Protection measurement and control device power supply fault early warning method and system, and storage medium

Abstract

The invention discloses a protection measurement and control device power supply fault early warning method and system and a storage medium. The invention aims to solve the problem that fault early warning and detection cannot be accurately carried out when power supply fault early warning is carried out by monitoring ripple change in the prior art. The method comprises the following steps: establishing a transition mapping relation from a normal state to a fault state according to physical parameter characteristic changes in a historical fault scene; respectively acquiring each physical parameter characteristic according to a set sampling frequency, and respectively adjusting each sampling frequency according to the sampled physical parameter characteristics; the sampled physical parameter characteristics are imported into a power supply fault early warning model, and a fault scene of the power supply of the measurement and control device is matched; and executing preset processing preparation according to the matched and predicted fault scene. And a power supply fault early warning and evaluation model is established according to the transition mapping relationship, so that the fault occurrence precursor can be found as early as possible, and the power supply fault early warning and detection cannot be accurately carried out.

Description

Technical Field

[0001] The present invention relates to the field of power supply status monitoring, and in particular to a power supply failure early warning method, system and storage medium for a protection measurement and control device. Background Art

[0002] The protection and measurement device (i.e., protection and measurement and control device) is an important barrier to ensure the safe and stable operation of the power system. "Abnormal alarm of the protection and measurement device" is a common secondary emergency defect. When an abnormal alarm occurs in the protection and measurement device of the 10kV outgoing line of the substation, the operator needs to rush to the scene quickly, turn the line to the opposite side for power supply, and shut down the 10kV switch on this side. It takes about 2-3 hours. Once a fault occurs on the line during this period, the switch on the 10kV side of the main transformer will trip, causing the 10kV busbar and its outgoing lines to all be out of power, seriously expanding the scope of the power outage.

[0003] According to statistics, most of the abnormal alarms of the protection device are caused by the failure of the power module of the protection device. At present, the power supply of domestic protection manufacturers adopts flyback converter, forward converter, push-pull, half-bridge, full-bridge and other topological structures. The filter electrolytic capacitors on the input and output sides of the power module are the weakest link of the entire power module. If the operating temperature of the power module is too high, the electrolytic capacitor will bulge, thereby reducing the capacitance value, and eventually causing the output voltage (24V, 5V) of the power board to be unstable, causing the protection device to freeze and fail to operate normally, and even causing the capacitor to explode, threatening the overall operation safety of the protection device. In addition, the MOSFET switch, inductor, transformer, PWM drive module and other parts in various power modules may fail. Therefore, it is necessary to monitor the power supply operation status of the protection device in real time, warn the protection device power supply operation abnormality, realize the proactive judgment of defects in advance, and avoid the shutdown of primary equipment due to abnormal protection device.

[0004] At present, power supply fault warning is mostly achieved by monitoring ripple changes. For example, a "fault warning method for controlling power supply system" disclosed in Chinese patent literature, with announcement number CN113433480A, includes: sampling the output voltage ripple data of the control power supply system; analyzing the sampled historical voltage ripple data to predict the switch pulse in the current switching cycle; judging whether the switch device has a fault and determining the fault type based on the switch pulse of the current switching cycle and the voltage ripple data actually sampled in the current switching cycle. However, the ripple amplitude is small and the frequency is high. At the same time, it will be affected by electromagnetic interference, which is extremely difficult to detect accurately, making it impossible to accurately perform fault warning and detection. Summary of the invention

[0005] The present invention mainly solves the problem that the prior art cannot accurately warn and detect power supply faults by monitoring ripple changes; and provides a power supply fault warning method, system and storage medium for a protection measurement and control device.

[0006] The above technical problems of the present invention are mainly solved by the following technical solutions: A protection and measurement control device power failure early warning method comprises the following steps: S1: Establish a transition mapping relationship from normal state to fault state based on the physical parameter characteristic changes of the power supply of the protection device under the historical fault scenarios, so as to construct a power supply fault warning model for each fault scenario; S2: obtaining physical parameter characteristics of the power supply of the measurement and control device at a set sampling frequency, and adjusting the sampling frequencies according to the physical parameter characteristics obtained by sampling; S3: Import the sampled physical parameter characteristics into the power supply fault warning model to match the fault scenario of the power supply of the measurement and control device; S4: Execute preset processing preparation according to the matching predicted fault scenario.

[0007] This solution collects statistics on the fault scenarios that are prone to occur in the power supplies of different manufacturers and models of security and testing devices, and extracts the main faults and characteristics of the power supplies of security and testing devices. The electrical characteristic quantities, temperature changes, noise spectra, etc. under each fault scenario are analyzed to explore the nonlinear transfer law of the physical parameter characteristics of the power supply of the security and testing device under fault conditions, so as to establish a "fault transition-parameter characteristic" mapping relationship from normal to fault to form a mapping set. Based on the mapping relationship, a power supply fault warning evaluation model is established, which can detect the precursors of faults early, and the power supply fault warning cannot accurately carry out fault warning and detection.

[0008] Preferably, the step S1 specifically includes the following process: S101: Counting the changes in physical parameter characteristics of different power supplies of different protection and measurement devices under various fault scenarios, and establishing a corresponding "fault transition-parameter characteristic" mapping relationship from a normal state to a fault state; S102: According to the “fault transition-parameter feature” mapping relationship under different fault scenarios, the change characteristics of the physical parameter features are extracted, and a power supply fault warning evaluation function is designed.

[0009] Preferably, the “fault transition-parameter characteristic” mapping relationship is a change curve of the physical parameter characteristic during the fault transition, with the horizontal axis being time and the vertical axis being the physical parameter characteristic.

[0010] The "fault transition-parameter feature" mapping relationship can be used to assist in quickly and accurately identifying the fault type of the power supply of the test device, and taking corresponding measures to repair or maintain it. At the same time, this mapping relationship can also provide a valuable reference for the design and manufacture of the power supply of the test device, thereby improving the stability and reliability of the power supply.

[0011] Preferably, the physical parameter characteristics include electrical characteristics, temperature characteristics and operating noise spectrum; For different fault scenarios, the physical parameter features corresponding to the occurrence or impending occurrence of a fault are selected according to the nonlinear change characteristics of the physical parameter features in the mapping relationship; According to the selected physical parameter characteristics and the order of their nonlinear changes as the function starting conditions, the change rate, amplitude characteristics and change trend shape of the physical parameter characteristics are weighted calculated to design the power supply fault warning evaluation function.

[0012] The physical parameter characteristics that can be used to characterize the fault characteristics are different in each fault scenario. Therefore, for each different fault scenario, the corresponding physical parameter characteristics that can characterize the occurrence or impending occurrence of a fault are selected according to the nonlinear change characteristics of the physical parameter characteristics in the "fault transition-parameter characteristic" mapping relationship.

[0013] Preferably, in the power supply fault evaluation function, a Boolean nonlinear marker is used to determine whether there is a nonlinear change in the physical parameter characteristics selected for the corresponding fault scenario; a Boolean occurrence order marker is used to determine whether the order of the nonlinear changes in each physical parameter characteristic is consistent with the fault characteristics.

[0014] Preferably, the trend shape is compared with each physical parameter feature in the mapping relationship, and the shape coincidence rate is used as an evaluation value, which is weighted to evaluate the matching fault scenario. The higher the shape coincidence rate, the greater the possibility of the fault.

[0015] Preferably, for the change rate and amplitude, the maximum and minimum values ​​of the change rate and amplitude of the physical parameter characteristics in the mapping relationship are respectively selected as the range of the corresponding physical parameter characteristic change rate and amplitude; it is determined whether the change rate and amplitude of the sampled physical parameter characteristics are within the range of the physical parameter characteristic change rate and amplitude. If so, they are weighted and used to evaluate the matching fault scenario, otherwise they are directly taken as 0.

[0016] Preferably, the change rate and amplitude average value of each physical parameter feature in the mapping relationship are used as the standard, and the difference between the change rate and amplitude of the sampled physical parameter feature and the change rate and amplitude average value of the physical parameter feature in the mapping relationship is calculated as the evaluation calculation. The smaller the difference between the change rate of the physical parameter feature and the average value of the change rate of the physical parameter feature in the mapping set, the greater the possibility that this is the fault.

[0017] Preferably, different sampling frequencies are set for different physical parameter characteristics, and upper and lower limits of the sampling frequencies are set respectively.

[0018] Since different physical parameter characteristics have different rates and magnitudes of change during actual use, in order to maximize data sampling efficiency, that is, to sample as many valid data points as possible, the historical sampling data is analyzed through a big data model, and appropriate sampling frequencies and corresponding upper and lower limits are set for each physical parameter characteristic.

[0019] Preferably, the electrical characteristic sampling frequency is coupled and adjusted according to the sampled temperature data, noise spectrum data and electrical characteristic data.

[0020] Preferably, when the detected electrical characteristic data does not match all electrical characteristic change curves in the “fault transition-parameter characteristic” mapping relationship, the expression for adjusting the electrical characteristic sampling frequency is: f E =f Emin +(ε T γ T +ε Q γ Q )·Δf E When the detected electrical characteristic data matches any electrical characteristic change curve in the "fault transition-parameter characteristic" mapping relationship, the expression for adjusting the electrical characteristic sampling frequency is: f E =f Emax Furthermore, set the limiting conditions: f Emin +(γ T +γ Q )·Δf E ≤f Emax Among them, f E is the electrical characteristics sampling frequency; f Emax is the preset maximum value of the electrical characteristic sampling frequency; f Emin is the preset minimum value of the electrical characteristic sampling frequency; ε T It is a temperature characteristic fault precursor mark. When the detected temperature characteristic matches the temperature characteristic change curve in the "fault transition-parameter characteristic" mapping relationship, it takes 1; otherwise, it takes 0; ε QIt is a noise spectrum fault precursor marker. When the detected noise spectrum matches the noise spectrum change curve in the "fault transition-parameter feature" mapping relationship, it takes 1; otherwise, it takes 0. γ T is the correlation coefficient between temperature characteristics and electrical characteristics; γ Q is the correlation coefficient between the noise spectrum and the electrical characteristics; Δf E It is the minimum adjustment frequency of the electrical characteristic sampling frequency.

[0021] Preferably, the temperature sampling frequency is corrected according to the collector current rising rate of the switching device, and the expression for correcting the temperature sampling frequency is: The constraints are: Among them, f T is the sampling frequency of temperature characteristics; f Tmin is the lower limit of the set temperature characteristic sampling frequency; f Tmax The upper limit of the set temperature characteristic sampling frequency; Δf T The minimum sampling adjustment frequency for temperature characteristics; R IS is the collector current rise rate of the switching device obtained by sampling; R IA It is the standard value of the collector current rise rate of the set switching device.

[0022] The module temperature sampling and correction mechanism is integrated with the collector current rise rate of the switching device in order to more accurately reflect the internal junction temperature of the device and obtain the precursor of device failure.

[0023] Preferably, for the running noise spectrum characteristics, the noise spectrum sampling frequency is adjusted according to the comparison between the width of the sampled noise spectrum and the width of the noise spectrum during normal operation.

[0024] A protection and control device power failure early warning system, comprising A historical fault database stores the physical parameter characteristic data of each fault scenario of the power supply of the measurement and control device in the time period before and after the historical fault; A mapping set storing a corresponding "fault transition-parameter characteristic" mapping relationship from a normal state to a fault state established according to the physical parameter characteristic changes of the power supply of different protection and measurement devices under various fault scenarios; The sampling module sets different sampling frequencies for different physical parameter characteristics, and sets upper and lower limits of the sampling frequencies respectively; and adjusts the sampling frequency of the physical parameter characteristics according to the data coupling obtained by sampling; The analysis and calculation module constructs a power supply fault warning model for each fault scenario based on the mapping relationship, imports the sampled physical parameter characteristics into the power supply fault warning model, and matches the fault scenario of the power supply of the measurement and control device; The processing module sets different processing methods for different power failure scenarios of the measurement and control device.

[0025] Preferably, the analysis and calculation module includes a deep convolutional neural network, and the change curves of each physical parameter characteristic obtained by sampling are input into the trained deep convolutional neural network to calculate the overlap rate of the change curves of each physical parameter characteristic and the trend shape of the physical parameter characteristic change curve corresponding to the mapping relationship in the mapping set.

[0026] Preferably, the analysis and calculation module sorts the average values ​​of the coincidence rates of the characteristic change curves of the physical parameters from large to small, and respectively introduces them into the power supply fault warning evaluation function of the corresponding fault scenario to obtain the corresponding fault warning evaluation value.

[0027] A storage medium is used for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of a power failure early warning method for a protection measurement and control device.

[0028] The beneficial effects of the present invention are: 1. Statistics are collected on the fault scenarios that are prone to occur in the power supplies of different manufacturers and models of security devices, and the main faults and their characteristics of the power supplies of security devices are extracted to establish the "fault transition-parameter characteristics" mapping relationship from normal to faulty. The "fault transition-parameter characteristics" mapping relationship can be used to assist in quickly and accurately identifying the fault type of the power supply of the security device, and take corresponding measures to repair or maintain it. At the same time, this mapping relationship can also provide a valuable reference for the design and manufacture of the power supply of the security device, thereby improving the stability and reliability of the power supply.

[0029] 2. A power supply fault warning evaluation model is established based on the mapping relationship, which can detect the precursors of faults as early as possible, and the power supply fault warning cannot accurately carry out fault warning and detection.

[0030] 3. The module temperature sampling and correction mechanism is integrated with the collector current rise rate of the switching device to more accurately reflect the internal junction temperature of the device and obtain the precursor of device failure.

