A digital transformer fault simulation generator and a fault simulation method
Patent Information
- Application Number
- CN202611156151.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-31
- Publication Date
- 2026-10-09
AI Technical Summary
[0003]但是，变压器实际故障信号具备典型的非线性变化特征，且伴随随机扰动特性，故障工况处于动态变化过程，而非固定不变的波形状态
本申请考虑到变压器出现故障后，变压器的所有种类的运行信号都会受到影响，导致正常情况下呈现稳定周期性变化的运行信号的周期与振幅均出现改变，根据运行信号序列在频域内的特征，提取特征频率集合、扰动频率集合、必要故障频率集合和非必要故障频率集合，其中，必要故障频率集合内频率对应同一故障类型的相同故障机理在频域信号中存在的共同的特征频率；由于同一类型的不同严重程度的故障的各频率分量的幅值波动区间存在明显差异，对于必要故障频率集合中的频率，计算故障样本的故障发生时刻的必要信号强度，必要信号强度用于表征故障在发生时必定出现的信号强度，同时，根据非必要故障频率集合中选取的部分频率的幅值和对应的相位，计算对应故障发生时刻的非必要信号强度；考虑到真实的故障受到环境因素的影响，运行信号会出现噪声扰动，根据故障样本的扰动频率集合中所有频率的幅值，计算故障样本的故障发生时刻的扰动信号强度，根据扰动信号强度添加扰动信号，并结合必要信号强度和非必要信号强度，确定故障样本在故障发生时刻的故障模拟信号，解决故障信号模拟无法适配变压器故障信号的非线性变化与随机扰动特性，模拟结果与真实故障情况偏差较大的问题。
Smart Images
Figure CN122882864A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of digital mapping, and in particular to a digital transformer fault simulation generator and a fault simulation method. Background Art
[0002] A transformer is a core electrical device that undertakes the responsibility of conversion and transmission of high-voltage and low-voltage electric energy in the power transmission and distribution links of a power system. During the long-term operation of a transformer, affected by various factors such as insulation aging, winding damage, abnormal operating conditions of the iron core, and line short circuits, various types of faults are prone to occur, causing problems such as equipment outage and power supply interruption. Therefore, accurately identifying and analyzing various faults of transformers is the core work to ensure the operation and maintenance safety of power equipment. In order to avoid the safety risks of real-machine fault tests, reduce test costs, and improve experimental repeatability, fault signal simulation is generally used to replace real-machine tests. By simulating real fault signals to reproduce fault working conditions, fault mechanism analysis and detection verification are completed. Among them, the historical fault waveform playback technology can collect on-site real fault signals or artificially simulated fault signals, and realize fault simulation and replication through waveform playback.
[0003] However, actual transformer fault signals have typical nonlinear variation characteristics, accompanied by random disturbance characteristics, and the fault working condition is in a dynamic change process rather than a fixed waveform state. Traditional historical fault waveform playback can only replicate fixed collected waveforms, which is difficult to cover the full-dimensional dynamic evolution process of faults, and cannot simulate the nonlinear fluctuation and random disturbance changes of fault signals. This results in a large deviation between the simulated fault working conditions and real fault scenarios, and the problems of low simulation accuracy and poor coverage. Summary of the Invention
[0004] In order to solve the above technical problems, the present application provides a digital transformer fault simulation generator and a fault simulation method, and the specific technical solution adopted is as follows: In a first aspect, an embodiment of the present application provides a fault simulation method for a digital transformer fault simulation generator, the method comprising the following steps: collecting different types of operating signals from a preset number of fault samples of each fault type, establishing an operating signal sequence, and calculating the fault degree of the fault samples; for the operating signal sequence of each type of operating signal of the fault samples, determining a characteristic frequency set and a disturbance frequency set according to the amplitudes corresponding to different frequency components and the frequencies corresponding to the amplitudes of the operating signal sequence in the frequency domain, and determining a necessary fault frequency set and a non-essential fault frequency set according to all characteristic frequency sets corresponding to the same type of operating signal determined by all fault samples of the same fault type; For frequencies in the necessary fault frequency set, a linear fit is performed based on the fault severity of the fault sample corresponding to the frequency and the amplitude corresponding to the frequency. Based on the result of the linear fit, the necessary signal strength at the time of fault occurrence of the fault sample is calculated. Based on the amplitude and corresponding phase of a subset of frequencies selected from the non-necessary fault frequency set, the non-necessary signal strength at the time of fault occurrence is calculated. Based on the amplitude of all frequencies in the disturbance frequency set of the fault sample, the disturbance fault coefficient and disturbance phase coefficient of the frequencies in the disturbance frequency set are randomly generated. The disturbance signal strength at the time of fault occurrence of the fault sample is calculated. Combining the necessary signal strength and the unnecessary signal strength, the fault simulation signal of the fault sample at the time of fault occurrence is determined.
