this embodiment, the preset aging deviation index is one of three parallel judgment methods for identifying latent manufacturing defects. When the second characteristic parameter meets this index, it may satisfy one of the judgment conditions of the preset second defect index, and the failure cause type may be determined to be a latent manufacturing defect.

[0104] The third characteristic parameter refers to a specific characteristic parameter in the condition detection data used in the third determination method of latent manufacturing defects. In this embodiment, the selection of the third characteristic parameter also depends on the type of target equipment. For oil-immersed transformers, the third characteristic parameter can be the concentration value of a certain characteristic gas; for other types of power equipment, the third characteristic parameter can be the corresponding equipment condition characteristic quantity.

[0105] The preset abnormal fluctuation index refers to the preset judgment condition used to determine whether the fluctuation of condition monitoring data belongs to abnormal fluctuation. Under normal equipment operation, the change of condition monitoring data over time is usually stable; however, when the equipment has hidden manufacturing defects (such as loose bolts, poor contact, etc.), the condition monitoring data will be affected by mechanical instability and exhibit violent fluctuation characteristics. In this embodiment, the preset abnormal fluctuation index is one of the three parallel judgment methods for identifying hidden manufacturing defects. When the fluctuation of the third characteristic parameter meets the index, it may meet one of the judgment conditions of the preset second defect index, and the failure cause type may be determined to be a hidden manufacturing defect.

[0106] Step S024: If the runtime is greater than or equal to a preset first duration threshold and less than a preset second duration threshold, or if the state detection data does not meet the preset first defect index and the preset second defect index, the failure cause type is determined to be runtime stress.

[0107] It should be noted that operational stress refers to the accumulated damage to the target equipment due to external operating conditions during operation, and is a separate category of failure causes distinct from manufacturing defects. Unlike manufacturing defects, operational stress refers to the cumulative damage caused by operating conditions such as load, temperature, electric field strength, and environmental conditions during normal use, such as thermal aging caused by long-term overload and insulation degradation caused by voltage fluctuations. In this embodiment, the identification method of operational stress is used as a fallback; that is, when the target equipment does not meet the identification conditions of the above-mentioned explicit and implicit manufacturing defects, the failure cause type is determined to be operational stress. After operational stress is identified, a third calculation strategy is adopted in subsequent steps to control the calculation method of the second probability distribution, that is, to control the dynamic change of the second probability distribution with the real-time operating parameters of the target equipment, so as to reflect the dynamic influence of operational stress on the PF interval time distribution.

[0108] It is understood that this embodiment further obtains the running time and load current of the target device, and uses the running time as the initial judgment condition to divide different identification branches. When the running time is less than a preset first time threshold and the status detection data meets the preset first defect index, the failure cause type is determined to be an explicit manufacturing defect. When the running time is greater than or equal to a preset second time threshold and the status detection data meets the preset second defect index, the failure cause type is determined to be a latent manufacturing defect. The preset second defect index is further refined into three parallel judgment methods. When the running time is between the two or the status detection data does not meet the defect index, the failure cause type is determined to be operating stress.

[0109] Therefore, by introducing specific judgment parameters and judgment logic such as running time, load current, and preset first time threshold, preset second time threshold, preset first defect index, and preset second defect index, this embodiment enables the system to automatically identify the three failure cause types—explicit manufacturing defects, implicit manufacturing defects, and operating stress—based on clear numerical values. This ensures the objectivity and consistency of the failure cause type identification results and provides an accurate and reliable input basis for subsequent determination of the corresponding probability distribution calculation strategy based on the failure cause type.