[0031] 4. By adjusting the sampling frequency of each physical parameter characteristic, we can avoid obtaining too much useless data due to too high sampling frequency, which will waste storage and processing resources; at the same time, we can adjust the sampling frequency according to the feedback coupling of the sampling data to avoid missing important changes in physical parameter characteristics and detect fault precursors early. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flow chart of a power failure early warning method for a protection and measurement control device of the present invention.

[0033] Figure 2 It is a fault warning evaluation flow chart of the present invention.

[0034] Figure 3 It is a structural block diagram of a power failure early warning system of a protection and measurement and control device of the present invention.

[0035] In the figure, 1. Historical fault library, 2. Mapping set, 3. Analysis and calculation module, 3-1. Deep convolutional neural network, 3-2. Fault warning evaluation function, 3-3. Shape trend matching unit, 3-4. Change rate matching unit, 3-5 Amplitude matching unit, 4. Sampling module, 4-1. Electrical sampling unit, 4-2. Temperature sampling unit, 4-3. Spectrum sampling unit, 5. Processing module. DETAILED DESCRIPTION

[0036] The technical solution of the present invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.

[0037] Embodiment 1: A power failure early warning method for a protection and control device of this embodiment is as follows: Figure 1 As shown, the following steps are included: S1: Construct a power supply failure warning model based on the parameter characteristic changes under the historical failure scenarios of the power supply of the protection device.

[0038] S101: Count the changes in physical parameter characteristics of power supplies of different security devices under various fault scenarios, and establish a corresponding “fault transition-parameter characteristic” mapping relationship from a normal state to a fault state.

[0039] Statistics are collected on the fault scenarios that are prone to occur in the power supplies of different manufacturers and models of security devices, and the main faults and their characteristics of the power supplies of security devices are extracted. The electrical characteristic quantities, temperature changes, noise spectrum, etc. under the fault scenarios are analyzed to explore the nonlinear transfer law of the physical parameter characteristics of the power supply of the security device under fault conditions, so as to establish the "fault transition-parameter characteristics" mapping relationship from normal to fault.

[0040] Specifically, due to the differences in the design, manufacturing, use and maintenance of power supplies for test and measurement devices from different manufacturers and models, their failure scenarios also present diverse characteristics.

[0041] First, the fault scenarios are collected to obtain all possible fault scenarios of the power supply of the test device.

[0042] Fault data of power supplies of different models of security devices from multiple manufacturers are collected, including detailed information such as the time, location, environmental conditions, and operation records of the fault. In this embodiment, through the collation and analysis of these data, the faults of the power supplies of security devices are mainly concentrated in the following aspects: 1) Power module failure: This type of failure is often caused by aging, damage or design defects of components inside the power module. When a failure occurs, the electrical parameters of the power module will become abnormal, such as output voltage fluctuations, current instability, etc.

[0043] 2) Power line failure: Power line faults include poor line contact, short circuit, open circuit, etc. Such faults usually cause the power supply to not work properly, and even cause serious consequences such as fire.

[0044] 3) Cooling system failure: Since the power supply of the test device generates a large amount of heat during operation, if the heat dissipation system fails, the internal temperature of the power supply will be too high, causing damage to components or performance degradation.

[0045] 4) Electromagnetic interference failure: In a complex electromagnetic environment, the power supply of the protection and testing device is easily affected by external electromagnetic interference, resulting in unstable operation or failure of the power supply.

[0046] After the failure scenarios are statistically analyzed, the main physical characteristic changes when the power supply of the protection device fails are further refined. In this embodiment, these physical characteristic changes include: Abnormal electrical characteristics: Such as abnormal changes in parameters such as output voltage, current, power factor, etc., these changes can directly reflect the working status of the power supply.

[0047] Abnormal temperature change: By monitoring the temperature changes inside the power supply, it is possible to detect cooling system failures or component overheating in a timely manner.

[0048] Abnormal noise spectrum: The power supply will generate a certain amount of noise during operation. By analyzing the changes in the noise spectrum, it is possible to determine whether the power supply has problems such as electromagnetic interference or internal component failure.

[0049] For the analysis of physical parameter characteristics under fault conditions, in the actual measurement and control device power supply fault scenario or in the simulation of various fault scenarios, record the data such as the electrical characteristic quantity, temperature change and noise spectrum of the power supply, and analyze the change law of the physical characteristic data under different fault scenarios. For example, when the power module fails, the output voltage and current will fluctuate nonlinearly; when the cooling system fails, the internal temperature of the power supply will show a nonlinear upward trend; when there is an electromagnetic interference fault, the noise spectrum will change nonlinearly, etc. Establish the "fault transition-parameter characteristic" mapping relationship between the power supply of different measurement and control devices from normal to fault in various fault scenarios.

[0050] Specifically, the “fault transition-parameter feature” mapping relationship of this embodiment is a change curve of the physical parameter feature during the fault transition, where the abscissa is time and the ordinate is the physical parameter feature.

[0051] Furthermore, for the measurement and control devices of the same manufacturer and the same model, the “fault transition-parameter feature” mapping relationship under the same power supply fault scenario is stored in the same mapping set.

[0052] For each different type of measurement and control device produced by different manufacturers, the physical characteristic changes in the time period before and after the fault occurs are captured in each possible fault scenario, the physical characteristic quantities of nonlinear changes are retained, and a "fault transition-parameter characteristic" mapping relationship is established.

[0053] The "fault transition-parameter feature" mapping relationship can be used to assist in quickly and accurately identifying the fault type of the power supply of the test device, and taking corresponding measures to repair or maintain it. At the same time, this mapping relationship can also provide a valuable reference for the design and manufacture of the power supply of the test device, thereby improving the stability and reliability of the power supply.

[0054] S102: Construct a fault warning model according to the “fault transition-parameter feature” mapping relationship under different fault scenarios.

[0055] For each different fault scenario, the change characteristics of the physical parameter characteristics in the "fault transition-parameter characteristics" mapping relationship are extracted, and the evaluation function is designed.

[0056] The physical parameter characteristics that can be used to characterize the fault characteristics are different in each fault scenario. Therefore, for each different fault scenario, the corresponding physical parameter characteristics that can characterize the occurrence or impending occurrence of a fault are selected according to the nonlinear change characteristics of the physical parameter characteristics in the "fault transition-parameter characteristic" mapping relationship. For example, for the filter electrolytic capacitors on the input and output sides of the power module, if the operating temperature of the power module is too high, the electrolytic capacitor will swell, thereby reducing the capacitance value, and eventually causing the output voltage (24V, 5V) of the power board to be unstable, causing the security device to freeze and fail to operate normally, and even causing the capacitor to explode, threatening the overall operation safety of the security device; therefore, the physical parameter characteristics selected for the electrolytic capacitor are temperature and output voltage, and the change matching factors of the selected physical parameter characteristics are used to evaluate whether a fault has occurred.

[0057] Specifically, the change matching factors of the physical parameter characteristics include the order in which the nonlinear changes occur, the rate of change of the parameters, the amplitude of the parameter changes, and the overall change trend shape.

[0058] Taking the change matching factor as the evaluation dimension, a fault warning evaluation function is established to evaluate the probability of the corresponding fault. Specifically, for a certain model of measurement and control device produced by a certain manufacturer, the expression of its power supply fault warning evaluation function is: Among them, G Mk is the fault warning evaluation function of the kth fault scenario; kn is the total number of physical parameter features selected under the kth fault scenario; σ i is the existence mark of the i-th selected physical parameter feature. If the corresponding physical parameter feature has nonlinear changes, it is marked as 1, otherwise it is marked as 0; Ord i is the occurrence order of the nonlinear change of the physical parameter characteristics of the i-th selected type. If it is consistent with the occurrence order of the nonlinear change of the physical parameter characteristics of the "fault transition-parameter characteristics" mapping relationship in the mapping set, it is marked as 1, otherwise it is marked as 0; α1 is the weighting coefficient of the change trend shape; α2 is the weighted coefficient of the rate of change; α3 is the amplitude weighting coefficient; The sum of α1, α2 and α3 is 1; S ki is the coincidence rate of the trend shape of the detected physical parameter characteristic change curve and the corresponding physical parameter change characteristic curve in the “fault transition-parameter characteristic” mapping relationship of the kth fault scenario; R kiis a change rate matching mark. If the detected change rate of the ith physical parameter characteristic is within the corresponding range of the “fault transition-parameter characteristic” mapping relationship of the kth fault scenario, it takes 1; otherwise, it takes 0. A ki is the amplitude matching mark. If the detected characteristic amplitude of the ith physical parameter is within the corresponding range in the “fault transition-parameter characteristic” mapping relationship of the kth fault scenario, it takes 1; otherwise, it takes 0. R ai is the characteristic change rate of the ith physical parameter detected in real time; R vi is the average value of the change rate of the ith physical parameter characteristic in the “fault transition-parameter characteristic” mapping relationship; A ai is the characteristic amplitude of the ith physical parameter detected in real time; A vi is the average value of the amplitude of the i-th physical parameter feature in the “fault transition-parameter feature” mapping relationship.

[0059] Different physical parameter features are selected for evaluation and matching for different fault scenarios; in the above expression, the existence marker σ is used i It is determined whether the physical parameter characteristics selected for the corresponding fault scenario have nonlinear changes. If so, the corresponding fault scenario may occur, otherwise it is determined that the fault scenario will not occur.

[0060] Furthermore, using Ord i Determine the occurrence order of each physical parameter feature. For some fault scenarios, the nonlinear changes of each physical parameter feature occur in a certain order. For example, when an electrolytic capacitor fails, the temperature rises first, and then the output voltage changes. The order in which the nonlinear changes of each physical parameter feature occur can be used to further determine whether the corresponding fault may occur. For physical parameter features with a strict order of occurrence, the Ord i To determine the corresponding order of occurrence, and for physical parameter characteristics that do not have a strict order of occurrence, Ord i The default is 1.

[0061] First, determine whether the selected physical parameter characteristics undergo nonlinear changes, and then further determine whether the order in which the nonlinear changes of the physical parameter characteristics occur conforms to the fault characteristics, and match the fault scenarios in turn.

[0062] Furthermore, the power supply fault warning evaluation function in the above formula also matches the rate of change, amplitude and trend shape of the physical parameter characteristics with the curve in the corresponding "fault transition-parameter characteristic" mapping relationship to determine the degree of matching with the corresponding fault.

[0063] For the matching of trend shapes, each physical parameter feature is compared separately, and the shape overlap rate is used as the evaluation value. After weighting, it is used to evaluate the fault matching degree. The higher the shape overlap rate, the greater the possibility of the fault.

[0064] For the matching of the change rate, for the mapping set of "fault transition-parameter feature" mapping of each fault scenario, the maximum and minimum values ​​of the physical parameter feature change rate are taken as the range of the corresponding physical parameter feature change rate; first, it is determined whether the detected physical parameter feature change rate is within the range of the physical parameter feature change rate. If so, its matching is evaluated again, otherwise it is directly taken as 0, and there is no need to further evaluate the matching of the change rate.

[0065] Taking the average value of the change rate of each physical parameter feature in the mapping set as the standard, the difference between the change rate of the detected physical parameter feature and the average value of the change rate of the physical parameter feature in the mapping set is calculated as the basis for the matching degree; the smaller the difference between the change rate of the physical parameter feature and the average value of the change rate of the physical parameter feature in the mapping set, the greater the possibility that this is the fault.

[0066] Similarly, for the matching of amplitudes, for the mapping set of "fault transition-parameter feature" mapping for each fault scenario, the maximum and minimum values ​​of the physical parameter feature amplitude are taken as the range of the corresponding physical parameter feature amplitude; first, it is determined whether the amplitude of the detected physical parameter feature is within the range of the physical parameter feature amplitude. If so, its matching is evaluated again, otherwise it is directly taken as 0, and there is no need to further evaluate the matching of the amplitude.

[0067] Taking the average amplitude value of each physical parameter feature in the mapping set as the standard, the difference between the amplitude of the detected physical parameter feature and the average amplitude value of the physical parameter features in the mapping set is calculated as the basis for the matching degree; the smaller the difference between the amplitude of the physical parameter feature and the average amplitude value of the physical parameter features in the mapping set, the greater the possibility that this is the fault.

[0068] When establishing the early warning model, multiple change matching factor dimensions are evaluated to eliminate the impact of cross-coupling and improve the accuracy of early warning.

[0069] S2: sampling the physical parameter characteristics of the power supply of the measurement and control device according to the set sampling frequency, and adjusting the corresponding sampling frequency according to the physical parameter characteristics obtained by sampling.

[0070] Different sampling frequencies are set for different physical parameter characteristics, and the upper and lower limits of the sampling frequencies are set respectively.

[0071] Specifically, since different physical parameter characteristics have different rates of change and magnitudes of change during actual use, in order to improve the data sampling efficiency as much as possible, that is, to sample as many valid data points as possible. The historical sampling data is analyzed by a big data model, and appropriate sampling frequencies and corresponding upper and lower limits are set for each physical parameter characteristic. Specific numerical values ​​can be obtained by a person skilled in the art through a limited number of experiments, which will not be repeated in this embodiment.

[0072] In this embodiment, the electrical characteristics (voltage, current, etc.), temperature characteristics and operating noise spectrum of the power supply of the measurement and control device are sampled, and the sampling frequency and the corresponding upper and lower limits of the sampling frequency are set respectively.

[0073] Furthermore, for the three physical parameter characteristics of this embodiment, the physical parameter characteristics are adjusted according to the sampled data coupling. For the sampling frequencies of the three physical parameter characteristics, the minimum sampling frequency is used by default to avoid wasting resources caused by sampling too much useless data.