[0005] Furthermore, the method for calculating the fault severity of the fault sample is as follows: The root mean square deviation between the operating signal sequence of the fault sample and the preset standard operating signal corresponding to the same type of operating signal is denoted as the fault deviation of the same type of operating signal. The mean of the normalized values of the fault deviations of all types of operating signals corresponding to the fault sample is denoted as the fault degree of the fault sample.
[0006] Furthermore, the method for determining the characteristic frequency set and the perturbation frequency set is as follows: The operating signal sequence of the fault sample is converted to the frequency domain to obtain the amplitude spectrum sequence of the fault sample. The amplitude spectrum sorting sequence of the fault sample is established. The cumulative energy ratio of the absolute value of the amplitude within the amplitude spectrum sorting sequence is calculated. The maximum value of all amplitudes with a cumulative energy ratio greater than the preset main energy threshold is selected and recorded as the boundary amplitude. The set of frequencies corresponding to the absolute values of all amplitudes greater than the threshold amplitude in the amplitude spectrum sorting sequence of the fault sample is denoted as the characteristic frequency set, and the set of frequencies corresponding to the absolute values of all amplitudes less than or equal to the threshold amplitude in the amplitude spectrum sorting sequence of the fault sample is denoted as the disturbance frequency set.
[0007] Furthermore, the method for determining the necessary fault frequency set and the unnecessary fault frequency set is as follows: The necessary fault frequency set is obtained by taking the intersection of all characteristic frequency sets corresponding to the same type of operating signal determined by all fault samples of the same fault type. The set of frequencies that are not in the necessary fault frequency set is defined as the set of frequencies contained in the characteristic frequency set corresponding to the same type of operating signal determined by all fault samples of the same fault type.
[0008] Furthermore, the linear fitting specifically refers to: For any frequency in the necessary fault frequency set, the fault degree of the fault sample corresponding to the frequency is used as the independent variable, and the amplitude corresponding to the frequency is used as the dependent variable. Linear fitting is performed to obtain the goodness of fit of the fitted line corresponding to the frequency.
[0009] Furthermore, the method for calculating the necessary signal strength at the time of fault occurrence of the fault sample is as follows: When the goodness of fit of the fitted line is greater than or equal to the preset goodness of fit threshold, the difference between the fitted value of the amplitude and the corresponding amplitude is recorded as the fitting fluctuation. The mean of all fitting fluctuations greater than 0 is recorded as the upper limit of fluctuation, and the mean of all fitting fluctuations less than 0 is recorded as the lower limit of fluctuation. A random number greater than or equal to the lower limit of fluctuation and less than or equal to the upper limit of fluctuation is generated. The generated random number is recorded as the frequency fluctuation amplitude of the corresponding frequency. The sum of the frequency fluctuation amplitude of the frequency and the fitted value of the amplitude corresponding to the frequency is recorded as the necessary fault coefficient of the frequency. When the goodness of fit of the fitted line is less than the preset goodness of fit threshold, the range of values between the minimum and maximum values of the amplitude of the fault sample corresponding to the frequency is denoted as the amplitude range. The amplitude range is divided into sub-ranges. Random numbers are generated randomly in the sub-ranges where the amplitude of the frequency in the necessary fault frequency set is located using a normal distribution. The generated random numbers are denoted as the necessary fault coefficients of the frequency in the necessary fault frequency set. Calculate the necessary signal strength at the time of the corresponding fault occurrence based on the necessary fault coefficients of the frequencies in the necessary fault frequency set.
[0010] Furthermore, the method for calculating the unnecessary signal strength at the moment of the fault occurrence is as follows: Randomly generate an integer greater than or equal to 1 and less than or equal to the total number of frequencies in the set of unnecessary fault frequencies. Randomly select from the set of unnecessary failure frequencies For each frequency in the selected set of unnecessary fault frequencies, the amplitude and corresponding phase of each frequency are obtained. For any frequency in the selected set of unnecessary fault frequencies, the maximum and minimum values of all amplitudes of the frequency are used as the upper and lower limits of the interval, respectively. Random numbers are generated within the interval range and recorded as the unnecessary fault coefficient of the frequency. The maximum and minimum values of all phases corresponding to the frequency are used as the upper and lower limits of the interval, respectively. Random numbers are generated within the interval range and recorded as the unnecessary phase coefficient of the frequency. Based on the unnecessary fault coefficients and unnecessary phase coefficients of the frequencies in the unnecessary fault frequency set of the fault sample, calculate the unnecessary signal strength of the fault sample at the time of fault occurrence.