[0110] For example, taking a transformer in a substation as the target device, the system first obtains the transformer's operating time and load current. When the operating time is less than a preset first time threshold (e.g., 72 hours) and the detected value of a certain characteristic parameter in the status detection data exceeds a preset threshold (e.g., acetylene concentration greater than 0.1 μL / L or hydrogen concentration greater than 50 μL / L), it indicates that the transformer has a serious abnormality in the early stage of operation, and the computer system determines the failure cause type as an explicit manufacturing defect. When the runtime is greater than or equal to a preset second duration threshold (e.g., 1 year), the system further performs at least one of the following determinations: Calculate the power law exponent between the rate of change of the first characteristic parameter (e.g., total hydrocarbons) in the condition detection data and the load current. If the power law exponent is greater than or equal to the first preset exponent (e.g., 2.0), it indicates that the characteristic parameter grows non-linearly with the load, conforming to the preset non-linear relationship, and the failure cause type is determined to be a latent manufacturing defect; or, when the detected value of the second characteristic parameter (e.g., total hydrocarbon content) in the condition detection data is greater than a preset multiple (e.g., 2 times) of the standard baseline and the detected value of the third characteristic parameter (e.g., furfural content) is less than the third preset threshold (e.g., 0.1 mg / L), it indicates that the aging state of the oil and paper deviates, conforming to the preset aging deviation index, and the failure cause type is determined to be a latent manufacturing defect; or, calculate the coefficient of variation of the detected value of a certain characteristic parameter in the condition detection data within a preset time period (e.g., within 30 days). If the coefficient of variation is greater than the first preset coefficient (e.g., 0.5), it indicates that the data fluctuates drastically, conforming to the preset abnormal fluctuation index, and the failure cause type is determined to be a latent manufacturing defect. When the runtime is between a preset first runtime threshold and a preset second runtime threshold, or when the status detection data does not meet the above-mentioned identification conditions for explicit manufacturing defects and implicit manufacturing defects, the system will determine the failure cause type as operating stress.

[0111] In one feasible implementation, step S03 may include steps S031 to S033: Step S031: When the failure cause type is the explicit manufacturing defect, the probability distribution calculation strategy is determined to be the first calculation strategy, wherein the first calculation strategy is used to control the probability weight of the first probability distribution to be greater than the preset probability weight in a preset early stage. It should be noted that the first calculation strategy refers to the probability distribution calculation strategy corresponding to the failure cause type being identified as an explicit manufacturing defect. Its core function is to control the relative magnitude between the probability weight of the first probability distribution in the preset early stage and the preset probability weight. In this embodiment, the characteristic of an explicit manufacturing defect is that the equipment exhibits obvious abnormal conditions within a very short time after commissioning. Therefore, its failure probability should be assigned a higher weight in the preset early stage to accurately reflect the physical characteristic that the explicit manufacturing defect leads to a significant increase in the early failure probability. By controlling the probability weight of the first probability distribution in the preset early stage to be greater than the preset probability weight, the first calculation strategy enables the first probability distribution calculated based on this strategy to more accurately reflect the actual distribution pattern of potential failure events concentrated in the early stage under the explicit manufacturing defect situation, thereby providing an accurate first probability distribution input for subsequent fusion and maintenance suggestion generation.

[0112] The preset early stage refers to a specific time interval starting from the commissioning time of the target equipment. The length of this time interval is preset by the operation and maintenance personnel based on the type of target equipment, historical fault statistics, and engineering experience. In this embodiment, the preset early stage is a time dimension that distinguishes between explicit manufacturing defects and implicit manufacturing defects. That is, explicit manufacturing defects usually manifest as obvious abnormalities within this stage, while implicit manufacturing defects usually do not show obvious abnormalities within this stage.

[0113] The preset probability weight refers to a pre-set probability weight reference value used as a comparison benchmark in this technical solution. In this embodiment, the preset probability weight is used to compare with the actual probability weight of the preset early stage to distinguish the calculation strategies corresponding to different failure cause types. That is, when there is an explicit manufacturing defect, the probability weight of the preset early stage needs to be greater than the preset probability weight; when there is a latent manufacturing defect, the probability weight of the preset early stage needs to be less than or equal to the preset probability weight. The specific value of the preset probability weight depends on the type of target equipment, the benchmark probability weight obtained by fitting the Weibull distribution, and the settings made by the operation and maintenance personnel based on engineering experience and historical fault data. Its function is to enable the probability weights of explicit and latent manufacturing defects in the preset early stage to be quantitatively and comparatively distinguished.