[0074] Specifically, for electrical characteristics, the operating state of the power supply of the measurement and control device is generally most intuitively reflected through electrical characteristics. Therefore, the electrical characteristic sampling frequency in this embodiment is coupled and adjusted according to the sampled temperature data, noise spectrum data and electrical characteristic data.

[0075] Specifically, when the detected electrical characteristic data does not match all electrical characteristic change curves in the “fault transition-parameter characteristic” mapping relationship, the expression for adjusting the electrical characteristic sampling frequency is: f E =f Emin +(ε T γ T +ε Q γ Q )·Δf E When the detected electrical characteristic data matches any electrical characteristic change curve in the "fault transition-parameter characteristic" mapping relationship, the expression for adjusting the electrical characteristic sampling frequency is: f E =f Emax Furthermore, set the limiting conditions: f Emin +(γ T +γ Q )·Δf E ≤f Emax Among them, f E is the electrical characteristics sampling frequency; f Emax is the preset maximum value of the electrical characteristic sampling frequency; f Emin is the preset minimum value of the electrical characteristic sampling frequency; ε T It is a temperature characteristic fault precursor mark. When the detected temperature characteristic matches the temperature characteristic change curve in the "fault transition-parameter characteristic" mapping relationship, it takes 1; otherwise, it takes 0; ε Q It is a noise spectrum fault precursor marker. When the detected noise spectrum matches the noise spectrum change curve in the "fault transition-parameter feature" mapping relationship, it takes 1; otherwise, it takes 0. γ T It is the correlation coefficient between temperature characteristics and electrical characteristics, which can be obtained by artificially setting and analyzing historical faults in the database; γ Q is the correlation coefficient between the noise spectrum and the electrical characteristics, which can be obtained by artificially setting and analyzing the historical faults in the database; Δf E It is the minimum adjustment frequency of the electrical characteristic sampling frequency.

[0076] Specifically, with regard to temperature characteristics, the junction temperature has an important influence on the safe operating area of ​​the switching device. Various information (including the junction temperature) on the operating status of the switching device (MOSFET) in the power supply of the protection and control device are reflected in the electrical characteristics of the transient process.

[0077] The temperature sampling frequency is corrected according to the collector current rise rate of the switching device. Specifically, the expression for the temperature sampling frequency correction is: The constraints are: Among them, f T is the sampling frequency of temperature characteristics; f Tmin is the lower limit of the set temperature characteristic sampling frequency; f Tmax The upper limit of the set temperature characteristic sampling frequency; Δf T The minimum sampling adjustment frequency for temperature characteristics; R IS is the collector current rise rate of the switching device obtained by sampling; R IA It is the standard value of the collector current rise rate of the set switching device, which can be obtained through multiple experiments.

[0078] The module temperature sampling and correction mechanism is integrated with the collector current rise rate of the switching device in order to more accurately reflect the internal junction temperature of the device and obtain the precursor of device failure.

[0079] Specifically, for the running noise spectrum characteristics, the noise spectrum sampling frequency is adjusted according to the comparison between the width of the sampled noise spectrum and the width of the noise spectrum during normal operation. The specific expression for the noise spectrum sampling frequency correction is: Among them, f Q is the sampling frequency of the running noise spectrum; f Qmin is the minimum value of the sampling frequency of the running noise spectrum; f Qmax is the maximum value of the sampling frequency of the running noise spectrum; ΔN is the sampling frequency adjustment constant, which is set by the staff and is used to set the minimum adjustment frequency of the operating noise spectrum; β is the sampling frequency adjustment coefficient, and the expression of the sampling frequency adjustment coefficient is: The constraints are: β≤ΔN Among them, Q smax and Q smin are the maximum and minimum values ​​of the noise spectrum obtained by sampling, and the difference between them is the width of the noise spectrum obtained by sampling; Q bmax and Q bmin They are the maximum and minimum values ​​of the noise spectrum during normal operation, respectively. The difference between them is the width of the noise spectrum during normal operation, obtained through the analysis of historical data. Indicates rounding down.

[0080] When the width of the sampled noise spectrum is within the noise spectrum width range during normal operation, it indicates that the power supply of the current measurement and control device is working normally and there is no need to adjust the sampling frequency of the noise spectrum; When the width of the sampled noise spectrum is outside the noise spectrum width range during normal operation, the sampling frequency of the noise spectrum is adjusted according to the ratio of the sampled noise spectrum width to the noise spectrum width during normal operation.

[0081] By adjusting the sampling frequency of each physical parameter characteristic, we can avoid obtaining too much useless data due to too high sampling frequency, which will waste storage and processing resources; at the same time, the sampling frequency is adjusted according to the feedback coupling of the sampling data to avoid missing important changes in physical parameter characteristics and detect fault precursors early.

[0082] The detected physical parameter characteristics are used through data fitting, interpolation and prediction methods to obtain the change curves of each physical parameter characteristic.

[0083] S3: Import the sampled physical parameter characteristics into the fault warning model to match the fault type of the power supply of the measurement and control device.

[0084] In this embodiment, the change curves of each physical parameter characteristic obtained by sampling are input into the trained neural network to calculate the overlap rate between the change curves of each physical parameter characteristic and the trend shape of the corresponding physical parameter characteristic change curve in the "fault transition-parameter characteristic" mapping relationship.

[0085] In this embodiment, the neural network used is a deep convolutional neural network, which is a multi-layer feedforward neural network mainly composed of three parts: input layer, feature extraction layer and classifier. Each neuron in the feature extraction layer is only connected to some neurons in the adjacent network layer, and only acts on image data with spatial correlation, so as to capture local features.

[0086] According to the average value of the coincidence rate of each physical parameter characteristic change curve, sort them from large to small, and bring them into the power supply fault warning evaluation function of the corresponding fault scenario to obtain the corresponding fault warning evaluation value. Figure 2 As shown in the figure, the power failure warning evaluation process is: First, it is determined whether the selected physical parameter feature undergoes nonlinear changes. If so, the order of occurrence of the nonlinear changes of the feature is further determined. Otherwise, it returns to step S2 to resample the physical parameter feature.

[0087] After determining that the selected physical parameter characteristics undergo nonlinear changes, determine whether the order in which the nonlinear changes of the physical parameter characteristics occur conforms to the fault characteristics; if so, calculate the fault warning evaluation value under each fault scenario respectively to match the fault scenarios in turn.

[0088] Specifically, the expression of the fault warning evaluation function is: Furthermore, the power supply fault warning evaluation function in the above formula also matches the rate of change, amplitude and trend shape of the physical parameter characteristics with the curve in the corresponding "fault transition-parameter characteristic" mapping relationship to determine the degree of matching with the corresponding fault.

[0089] For the matching of trend shapes, each physical parameter feature is compared separately, and the shape overlap rate is used as the evaluation value. After weighting, it is used to evaluate the fault matching degree. The higher the shape overlap rate, the greater the possibility of the fault.

[0090] For the matching of the change rate, for the mapping set of "fault transition-parameter feature" mapping of each fault scenario, the maximum and minimum values ​​of the physical parameter feature change rate are taken as the range of the corresponding physical parameter feature change rate; first, it is determined whether the detected physical parameter feature change rate is within the range of the physical parameter feature change rate. If so, its matching is evaluated again, otherwise it is directly taken as 0, and there is no need to further evaluate the matching of the change rate.

[0091] Taking the average value of the change rate of each physical parameter feature in the mapping set as the standard, the difference between the change rate of the detected physical parameter feature and the average value of the change rate of the physical parameter feature in the mapping set is calculated as the basis for the matching degree; the smaller the difference between the change rate of the physical parameter feature and the average value of the change rate of the physical parameter feature in the mapping set, the greater the possibility that this is the fault.

[0092] Similarly, for the matching of amplitudes, for the mapping set of "fault transition-parameter feature" mapping for each fault scenario, the maximum and minimum values ​​of the physical parameter feature amplitude are taken as the range of the corresponding physical parameter feature amplitude; first, it is determined whether the amplitude of the detected physical parameter feature is within the range of the physical parameter feature amplitude. If so, its matching is evaluated again, otherwise it is directly taken as 0, and there is no need to further evaluate the matching of the amplitude.

[0093] Taking the average amplitude value of each physical parameter feature in the mapping set as the standard, the difference between the amplitude of the detected physical parameter feature and the average amplitude value of the physical parameter features in the mapping set is calculated as the basis for the matching degree; the smaller the difference between the amplitude of the physical parameter feature and the average amplitude value of the physical parameter features in the mapping set, the greater the possibility that this is the fault.

[0094] The fault warning evaluation values ​​obtained by the calculation are screened, and power supply failure scenarios that are greater than a preset warning evaluation threshold are retained. Specifically, in this embodiment, the preset warning evaluation threshold is 85%; the fault scenario with the largest fault warning evaluation value is selected as the estimated power supply failure type of the measurement and control device.

[0095] S4: Process according to the predicted fault type.

[0096] Different processing methods are set for different power failures of measurement and control devices, including switching to backup power, automatic restart, and active alarm. According to the predicted fault type, the alarm information is sent to the corresponding staff, and corresponding processing preparations are made. When a fault occurs and there is no staff to handle it, a corresponding processing response is made.

[0097] The scheme of this embodiment counts the fault scenarios that are prone to occur in the power supplies of different manufacturers and different models of security devices, extracts the main faults and characteristics of the power supplies of security devices, analyzes the electrical characteristic quantities, temperature changes, noise spectrum, etc. under the fault scenarios, and explores the nonlinear transfer law of the parameters of the power supplies of security devices under fault conditions, thereby establishing a "fault transition-parameter characteristics" mapping relationship from normal to fault. According to the analysis of the fault mechanism, a multi-dimensional security device power supply early warning indicator system is constructed. On the basis of the fusion of multi-dimensional information of fault characteristic quantities, combined with artificial intelligence algorithms, an online early warning method for security device power supply failures suitable for embedded systems is proposed, and the life of the device is further estimated. The relevant modules are replaced early during periodic maintenance to avoid the shutdown of related primary equipment due to the failure of the security device power supply.

[0098] Embodiment 2: A protection and control device power failure early warning system of this embodiment, such as Figure 3 As shown, it includes a historical fault library 1, a mapping set 2, an analysis and calculation module 3, a sampling module 4 and a processing module 5.

[0099] Specifically, the historical fault database 1 stores physical parameter characteristic data of various fault types of the power supply of the measurement and control device in the time period before and after the historical fault.

[0100] The analysis and calculation module 3 collects statistics on the fault scenarios that are prone to occur in the power supplies of the security and measurement devices of different manufacturers and models in the historical fault library 1, and extracts the main faults and their characteristics of the power supplies of the security and measurement devices. The electrical characteristic quantities, temperature changes, noise spectra, etc. under each fault scenario in the historical fault library 1 are analyzed to explore the nonlinear transfer law of the physical parameter characteristics of the power supply of the security and measurement device under fault conditions, thereby establishing a "fault transition-parameter characteristic" mapping relationship from normal to fault to form a mapping set 2.

[0101] The analysis and calculation module 3 analyzes the physical parameter characteristics under fault conditions. In the actual measurement and control device power supply fault scenario or in the simulation of various fault scenarios, the data such as the electrical characteristic quantity, temperature change and noise spectrum of the power supply are recorded, and the change law of the physical characteristic data under different fault scenarios is analyzed. For example, when the power module fails, the output voltage and current will fluctuate nonlinearly; when the heat dissipation system fails, the internal temperature of the power supply will show a nonlinear upward trend; when the electromagnetic interference fails, the noise spectrum will change nonlinearly, etc. A "fault transition-parameter characteristic" mapping relationship between normal and fault of the power supply of different measurement and control devices under various fault scenarios is established to form a mapping set 2.

[0102] Specifically, the "fault transition-parameter feature" mapping relationship of this embodiment is a change curve of the physical parameter feature during the fault transition, with the horizontal axis being time and the vertical axis being the physical parameter feature. For the measurement and control devices of the same manufacturer and the same model, the "fault transition-parameter feature" mapping relationship under the same power supply fault scenario is stored in the same mapping set 2.

[0103] For each different type of measurement and control device produced by different manufacturers, the physical characteristic changes in the time period before and after the fault occurs are captured in each possible fault scenario, the physical characteristic quantities of nonlinear changes are retained, and a "fault transition-parameter characteristic" mapping relationship is established.

[0104] The "fault transition-parameter feature" mapping relationship can be used to assist in quickly and accurately identifying the fault type of the power supply of the test device, and taking corresponding measures to repair or maintain it. At the same time, this mapping relationship can also provide a valuable reference for the design and manufacture of the power supply of the test device, thereby improving the stability and reliability of the power supply.

[0105] The analysis and calculation module 3 extracts the change characteristics of the physical parameter characteristics in the "fault transition-parameter characteristic" mapping relationship in the mapping set 2 for each different fault scenario, and designs the fault warning evaluation function 3-2.

[0106] The physical parameter characteristics that can be used to characterize the fault characteristics are different in each fault scenario. Therefore, for each different fault scenario, the corresponding physical parameter characteristics that can characterize the occurrence or impending occurrence of a fault are selected according to the nonlinear change characteristics of the physical parameter characteristics in the "fault transition-parameter characteristic" mapping relationship. For example, for the filter electrolytic capacitors on the input and output sides of the power module, if the operating temperature of the power module is too high, the electrolytic capacitor will swell, thereby reducing the capacitance value, and eventually causing the output voltage (24V, 5V) of the power board to be unstable, causing the security device to freeze and fail to operate normally, and even causing the capacitor to explode, threatening the overall operation safety of the security device; therefore, the physical parameter characteristics selected for the electrolytic capacitor are temperature and output voltage, and the change matching factors of the selected physical parameter characteristics are used to evaluate whether a fault has occurred.