[0011] Furthermore, the method for generating the frequency disturbance fault coefficient and disturbance phase coefficient is as follows: Extract the maximum amplitude of all frequencies in the disturbance frequency set of the fault sample, and record the maximum of the maximum amplitude values corresponding to all fault samples as the upper limit of amplitude. Randomly generate random numbers that are greater than or equal to 0 and less than or equal to the upper limit of amplitude. The number of random numbers generated is equal to the number of frequencies in the disturbance frequency set. The generated random numbers are recorded as the disturbance fault coefficients of each frequency in the disturbance frequency set of the fault sample. Randomly generate greater than or equal to 0 and less than or equal to The number of random numbers generated is equal to the number of frequencies in the disturbance frequency set. The generated random numbers are then recorded as the disturbance phase coefficients of each frequency in the disturbance frequency set of the fault sample.
[0012] Furthermore, the method for determining the fault simulation signal of the fault sample at the time of fault occurrence is as follows: The sum of the necessary signal strength, unnecessary signal strength, and disturbance signal strength of the fault sample at the time of fault occurrence is used as the fault simulation signal of the fault sample at the time of fault occurrence.
[0013] Secondly, another embodiment of this application provides a digital transformer fault simulation generator, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the fault simulation method of the digital transformer fault simulation generator described above.
[0014] The embodiments of this application have at least the following beneficial effects: This application considers that when a transformer fails, all types of operating signals of the transformer are affected, causing changes in the period and amplitude of the normally stable and periodic operating signals. Based on the characteristics of the operating signal sequence in the frequency domain, a set of characteristic frequencies, a set of disturbance frequencies, a set of necessary fault frequencies, and a set of unnecessary fault frequencies are extracted. The frequencies in the necessary fault frequency set correspond to common characteristic frequencies of the same fault mechanism of the same fault type in the frequency domain signal. Since the amplitude fluctuation ranges of each frequency component of faults of different severity of the same type differ significantly, for the frequencies in the necessary fault frequency set, the necessary signal strength at the time of fault occurrence of the fault sample is calculated. The necessary signal strength is used to represent... The system identifies the signal strength that will inevitably occur when a fault occurs. Simultaneously, based on the amplitude and corresponding phase of selected frequencies from the set of unnecessary fault frequencies, the system calculates the unnecessary signal strength at the time of fault occurrence. Considering that real faults are affected by environmental factors, resulting in noise disturbances in the operating signal, the system calculates the disturbance signal strength at the time of fault occurrence based on the amplitude of all frequencies in the disturbance frequency set of the fault sample. A disturbance signal is added based on this disturbance signal strength, and combined with both necessary and unnecessary signal strengths, the system determines the simulated fault signal of the fault sample at the time of fault occurrence. This addresses the problem that fault signal simulation cannot adapt to the nonlinear changes and random disturbance characteristics of transformer fault signals, resulting in significant deviations between simulation results and actual fault conditions. Attached Figure Description
[0015] Figure 1 This is a flowchart illustrating the steps of a fault simulation method using a digital transformer fault simulator, as provided in one embodiment of this application. Detailed Implementation
[0016] Please see Figure 1 The diagram illustrates a flowchart of a digital transformer fault simulation generator fault simulation method according to an embodiment of this application. The method includes the following steps: Step S001: Collect different types of operating signals from a preset number of fault samples for each fault type, establish an operating signal sequence, and calculate the fault degree of the fault samples.
[0017] The system collects the transformer's operating signals within a preset time period starting from the moment the transformer fault occurs. These operating signals include current, voltage, and vibration signals.
[0018] In this embodiment, the duration of the preset time period is set to 5 seconds, and the sampling frequency of the running signal is set to 10kHz.
[0019] Specifically, merging units are installed at the output terminals of the high-voltage and low-voltage bushings of the transformer to collect the transformer's current and voltage. Piezoelectric accelerometers are installed on the tank wall opposite the transformer core to collect the transformer's vibration signals.
[0020] The specific fault types covered in this application include: inter-turn short circuit, winding open circuit, winding-to-ground short circuit, partial discharge, lead wire fault, cooling system fault, and core fault. For each type of fault, this embodiment collects the transformer's operating signals within a preset time period starting from the fault occurrence time of 1000 fault samples.
[0021] For each fault sample, the same type of operating signal is arranged in the order of acquisition to obtain the operating signal sequence of the fault sample.
[0022] The fault type corresponding to the fault sample is determined by a fault identification model, which employs a one-dimensional convolutional neural network (1D-CNN). The training process of a one-dimensional convolutional neural network is a well-known technique and will not be elaborated further. Specifically, frequency domain analysis is performed on the operating signal sequence of the fault sample to extract a fault feature vector composed of fault feature frequencies, amplitudes, and energy. This fault feature vector is then input into the fault identification model to obtain the fault type corresponding to the fault sample.
[0023] The root mean square deviation between the operating signal sequence of the fault sample and the preset standard operating signal corresponding to the same type of operating signal is denoted as the fault deviation of the same type of operating signal. The mean of the normalized values of the fault deviations of all types of operating signals corresponding to the fault sample is denoted as the fault degree of the fault sample.