[0114] Step S032, when the failure cause type is the latent manufacturing defect, the probability distribution calculation strategy is determined to be the second calculation strategy, wherein the second calculation strategy is used to control the probability weight of the first probability distribution in a preset early stage to be less than or equal to the preset probability weight. It should be noted that the second calculation strategy refers to the probability distribution calculation strategy corresponding to the failure cause type being identified as a latent manufacturing defect. Its core function is to control the relative magnitude between the probability weight of the first probability distribution in the preset early stage and the preset probability weight. In this embodiment, the characteristic of a latent manufacturing defect is that the equipment does not immediately exhibit an obvious abnormal state after commissioning, but gradually reveals itself after a relatively long period of operation. Therefore, although its failure probability may still be higher than that of the normal operating state in the preset early stage, it should not be given the same high weight as an explicit manufacturing defect. The second calculation strategy controls the probability weight of the first probability distribution in the preset early stage to be less than or equal to the preset probability weight, so that the first probability distribution calculated based on this strategy can accurately reflect the actual distribution law that the failure probability in the case of a latent manufacturing defect is relatively low in the early stage, but gradually increases with the passage of operating time.

[0115] Step S033: When the failure cause type is the operating stress, determine the probability distribution calculation strategy as a third calculation strategy, wherein the third calculation strategy is used to control the second probability distribution to dynamically change with the real-time operating parameters of the target device.

[0116] It should be noted that the third calculation strategy refers to the probability distribution calculation strategy corresponding to the failure cause type being identified as operating stress. Its core function is to control the dynamic change of the second probability distribution with the real-time operating parameters of the target equipment. In this embodiment, the characteristic of operating stress is that the failure does not originate from the inherent defects of the equipment itself, but is caused by the cumulative effect of operating conditions. Therefore, the distribution of the PF interval time should not be fixed, but should change dynamically with the real-time operating parameters of the equipment (such as the current load size, operating temperature, etc.). By associating the second probability distribution with the real-time operating parameters, the third calculation strategy enables the second probability distribution subsequently calculated based on this strategy to adaptively adjust its distribution shape according to the current operating state of the target equipment, thereby accurately reflecting the actual physical law of the dynamic change of the PF interval time with the operating conditions under operating stress.

[0117] Real-time operating parameters refer to physical quantities that reflect the immediate operating conditions of the target equipment under its current operating state. These parameters change dynamically with the operating state of the equipment. In this embodiment, the specific type of real-time operating parameters depends on the type of target equipment. For transformers, real-time operating parameters may include the current load current value, top oil temperature value, ambient temperature value, etc.; for other types of power equipment, real-time operating parameters may include corresponding operating condition parameters such as current, voltage, and temperature. The role of real-time operating parameters in the third calculation strategy is to serve as the basis for the dynamic adjustment of the second probability distribution, enabling the second probability distribution to adaptively adjust with changes in the actual operating state of the target equipment, thereby reflecting the different impacts of operating stress on the PF interval time distribution under different operating conditions.

[0118] It is understood that this embodiment matches a specific calculation strategy for each failure cause type, that is, transforms the failure cause type category label into an executable probability distribution calculation logic. This includes different probability weight control rules for explicit and implicit manufacturing defects to adapt to the different performance intensities of manufacturing defects in the early stages, and dynamic adjustment rules for operating stress to adapt to the characteristics of PF interval time changing with operating conditions. As a result, the identification results of failure cause types can be directly used to guide the probability distribution calculation process, ensuring that the appropriate probability distribution calculation strategy can be automatically and accurately selected after the failure cause type is identified, forming a complete automated decision chain from failure cause identification to probability distribution calculation to maintenance suggestion generation.