[0107] Specifically, the change matching factors of physical parameter characteristics include the order of nonlinear changes, the rate of change of parameters, the amplitude of parameter changes, and the overall change trend shape, etc. Taking the change matching factors as the evaluation dimension, a fault warning evaluation function is established to evaluate the probability of the corresponding fault.

[0108] When establishing the early warning model, multiple change matching factor dimensions are evaluated to eliminate the impact of cross-coupling and improve the accuracy of early warning.

[0109] The sampling module 4 sets different sampling frequencies for different physical parameter characteristics, and sets upper and lower limits of the sampling frequencies respectively.

[0110] Specifically, since different physical parameter characteristics have different rates of change and magnitudes of change during actual use, in order to improve the data sampling efficiency as much as possible, that is, to sample as many valid data points as possible. The sampling module 4 analyzes the historical sampling data through the big data model, and sets a suitable sampling frequency and corresponding upper and lower limits for each physical parameter characteristic. The specific numerical values ​​can be obtained by a limited number of experiments by those skilled in the art, and will not be repeated in this embodiment.

[0111] In this embodiment, the sampling module 4 includes an electrical sampling unit 4-1 for sampling the electrical characteristics (voltage, current, etc.) of the power supply of the measurement and control device, a temperature sampling unit 4-2 for sampling the temperature characteristics, and a spectrum sampling unit 4-3 for sampling the operating noise spectrum; each sampling unit in the sampling module 4 is respectively set with a sampling frequency and corresponding upper and lower limits of the sampling frequency.

[0112] Furthermore, for the three physical parameter features of this embodiment, the sampling frequency of the physical parameter features is adjusted according to the sampled data coupling. For the sampling frequency of the three physical parameter features, the minimum sampling frequency is used by default to avoid wasting resources caused by sampling too much useless data.

[0113] Specifically, for electrical characteristics, the operating state of the power supply of the measurement and control device is generally most intuitively reflected through electrical characteristics. Therefore, the electrical characteristic sampling frequency in this embodiment is coupled and adjusted according to the sampled temperature data, noise spectrum data and electrical characteristic data.

[0114] When the electrical characteristic data sampled by the electrical sampling unit 4-1 do not match all the electrical characteristic change curves in the "fault transition-parameter characteristic" mapping relationship in the mapping set 2, the electrical sampling unit 4-1 adjusts the electrical characteristic sampling frequency as follows: f E =f Emin +(ε T γ T +ε Q γ Q )·Δf E

[0115] When the electrical characteristic data sampled by the electrical sampling unit 4-1 matches any electrical characteristic change curve in the "fault transition-parameter characteristic" mapping relationship in the mapping set 2, the electrical sampling unit 4-1 adjusts the electrical characteristic sampling frequency as follows: f E =f Emax .

[0116] Furthermore, the electrical sampling unit 4-1 sets limiting conditions: f Emin +(γ T +γ Q )·Δf E ≤f Emax Among them, f E is the electrical characteristics sampling frequency; f Emax is the preset maximum value of the electrical characteristic sampling frequency; f Emin is the preset minimum value of the electrical characteristic sampling frequency; ε T It is a temperature characteristic fault precursor mark. When the temperature characteristic sampled by the temperature sampling unit 4-2 matches the temperature characteristic change curve in the "fault transition-parameter characteristic" mapping relationship in the mapping set 2, it takes 1; otherwise, it takes 0; ε Q It is a noise spectrum fault precursor mark. When the noise spectrum sampled by the spectrum sampling unit 4-3 matches the noise spectrum change curve in the "fault transition-parameter feature" mapping relationship in the mapping set 2, it takes 1; otherwise, it takes 0; γ T It is the correlation coefficient between temperature characteristics and electrical characteristics, which can be obtained by artificially setting and analyzing historical faults in the database; γ Q is the correlation coefficient between the noise spectrum and the electrical characteristics, which can be obtained by artificially setting and analyzing the historical faults in the database; Δf E It is the minimum adjustment frequency of the electrical characteristic sampling frequency.

[0117] Specifically, with regard to temperature characteristics, the junction temperature has an important influence on the safe operating area of ​​the switching device. Various information (including the junction temperature) on the operating status of the switching device (MOSFET) in the power supply of the protection and control device are reflected in the electrical characteristics of the transient process.

[0118] The temperature sampling frequency is corrected according to the collector current rising rate of the switching device. Specifically, the expression for the temperature sampling frequency correction of the temperature sampling unit 4-2 is: The constraints are: Among them, f T is the sampling frequency of temperature characteristics; f Tmin is the lower limit of the set temperature characteristic sampling frequency; fTmax The upper limit of the set temperature characteristic sampling frequency; Δf T The minimum sampling adjustment frequency for temperature characteristics; R IS is the collector current rise rate of the switching device obtained by sampling; R IA It is the standard value of the collector current rise rate of the set switching device, which can be obtained through multiple experiments.

[0119] The module temperature sampling and correction mechanism is integrated with the collector current rise rate of the switching device in order to more accurately reflect the internal junction temperature of the device and obtain the precursor of device failure.

[0120] Specifically, for the running noise spectrum characteristics, the noise spectrum sampling frequency is adjusted according to the comparison between the width of the noise spectrum sampled by the spectrum sampling unit 4-3 and the width of the noise spectrum during normal operation. The specific expression for adjusting the noise spectrum sampling frequency by the spectrum sampling unit 4-3 is: Among them, f Q is the sampling frequency of the running noise spectrum; f Qmin is the minimum value of the sampling frequency of the running noise spectrum; f Qmax is the maximum value of the sampling frequency of the running noise spectrum; ΔN is the sampling frequency adjustment constant, which is set by the staff and is used to set the minimum adjustment frequency of the operating noise spectrum; β is the sampling frequency adjustment coefficient, and the expression of the sampling frequency adjustment coefficient is: The spectrum sampling unit 4-3 adjusts the constraint condition as follows: β≤ΔN Among them, Q smax and Q smin are the maximum and minimum values ​​of the noise spectrum obtained by sampling, and the difference between them is the width of the noise spectrum obtained by sampling; Q bmax and Q bmin They are the maximum and minimum values ​​of the noise spectrum during normal operation, respectively. The difference between them is the width of the noise spectrum during normal operation, obtained through the analysis of historical data. Indicates rounding down.

[0121] When the width of the noise spectrum sampled by the spectrum sampling unit 4-3 is within the noise spectrum width range during normal operation, it indicates that the power supply of the current measurement and control device is working normally and there is no need to adjust the sampling frequency of the noise spectrum; When the width of the noise spectrum sampled by the spectrum sampling unit 4-3 is outside the noise spectrum width range during normal operation, the sampling frequency of the noise spectrum is adjusted according to the ratio of the noise spectrum width sampled by the spectrum sampling unit 4-3 to the noise spectrum width during normal operation.

[0122] The sampling module 4 adjusts the sampling frequency of each physical parameter feature to avoid excessive useless data due to too high sampling frequency, thereby wasting storage and processing resources. At the same time, the sampling frequency is adjusted according to the feedback coupling of the sampling data to avoid missing important changes in physical parameter features and to detect fault precursors early.

[0123] The sampling module 4 obtains the variation curve of each physical parameter characteristic through data fitting, interpolation and prediction methods, and transmits the variation curve to the analysis and calculation unit 3 for fault warning evaluation.

[0124] The analysis and calculation module 3 includes a trained neural network. In this embodiment, the change curves of the sampled physical parameter characteristics are input into the trained neural network to calculate the overlap rate between the change curves of the physical parameter characteristics and the trend shapes of the physical parameter characteristic change curves corresponding to the "fault transition-parameter characteristic" mapping relationship in the mapping set 2.

[0125] In this embodiment, the neural network used is a deep convolutional neural network 3-1. The deep convolutional neural network (CNN) is a multi-layer feedforward neural network, which mainly consists of three parts: input layer, feature extraction layer and classifier. Each neuron in the feature extraction layer is only connected to some neurons in the adjacent network layer, and only acts on image data with spatial correlation, so as to capture local features.

[0126] The analysis and calculation module 3 sorts the average values ​​of the coincidence rates of the characteristic change curves of the physical parameters from large to small, and respectively brings them into the power supply fault warning evaluation function 3-2 of the corresponding fault scenario to obtain the corresponding fault warning evaluation value.

[0127] Specifically, the expression of the fault warning evaluation function is: Among them, G Mk is the fault warning evaluation function of the kth fault scenario; kn is the total number of physical parameter features selected under the kth fault scenario; σ i is the existence mark of the i-th selected physical parameter feature. If the corresponding physical parameter feature has nonlinear changes, it is marked as 1, otherwise it is marked as 0; Ord i is the occurrence order of the nonlinear change of the physical parameter characteristics of the i-th selected type. If it is consistent with the occurrence order of the nonlinear change of the physical parameter characteristics of the "fault transition-parameter characteristics" mapping relationship in the mapping set, it is marked as 1, otherwise it is marked as 0; α1 is the weighted coefficient of the change trend shape; α2 is the weighted coefficient of the rate of change; α3 is the amplitude weighting coefficient; The sum of α1, α2 and α3 is 1.

[0128] S ki is the coincidence rate of the trend shape of the detected physical parameter characteristic change curve and the corresponding physical parameter change characteristic curve in the “fault transition-parameter characteristic” mapping relationship of the kth fault scenario; R ki is a change rate matching mark. If the detected change rate of the ith physical parameter characteristic is within the corresponding range of the “fault transition-parameter characteristic” mapping relationship of the kth fault scenario, it takes 1; otherwise, it takes 0. A ki is the amplitude matching mark. If the detected characteristic amplitude of the ith physical parameter is within the corresponding range in the “fault transition-parameter characteristic” mapping relationship of the kth fault scenario, it takes 1; otherwise, it takes 0. R ai is the characteristic change rate of the ith physical parameter detected in real time; R vi is the average value of the change rate of the ith physical parameter characteristic in the “fault transition-parameter characteristic” mapping relationship; A ai is the characteristic amplitude of the ith physical parameter detected in real time; A vi is the average value of the amplitude of the i-th physical parameter feature in the “fault transition-parameter feature” mapping relationship.

[0129] The power failure warning evaluation function 3-2 in the above formula includes a shape trend matching unit 3-3, a change rate matching unit 3-4 and an amplitude matching unit 3-5.

[0130] The shape trend matching unit 3-3, the change rate matching unit 3-4 and the amplitude matching unit 3-5 respectively match the shape trend, change rate and amplitude of the physical parameter characteristics with the curves in the "fault transition-parameter characteristic" mapping relationship in the corresponding mapping set 2 to determine the degree of matching with the corresponding fault.

[0131] The shape trend matching unit 3 - 3 compares each physical parameter feature respectively, takes the shape coincidence rate as an evaluation value, and uses it to evaluate the fault matching degree after weighting. The higher the shape coincidence rate, the greater the possibility of the fault.

[0132] For the change rate matching unit 3-4, for the "fault transition-parameter feature" mapping relationship of each fault scenario, the maximum and minimum values ​​of the physical parameter feature change rate are taken as the range of the corresponding physical parameter feature change rate; first, it is determined whether the detected physical parameter feature change rate is within the range of the physical parameter feature change rate. If so, its matching is evaluated again, otherwise it is directly taken as 0, and there is no need to further evaluate the matching of the change rate.

[0133] Taking the average value of the change rate of each physical parameter feature in the mapping set as the standard, the difference between the change rate of the detected physical parameter feature and the average value of the change rate of the physical parameter feature in the mapping set is calculated as the basis for the matching degree; the smaller the difference between the change rate of the physical parameter feature and the average value of the change rate of the physical parameter feature in the mapping set, the greater the possibility that this is the fault.

[0134] Similarly, for the "fault transition-parameter feature" mapping relationship of the amplitude matching unit 3-5 for each fault scenario, the maximum and minimum values ​​of the physical parameter feature amplitude are taken as the range of the corresponding physical parameter feature amplitude; first, it is determined whether the amplitude of the detected physical parameter feature is within the range of the physical parameter feature amplitude. If so, its matching is evaluated again, otherwise it is directly taken as 0, and there is no need to further evaluate the matching of the amplitude.

[0135] Taking the average amplitude value of each physical parameter feature in mapping set 2 as the standard, the difference between the amplitude of the detected physical parameter feature and the average amplitude value of the physical parameter features in the mapping set is calculated as the basis for the matching degree; the smaller the difference between the amplitude of the physical parameter feature and the average amplitude value of the physical parameter features in the mapping set, the greater the possibility that this is the fault.

[0136] The analysis and calculation module 3 screens the fault warning evaluation values ​​obtained by calculation, and retains the power supply fault scenarios that are greater than the preset warning evaluation threshold. Specifically, in this embodiment, the preset warning evaluation threshold is 85%; the fault scenario with the largest fault warning evaluation value is selected as the estimated power supply fault type of the measurement and control device.

[0137] The processing module 5 sets different processing methods for different power failures of the measurement and control device, including switching to backup power, automatic restart, and active alarm, etc. The processing module 5 sends alarm information to the corresponding staff according to the fault type predicted by the calculation and analysis module 3, and makes corresponding processing preparations.

[0138] Embodiment three: A storage medium of this embodiment is used for computer-readable storage, and the storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of a power failure warning method for a protection and measurement and control device in the above embodiment.

