Dynamic risk assessment method for transformer dc bias caused by metro stray current

Abstract

The present application provides a dynamic risk assessment method of transformer DC bias caused by subway stray current, relates to the technical field of power system, and acquires train dynamic operation data of a current period, simulates the train dynamic operation data based on a constructed subway and regional power grid coupling model, so as to obtain DC bias current of each substation area at each time, and determine the risk level of absolute bias current at each time; for each substation area, the risk type thereof is determined based on the risk level distribution of the absolute bias current in the current period, a corresponding processing strategy is set, so as to obtain the comprehensive absolute bias current of each risk type, the comprehensive absolute bias current of each non-substation area is determined based on a partition weighted interpolation algorithm, the risk condition of the simulation period is determined based on the comprehensive absolute bias current, and a risk assessment diagram is exported; through the specific processing of different risk types, the accurate assessment of the period risk is realized.

Description

Technical Field

[0001] This invention relates to the field of power system technology, specifically to a dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways. Background Technology

[0002] With the rapid development of urban rail transit, stray currents generated by subway operation pose a potential threat to power system safety, potentially leading to DC bias in transformers and subsequently causing problems such as increased transformer vibration, localized overheating, and insulation aging. While existing technologies exist for monitoring stray currents, most are static assessments, failing to reflect the real-time changes in stray current distribution during the dynamic operation of subway trains, and lacking methods for differentiated risk assessment between substation and non-substation areas. Furthermore, existing methods often employ a uniform calculation logic when dealing with different risk level distribution characteristics, failing to adopt appropriate comprehensive current determination methods for different situations such as distributions of the same level, adjacent levels, and irregular levels, resulting in insufficient accuracy and specificity in risk assessment. In addition, for risk scenarios with irregular and drastic fluctuations, traditional time-domain analysis methods struggle to effectively extract risk characteristics and cannot accurately reconstruct the complex electromagnetic disturbances experienced by transformers.

[0003] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A dynamic risk assessment method for transformer DC bias magnetization caused by stray currents in subways, including the following steps: Step 1: Obtain the train dynamic operation data for the current time period and input it into the subway and regional power grid coupling model to simulate the absolute bias current of each substation area at each moment in the current time period, and determine the risk level of each absolute bias current. Step 2: For each substation area, based on the risk level distribution of its absolute bias current in the current time period, determine its risk type in the current time period. The risk types include same level distribution, adjacent level distribution and irregular level distribution. Step 3: For substation areas belonging to the same level distribution, the average absolute bias current at each time moment is taken as its comprehensive absolute bias current. For substation areas belonging to adjacent level distribution, the transition of its absolute bias current and the average absolute bias current at the current time period are analyzed to determine its comprehensive absolute bias current. For substation areas belonging to irregular level distribution, the spectrum phase reconstruction analysis of its absolute bias current at the current time period is performed to determine its comprehensive absolute bias current. Step 4: For non-substation areas, set weighting factors based on their distance from substation areas, determine the comprehensive absolute bias current of each non-substation area based on the partition weighted interpolation algorithm, determine the risk situation of the current period based on the comprehensive absolute bias current, and derive the risk assessment map.

[0006] Furthermore, the train dynamic operation data for the current time period includes the train's position and traction current at various times during the current time period; The absolute bias current is the absolute value of the DC bias current.

[0007] Furthermore, the logic for determining the risk level of each absolute bias current is as follows: by using a preset threshold, the risk level of each absolute bias current is determined, which includes four risk levels: Level 1, Level 2, Level 3 and Level 4.

[0008] Furthermore, the logic for determining the risk type of a substation area in the current time period is as follows: for any substation area, if the absolute bias current at all times in the current time period belongs to the same risk level, then the risk type of this substation area is defined as the same level distribution; otherwise, the following logical judgment is made. The number of times the absolute bias current falls into each risk level in the substation area during the current time period is counted. A preset threshold is set, and the risk level with a number of times exceeding the threshold is called the primary risk level. If there is only one primary risk level in the substation area, and the risk level to which the absolute bias current belongs at other times is adjacent to the primary risk level, then the risk type of this substation area is defined as adjacent level distribution. If there are two or more primary risk levels in the substation area, and the primary risk levels are continuously distributed, then the risk type of this substation area is also defined as adjacent level distribution. For any substation area, if its risk type is neither distributed at the same level nor at adjacent levels, then its risk type is defined as an irregular level distribution.