[0119] For example, based on the aforementioned scheme having already identified the failure cause type as any one of explicit manufacturing defects, implicit manufacturing defects, or operating stress, the system further determines the probability distribution calculation strategy corresponding to that failure cause type: When the failure cause is an explicit manufacturing defect, the system determines the probability distribution calculation strategy as the first calculation strategy. This strategy controls the probability weight of the first probability distribution in the preset early stage to be greater than the preset probability weight. Since explicit manufacturing defects exhibit obvious anomalies in the early stages of equipment operation, by setting the probability weight of the preset early stage to be higher than the preset probability weight during the calculation of the first probability distribution, the calculated first probability distribution has a higher probability density in the preset early stage, thereby accurately reflecting the actual distribution pattern of potential failure events concentrated in the early stage under the case of explicit manufacturing defects. When the failure cause is a latent manufacturing defect, the system determines the probability distribution calculation strategy as the second calculation strategy. This strategy is used to control the probability weight of the first probability distribution in the preset early stage to be less than or equal to the preset probability weight. Since latent manufacturing defects do not immediately show obvious abnormalities in the early stage of equipment operation, by setting the probability weight of the preset early stage to no higher than the preset probability weight during the calculation of the first probability distribution, the calculated first probability distribution maintains a relatively low weight in the preset early stage, thereby accurately reflecting the actual distribution law that the failure probability is relatively low in the early stage and gradually increases with the passage of time in the case of latent manufacturing defects. When the failure cause is operational stress, the system determines the probability distribution calculation strategy as the third calculation strategy. This strategy is used to control the dynamic changes of the second probability distribution with the real-time operating parameters of the target equipment. Since failure caused by operational stress is not due to inherent defects in the equipment, but is dynamically affected by operating conditions, by introducing real-time operating parameters as dynamic variables in the calculation process of determining the second probability distribution, the calculated second probability distribution can adaptively adjust its distribution shape according to changes in current operating parameters such as load and temperature. This accurately reflects the actual physical law of the dynamic changes of the PF interval time with operating conditions under operational stress.

[0120] For example, to help understand the implementation flow of the reliability maintenance suggestion generation method obtained by combining this embodiment with the above embodiment one, please refer to... Figure 3 , Figure 3 A simplified flowchart of a method for generating reliability maintenance recommendations is provided, specifically: First, condition monitoring data is acquired and failure cause types are identified, thereby determining the corresponding probability distribution calculation strategy. Then, a first probability distribution and a second probability distribution are determined based on the first and second monitoring data, respectively. The first probability distribution describes the distribution pattern of potential failure times of the target equipment, while the second probability distribution describes the distribution pattern of the time interval between potential failure times and functional failure times. The first and second probability distributions are convolved and fused to obtain a third probability distribution describing the distribution pattern of functional failure times. Finally, the current reliability evaluation index is determined based on the third probability distribution, and reliability maintenance recommendations are generated based on the current reliability evaluation index. The identification result of the failure cause type determines the selection of the probability distribution calculation strategy; that is, explicit manufacturing defects and implicit manufacturing defects correspond to different calculation strategies applied to the first probability distribution, and operating stress corresponds to the calculation strategy applied to the second probability distribution, ensuring that the calculated probability distribution reflects the impact of the failure cause type on the equipment failure behavior. The convolutional fusion is based on the physical logic that the potential failure time and the PF interval are independent, and the functional failure time is the sum of the two. The current reliability evaluation index includes the current failure rate and / or the current reliability, which are used to determine the corresponding maintenance recommendations. The entire process realizes a complete link from data collection, cause identification, distribution modeling to maintenance suggestion output.

[0121] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the reliability maintenance suggestion generation method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0122] This application also provides a reliability maintenance suggestion generation device, please refer to... Figure 4 The reliability maintenance recommendation generation device includes: The data acquisition module 10 is used to acquire first monitoring data and second monitoring data of the target device, wherein the first monitoring data is used to characterize the potential failure time of the target device, and the second monitoring data is used to characterize the time interval between the potential failure time and the functional failure time of the target device. The probability calculation module 20 is used to determine a first probability distribution of the potential failure time based on the first monitoring data, and to determine a second probability distribution of the time interval based on the second monitoring data. The probability fusion module 30 is used to fuse the first probability distribution and the second probability distribution to obtain the third probability distribution at the moment of functional failure. The suggestion generation module 40 is used to determine the current reliability evaluation index of the target device based on the third probability distribution, and generate a reliability maintenance suggestion for the target device based on the current reliability evaluation index.