[0139] If the system-integrated module / unit of the power failure early warning system of the protection and control device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0140] A terminal of the present embodiment includes a processor, a memory, a program stored in the memory and executable on the processor, and a data bus for realizing connection and communication between the processor and the memory. When the program is executed by the processor, the steps of a power failure warning method for a protection and measurement and control device of the above embodiment are realized, as described above in detail and will not be repeated here.

[0141] It should be noted that the terminal of the third embodiment of the present invention and the method embodiment 1 belong to the same concept, and its specific implementation process is detailed in the method embodiment 1, and the technical features in the method embodiment 1 are applicable in the terminal embodiment, which will not be repeated here.

[0142] The processor is the control center of the terminal, which uses various interfaces and lines to connect various parts of the entire system. It executes various functions of the electronic terminal and processes data by running or executing software programs and / or modules stored in the memory, and calling data stored in the memory, thereby monitoring the terminal as a whole. Optionally, the processor may include one or more processing cores; in some embodiments, the processor may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It can be understood that the above-mentioned modem processor may not be integrated into the processor.

[0143] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0144] The memory can be used to store software programs and modules, such as the corresponding program instructions / modules in the above embodiments, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0145] Embodiment 4: A power failure early warning method for a protection and control device of this embodiment is as follows: Figure 1 As shown, the main steps are: Step 1: Construct a "fault transition-parameter feature" mapping relationship and set a power supply fault warning model under each fault scenario; Step 2: Collect physical parameter characteristics at the set sampling frequency. If the sampling frequency needs to be adjusted, go to step 3, otherwise go to step 4; Step 3: Adjust the sampling frequency, re-collect the physical parameter characteristics at the set sampling frequency and then go to step 2; Step 4: Input the physical parameter characteristics with the set sampling frequency into the neural network, calculate the coincidence rate of the trend shape of the physical parameter characteristic change curve, sort the fault situations from large to small according to the coincidence rate, and respectively bring them into the power supply fault warning evaluation for function calculation, and then go to step 5; Step 5: Determine whether the obtained function value is greater than the evaluation threshold. If so, proceed to step 6; otherwise, proceed to step 7. Step 6: Determine if there are signs of a fault, sort and match the fault type, and notify the staff to prepare for fault elimination; Step 7: Eliminate the corresponding data, determine that there are no signs of a fault, and go to step 2 to start a new round of cyclic judgment.

[0146] In the above steps, step 1 mainly involves several basic steps such as fault data collection and analysis, fault scenario statistics and feature extraction, physical parameter feature analysis, establishment of necessary mapping relationships, data storage and management, etc.

[0147] In the fault data collection and analysis step, it is necessary to collect fault data of the power supplies of the test devices from multiple manufacturers and different models, including detailed information such as the time, location, environmental conditions, and operation records of the fault. This is the basis for building a data-driven fault warning model, which requires analysis and organization of a large amount of fault data.

[0148] In the process of fault scenario statistics and feature extraction, after collecting fault data, it is necessary to count the fault scenarios and extract the main faults and their characteristics of the power supply of the protection device. This usually involves the classification and summary of fault data to identify the common characteristics and laws of different fault types.

[0149] The physical parameter characteristics under fault conditions, such as electrical characteristic quantities, temperature changes, noise spectrum, etc., are analyzed to explore the nonlinear transfer law of the physical parameter characteristics of the power supply of the protection device under fault conditions. This may require the use of signal processing technology, statistical analysis methods, or machine learning algorithms to analyze the nonlinear transfer law of physical parameter characteristics. This may require the use of signal processing technology, statistical analysis methods, or machine learning algorithms to analyze the changing trends and patterns of physical parameters.

[0150] In the process of establishing the necessary mapping relationship, based on the analysis results of the physical parameter characteristics, a "fault transition-parameter characteristic" mapping relationship from the normal state to the fault state is established. This can be done by drawing a curve chart of the parameter characteristics changing over time or using a mathematical model to describe the changing law of the parameter characteristics during the fault transition process.

[0151] In the process of data storage and management, for the measurement and control devices of the same manufacturer and model, the "fault transition-parameter feature" mapping relationship under the same power failure scenario needs to be stored in the same mapping set. This involves database technology or file storage technology to manage a large amount of mapping relationship data.

[0152] Among them, "fault transition-parameter characteristics" is a concept used to describe the characteristics or changes of the physical parameters or performance indicators of a device or system during the transition from a normal state to a faulty state. This process involves the gradual deterioration of the internal state of the device or system, which may eventually lead to device failure or malfunction. Specifically, "fault transition" refers to the entire process in which a device or system starts from a normal working state and gradually shows signs of performance degradation, reduced stability, etc. until complete failure or malfunction occurs. This process may be sudden or gradual, depending on the characteristics of the device or system and the nature of the fault.

[0153] "Parameter characteristics" refer to the physical parameters or performance indicators that can reflect the changes in the state of the equipment or system during this process. These parameters or indicators may include electrical parameters (such as voltage, current, power, etc.), mechanical parameters (such as speed, vibration, displacement, etc.), thermal parameters (such as temperature, thermal resistance, etc.), chemical parameters (such as concentration, pH value, etc.), etc. The changes in these parameters or indicators can directly or indirectly reflect the changes in the internal state of the equipment or system, thereby providing an important basis for fault warning and diagnosis. When describing "fault transition-parameter characteristics", it is usually necessary to analyze them in combination with specific equipment or systems, as well as possible fault types.

[0154] For example, for the power supply of the test device, its fault transition may be manifested as fluctuations in the output voltage of the power module, a gradual increase in the temperature inside the power supply, abnormal changes in the noise spectrum, etc. The changing trends and patterns of these parameter characteristics can be described and analyzed by drawing curve graphs and establishing mathematical models, so as to better understand and predict the fault behavior of the equipment. When building a fault warning model, "fault transition-parameter characteristics" is an important input information, which can help the model accurately identify the fault state of the equipment or system and issue a warning signal in advance, thereby avoiding losses caused by equipment failure.

[0155] In step 2 and step 3, the technical core involved lies in the construction of the fault warning evaluation function and the selection and matching of the physical parameter characteristic values. In terms of the construction of the fault warning evaluation function, a comprehensive fault warning evaluation function (G_Mk) is constructed based on the "fault transition-parameter characteristic" mapping relationship for different fault scenarios. This function evaluates the probability of the corresponding fault through multiple dimensions (including the existence, occurrence order, change trend shape, change rate and amplitude of physical parameter characteristics).

[0156] In the step of selecting and matching the physical parameter characteristic values, the physical parameter characteristics that can characterize the occurrence or impending occurrence of a fault are selected according to the nonlinear change characteristics of the physical parameter characteristics in the "fault transition-parameter characteristic" mapping relationship under different fault scenarios. The change characteristics of these physical parameter characteristics are evaluated, including the order of nonlinear changes, the rate of change of the parameters, the amplitude of the parameter changes, and the overall change trend shape.

[0157] The selection and matching of the characteristic values ​​of physical parameters require a series of processing methods, including dynamic sampling frequency adjustment of parameters, sampling frequency optimization based on big data, parameter coupling adjustment, setting of upper and lower limits of sampling frequency, and determination of sampling parameters through limited experiments.

[0158] When dynamically adjusting the sampling frequency, the core technology lies in dynamically adjusting the corresponding sampling frequency according to the actual change rate and amplitude of each physical parameter characteristic (such as electrical characteristics, temperature characteristics, and operating noise spectrum). This ensures that enough data points are collected when key parameters change rapidly, while avoiding oversampling when parameters are stable, thereby improving data sampling efficiency.

[0159] When optimizing the sampling frequency based on big data, the appropriate sampling frequency and upper and lower limits of each physical parameter feature are determined by analyzing the historical sampling data and using the big data model. This method can optimize the sampling strategy based on actual operating conditions and historical data, and improve the accuracy and effectiveness of the sampling data.

[0160] When adjusting the parameter coupling, the core technology is to use the coupling relationship between different physical parameter characteristics to adjust the sampling frequency. For example, the sampling frequency of electrical characteristics can be adjusted according to the changes in temperature data and noise spectrum data to reflect the overall operating status of the power supply of the measurement and control device. This coupling adjustment method can more comprehensively consider the system status and improve the accuracy and reliability of fault warning.

[0161] In the process of setting the upper and lower limits of the sampling frequency, the upper and lower limits of the sampling frequency are set to avoid problems caused by too high or too low sampling frequencies. Too high sampling frequencies may waste resources, while too low sampling frequencies may not capture key data. Therefore, it is very important to set appropriate upper and lower limits according to the actual needs of each physical parameter characteristic. The upper and lower limits of the sampling frequency are set to avoid problems caused by too high or too low sampling frequencies. Too high sampling frequencies may waste resources, while too low sampling frequencies may not capture key data. Therefore, it is very important to set appropriate upper and lower limits according to the actual needs of each physical parameter characteristic.

[0162] Determining sampling parameters through limited number of experiments is an important means to determine the appropriate sampling frequency and upper and lower limits of each physical parameter characteristic. Through actual operation and experimental verification, the sampling strategy can be continuously optimized to improve the quality and accuracy of sampling data.

[0163] The relevant process of adjusting the sampling frequency of electrical characteristics is described in the following example in this embodiment. It is assumed that the electrical characteristic data (such as voltage and current) monitored in a power supply system of a measurement and control device does not match any known electrical characteristic change curve in the "fault transition-parameter characteristic" mapping relationship. The sampling frequency of the electrical characteristics can be adjusted according to the temperature characteristics and noise spectrum characteristics of the current system.

[0164] Assume f_Emin is 100Hz, f_Emax is 1000Hz, and [Δf]_E is 50Hz. If the system detects that both the temperature signature (ε_T) and the noise spectrum signature (ε_Q) do not match the fault precursor (i.e. both are 0), then the sampling frequency f_E will be f_Emin, i.e. 100Hz. However, if the temperature signature matches (ε T =1) and the noise spectrum does not match (ε Q =0), and assuming that γ T =0.8, then the sampling frequency f_E will be f_Emin+0.8*50Hz=140Hz.

[0165] According to the formula f E =f Emin +(ε T γ T +ε Q γ Q )·[Δf]_E, we can deduce: When ε_T and ε_Q are both 0, f_E=f_Emin; When ε T =1 and ε Q = 0, f E =f Emin +γ T ·[Δf]_E; When ε T = 0 and ε Q =1, f E =f Emin +γ Q ·[Δf]_E; When ε_T and ε_Q are both 1, f E =f Emin +(γ T +γ Q )·[Δf]_E.

[0166] The following example is used to describe the process of correcting the temperature characteristic sampling frequency in this embodiment. Assume that the collector current rise rate R_IS of the switching device is 1.2 times the standard value R_IA. Set f_Tmin to 5Hz, f_Tmax to 50Hz, and [Δf]_T to 5Hz. Then the temperature sampling frequency f_T will be calculated according to the formula Adjust to

[0167] By formula It can be seen that: When R IS =R_IA, f_T = f_Tmin; When R IS >R_IA, f_T will increase with the increase of R_IS / R_IA, but the increase is an integer multiple of [Δf]_T.

[0168] The relevant process of noise spectrum sampling frequency correction is described in the following example in this embodiment. It is assumed that the sampled noise spectrum width (Q_smax-Q_smin) is 1.5 times the normal working noise spectrum width (Q_bmax-Q_bmin). Set f_Qmin to 10Hz, f_Qmax to 100Hz, and ΔN to 5. According to the formula, the sampling frequency adjustment coefficient β will be Therefore, the sampling frequency f_Q of the noise spectrum will be f_Q=f_Qmin+5*(f_Qmax-f_Qmin) / 5=50 Hz.

[0169] The sampled physical parameter characteristics are imported into the fault warning model to match the fault type of the power supply of the measurement and control device.

[0170] Steps 4 and 5 mainly involve dimensional analysis of the evaluation function and fine calculation of the matching degree. In the dimensional analysis step of the evaluation function, the evaluation function includes the consideration of the existence (σ_i), occurrence order (Ord_i), change trend shape (S_ki), change rate (R_ki) and amplitude (A_ki) of the physical parameter characteristics. For the matching of trend shape, the shape overlap rate is used as the evaluation value. By comparing the real-time detected physical parameter characteristic change curve with the curve in the "fault transition-parameter characteristic" mapping relationship, the overlap degree of the two is evaluated. For the matching of the change rate, first determine whether the detected physical parameter characteristic change rate is within the preset range, and then calculate the difference between it and the average value of the physical parameter characteristic change rate in the mapping relationship, so as to evaluate the matching degree of the change rate.

[0171] Steps 6 and 7 mainly involve the identification and judgment of fault scenarios. By comprehensively evaluating the calculation results of the function, it is possible to accurately judge whether the current device or system is in the transition stage of a certain fault scenario and the probability of a fault. This provides an important basis for taking preventive measures in a timely manner to avoid faults or reduce the impact of faults.

[0172] The fault warning process mentioned after step 4 can be divided into three main implementation steps: fault prediction process, fault type processing, and technical implementation and optimization.

[0173] The fault prediction process is carried out in the following steps: First, data sampling and input are performed to sample the physical parameter characteristics of the power supply of the measurement and control device, such as voltage, current, temperature, etc. The change curve obtained by sampling is input into the trained deep convolutional neural network (CNN).

[0174] Then, feature extraction and trend analysis are carried out. CNN is used to extract the changing trends and key features of physical parameter characteristics, and the coincidence rate between the changing curves of each physical parameter characteristic and the corresponding curve in the “fault transition-parameter characteristic” mapping relationship is evaluated.