[0024] The preset standard operating signals are extracted in advance by those skilled in the art and are a sequence of fault-free operating signals.
[0025] It should be noted that this embodiment uses the maximum-minimum normalization method. Based on the fault deviations of all operating signals of the same type, the normalized value of the fault deviation of the same type of operating signal is calculated. In the maximum-minimum normalization method, when the difference between the maximum and minimum values is 0, the sum of the difference between the maximum and minimum values and a preset non-zero small constant is used as the denominator to calculate the normalized value. In this embodiment, the preset non-zero small constant is set to a value of [value missing]. In practical applications, implementers may use other methods from existing technologies, such as the tanh function or the sigmoid function, to calculate the normalized value; no restrictions are imposed here.
[0026] The greater the degree of fault in a fault sample, the more significant the deviation of the transformer's operating state from the normal state, and the more severe the fault in the transformer corresponding to the fault sample.
[0027] At this point, a preset number of fault samples for each fault type are obtained, along with different types of operating signals, operating signal sequences, and the fault severity of the fault samples.
[0028] Step S002: For the operating signal sequence of each type of fault sample, determine the characteristic frequency set and the disturbance frequency set based on the amplitude and frequency corresponding to different frequency components in the frequency domain of the operating signal sequence. Determine the necessary fault frequency set and the unnecessary fault frequency set based on the characteristic frequency set corresponding to the same type of operating signal determined by all fault samples of the same fault type.
[0029] When a transformer malfunctions, all types of operating signals of the transformer will be affected, causing changes in the period and amplitude of the operating signals that normally exhibit stable periodic changes.
[0030] Fourier transform can decompose time-series signals and extract their frequency domain features. For each type of operating signal, a Fourier transform is performed on the operating signal sequence of the fault sample to obtain a complex frequency domain sequence. Based on the relationship between complex amplitude and phase, the amplitude and phase corresponding to each frequency component of the operating signal are calculated. The amplitude and phase are then sorted in ascending order according to the frequency components corresponding to the amplitude and phase, thus obtaining the amplitude spectrum sequence and phase spectrum sequence of the fault sample.
[0031] The transformer's on-site operating environment is subject to high voltage and strong magnetic field interference, which often causes noise to be mixed in the timing signal. After frequency domain decomposition, the noise will generate several additional interference frequency components, and the energy proportion corresponding to the interference frequency components is generally low.
[0032] Arrange all amplitudes in the amplitude spectrum sequence of the fault sample in descending order of absolute value to establish the amplitude spectrum sorting sequence of the fault sample. Calculate the cumulative energy percentage of the absolute values of the amplitudes within the amplitude spectrum sorting sequence. The formula for calculating the cumulative energy percentage is: in, Indicates the first value in the amplitude spectrum sorting sequence The cumulative energy percentage of the absolute value of each amplitude; Indicates the first value in the amplitude spectrum sorting sequence The absolute value of each amplitude; This indicates the number of absolute values of amplitudes within the amplitude spectrum sorting sequence.
[0033] The maximum value of all amplitudes whose cumulative energy percentage is greater than the preset main energy threshold is selected and denoted as the boundary amplitude. The set of frequencies corresponding to the absolute values of all amplitudes greater than the boundary amplitude in the amplitude spectrum sorting sequence of the fault sample is denoted as the characteristic frequency set, and the set of frequencies corresponding to the absolute values of all amplitudes less than or equal to the boundary amplitude in the amplitude spectrum sorting sequence of the fault sample is denoted as the disturbance frequency set.
[0034] The main energy threshold should be greater than or equal to 90% and less than or equal to 95%. When the transformer is in normal operation and single fault condition, the core characteristic frequencies such as the fundamental wave and low harmonics can gather more than 80% of the total signal energy. Therefore, in this embodiment, the main energy threshold is set to 90%. At this time, the main energy threshold can completely screen out all core characteristic frequencies. At the same time, the main energy threshold can effectively eliminate noise interference frequencies with extremely low energy proportion.
[0035] It is understandable that for each type of operating signal in a fault sample, there is a corresponding set of characteristic frequencies and a set of disturbance frequencies. That is to say: the current of a fault sample corresponds to a set of characteristic frequencies and a set of disturbance frequencies, the voltage of a fault sample corresponds to a set of characteristic frequencies and a set of disturbance frequencies, and the vibration signal of a fault sample corresponds to a set of characteristic frequencies and a set of disturbance frequencies.
[0036] The operating signals of faulty transformers have nonlinear variation characteristics and random disturbance characteristics. Therefore, the characteristic frequency sets of the same type of operating signals determined by all fault samples of the same fault type are not completely the same. Based on the characteristic frequency set and the disturbance frequency set, we analyze the same and different frequencies appearing in the characteristic frequency set corresponding to the same type of operating signals of the same fault type.