[0009] Furthermore, the logic for determining the transition status in the current time period is as follows: For any substation area with adjacent risk levels, the logic for determining its comprehensive absolute bias current is as follows: For any time starting from the second time point in the current time period, if the risk level of the absolute bias current changes relative to the previous time point, it is recorded as a transition; otherwise, it is recorded as a translation. If the risk level of the absolute bias current at this time point is greater than the risk level at the previous time point, the transition is called a positive transition; otherwise, the transition is called a negative transition. Furthermore, the logic for analyzing the transition of the absolute bias current in the current time period is as follows: count the number of transitions and translations occurring in the substation area, calculate the sum of the number of transitions and translations, which is called the total number of statistics, divide the number of transitions by the total number of statistics, which is called the transition density, preset a transition amplification factor for each risk level, calculate the average absolute bias current of the substation area at each moment in the current time period, and determine the risk level to which the average absolute bias current belongs as the target risk level, extract the transition amplification factor corresponding to the target risk level as the target transition amplification factor, and calculate the product of the transition density and the target transition amplification factor to obtain the transition adjustment ratio.

[0010] For substation areas belonging to adjacent grade distributions, the logic for determining their comprehensive absolute magnetic bias current is as follows: if the number of positive transitions is greater than the number of negative transitions, add 1 to the transition adjustment ratio, and then multiply by the average absolute magnetic bias current to obtain the comprehensive absolute magnetic bias current for the substation area; if the number of positive transitions is equal to the number of negative transitions, then use the average absolute magnetic bias current as the comprehensive absolute magnetic bias current for the substation area; if the number of positive transitions is less than the number of negative transitions, subtract the transition adjustment ratio from 1, and then multiply by the average absolute magnetic bias current to obtain the comprehensive absolute magnetic bias current for the substation area.

[0011] Furthermore, for any substation area with an irregular risk level distribution, the logic for determining its comprehensive absolute bias current is as follows: For the absolute bias current of the substation area at each moment within the current time period, the risk level corresponding to the assignment is embedded into the absolute bias current in the form of a rotated phase to obtain the risk modulation current. The risk modulation current at each moment is used to form a risk modulation current sequence. The risk modulation current sequence is subjected to a Fourier transform to obtain the amplitude spectrum and phase spectrum of the risk modulation current sequence. The frequency is clustered based on the phase in the phase spectrum to obtain several cluster sets. Each cluster set includes several frequencies under the same cluster. An amplitude adjustment rule is set, and the amplitude corresponding to the frequency of each cluster set is adjusted according to the amplitude adjustment rule to obtain a new amplitude spectrum. The phase spectrum and the new amplitude spectrum are combined to form a new spectrum. The new spectrum is subjected to an inverse Fourier transform to obtain a new risk modulation current sequence. The rotated phase of the rotated risk modulation current sequence is removed to obtain the second absolute bias current sequence. The average value of the absolute bias current in the second absolute bias current sequence is calculated as the comprehensive absolute bias current of the substation area.

[0012] Furthermore, the logic for setting the amplitude adjustment rules is as follows: calculate the variance of the amplitude corresponding to each frequency in each cluster set, which is called amplitude dispersion. Set an amplitude dispersion threshold. If the amplitude dispersion is greater than the amplitude dispersion threshold, retain the amplitude value of each frequency in the cluster set. If the amplitude dispersion is not greater than the amplitude dispersion threshold, calculate the average value of the amplitude values ​​corresponding to all frequencies in the cluster set, which is called the comprehensive amplitude. Replace the amplitude value of each frequency in the cluster set with the comprehensive amplitude.