[0175] The evaluation function is calculated based on the obtained parameters and mapping relationships, and the fault warning evaluation function is used to calculate the fault warning evaluation value under each fault scenario. The evaluation function includes the consideration of the existence (σ_i), occurrence order (Ord_i), change trend shape (S_ki), change rate (R_ki) and amplitude (A_ki) of physical parameter characteristics.

[0176] Perform dimension analysis and match calculation. The specific calculation of each dimension in the evaluation function is as follows: Trend shape matching (S_ki): Calculate the overlap rate of the curve shape between the real-time detected physical parameter characteristic change curve and the "fault transition-parameter characteristic" mapping relationship. Change rate matching (R_ki): Determine whether the real-time detected physical parameter characteristic change rate is within the preset range, and calculate the difference between it and the average value of the physical parameter characteristic change rate in the mapping relationship. Amplitude matching (A_ki): Similar to the change rate matching, first determine whether the amplitude is within the preset range, and then calculate the difference between it and the average value of the physical parameter characteristic amplitude in the mapping relationship.

[0177] Finally, a comprehensive evaluation is performed to obtain the fault warning evaluation value by combining the calculation results of the matching degree of each dimension. The possible fault scenarios and their occurrence probabilities are identified according to the order of the warning evaluation values.

[0178] The process of fault type handling is as follows: Identify and judge the fault scenario, and based on the comprehensive evaluation results, accurately determine whether the power supply of the measurement and control device is in the transition stage of a certain fault scenario. Evaluate the probability of the fault and provide a basis for subsequent processing.

[0179] For alarm information sending and processing preparation, according to the predicted fault type, the alarm information is sent to the corresponding staff. The alarm information includes the fault type, predicted occurrence time, probability, and recommended processing method. The staff makes corresponding processing preparations based on the alarm information, such as preparing to switch to the backup power supply, formulating an automatic restart plan, or conducting equipment inspections.

[0180] Systematic and automatic fault handling: When a fault occurs and no response is received from the staff, the system can handle it accordingly according to the preset automatic handling strategy. The handling strategy includes automatic switching of backup power supply, automatic restart of equipment, isolation of faulty parts to prevent the impact from expanding, etc. During the handling process, the system continuously monitors the equipment status and feeds back the handling results to the staff.

[0181] Record the fault and fault type and report it to the relevant responsible department. The system records the time, type, handling method and handling results of the fault. Generate a fault report for the staff to analyze the cause of the fault, optimize the fault prediction model and improve the handling strategy.

[0182] The technical implementation and optimization process includes: real-time monitoring and feedback, model training and optimization, user interface and interaction, multi-source data fusion, security and reliability, etc. That is, the Internet of Things technology is used to realize real-time monitoring of the power supply of the measurement and control device, and the data is transmitted to the fault prediction system in real time. New fault data is collected regularly for training and optimizing the neural network model to improve the accuracy and efficiency of fault prediction. A friendly user interface is developed to facilitate staff to view alarm information, handle faults, and analyze fault reports. Combine multi-source data (such as historical data, environmental data, etc.) for fault prediction to improve the comprehensiveness and accuracy of the prediction. Safety and reliability factors are fully considered in system design to ensure that the system can operate stably and accurately predict faults in various complex environments.

[0183] In addition, the specific process of early warning evaluation is specified as follows in the method of this embodiment: Figure 2 The following steps are shown: Step 1: Collect physical parameter characteristics, input fault warning evaluation functions under various fault scenarios, and proceed to step 2; Step 2: Determine whether the function undergoes nonlinear changes. If so, proceed to step 3; otherwise, proceed to step 1 to re-collect physical parameter characteristics. Step 3: Determine whether the order in which the nonlinear changes in the physical parameter characteristics are consistent with the fault characteristics. If so, enter the matching stage and proceed to step 4. If not, proceed to step 1 to re-collect the physical parameter characteristics. Step 4: Match the trend shape and proceed to step 5, match the rate of change and proceed to step 6, match the amplitude and proceed to step 7; Step 5: Use the shape overlap ratio as an evaluation value as a calculation parameter in step 8; Step 6: Determine whether the change rate is within the set change rate range. If so, calculate the change rate evaluation value as the parameter of step 8, otherwise, take 0; Step 7: Determine whether the amplitude is within the set amplitude change rate range. If so, calculate the amplitude evaluation value, otherwise take 0; Step 8: Based on the shape overlap rate, change rate evaluation value and amplitude evaluation value obtained in steps 5, 6 and 7 as basic parameters, weighted calculation is performed to obtain the evaluation value of the fault warning.

[0184] The above-mentioned early warning process is described in this embodiment by way of example. Assuming that there is a power supply of a measurement and control device, its electrical characteristics (voltage, current), temperature characteristics and operating noise spectrum are sampled. The change curves of these physical parameter characteristics have been input into the trained deep convolutional neural network to calculate the overlap rate with the corresponding curve in the "fault transition-parameter characteristic" mapping relationship.

[0185] First, the coincidence rate is calculated. The neural network outputs the coincidence rate of the curves in the mapping relationship between electrical characteristics, temperature characteristics and noise spectrum and the "fault transition-parameter characteristics", which are σ_E, σ_T and σ_Q respectively. Assume that these three values ​​are 0.8 (electrical characteristics), 0.7 (temperature characteristics) and 0.9 (noise spectrum).

[0186] Next, the fault warning evaluation value is calculated. Now the three coincidence rate values ​​and other parameters of the fault warning evaluation function are known. Assume that there are three fault scenarios (M1, M2, M3), and the corresponding physical parameter characteristic change rate, amplitude, etc. under each fault scenario have a certain range and standard value. For simplicity, only one fault scenario M1 is considered, and only the coincidence rate of electrical characteristics and temperature characteristics is considered (the noise spectrum is omitted in this example).

[0187] Assume that G_M1 is the fault warning evaluation value of fault scenario M1, and the following parameters are known: n = 2 (considering electrical characteristics and temperature characteristics); σ_1=0.8(electrical characteristic coincidence rate); σ_2=0.7(temperature characteristic overlap rate) Ord_1 and Ord_2 are the order weights of the features undergoing nonlinear changes (both assumed to be 1 since they occur simultaneously); α_1, α_2, α_3 are the weights of the rate of change, amplitude, and trend shape (assumed to be 1 / 3).

[0188] S_k1, R_k1, A_k1 are the rate of change, amplitude, and trend shape matching of the electrical characteristics under fault scenario M1 (assumed to be 0.9, 0.8, and 0.85, respectively); S_k2, R_k2, A_k2 are the rate of change, amplitude, and trend shape matching of the temperature characteristics under fault scenario M1 (assumed to be 0.8, 0.7, and 0.75, respectively); R_a1, R_v1, A_a1, A_v1 are the standard values ​​and actual values ​​(assuming they are known) of the electrical characteristics in the “fault transition-parameter characteristics” mapping relationship; R_a2, R_v2, A_a2, A_v2 are the standard values ​​and actual values ​​(assuming they are known) of the temperature characteristics in the "fault transition-parameter characteristics" mapping relationship.

[0189] According to the expression of the fault warning evaluation function, calculate G_M1: [G_{M1}=\prod_{i=1}^{2}(\sigma_i\cdot Ord_i)\cdot\left(\frac{1}{2}\sum_{i =1}^{2}\left[\frac{1}{3}(S_{ki}+R_{ki}(1-\frac{|R_{ai}-R_{vi}|}{R_{vi}})+A_{ki}(1-\frac{|A_{ai}-A_{vi}|}{A_{vi}}))\right]\right)] Substitute the specific values ​​into the above formula to obtain the calculation result of G_M1.

[0190] Then, the fault type is screened and selected, the G_Mk values ​​of all fault scenarios are compared, and the fault scenario with the largest value greater than the preset warning evaluation threshold (85%) is selected as the estimated power supply fault type of the measurement and control device.

[0191] Finally, the change rate matching evaluation is performed to determine whether the change rate of the detected physical parameter feature is within the range of the physical parameter feature change rate. If not, it is directly taken as 0. If yes, the gap value is calculated: (\text{Gap}R=|R{ki}-\text{Average}(R_k)|) Where (\text{Average}(R_k)) is the average value of the change rate of the physical parameter characteristics in the mapping set. The matching evaluation value can be expressed as: (R_{ki}\cdot(1-\frac{\text{Gap}_R}{\text{Range}_R})) Where (\text{Range}_R) is the range of the characteristic rate of change of the physical parameter.

[0192] The benefits and advantages of the above warning process settings are mainly reflected in the following aspects: Through the recognition and analysis of the physical parameter characteristic change curve by deep convolutional neural network (CNN), local characteristics and overall trends can be captured, thereby improving the accuracy of fault identification. The characteristics of CNN enable the model to automatically learn and extract complex feature patterns, thereby more accurately matching the corresponding curve in the "fault transition-parameter feature" mapping relationship.

[0193] The early warning evaluation function not only considers the trend shape overlap rate of the physical parameter characteristics, but also comprehensively considers factors such as the rate of change and amplitude, making the evaluation results more comprehensive and accurate. This multi-dimensional evaluation method can more comprehensively reflect the characteristics of power supply failures and reduce the possibility of false alarms and missed alarms. The early warning system can add or modify the mapping relationship between fault scenarios and corresponding physical parameter characteristics as needed to adapt to different application scenarios and fault types. In addition, due to the use of a neural network model, the system can continuously improve the accuracy and efficiency of early warnings through continuous learning and optimization.

[0194] The early warning process is automated, reducing the need for manual intervention and improving work efficiency. At the same time, the system can monitor the changes in physical parameter characteristics in real time and immediately trigger an early warning when an abnormality is found, ensuring that the fault can be responded to and handled in a timely manner. By comparing the early warning evaluation values ​​under different fault scenarios, the system can accurately locate and classify the type of power supply fault, providing valuable reference information for subsequent fault handling and repair. By setting and adjusting the early warning evaluation threshold, the system can flexibly adjust the sensitivity of the early warning according to actual needs and risk tolerance. This helps to reduce unnecessary false alarms and interference while ensuring the accuracy of the early warning.

[0195] Embodiment 5: In the power supply fault warning method of a protection and control device of this embodiment, in the prediction and processing of the power supply fault of the control device, the comprehensive consideration of the physical parameter characteristics is the key to improving the prediction accuracy and processing efficiency. This paper will start from the existence (σ_i), occurrence order (Ord_i), change trend shape (S_ki), change rate (R_ki) and amplitude (A_ki) of the physical parameter characteristics, and conduct a more detailed classification and analysis of the fault type, and propose corresponding processing scheme optimization.

[0196] Comprehensive consideration and fault judgment of physical parameter characteristics, including the separate and comprehensive judgment of the existence of physical parameter characteristics (σ_i), the order of occurrence of physical parameter characteristics (Ord_i), the shape of physical parameter change trend (S_ki), the rate of change of physical parameters (R_ki) and the amplitude of physical parameters (A_ki).

[0197] For the existence of physical parameter characteristics (σ_i), the first thing to consider in fault prediction is the existence of physical parameter characteristics. For example, when the power supply voltage of the measurement and control device suddenly disappears, it may mean that a power failure has occurred. Therefore, it is necessary to monitor key parameters such as power supply voltage and current in real time. Once a parameter value is found to be abnormal (such as zero or out of a reasonable range), a fault warning should be triggered immediately.

[0198] When setting thresholds related to the existence of physical characteristic parameters, first of all, it is necessary to understand the range of each physical parameter of the device or system during normal operation. These ranges can be determined by the device's specification, historical operating data, or expert experience. When setting the threshold, it should be ensured that the threshold is outside this normal operating range so that a fault warning can be triggered when the parameter value exceeds the normal range. A certain safety margin should also be considered. The safety margin is set to cope with short-term fluctuations or abnormal conditions that may occur in the device or system. When setting the threshold, a certain safety margin should be added outside the normal operating range to ensure that even if there is a short fluctuation in the parameter value, the fault warning will not be mistakenly triggered.

[0199] The setting of thresholds should also take into account the sensitivity to faults. If the threshold is set too high, some potential faults may not be discovered in time; if the threshold is set too low, it may cause false alarms and unnecessary fault warnings. Therefore, when setting the threshold, it is necessary to weigh the relationship between fault sensitivity and false alarm rate to find a suitable balance. As the device or system operates and data accumulates, the threshold may need to be adjusted dynamically. For example, if it is found that the fluctuation range of a certain parameter expands as the operating time of the device or system increases, the threshold needs to be adjusted accordingly to adapt to this change. In addition, algorithms such as machine learning can be used to automatically optimize and adjust the threshold.

[0200] In practical applications, it is often necessary to consider the comprehensive judgment of multiple physical parameters. For example, when the power supply voltage suddenly disappears, it may also be necessary to consider the changes in other parameters such as current and power. Therefore, when setting the threshold, it is necessary to comprehensively consider the relationship and mutual influence between multiple parameters to improve the accuracy and reliability of fault warning.

[0201] For the order of occurrence of physical parameter characteristics (Ord_i), some faults will show a series of changes in the order of physical parameter characteristics before they occur. By analyzing the order of occurrence of these parameter characteristics, the fault type can be judged more accurately. For example, power supply overvoltage faults often occur in the process of power supply voltage first increasing and then decreasing. Therefore, corresponding logical judgment conditions can be set. When the parameter characteristics appear in a specific order, it is considered that a certain fault may occur.