[0037] For all fault samples of the same type that determine the same type of operating signal, take the intersection of all characteristic frequency sets corresponding to the same type of operating signal determined by all fault samples of the same type to obtain the necessary fault frequency set; the set of frequencies that are not in the necessary fault frequency set but are included in all characteristic frequency sets corresponding to the same type of operating signal determined by all fault samples of the same type is denoted as the non-necessary fault frequency set.
[0038] Specifically: Since the power grid allows for a certain degree of deviation during operation, in this embodiment, when determining the necessary set of fault frequencies, frequencies whose absolute difference is less than a preset frequency threshold are considered as the same frequency. The value of the same frequency is the average of all frequencies whose absolute difference is less than the preset frequency threshold. In this embodiment, the preset frequency threshold is set to 0.2Hz.
[0039] For example, for 50.01Hz and 50.03Hz, the absolute value of the difference between 50.01Hz and 50.03Hz is 0.02Hz, which is less than 0.2Hz. Therefore, 50.01Hz and 50.03Hz are determined to be the same frequency, and the value of the same frequency is 50.02Hz, which is the value corresponding to the necessary fault frequency set.
[0040] It is important to note that faults of the same type share the same fault mechanism and common characteristic frequencies in the frequency domain signal. Therefore, the necessary fault frequency set must contain at least one frequency, and it cannot be empty. By using the necessary and unnecessary fault frequency sets, fault signals can be effectively digitally mapped.
[0041] At this point, the necessary fault frequency set and the unnecessary fault frequency set are determined.
[0042] Step S003: For frequencies in the necessary fault frequency set, perform linear fitting based on the fault degree of the fault sample corresponding to the frequency and the amplitude corresponding to the frequency. Based on the result of the linear fitting, calculate the necessary signal strength at the time of fault occurrence of the fault sample. Based on the amplitude and corresponding phase of some frequencies selected from the non-necessary fault frequency set, calculate the non-necessary signal strength at the time of fault occurrence.
[0043] For faults of the same type but different severities, the amplitude fluctuation ranges of each frequency component differ significantly. In the process of characteristic analysis of transformer fault signals, it is necessary to focus on analyzing the inherent characteristic frequencies under fault conditions, including the fundamental frequency, second harmonic, and third harmonic.
[0044] Extract the amplitude corresponding to each frequency in the necessary fault frequency set. For any frequency in the necessary fault frequency set, take the fault degree of the fault sample corresponding to the frequency as the independent variable and the amplitude corresponding to the frequency as the dependent variable, perform linear fitting, and obtain the goodness of fit of the fitted line corresponding to the frequency.
[0045] When the goodness of fit of the fitted line is greater than or equal to the preset goodness of fit threshold, the fitted value of the fault degree corresponding to the frequency in the necessary fault frequency set is calculated based on the fitted line. The difference between the fitted value of the amplitude and the corresponding amplitude is recorded as the fitted fluctuation. The mean of all fitted fluctuations greater than 0 is recorded as the upper limit of fluctuation, and the mean of all fitted fluctuations less than 0 is recorded as the lower limit of fluctuation. Random numbers greater than or equal to the lower limit of fluctuation and less than or equal to the upper limit of fluctuation are generated by a random function. The generated random numbers are recorded as the frequency fluctuation amplitude of the corresponding frequency. The sum of the frequency fluctuation amplitude of the frequency and the fitted value of the amplitude corresponding to the frequency is recorded as the necessary fault coefficient of the frequency.
[0046] In this embodiment, the least squares algorithm is used to achieve linear fitting; the preset goodness-of-fit threshold value in this embodiment is 0.8.
[0047] When the goodness of fit of the fitted line is less than the preset goodness of fit threshold, the range of values between the minimum and maximum values of the amplitude of the fault sample corresponding to the frequency is denoted as the amplitude range. The amplitude range is divided into sub-ranges with a step size of 10% of the length of the amplitude range. The mean and standard deviation of the fault degree in each sub-range are calculated. Random numbers are randomly generated in the sub-range where the amplitude corresponding to the frequency in the amplitude set is located using a normal distribution. The generated random numbers are denoted as the necessary fault coefficients of the frequency in the necessary fault frequency set. The mean and standard deviation of the normal distribution are the mean and standard deviation of the fault degree in the sub-range where the fault degree corresponding to the frequency is located.
[0048] Furthermore, based on the necessary fault coefficients of the frequencies in the necessary fault frequency set, the necessary signal strength at the corresponding fault occurrence time is calculated.