[0123] Optionally, the probability calculation module 20 is further configured to: Identify potential failure events of the target device and the first occurrence time of the potential failure events from the first monitoring data, and use the first occurrence time as the potential failure time; A first time series is generated based on the potential failure time and the historical potential failure time corresponding to the target device, wherein the historical potential failure time is the occurrence time of at least one historical potential failure event under the device category to which the target device belongs; The first probability distribution is obtained by fitting the first time series with a Weibull distribution.

[0124] Optionally, the probability calculation module 20 is further configured to: From the second monitoring data, identify potential failure events, functional failure events, the first occurrence time of the potential failure event and the second occurrence time of the functional failure event of the target device, and use the duration between the first occurrence time and the second occurrence time as the time interval. A second time series is generated based on the time interval and the historical time interval corresponding to the target device, wherein the historical time interval is the duration between at least one historical potential failure event and the corresponding historical functional failure event under the device category to which the target device belongs; The second probability distribution is obtained by fitting the second time series with a Weibull distribution.

[0125] Optionally, the probability fusion module 30 is further configured to: Convolve the first probability density function corresponding to the first probability distribution and the second probability density function corresponding to the second probability distribution to obtain the third probability density function. The third probability density function is taken as the third probability distribution.

[0126] Optionally, the current reliability evaluation indicators include the current failure rate and / or the current reliability, and the suggestion generation module 40 is further configured to: Extract the shape parameters of the third probability distribution and use the shape parameters as the current failure rate; and / or, Extract the shape parameter and scale parameter of the third probability distribution, and calculate the current reliability based on the shape parameter and the scale parameter, wherein the shape parameter is negatively correlated with the current reliability, and the scale parameter is positively correlated with the current reliability.

[0127] Optionally, the suggestion generation module 40 is further configured to: When the current reliability evaluation index includes the current failure rate, the shape parameter of the current failure rate is obtained, and the first reference suggestion corresponding to the target parameter interval in multiple preset parameter intervals is used as the reliability maintenance suggestion; When the current reliability evaluation index includes the current reliability, the target reliability of the target device is obtained, the reliability difference between the target reliability and the current reliability is determined, and the second reference suggestion corresponding to the target difference interval in multiple preset difference intervals is used as the reliability maintenance suggestion.

[0128] Optionally, the reliability maintenance recommendation generation device further includes a strategy determination module 50, which is used to: Acquire the status detection data of the target device, wherein the status detection data is used to characterize the failure cause of the target device; Based on the status detection data, identify the failure cause type of the target device; A probability distribution calculation strategy corresponding to the failure cause type is determined, and the steps of determining the first probability distribution of the potential failure time based on the first monitoring data and determining the second probability distribution of the time interval based on the second monitoring data are executed based on the probability distribution calculation strategy.

[0129] Optionally, the strategy determination module 50 is further configured to: Obtain the running time and load current of the target device; If the running time is less than a preset first time threshold and the status detection data meets the preset first defect index, the failure cause type is determined to be an explicit manufacturing defect. If the runtime is greater than or equal to a preset second duration threshold and the status detection data meets a preset second defect index, the failure cause type is determined to be a latent manufacturing defect. Specifically, if the relationship between the rate of change of the first characteristic parameter and the load current in the status detection data meets a preset nonlinear relationship, and / or the second characteristic parameter in the status detection data meets a preset aging deviation index, and / or the fluctuation of the third characteristic parameter in the status detection data meets a preset abnormal fluctuation index, the status detection data is determined to meet the preset second defect index. If the runtime is greater than or equal to a preset first duration threshold and less than a preset second duration threshold, or if the state detection data does not meet the preset first defect index and the preset second defect index, the failure cause type is determined to be runtime stress.