[0202] The key to determining the order of occurrence of physical parameter characteristics (Ord_i) is to define a series of logical conditions and order rules so that when the parameter characteristics appear according to these rules, the potential fault type can be accurately identified. First, it is necessary to clarify the order of changes in the physical parameter characteristics before the fault occurs. This is usually based on historical data, expert knowledge, or experimental observations. For example, the characteristic order of a power supply overvoltage fault may be that the power supply voltage first increases above the upper threshold and then decreases to the normal range or lower.

[0203] For each parameter feature, set a corresponding threshold value to determine whether the parameter is in an abnormal state. These thresholds should be determined based on the normal operating range of the equipment, safety standards, or statistical data. Record the timestamp of each parameter feature exceeding its threshold. Timestamp information is critical to determine the order in which the parameter features occur. Write logic code or algorithms to check whether the parameter features occur in the predetermined order. This can be achieved by comparing timestamps to ensure that one parameter feature occurs before or after another.

[0204] Considering the noise, interference or abnormal conditions that may exist in the actual environment, the logical judgment should have a certain fault tolerance. For example, a certain time window can be set to allow parameter characteristics to occur within a certain time range in a predetermined order. When the parameter characteristics occur in a predetermined order, the logical judgment should be able to identify the corresponding fault type. This can be achieved by triggering an alarm, recording a log or sending a notification. Use historical data or simulation data to verify the logical judgment to ensure its accuracy and reliability. According to the actual operation situation and feedback, the conditions and parameters of the logical judgment are continuously adjusted and optimized.

[0205] In some cases, parameter features may appear in a different order but still indicate the same fault type. Logical judgment should be able to handle these abnormal situations and avoid false positives or negatives. Monitor the operation of the system and continuously collect and analyze new data. Continuously update and optimize the conditions and parameters of logical judgment based on new data and experience. Provide operators with intuitive user interfaces and visualization tools so that they can easily view and analyze the order of changes in parameter features and the types of faults identified.

[0206] For the shape of the physical parameter change trend (S_ki), by comparing the real-time detected physical parameter change curve with the curve in the "fault transition-parameter feature" mapping relationship, the degree of overlap between the two can be evaluated to determine the possibility of the fault. In order to identify faults more accurately, more complex shape matching algorithms, such as the dynamic time warping (DTW) algorithm, can be introduced to process curves with similar shapes at different speeds.

[0207] When judging the degree of overlap between the shape of the physical parameter change trend (S_ki) and the curve in the "fault transition-parameter feature" mapping relationship, introducing the dynamic time warping (DTW) algorithm or other shape matching algorithm is a good method because it can handle curves with similar shapes at different speeds. First, it is necessary to collect and analyze historical data to establish a "fault transition-parameter feature" mapping relationship. This usually includes a series of physical parameter change curves associated with different fault types.

[0208] Before the comparison, the real-time detected physical parameter change curve and the curve in the mapping relationship are preprocessed, such as normalization, smoothing, etc., to eliminate the influence of factors such as scale and noise. Select a suitable shape matching algorithm, such as the DTW algorithm. The DTW algorithm can measure the similarity between two time series by nonlinearly stretching or compressing the time axis. The shape matching algorithm is applied to calculate the similarity between the real-time detected physical parameter change curve and the curve in the mapping relationship. This is usually achieved by the algorithm outputting a distance metric or similarity score.

[0209] One or more thresholds are set based on historical data and actual needs. These thresholds can be determined based on the statistical distribution of similarity scores (such as mean, standard deviation), expert advice, or a trade-off between the sensitivity of fault detection and the false alarm rate. When the similarity between the real-time detected physical parameter change curve and a curve in the mapping relationship exceeds the set threshold, it is considered that a fault associated with the curve may have occurred. In some cases, a change in a physical parameter may be associated with multiple fault types at the same time. Therefore, this situation needs to be considered when making judgments, and it may be necessary to set multiple thresholds or use more complex classification algorithms.

[0210] Use historical data of known faults for verification to check the rationality of threshold settings and the accuracy of fault identification. Adjust and optimize the thresholds and parameters of the shape matching algorithm based on the verification results. As the system runs and data accumulates, it may be necessary to regularly update the "fault transition-parameter feature" mapping relationship and threshold settings to adapt to changes in system status. Provide an intuitive user interface and visualization tools to display the real-time detected physical parameter change curves, the curves in the mapping relationship, and the similarity calculation results, so that operators can monitor and analyze fault conditions.

[0211] For the rate of change of physical parameters (R_ki), the rate of change is an important indicator that describes how fast the physical parameters change. In fault prediction, the urgency of the fault can be judged by calculating the rate of change of physical parameters. For example, when the rate of change of power supply voltage suddenly increases, it may mean that a power failure is about to occur. In order to more accurately evaluate the rate of change, the sliding window technology can be used to calculate the average rate of change over a period of time in real time and set the corresponding threshold for judgment.

[0212] When using the rate of change of physical parameters (R_ki) for fault prediction, sliding window technology and threshold setting are key steps. First, it is necessary to clarify how to calculate the rate of change of physical parameters. The rate of change is usually defined as the amount of change in the parameter value per unit time, that is, R_ki = (V_t-V_{t-1}) / Δt, where V_t and V_{t-1} are the parameter values ​​at the current and previous moments, respectively, and Δt is the time interval. The size of the sliding window (that is, the time interval Δt) needs to be selected according to the specific application scenario. Shorter windows can respond to rapid changes in parameter values ​​more quickly, but may be affected by noise; longer windows are more stable, but respond more slowly.

[0213] Within the sliding window, the average rate of change of the parameter value can be calculated to reduce the impact of noise. This can be achieved by averaging all the rates of change within the window. The setting of the threshold is usually based on historical data, expert experience or safety standards. The level of parameter change rate that is usually reached before the failure can be determined by analyzing historical fault data, and the threshold is set accordingly. The threshold can be a fixed value or a value that is dynamically adjusted according to the system status.

[0214] The average rate of change of a physical parameter is calculated in real time using a sliding window technique and compared to a set threshold. If the average rate of change exceeds the threshold, it may mean that a failure is imminent and that appropriate action needs to be taken. In some cases, the rate of change of a parameter may have different importance over time. For example, the rate of change of a parameter may be naturally higher when the equipment is starting up or shutting down, and lower during stable operation. Therefore, this time dependency needs to be taken into account when setting the threshold.

[0215] Use historical data to verify the threshold to ensure its accuracy and reliability under different fault conditions. Adjust and optimize the threshold and the size of the sliding window based on the verification results. In complex systems, it may be necessary to consider the change rates of multiple physical parameters at the same time. In this case, a multi-parameter comprehensive judgment method can be used, such as combining the change rates of multiple parameters by weighting or voting.

[0216] For the physical parameter amplitude (A_ki), the amplitude reflects the size of the physical parameter. In fault prediction, the type and severity of the fault can be determined by analyzing the amplitude change of the physical parameter. For example, in a power undervoltage fault, the amplitude of the power supply voltage will be lower than the normal range; while in a power overvoltage fault, the amplitude of the power supply voltage will exceed the normal range. Therefore, it is necessary to set a reasonable amplitude threshold and judge the possibility of the fault based on the amplitude change detected in real time.

[0217] In fault prediction, it is an important and common method to judge the type and severity of faults by the amplitude of physical parameters (A_ki). Collect historical data, including the amplitude of physical parameters during normal operation and the amplitude data when the fault occurs. Perform statistical analysis on the data to determine the amplitude range and distribution of physical parameters under normal conditions.

[0218] Based on historical data, set one or more amplitude thresholds for each fault type. These thresholds should be able to distinguish between normal operation and fault conditions. Thresholds can be set based on statistical methods (such as mean plus or minus standard deviation) or based on expert knowledge. Different thresholds may need to be set for different fault types. Use a portion of historical data to verify the set thresholds to check their accuracy and reliability in distinguishing between normal and fault conditions. Adjust and optimize the thresholds based on the verification results to ensure that they work effectively in various situations.

[0219] During system operation, the amplitude of the physical parameter is detected in real time. The amplitude detected in real time is compared with the set threshold to determine whether the current state is normal or faulty. If the amplitude exceeds or falls below a certain threshold, it may mean that a fault associated with the threshold has occurred. In some cases, the amplitude of the physical parameter may change over time or with changes in the system state. Therefore, dynamic thresholds can be considered to adjust the thresholds in real time according to the system state. Dynamic thresholds can be implemented through machine learning or adaptive algorithms to better adapt to changes in system states.

[0220] In complex systems, it may be necessary to consider the amplitude changes of multiple physical parameters at the same time. In this case, a multi-parameter comprehensive analysis method can be used to comprehensively consider the changes in multiple parameters to determine the possibility of a fault. For example, a weight vector can be set to assign different weights to each parameter, and the fault type can be determined based on the weighted amplitude changes. In some cases, the amplitude changes of physical parameters may be time-dependent. For example, during startup or shutdown, the parameter amplitude may naturally fluctuate greatly. This time dependency needs to be considered when setting thresholds and judging faults to avoid false positives or false negatives.

[0221] In addition to determining whether a fault has occurred, the severity of the fault can also be assessed based on the change in the amplitude of the physical parameter. Different amplitude ranges can be set to correspond to different fault severities, and the severity of the current fault can be assessed based on the amplitude detected in real time. An intuitive user interface and visualization tool are provided to display the real-time detected physical parameter amplitude, threshold, fault type and severity information, making it easier for operators to monitor and analyze fault conditions.

[0222] Fault type classification and processing scheme optimization, including analysis and processing of different fault types such as power undervoltage fault, power overvoltage fault, power fluctuation fault, power ripple fault, etc.

[0223] When the power supply voltage is detected to be lower than the preset threshold, it is judged as a power supply undervoltage fault. At this time, the following treatment solutions can be adopted: Backup power supply switching: Immediately start the backup power supply system to ensure that the measurement and control device can continue to work. Fault warning notification: Send a fault warning notification to relevant personnel to remind them to pay attention to and deal with the fault. Fault cause analysis: Analyze the cause of the fault, find out the cause of the undervoltage (such as power line failure, power module damage, etc.), and take corresponding repair measures.

[0224] For power supply undervoltage faults, in addition to the already mentioned solutions of backup power supply switching, fault warning notification and fault cause analysis, further improvements can be made. By introducing advanced monitoring systems and intelligent diagnostic algorithms, real-time monitoring and rapid diagnosis of power supply voltage can be achieved. When an undervoltage fault is detected, the system can automatically analyze the cause of the fault and give corresponding treatment suggestions. According to the real-time status and operating environment of the power grid, the preset threshold of the power supply voltage is dynamically adjusted. This can more accurately determine whether the power supply is in an undervoltage state and reduce false alarms and missed alarms.

[0225] Install a voltage stabilizer in the power supply system to stabilize the output voltage. The voltage stabilizer can automatically adjust the output voltage to ensure that it fluctuates within a preset range, thereby avoiding the occurrence of power supply undervoltage faults. Reduce the load pressure on the power grid through reasonable load management strategies. For example, limit the power consumption of certain non-critical equipment during peak hours, or adopt a time-sharing power system to reduce the load fluctuation of the power grid. If the aging or overloaded power grid causes frequent undervoltage faults, consider upgrading or optimizing the power grid. For example, increase the capacity of the transformer, replace thicker wires, optimize the power grid structure, etc. to improve the power supply capacity and stability of the power grid.

[0226] Regularly maintain and inspect the power supply system and related equipment to promptly identify and handle potential problems. For example, check whether the power supply line is aging, whether the power module is damaged, whether the contact point is loose, etc., to ensure the normal operation of the power supply system. Strengthen the training and awareness of relevant personnel to make them understand the hazards and countermeasures of power undervoltage failure. This can improve their ability to respond to failures and reduce the impact of failures on measurement and control devices and systems.

[0227] When the power supply voltage is detected to exceed the preset threshold, it is judged as a power supply overvoltage fault. At this time, the following treatment solutions can be adopted: Overvoltage protection mechanism activation: Immediately activate the overvoltage protection mechanism, such as disconnecting the power supply, reducing the output voltage, etc., to prevent equipment damage or safety accidents. Fault recording and analysis: Record the time and voltage value of the fault to provide a basis for subsequent fault analysis and processing. Power supply system optimization: Optimize and upgrade the power supply system to improve the stability and reliability of the power supply.

[0228] When the power supply voltage is detected to exceed the preset threshold and it is judged as a power supply overvoltage fault, in addition to the above-mentioned overvoltage protection mechanism activation, fault recording and analysis, and power supply system optimization, further improvements can be made. Introducing an intelligent overvoltage protection and recovery mechanism, when overvoltage is detected, not only will the power supply be immediately disconnected or the output voltage reduced to protect the equipment, but the power supply will also be automatically restored after the overvoltage problem is resolved. This can reduce human intervention and improve the automation of the system.

[0229] Implement dynamic voltage regulation in the power supply system to adjust the output voltage in real time according to the load and grid status to ensure that it is always within the preset safety range. This helps reduce the occurrence of overvoltage faults and improves the stability of the power supply system. Use power redundancy design to provide additional voltage stability and reliability by connecting multiple power modules in parallel or series. When a power module fails or overvoltage occurs, other modules can take over the power supply task to ensure that the measurement and control device can continue to work.

[0230] Combined with the results of fault recording and analysis, predictive maintenance technology is used to predict potential problems in the power supply system. By regularly checking and replacing wearing parts, abnormal conditions of the power supply system are detected in advance, and appropriate measures are taken before an overvoltage fault occurs, the fault can be avoided or its impact can be mitigated. A backup power supply is set up in the power supply system, and it is automatically switched to the backup power supply when an overvoltage fault is detected. This ensures that the measurement and control device can continue to work when an overvoltage fault occurs, while avoiding damage to the equipment caused by overvoltage of the main power supply.