[0049] in, Indicates the first The necessary signal strength at the moment a fault occurs; Represents the set of necessary fault frequencies. The necessary failure factor for each frequency; Indicates the number of frequencies in the necessary fault frequency set; Represents the set of necessary fault frequencies. One frequency; Represents the set of necessary fault frequencies. The phase corresponding to each frequency, specifically, is generated by a random number generator, with a value greater than or equal to 1 and less than or equal to... The integer between the two ranges will be denoted as . , Indicates the number of fault samples, for the th The operating signal sequence of the fault sample is decomposed to obtain the sequence decomposition result that matches the set of necessary fault frequencies. The phases of the same frequency are denoted as the i-th frequency in the set of necessary fault frequencies. The phase corresponding to each frequency; Represents pi; This represents the sine function.
[0050] The necessary signal strength at the moment of fault occurrence is used to characterize the signal strength that must occur when the fault occurs.
[0051] When a transformer experiences a similar fault, the frequency domain characteristics of the fault signal are not completely fixed due to factors such as operating conditions and random interference. The fault signal may exhibit both regular frequency changes and subtle frequency domain fluctuations.
[0052] The frequencies in the set of unnecessary fault frequencies correspond to frequency domain components that do not exhibit fixed changing patterns and are not necessarily subject to characteristic shifts under fault conditions. Therefore, based on the nonlinear and random disturbance characteristics of transformer fault signals, a portion of frequencies can be randomly selected from the set of unnecessary fault frequencies to participate in signal synthesis during fault signal simulation. This reproduces the random frequency domain disturbance characteristics of real fault scenarios, effectively improving the fit and realism of the synthesized fault signal. Simultaneously, the number of frequencies participating in the synthesis from the set of unnecessary fault frequencies must not exceed the total number of frequencies in the set to avoid introducing invalid redundant frequency components, which could distort the synthesized signal and reduce the accuracy of the fault simulation.
[0053] Randomly generate an integer greater than or equal to 1 and less than or equal to the total number of frequencies in the set of unnecessary fault frequencies. Denote the randomly generated integer as . Randomly select from the set of unnecessary failure frequencies For each frequency in the selected set of unnecessary fault frequencies, the amplitude and corresponding phase of each frequency are obtained. For any frequency in the selected set of unnecessary fault frequencies, the maximum and minimum values of all amplitudes of the frequency are used as the upper and lower limits of the interval, respectively. Random numbers are generated within the interval range and recorded as the unnecessary fault coefficient of the frequency. The maximum and minimum values of all phases corresponding to the frequency are used as the upper and lower limits of the interval, respectively. Random numbers are generated within the interval range and recorded as the unnecessary phase coefficient of the frequency.
[0054] It is understandable that the corresponding frequency can be obtained for each of the 1000 fault samples. Therefore, each frequency in the set of unnecessary fault frequencies corresponds to 1000 amplitudes, with the extreme value of the amplitude being 0.
[0055] Unnecessary phase coefficients are determined by randomly selecting a subset of frequencies from the set of unnecessary fault frequencies and randomly generating their values. This simulates random changes in the phase of unnecessary frequencies, avoiding a fixed phase in the synthesized signal and improving the realism of the fault signal's time-domain waveform. Unnecessary fault coefficients and unnecessary phase coefficients are used to simulate changes in the amplitude and phase of the fault signal.
[0056] Furthermore, based on the unnecessary fault coefficients and unnecessary phase coefficients of the frequencies in the unnecessary fault frequency set of the fault samples, the unnecessary signal strength of the fault samples at the time of fault occurrence is calculated.
[0057] in, Indicates the fault sample in the first... Unnecessary signal strength at the moment of fault occurrence; An integer representing the total number of frequencies in a randomly generated set of unnecessary fault frequencies that is greater than or equal to 1 and less than or equal to 1. The first fault sample is randomly selected from the set of unnecessary fault frequencies. Unnecessary failure coefficients at each frequency; The first fault sample is randomly selected from the set of unnecessary fault frequencies. Unnecessary phase coefficients at each frequency; The first fault sample is randomly selected from the set of unnecessary fault frequencies. Each frequency.
[0058] The unnecessary signal strength at the moment of fault occurrence is used to characterize the random frequency changes that occur during the fault process, thereby improving the realism of the main part of the fault in digital mapping.
[0059] Thus, the necessary and unnecessary signal strengths at the moment the fault occurred are obtained.
[0060] Step S004: Based on the amplitude of all frequencies in the disturbance frequency set of the fault sample, randomly generate the disturbance fault coefficient and disturbance phase coefficient of the frequencies in the disturbance frequency set, calculate the disturbance signal strength at the time of fault occurrence of the fault sample, and combine the necessary signal strength and the unnecessary signal strength to determine the fault simulation signal of the fault sample at the time of fault occurrence.
[0061] Real faults are affected by environmental factors, and operating signals will be subject to noise disturbances. Therefore, disturbance signals need to be added during the process of synthesizing fault signals.