[0130] Optionally, the strategy determination module 50 is further configured to: When the failure cause type is the explicit manufacturing defect, the probability distribution calculation strategy is determined to be a first calculation strategy, wherein the first calculation strategy is used to control the probability weight of the first probability distribution to be greater than the preset probability weight in a preset early stage. When the failure cause type is the latent manufacturing defect, the probability distribution calculation strategy is determined to be the second calculation strategy, wherein the second calculation strategy is used to control the probability weight of the first probability distribution in a preset early stage to be less than or equal to the preset probability weight. When the failure cause type is the operating stress, the probability distribution calculation strategy is determined to be a third calculation strategy, wherein the third calculation strategy is used to control the second probability distribution to dynamically change with the real-time operating parameters of the target device.

[0131] The reliability maintenance suggestion generation device provided in this application, employing the reliability maintenance suggestion generation method in the above embodiments, can solve the technical problem of poor efficiency and accuracy in the maintenance and management of power equipment under the current RCM theory. Compared with the prior art, the beneficial effects of the reliability maintenance suggestion generation device provided in this application are the same as those of the reliability maintenance suggestion generation method provided in the above embodiments, and other technical features in the reliability maintenance suggestion generation device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0132] This application provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the reliability maintenance suggestion generation method in Embodiment 1 above.

[0133] The following is for reference. Figure 5 The diagram illustrates a structural schematic of an electronic device suitable for implementing embodiments of this application. The electronic devices in these embodiments may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), and fixed terminals such as desktop computers. Figure 5 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0134] like Figure 5As shown, the electronic device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the electronic device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. The communication device 1009 allows the electronic device to communicate wirelessly or wiredly with other devices to exchange data. Although the diagrams show electronic devices with various systems, it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems may be implemented alternatively.

[0135] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0136] The electronic device provided in this application, employing the reliability maintenance suggestion generation method described in the above embodiments, can solve the technical problem of poor efficiency and accuracy in the maintenance and management of power equipment under the current RCM theory. Compared with the prior art, the beneficial effects of the electronic device provided in this application are the same as those of the reliability maintenance suggestion generation method provided in the above embodiments, and other technical features of this electronic device are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0137] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0138] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0139] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for generating reliability maintenance recommendations, characterized in that, The method for generating reliability maintenance recommendations includes: Acquire first monitoring data and second monitoring data of the target device, wherein the first monitoring data is used to characterize the potential failure time of the target device, and the second monitoring data is used to characterize the time interval between the potential failure time and the functional failure time of the target device; A first probability distribution of the potential failure time is determined based on the first monitoring data, and a second probability distribution of the time interval is determined based on the second monitoring data; By combining the first probability distribution and the second probability distribution, a third probability distribution is obtained for the moment when the function fails. Based on the third probability distribution, the current reliability evaluation index of the target equipment is determined, and based on the current reliability evaluation index, a reliability maintenance suggestion for the target equipment is generated.

2. The reliability maintenance suggestion generation method as described in claim 1, characterized in that, The step of determining the first probability distribution of the potential failure time based on the first monitoring data includes: Identify potential failure events of the target device and the first occurrence time of the potential failure events from the first monitoring data, and use the first occurrence time as the potential failure time; A first time series is generated based on the potential failure time and the historical potential failure time corresponding to the target device, wherein the historical potential failure time is the occurrence time of at least one historical potential failure event under the device category to which the target device belongs; The first probability distribution is obtained by fitting the first time series with a Weibull distribution.

3. The reliability maintenance suggestion generation method as described in claim 1, characterized in that, The step of determining the second probability distribution of the time interval based on the second monitoring data includes: From the second monitoring data, identify potential failure events, functional failure events, the first occurrence time of the potential failure event and the second occurrence time of the functional failure event of the target device, and use the duration between the first occurrence time and the second occurrence time as the time interval. A second time series is generated based on the time interval and the historical time interval corresponding to the target device, wherein the historical time interval is the duration between at least one historical potential failure event and the corresponding historical functional failure event under the device category to which the target device belongs; The second probability distribution is obtained by fitting the second time series with a Weibull distribution.