[0231] Through the remote monitoring and management system, the operating status and voltage value of the power supply system are monitored in real time. When an overvoltage fault is detected, the power supply can be disconnected, the voltage can be adjusted, or other emergency measures can be taken immediately through remote operation to reduce the impact of the fault on the system. The power quality is regularly tested, including indicators such as voltage stability and waveform distortion. By timely discovering and solving power quality problems, the occurrence of overvoltage faults can be reduced and the reliability and stability of the power supply system can be improved.

[0232] When the power supply voltage is monitored to fluctuate violently in a short period of time, it is judged as a power supply fluctuation fault. At this time, the following treatment solutions can be adopted: Power supply filtering: Use power supply filters to filter the power supply to reduce the impact of power supply fluctuations on the equipment; Stability optimization: Optimize the stability of the power supply system, such as adding voltage stabilization circuits, optimizing power supply design, etc. Monitoring and alarm: Real-time monitoring of power supply fluctuations, once exceeding the preset threshold, trigger an alarm to notify relevant personnel to handle.

[0233] When the power supply voltage is monitored to fluctuate violently in a short period of time and it is judged to be a power supply fluctuation fault, in addition to the power supply filtering processing, stability optimization, monitoring and alarm processing solutions you mentioned, further improvements can be made. Introduce intelligent regulation and control technology, such as using advanced control algorithms and microprocessors to analyze the causes of power supply fluctuations in real time and automatically adjust power supply parameters. This technology can predict and respond to changes in the power system, thereby quickly stabilizing voltage fluctuations. Add dynamic reactive power compensation equipment (such as SVC, SVG, etc.) to the power system. These devices can compensate for the reactive power in the system in real time and improve the power factor of the system, which helps to stabilize the voltage and reduce fluctuations.

[0234] Supercapacitors and energy storage systems (such as battery energy storage systems) can provide rapid energy support when power fluctuates, thereby smoothing voltage fluctuations. These devices can absorb or release a large amount of energy in a short period of time, effectively alleviating the impact of power fluctuations on equipment. Optimize and transform the lines in the power system, such as reducing line length, using thicker cables, optimizing line layout, etc., to reduce line impedance and resistance, reduce voltage loss and fluctuations. Strengthen the construction of power monitoring systems, and collect and analyze data on power fluctuations. Through big data analysis technology, the laws and trends of power fluctuations can be identified, providing data support for optimizing the power system.

[0235] When designing equipment, improve its tolerance to power fluctuations. For example, by optimizing circuit design and using components with higher voltage resistance, the equipment can resist the impact of power fluctuations to a certain extent. Distributed power systems provide additional power support to the system by connecting multiple small power generation devices (such as solar energy, wind energy, etc.) to the power grid. When the power supply fluctuates, these power generation devices can automatically adjust their output power to help stabilize the voltage and reduce fluctuations. Strengthen the training and education of relevant personnel so that they understand the impact of power fluctuations on equipment and how to deal with power fluctuation failures. By improving the skill level and awareness level of personnel, power fluctuation failures can be better handled and their impact on the system can be reduced.

[0236] When high-frequency noise or clutter is detected in the power supply output, it is judged as a power supply ripple fault. At this time, the following treatment solutions can be adopted: Ripple suppression technology: Use ripple suppression technology to reduce the ripple level in the power supply output, such as adding filter capacitors, optimizing power supply design, etc. Power supply quality improvement: Select high-quality power supply components and modules to improve the quality and reliability of the power supply. Monitoring and evaluation: Regularly perform ripple detection on the power supply output to evaluate whether the ripple level exceeds the standard and take corresponding measures to deal with it.

[0237] When high-frequency noise or clutter is detected in the power supply output, that is, when the power supply ripple fault occurs, in addition to the ripple suppression technology, power quality improvement, monitoring and evaluation solutions you mentioned, further optimization can be performed. In addition to ordinary filter capacitors, more advanced filters such as LC filters, π-type filters or active filters can be used. These filters can more effectively remove high-frequency noise and clutter in the power supply output. By introducing a feedback control loop, the ripple level of the power supply output can be monitored in real time, and the power supply output can be adjusted according to the feedback signal to keep the ripple within an acceptable range.

[0238] During the design and installation of the power supply, pay attention to the rationality of the power supply layout and wiring. Reasonable layout and wiring can reduce the propagation of electromagnetic interference and noise, thereby reducing the power supply ripple level. Apply soft switching technology in power supply design, such as PWM (pulse width modulation) or PFM (pulse frequency modulation). These technologies can reduce the noise and clutter generated during the power switching process, thereby improving the power supply ripple performance. Choose low-noise power modules, which are designed to reduce the generation of noise and clutter, so they can provide higher quality power output.

[0239] Integrate multiple power modules or components together to form an integrated power system. This design can reduce the propagation path of electromagnetic interference and noise and improve the power ripple performance of the entire system. Add shielding and isolation measures to the power design, such as using metal shielding covers, isolation transformers, etc. These measures can effectively isolate the electromagnetic interference between the power supply and the load, thereby reducing the power ripple level. Digital power technology uses digital control algorithms to precisely control the power supply, which can achieve more efficient power management and lower ripple levels. In addition to regular ripple detection, the real-time monitoring and fault diagnosis functions of the power supply can also be enhanced. By real-time monitoring of the ripple level and other performance indicators of the power supply, potential power supply problems can be discovered and resolved in a timely manner.

[0240] It should be understood that the embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope limited by the appended claims of the application.

Claims

1. A power failure early warning method for a protection and measurement control device, characterized in that: The following steps are involved: S1: Establish a transition mapping relationship from normal state to fault state based on the physical parameter characteristic changes of the power supply of the protection device under the historical fault scenarios, so as to construct a power supply fault warning model for each fault scenario; S2: obtaining physical parameter characteristics of the power supply of the measurement and control device at a set sampling frequency, and adjusting the sampling frequencies according to the physical parameter characteristics obtained by sampling; S3: Import the sampled physical parameter characteristics into the power supply fault warning model to match the fault scenario of the power supply of the measurement and control device; S4: Execute preset processing preparation according to the matching predicted fault scenario.

2. A power failure early warning method for a protection and measurement control device according to claim 1, characterized in that: The step S1 specifically includes the following process: S101: Count the changes in physical parameter characteristics of different power supplies of the test and protection devices under various fault scenarios, and establish a corresponding "fault transition-parameter characteristic" mapping relationship from a normal state to a fault state; S102: According to the "fault transition-parameter feature" mapping relationship under different fault scenarios, the change characteristics of the physical parameter features are extracted, and a power supply fault early warning evaluation function is designed.

3. A power failure early warning method for a protection and measurement control device according to claim 2, characterized in that: The "fault transition-parameter feature" mapping relationship is a change curve of the physical parameter feature during the fault transition, where the horizontal axis is time and the vertical axis is the physical parameter feature.

4. A power failure early warning method for a protection and measurement control device according to claim 1, 2 or 3, characterized in that: The physical parameter characteristics include electrical characteristics, temperature characteristics and operating noise spectrum; For different fault scenarios, the physical parameter features corresponding to the occurrence or impending occurrence of a fault are selected according to the nonlinear change characteristics of the physical parameter features in the mapping relationship; According to the selected physical parameter characteristics and the order of their nonlinear changes as the function starting conditions, the change rate, amplitude characteristics and change trend shape of the physical parameter characteristics are weighted calculated to design the power supply fault warning evaluation function.

5. A power failure early warning method for a protection and measurement control device according to claim 4, characterized in that: In the power supply fault evaluation function, a Boolean nonlinear marker is used to determine whether there is a nonlinear change in the physical parameter characteristics selected for the corresponding fault scenario; a Boolean occurrence order marker is used to determine whether the order of the nonlinear changes in the physical parameter characteristics conforms to the fault characteristics.

6. A power failure early warning method for a protection and measurement control device according to claim 4, characterized in that: The trend shape is compared with each physical parameter feature in the mapping relationship, and the shape overlap rate is used as the evaluation value, which is weighted and used to evaluate the matching fault scenario.

7. A power failure early warning method for a protection and measurement control device according to claim 5 or 6, characterized in that: For the change rate and amplitude, the maximum and minimum values ​​of the change rate and amplitude of the physical parameter characteristics in the mapping relationship are respectively selected as the range of the corresponding physical parameter characteristic change rate and amplitude; it is determined whether the change rate and amplitude of the sampled physical parameter characteristics are within the range of the physical parameter characteristic change rate and amplitude. If so, they are weighted and used to evaluate the matching fault scenario, otherwise they are directly taken as 0.

8. A power failure early warning method for a protection and measurement control device according to claim 7, characterized in that: Taking the change rate and amplitude average value of each physical parameter feature in the mapping relationship as the standard, the difference between the change rate and amplitude of the sampled physical parameter feature and the change rate and amplitude average value of the physical parameter feature in the mapping relationship is calculated as an evaluation calculation.

9. A power failure early warning method for a protection and measurement control device according to claim 4, characterized in that: Different sampling frequencies are set for different physical parameter characteristics, and the upper and lower limits of the sampling frequencies are set respectively.

10. A power failure early warning method for a protection and measurement control device according to claim 1 or 9, characterized in that: The electrical characteristic sampling frequency is coupled and adjusted according to the sampled temperature data, noise spectrum data and electrical characteristic data.

11. A power failure early warning method for a protection and measurement control device according to claim 10, characterized in that: When the detected electrical characteristic data does not match all the electrical characteristic change curves in the "fault transition-parameter characteristic" mapping relationship, the expression for adjusting the electrical characteristic sampling frequency is: f E =f Emin +(e T ·c T +e Q -c Q )·Δf E When the detected electrical characteristic data matches any electrical characteristic change curve in the "fault transition-parameter characteristic" mapping relationship, the expression for adjusting the electrical characteristic sampling frequency is: f E =f Emax Furthermore, set the limiting conditions: f Emin +(c T +g Q )·Δf E ≤f Emax Among them, f E is the electrical characteristics sampling frequency; f Emax is the preset maximum value of the electrical characteristic sampling frequency; f Emin is the preset minimum value of the electrical characteristic sampling frequency; ε T It is a temperature characteristic fault precursor mark. When the detected temperature characteristic matches the temperature characteristic change curve in the "fault transition-parameter characteristic" mapping relationship, it takes 1; otherwise, it takes 0; ε Q It is a noise spectrum fault precursor mark. When the detected noise spectrum matches the noise spectrum change curve in the "fault transition-parameter feature" mapping relationship, it takes 1; otherwise, it takes 0; γ T is the correlation coefficient between temperature characteristics and electrical characteristics; γ Q is the correlation coefficient between the noise spectrum and the electrical characteristics; Δf E It is the minimum adjustment frequency of the electrical characteristic sampling frequency.

12. A power failure early warning method for a protection and measurement control device according to claim 1 or 9, characterized in that: The temperature sampling frequency is corrected according to the collector current rise rate of the switching device. The expression for the temperature sampling frequency correction is: The constraints are: Among them, f T is the sampling frequency of temperature characteristics; f Tmin is the lower limit of the set temperature characteristic sampling frequency; f Tmax The upper limit of the set temperature characteristic sampling frequency; Δf T The minimum sampling adjustment frequency for temperature characteristics; R IS is the collector current rise rate of the switching device obtained by sampling; R IA It is the standard value of the collector current rise rate of the set switching device.

13. A power failure early warning method for a protection and measurement control device according to claim 1 or 9, characterized in that: For the running noise spectrum characteristics, the noise spectrum sampling frequency is adjusted according to the comparison between the width of the sampled noise spectrum and the width of the noise spectrum during normal operation.

14. A protection and measurement control device power failure early warning system, using a protection and measurement control device power failure early warning method as claimed in any one of claims 1 to 13, characterized in that: include A historical fault database (1), storing physical parameter characteristic data of each fault scenario of the power supply of the measurement and control device in the time period before and after the historical fault; A mapping set (2) stores a corresponding "fault transition-parameter characteristic" mapping relationship from a normal state to a fault state established according to the physical parameter characteristic changes of the power supply of different protection devices under various fault scenarios; A sampling module (4) sets different sampling frequencies for different physical parameter characteristics, and sets upper and lower limits of the sampling frequencies respectively; and adjusts the sampling frequency of the physical parameter characteristics according to the data coupling obtained by sampling; The analysis and calculation module (3) constructs a power supply fault warning model for each fault scenario according to the mapping relationship, imports the sampled physical parameter characteristics into the power supply fault warning model, and matches the fault scenario of the power supply of the measurement and control device; The processing module (5) is configured with different processing modes for different power failure scenarios of the measurement and control device.

15. A protection and measurement control device power failure early warning system according to claim 14, characterized in that: The analysis and calculation module (3) includes a deep convolutional neural network (3-1), which inputs the change curves of each physical parameter feature obtained by sampling into the trained deep convolutional neural network, and calculates the overlap rate between the change curves of each physical parameter feature and the trend shape of the physical parameter feature change curve corresponding to the mapping relationship in the mapping set (2).

16. A protection and measurement control device power failure early warning system according to claim 14 or 15, characterized in that: The analysis and calculation module (3) sorts the average values ​​of the coincidence rates of the characteristic change curves of the physical parameters from large to small, and respectively introduces them into the power supply fault warning evaluation function (3-2) of the corresponding fault scenario to obtain the corresponding fault warning evaluation value.

17. A storage medium for computer-readable storage, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of a power failure early warning method for a protection and measurement and control device as described in any one of claims 1-13.

Images