[0013] Compared with the prior art, the beneficial effects of the present invention are: This invention classifies risk levels in different regions into three types based on their distribution characteristics: same-level distribution, adjacent-level distribution, and irregular-level distribution. Differentiated comprehensive absolute bias current calculation strategies are adopted for each type. For the same-level distribution, the mean value can be directly used to concisely and effectively reflect the stable risk level. For adjacent-level distribution, by introducing transition density and transition direction to adjust the mean value, the cumulative effect of risk fluctuation trends on transformers can be captured, avoiding the simple mean value masking the actual harm caused by risk increases or decreases. For irregular-level distribution, a spectral phase reconstruction analysis method is used to embed the risk level into the current sequence in the form of a rotated phase to construct a risk-modulated current. Through frequency domain transformation, phase clustering, and amplitude adjustment, the frequency structure driven by risk is effectively identified, and the unstable spectrum is structurally reconstructed, weakening the spectral diffusion caused by random shocks. This allows the final comprehensive absolute bias current to more realistically restore the actual bias level under complex risk modes. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 This is a risk diagram of DC bias during the early departure period of this invention; Figure 3 This is a peak-period DC bias risk diagram for this invention. Figure 4 This is a DC bias risk diagram for off-peak periods according to the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0016] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0017] Example: Please see Figures 1-4 The present invention provides a technical solution: A dynamic risk assessment method for transformer DC bias magnetization caused by stray currents in subways, including the following steps: Step 1: Obtain the train dynamic operation data for the current time period and input it into the subway and regional power grid coupling model to simulate the absolute bias current of each substation area at each moment in the current time period, and determine the risk level of each absolute bias current. Furthermore, the train dynamic operation data for the current time period includes the train's position and traction current at various times during the current time period; The absolute bias current is the absolute value of the DC bias current.

[0018] The location at each time point in the current period can be obtained through the positioning system installed in the subway. If no positioning system is installed, it can also be calculated based on the actual subway departure timetable, the subway departure interval, stop time, station information of the line, and the subway's scheduled speed, combined with kinematic formulas. The traction current can be calculated directly using existing formulas, as follows: in, For the subway during the current time period The traction force (or braking force) at any given moment can be directly derived from the train's braking system. For the subway during the current time period Instantaneous velocity at a given moment This refers to the DC voltage of the overhead contact line bus. For the conversion efficiency of the subway traction system; In this embodiment, the current time period is half an hour, and the sampling frequency is once per minute. Specifically, it refers to the time corresponding to the sampling point.

[0019] Using the dynamic operation data of trains as input to the coupling model of subway and regional power grid, equivalent locomotive excitation conductors are set at the corresponding nodes of the contact network and rails in the coupling model. After batch calculation and simulation, the absolute bias current of the transformer can be output according to the changes in the operating conditions of the train.

[0020] The existing technology of subway-regional power grid coupling model is based on the equivalent circuitization of the "rail-ground-transformer" coupling network of the railway system. First, each train operating in the current time period is equivalent to a dynamic "excitation conductor". That is, the dynamic operating data such as the locomotive position, current draw power or regenerative braking power are mapped in real time to the controlled excitation source injected between the contact network and the rail electrical nodes. Then, these dynamic excitation sources are embedded into the pre-constructed subway-regional power grid coupling model, which includes the traction network, rail, ground current dissipation resistance and the neutral point grounding branch of the substation. By solving the node voltage and branch current of this large-scale subway-regional power grid coupling model at different times, the DC bias current of each substation transformer that changes dynamically due to train operation can be simulated and calculated.

[0021] It should be noted that the simulation calculation of the above subway and regional power grid coupling model outputs a DC bias current, which can be positive or negative. The positive or negative sign only indicates the direction. For risk assessment, the direction does not need to be considered. Therefore, the absolute value of the DC bias current (absolute bias current) is taken for subsequent risk analysis.

[0022] The logic for determining the risk level of each absolute bias current is as follows: by using a preset threshold, the risk level of each absolute bias current is determined, which includes four risk levels: Level 1, Level 2, Level 3 and Level 4. The threshold values ​​for division include 3A, 10A, and 15A. For specific division details, please refer to Table 1. In Table 1, I represents the absolute bias current.

[0023] Please refer to Table 1 for details; Table 1 Risk Level Analysis Table Step 2: For each substation area, based on the risk level distribution of its absolute bias current in the current time period, determine its risk type in the current time period. The risk types include same level distribution, adjacent level distribution and irregular level distribution. Furthermore, the logic for determining the risk type of a substation area within the current time period is as follows: For any substation area, if the absolute bias current at all times within the current time period belongs to the same risk level, then the risk type of this substation area is defined as a same-level distribution. This classification is for identifying steady-state operating conditions. If the absolute bias current never crosses the level threshold within the time period, it indicates that the impact of stray current from the subway on the substation area is relatively stable. It simplifies decision-making. Once it is determined to be a same-level distribution, maintenance personnel do not need to pay attention to its fluctuation details and can directly handle it according to the strategy corresponding to that level. For example, if it is always level 1, then continue monitoring; if it is always level 4, then the suppression device should be activated immediately to avoid misoperation caused by minor fluctuations.