4. The reliability maintenance suggestion generation method as described in claim 1, characterized in that, The step of fusing the first probability distribution and the second probability distribution to obtain the third probability distribution at the moment of functional failure includes: Convolve the first probability density function corresponding to the first probability distribution and the second probability density function corresponding to the second probability distribution to obtain the third probability density function. The third probability density function is taken as the third probability distribution.

5. The reliability maintenance suggestion generation method as described in claim 1, wherein the current reliability evaluation index includes the current failure rate and / or the current reliability, and the step of determining the current reliability evaluation index of the target equipment according to the third probability distribution includes: Extract the shape parameters of the third probability distribution, and use the shape parameters as the current failure rate; And / or, Extract the shape parameter and scale parameter of the third probability distribution, and calculate the current reliability based on the shape parameter and the scale parameter, wherein the shape parameter is negatively correlated with the current reliability, and the scale parameter is positively correlated with the current reliability.

6. The reliability maintenance suggestion generation method as described in claim 5, characterized in that, The step of generating reliability maintenance recommendations for the target equipment based on the current reliability evaluation indicators includes: When the current reliability evaluation index includes the current failure rate, the shape parameter of the current failure rate is obtained, and the first reference suggestion corresponding to the target parameter interval in multiple preset parameter intervals is used as the reliability maintenance suggestion; When the current reliability evaluation index includes the current reliability, the target reliability of the target device is obtained, the reliability difference between the target reliability and the current reliability is determined, and the second reference suggestion corresponding to the target difference interval in multiple preset difference intervals is used as the reliability maintenance suggestion.

7. The reliability maintenance suggestion generation method as described in claim 1, characterized in that, Before the steps of determining the first probability distribution of the potential failure time based on the first monitoring data and determining the second probability distribution of the time interval based on the second monitoring data, the method further includes: Acquire the status detection data of the target device, wherein the status detection data is used to characterize the failure cause of the target device; Based on the status detection data, identify the failure cause type of the target device; A probability distribution calculation strategy corresponding to the failure cause type is determined, and the steps of determining the first probability distribution of the potential failure time based on the first monitoring data and determining the second probability distribution of the time interval based on the second monitoring data are executed based on the probability distribution calculation strategy.

8. The reliability maintenance suggestion generation method as described in claim 7, characterized in that, The step of identifying the failure cause type of the target device based on the status detection data includes: Obtain the running time and load current of the target device; If the running time is less than a preset first time threshold and the status detection data meets the preset first defect index, the failure cause type is determined to be an explicit manufacturing defect. If the runtime is greater than or equal to a preset second duration threshold and the status detection data meets a preset second defect index, the failure cause type is determined to be a latent manufacturing defect. Specifically, if the relationship between the rate of change of the first characteristic parameter and the load current in the status detection data meets a preset nonlinear relationship, and / or the second characteristic parameter in the status detection data meets a preset aging deviation index, and / or the fluctuation of the third characteristic parameter in the status detection data meets a preset abnormal fluctuation index, the status detection data is determined to meet the preset second defect index. If the runtime is greater than or equal to a preset first duration threshold and less than a preset second duration threshold, or if the state detection data does not meet the preset first defect index and the preset second defect index, the failure cause type is determined to be runtime stress.

9. The reliability maintenance suggestion generation method as described in claim 8, characterized in that, The steps for determining the probability distribution calculation strategy corresponding to the failure cause type include: When the failure cause type is the explicit manufacturing defect, the probability distribution calculation strategy is determined to be a first calculation strategy, wherein the first calculation strategy is used to control the probability weight of the first probability distribution to be greater than the preset probability weight in a preset early stage. When the failure cause type is the latent manufacturing defect, the probability distribution calculation strategy is determined to be the second calculation strategy, wherein the second calculation strategy is used to control the probability weight of the first probability distribution in a preset early stage to be less than or equal to the preset probability weight. When the failure cause type is the operating stress, the probability distribution calculation strategy is determined to be a third calculation strategy, wherein the third calculation strategy is used to control the second probability distribution to dynamically change with the real-time operating parameters of the target device.

10. An electronic device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the reliability maintenance recommendation generation method as described in any one of claims 1 to 9.

Images