[0062] Extract the maximum amplitude of all frequencies from the disturbance frequency set of the fault samples. Record the maximum of the maximum amplitude values corresponding to all fault samples as the upper limit of amplitude. Randomly generate random numbers greater than or equal to 0 and less than or equal to the upper limit of amplitude. The number of generated random numbers is equal to the number of frequencies in the disturbance frequency set. Record these generated random numbers sequentially as the disturbance fault coefficients for each frequency in the disturbance frequency set of the fault samples. The number of random numbers generated is equal to the number of frequencies in the disturbance frequency set. The generated random numbers are then recorded as the disturbance phase coefficients of each frequency in the disturbance frequency set of the fault sample.
[0063] Based on the disturbance fault coefficient and disturbance phase coefficient of the frequency in the disturbance frequency set of the fault sample, calculate the disturbance signal strength at the time of fault occurrence of the fault sample.
[0064] in, Indicates the first The strength of the disturbance signal at the moment the fault occurs; This indicates the number of frequencies in the set of disturbance frequencies; Represents the set of disturbance frequencies. Disturbance fault coefficient at each frequency; Represents the set of disturbance frequencies. One frequency; Represents the set of disturbance frequencies. The perturbation phase coefficient at each frequency.
[0065] The intensity of the disturbance signal at the moment of fault occurrence is used to characterize the random noise variation in the fault signal.
[0066] The sum of the necessary signal strength, unnecessary signal strength, and disturbance signal strength of the fault sample at the time of fault occurrence is used as the fault simulation signal of the fault sample at the time of fault occurrence.
[0067] It is important to understand that for each type of operating signal sequence of a fault sample, the fault simulation signal at the time of the fault occurrence can be obtained.
[0068] This completes the simulation of transformer faults.
[0069] Secondly, embodiments of this application also provide a digital transformer fault simulation generator, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the above-described digital transformer fault simulation generator fault simulation method.
Claims
1. A fault simulation method for a digital transformer fault simulator, characterized in that, The method includes the following steps: Collect different types of operating signals from a preset number of fault samples for each fault type, establish an operating signal sequence, and calculate the fault degree of the fault samples; For each type of operating signal sequence of the fault sample, the characteristic frequency set and the disturbance frequency set are determined according to the amplitude and frequency corresponding to the different frequency components in the frequency domain of the operating signal sequence. The necessary fault frequency set and the unnecessary fault frequency set are determined according to the characteristic frequency set corresponding to the same type of operating signal determined by all fault samples of the same fault type. For frequencies in the necessary fault frequency set, a linear fit is performed based on the fault severity of the fault sample corresponding to the frequency and the amplitude corresponding to the frequency. Based on the result of the linear fit, the necessary signal strength at the time of fault occurrence of the fault sample is calculated. Based on the amplitude and corresponding phase of a subset of frequencies selected from the non-necessary fault frequency set, the non-necessary signal strength at the time of fault occurrence is calculated. Based on the amplitude of all frequencies in the disturbance frequency set of the fault sample, the disturbance fault coefficient and disturbance phase coefficient of the frequencies in the disturbance frequency set are randomly generated. The disturbance signal strength at the time of fault occurrence of the fault sample is calculated. Combining the necessary signal strength and the unnecessary signal strength, the fault simulation signal of the fault sample at the time of fault occurrence is determined.
2. The fault simulation method of a digital transformer fault simulator according to claim 1, characterized in that, The method for calculating the fault severity of the fault sample is as follows: The root mean square deviation between the operating signal sequence of the fault sample and the preset standard operating signal corresponding to the same type of operating signal is denoted as the fault deviation of the same type of operating signal. The mean of the normalized values of the fault deviations of all types of operating signals corresponding to the fault sample is denoted as the fault degree of the fault sample.
3. The fault simulation method of a digital transformer fault simulator according to claim 1, characterized in that, The method for determining the characteristic frequency set and the disturbance frequency set is as follows: The operating signal sequence of the fault sample is converted to the frequency domain to obtain the amplitude spectrum sequence of the fault sample. The amplitude spectrum sorting sequence of the fault sample is established. The cumulative energy ratio of the absolute value of the amplitude within the amplitude spectrum sorting sequence is calculated. The maximum value of all amplitudes with a cumulative energy ratio greater than the preset main energy threshold is selected and recorded as the boundary amplitude. The set of frequencies corresponding to the absolute values of all amplitudes greater than the threshold amplitude in the amplitude spectrum sorting sequence of the fault sample is denoted as the characteristic frequency set, and the set of frequencies corresponding to the absolute values of all amplitudes less than or equal to the threshold amplitude in the amplitude spectrum sorting sequence of the fault sample is denoted as the disturbance frequency set.