[0024] This classification provides a highly reliable steady-state benchmark for subsequent comprehensive bias current calculations. Due to the highly concentrated data distribution, the results are highly representative and stable when calculating the comprehensive bias current that reflects the overall level, without the need to consider dynamic correction factors, thus simplifying the process from risk identification to governance decision-making.

[0025] Conversely, the following logical judgment is made; The number of times the absolute bias current falls into each risk level in the substation area during the current time period is counted. A preset threshold is set, and the risk level with a number of times exceeding the threshold is called the primary risk level. If there is only one primary risk level in the substation area, and the risk level to which the absolute bias current belongs at other times is adjacent to the primary risk level, then the risk type of this substation area is defined as adjacent level distribution. If there are two or more primary risk levels in the substation area, and the primary risk levels are continuously distributed, then the risk type of this substation area is also defined as adjacent level distribution. The quantity threshold is typically set between 40% and 50% of the number of sampling moments within a time period.

[0026] This classification aims to identify the finite fluctuations in the system state and the continuity of the value space. It does not directly describe whether the current changes smoothly on the time axis, but emphasizes that the value range of the data is continuous and without discontinuities. On the contrary, if there are discontinuities in the data value space (such as the simultaneous existence of a large number of level 1 and level 4 data, but the lack of level 2 and level 3), it usually means that the system has experienced extreme and discontinuous state transitions.

[0027] For any substation area, if its risk type is neither distributed at the same level nor at adjacent levels, its risk type is defined as an irregular level distribution. Due to the discontinuity in the data distribution, the conventional statistical average or linear weighted value can no longer accurately reflect the severe impact that the transformer actually experienced during that period. Therefore, once this type is identified, the subsequent comprehensive bias current calculation model will be switched from a steady-state statistical model to an event impact model.

[0028] Step 3: For substation areas belonging to the same level distribution, the average absolute bias current at each time is taken as its comprehensive absolute bias current. For substation areas belonging to adjacent level distribution, the transition of absolute bias current and the average absolute bias current at the current time period are analyzed to determine its comprehensive absolute bias current. For substation areas belonging to irregular level distribution, the spectrum phase reconstruction analysis of absolute bias current at the current time period is performed to determine its comprehensive absolute bias current. Furthermore, for any substation area with adjacent risk levels, the logic for determining its comprehensive absolute bias current is as follows: for any moment starting from the second moment in the current time period, if the risk level of the absolute bias current changes relative to the previous moment, it is recorded as a transition; otherwise, it is recorded as a translation. If the risk level of the absolute bias current at this moment is greater than the risk level at the previous moment, the transition is called a positive transition; otherwise, the transition is called a negative transition. The number of jumps and translations that occur in the substation area is counted. The sum of the number of jumps and translations is called the total number of counts. The number of jumps is divided by the total number of counts, which is called the jump density. It counts the frequency of the current jumping back and forth between risk levels. For each risk level, a transition amplification factor is preset. The average absolute bias current of the substation area at each time point in the current period is calculated, and the risk level to which the average absolute bias current belongs is determined as the target risk level. The transition amplification factor corresponding to the target risk level is extracted as the target transition amplification factor.

[0029] The target transition amplification factor assigns different weights to fluctuations based on the risk level of the mean, reflecting the physical fact that the higher the current level, the greater the harm caused by the same fluctuation. This solves the problem of difficulty in uniformly measuring the impact of fluctuations under different baseline risks. The transition method coefficient is set based on experience, or an average absolute bias current control group and multiple random transition absolute bias current analysis groups can be set manually. The power of each group is counted within a time period, and the proportion of power is used to approximate the proportion of absolute bias current to estimate the mean of the absolute bias current of the random transition absolute bias current analysis group. Linear fitting is then performed to obtain the transition amplification factor at the corresponding level. This is existing technology and will not be elaborated here.

[0030] In this embodiment, the transition amplification factor for level 1 is 2%, for level 2 it is 7%, for level 3 it is 7%, and for level 4 it is 2%. Since levels 1 and 4 have only one adjacent risk level, the impact of the transition is relatively small, so the transition method coefficients are set to be relatively small. Since levels 2 and 3 have two risk levels in the vector and there may be situations where the risk level spans two levels, the transition method coefficients are set to be relatively large.

[0031] The product of the transition density and the target transition amplification factor is calculated to obtain the transition adjustment ratio. The transition adjustment ratio serves as the generator for the final correction amplitude. It is the product of the two factors and, combined with the trend information represented by the number of positive and negative transitions, dynamically determines whether to adjust the mean upward, downward, or keep it unchanged. If the number of positive transitions is greater than the number of negative transitions, 1 is added to the transition adjustment ratio and then multiplied by the mean absolute bias current to obtain the comprehensive absolute bias current for the substation area. If the number of positive transitions is equal to the number of negative transitions, the mean absolute bias current is used as the comprehensive absolute bias current for the substation area. If the number of positive transitions is less than the number of negative transitions, 1 is subtracted from the transition adjustment ratio and then multiplied by the mean absolute bias current to obtain the comprehensive absolute bias current for the substation area.

[0032] This step, by introducing the core concept of transitions, solves the technical problem of static averages masking dynamic hazards in traditional assessment methods. Specifically, when a transformer is subjected to DC bias, not only does the absolute value of the current (i.e., the average) affect the magnetic saturation depth, but the repeated changes in current (transitions) also bring additional mechanical stress, vibration shocks, and cumulative thermal effects. These dynamic characteristics may exacerbate equipment losses, and a simple average will smooth out these fluctuations and cannot truly reflect the actual impact during that period. By statistically analyzing the transition density, the frequency of current fluctuations is accurately quantified. By pre-setting a transition amplification factor tied to the risk level, the nonlinear hazard differences of the same transition under different current baselines are reflected. Combined with the overall trend revealed by the number of positive and negative transitions, a transition adjustment ratio that can be dynamically adjusted upwards or downwards is finally generated. The resulting comprehensive absolute bias current allows the dynamic impact of previously ignored "transitions" to be reinjected into the risk assessment.

[0033] The characteristic of irregular risk level distributions lies in the disordered and discontinuous jumps in the temporal variation of their absolute bias current. There are often gaps between risk levels (e.g., a jump directly from level 1 to level 4), and no obvious statistical trend or dominant risk level. This drastic and irregular fluctuation leads to random electromagnetic shocks and mechanical stresses on the transformer, making it impossible to accurately assess its comprehensive impact using simple averaging or linear corrections based on transition density. Therefore, previous static averaging methods for distributions within the same risk level and transition adjustment methods for distributions at adjacent risk levels are both ineffective—the former smooths out all dynamic information, and the latter relies on the assumption of continuous fluctuations, and the energy of transitions between adjacent risk levels is usually small, conditions that do not hold in irregular distributions. However, the risk-driven frequency structure can be identified using the spectral phase structure, and the unstable spectral structure can be structurally reconstructed to obtain a more stable and realistic characterization of the absolute bias current.

[0034] For any substation area with an irregular risk level distribution, the logic for determining its comprehensive absolute bias current is as follows: For the absolute bias current of the substation area at each moment within the current time period, the risk level corresponding to the assignment is embedded into the absolute bias current in the form of a rotating phase to obtain the risk-modulated current. The assignment corresponding to the risk level is its level value, for example, level 1 is assigned a value of 1. Embedding the rotating phase for any substation area at any moment within the time period is an existing technology and can be referred to the following formula: in, For risk modulation current, It is an absolute bias current. Assign values ​​to the corresponding risk levels.

[0035] During periods of irregular risk level distribution, the change in absolute bias current and the risk level often do not have a stable linear relationship. High-risk moments may manifest as transient impacts or local anomalies, which are easily masked by ordinary fluctuations when analyzing only the current amplitude. Therefore, by mapping the risk level to a rotating phase and embedding it into the absolute bias current, the risk information is incorporated into the complex representation structure of the current signal, allowing the signal, which originally only contained current amplitude information, to also carry risk level information. After this processing, high-risk moments not only affect the signal amplitude but also change the phase structure of the signal, thus forming identifiable structural features in subsequent frequency domain analysis. The effect is to embed risk information into the signal structure, providing a basis for subsequently distinguishing frequency components from different sources.