4. The fault simulation method of a digital transformer fault simulator according to claim 1, characterized in that, The method for determining the necessary fault frequency set and the unnecessary fault frequency set is as follows: The necessary fault frequency set is obtained by taking the intersection of all characteristic frequency sets corresponding to the same type of operating signal determined by all fault samples of the same fault type. The set of frequencies that are not in the necessary fault frequency set is defined as the set of frequencies contained in the characteristic frequency set corresponding to the same type of operating signal determined by all fault samples of the same fault type.
5. The fault simulation method of a digital transformer fault simulator according to claim 1, characterized in that, The linear fitting specifically refers to: For any frequency in the necessary fault frequency set, the fault degree of the fault sample corresponding to the frequency is used as the independent variable, and the amplitude corresponding to the frequency is used as the dependent variable. Linear fitting is performed to obtain the goodness of fit of the fitted line corresponding to the frequency.
6. The fault simulation method for a digital transformer fault simulator according to claim 5, characterized in that, The method for calculating the necessary signal strength at the time of fault occurrence for the fault sample is as follows: When the goodness of fit of the fitted line is greater than or equal to the preset goodness of fit threshold, the difference between the fitted value of the amplitude and the corresponding amplitude is recorded as the fitting fluctuation. The mean of all fitting fluctuations greater than 0 is recorded as the upper limit of fluctuation, and the mean of all fitting fluctuations less than 0 is recorded as the lower limit of fluctuation. A random number greater than or equal to the lower limit of fluctuation and less than or equal to the upper limit of fluctuation is generated. The generated random number is recorded as the frequency fluctuation amplitude of the corresponding frequency. The sum of the frequency fluctuation amplitude of the frequency and the fitted value of the amplitude corresponding to the frequency is recorded as the necessary fault coefficient of the frequency. When the goodness of fit of the fitted line is less than the preset goodness of fit threshold, the range of values between the minimum and maximum values of the amplitude of the fault sample corresponding to the frequency is denoted as the amplitude range. The amplitude range is divided into sub-ranges. Random numbers are generated randomly in the sub-ranges where the amplitude of the frequency in the necessary fault frequency set is located using a normal distribution. The generated random numbers are denoted as the necessary fault coefficients of the frequency in the necessary fault frequency set. Calculate the necessary signal strength at the time of the corresponding fault occurrence based on the necessary fault coefficients of the frequencies in the necessary fault frequency set.
7. The fault simulation method of a digital transformer fault simulator according to claim 1, characterized in that, The method for calculating the unnecessary signal strength at the time of the fault occurrence is as follows: Randomly generate an integer greater than or equal to 1 and less than or equal to the total number of frequencies in the set of unnecessary fault frequencies. Randomly select from the set of unnecessary failure frequencies For each frequency in the selected set of unnecessary fault frequencies, the amplitude and corresponding phase of each frequency are obtained. For any frequency in the selected set of unnecessary fault frequencies, the maximum and minimum values of all amplitudes of the frequency are used as the upper and lower limits of the interval, respectively. Random numbers are generated within the interval range and recorded as the unnecessary fault coefficient of the frequency. The maximum and minimum values of all phases corresponding to the frequency are used as the upper and lower limits of the interval, respectively. Random numbers are generated within the interval range and recorded as the unnecessary phase coefficient of the frequency. Based on the unnecessary fault coefficients and unnecessary phase coefficients of the frequencies in the unnecessary fault frequency set of the fault sample, calculate the unnecessary signal strength of the fault sample at the time of fault occurrence.
8. The fault simulation method of a digital transformer fault simulator according to claim 1, characterized in that, The method for generating the frequency disturbance fault coefficient and disturbance phase coefficient is as follows: Extract the maximum amplitude of all frequencies in the disturbance frequency set of the fault sample, and record the maximum of the maximum amplitude values corresponding to all fault samples as the upper limit of amplitude. Randomly generate random numbers that are greater than or equal to 0 and less than or equal to the upper limit of amplitude. The number of random numbers generated is equal to the number of frequencies in the disturbance frequency set. The generated random numbers are recorded as the disturbance fault coefficients of each frequency in the disturbance frequency set of the fault sample. Randomly generate greater than or equal to 0 and less than or equal to The number of random numbers generated is equal to the number of frequencies in the disturbance frequency set. The generated random numbers are then recorded as the disturbance phase coefficients of each frequency in the disturbance frequency set of the fault sample.
9. The fault simulation method of a digital transformer fault simulator according to claim 1, characterized in that, The method for determining the fault simulation signal of the fault sample at the time of fault occurrence is as follows: The sum of the necessary signal strength, unnecessary signal strength, and disturbance signal strength of the fault sample at the time of fault occurrence is used as the fault simulation signal of the fault sample at the time of fault occurrence.
10. A digital transformer fault simulation generator, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the fault simulation method of a digital transformer fault simulator as described in any one of claims 1 to 9.