[0036] The risk modulation current at each time moment is used to form a risk modulation current sequence. The risk modulation current sequence is subjected to Fourier transform to obtain the amplitude spectrum and phase spectrum of the risk modulation current sequence. The frequency is clustered based on the phase in the phase spectrum to obtain several cluster sets. Each cluster set includes several frequencies under the same cluster. In the frequency domain, signals from different sources typically exhibit different phase structures; that is, frequency components generated by the same physical mechanism often have similar phase distributions. Therefore, by clustering the phases in the phase spectrum, frequency components with similar phase structures can be grouped into the same cluster set, with each set corresponding to a class of frequency components with similar sources of change. For example, some sets might correspond to frequency components caused by changes in train operating load, while others might correspond to frequency components generated by risk shocks or random disturbances. In this way, complex spectral structures can be decomposed into multiple frequency sets with different source characteristics. The effect is to achieve source separation of the spectral structure, creating conditions for subsequent targeted adjustments to the spectral structure.

[0037] Clustering the phases of the phase spectrum can be performed using existing techniques, such as the K-means algorithm. Specifically, this involves: first, using the phase value corresponding to each frequency point in the phase spectrum as a data sample to be clustered, forming a one-dimensional phase dataset; then, determining the number of clusters K based on the spectral gap method; specifically: using frequencies as row and column indices (i.e., both rows and columns) to construct a similarity matrix, where the elements of the similarity matrix represent the phase difference between two frequencies; calculating the eigenvalues ​​of the similarity matrix and sorting them in ascending order; for any eigenvalue except the first one, the absolute difference between it and the previous eigenvalue is the spectral gap for that eigenvalue; and obtaining the sorting result corresponding to the eigenvalue with the largest spectral gap as the number of clusters K. K phase centers are randomly initialized. The absolute angular distance between each phase value and each cluster center is calculated and assigned to the nearest cluster center to form K initial clusters. Then, the mean of all phase values ​​in each cluster is recalculated as the new cluster center. The above assignment and update steps are iterated repeatedly until the change in the cluster center is less than a preset threshold or the maximum number of iterations is reached. Finally, frequency points with similar phase characteristics in the phase spectrum are grouped into the same cluster set.

[0038] Set amplitude adjustment rules, and adjust the amplitude corresponding to the frequency of each cluster set according to the amplitude adjustment rules to obtain a new amplitude spectrum; The amplitude adjustment rules are set as follows: Calculate the variance of the amplitude corresponding to each frequency in each cluster set, which is called amplitude dispersion. A preset amplitude dispersion threshold is set. If the amplitude dispersion is greater than the amplitude dispersion threshold, the amplitude value of each frequency in the cluster set is retained. If the amplitude dispersion is not greater than the amplitude dispersion threshold, the average value of the amplitude values ​​corresponding to all frequencies in the cluster set is calculated, which is called the comprehensive amplitude. The amplitude value of each frequency in the cluster set is replaced with the comprehensive amplitude.

[0039] After obtaining each phase cluster set, it is necessary to further determine whether the frequency components in the set have a stable energy structure. Therefore, the variance of the amplitude values ​​of all frequencies in the set is calculated and used as an index of amplitude dispersion. When the amplitude variance is large, it indicates that there are significant differences in the energy distribution of each frequency component in the set, which usually means that the set contains a true signal structure. Conversely, when the amplitude variance is small, it indicates that the amplitudes of each frequency are similar, which often means that the energy in the set is mainly formed by random disturbance diffusion. In this way, frequency sets with true structural characteristics can be distinguished from frequency sets formed by random disturbance diffusion, which provides a basis for subsequent spectrum adjustment.

[0040] When the amplitude dispersion of a cluster exceeds a preset threshold, it indicates that the cluster has a distinct energy structure. Therefore, the original amplitude value of each frequency in the cluster is retained to avoid disrupting the true signal structure. Conversely, when the amplitude dispersion is not greater than the threshold, it indicates that the energy distribution in the cluster is relatively uniform, possibly caused by random disturbances or noise diffusion. In this case, the average value of all amplitude values ​​in the cluster is calculated, and this average value is used as the composite amplitude. All frequency amplitudes in the cluster are then uniformly replaced with this composite amplitude. In this way, the spectral diffusion caused by random disturbances can be reduced while preserving the true signal structure, resulting in a more stable and regular spectral structure.

[0041] The preset amplitude dispersion threshold can be achieved using existing technology, specifically by calculating the amplitude dispersion values ​​of all current cluster sets, sorting them from smallest to largest, and selecting the upper quartile (i.e., the 75th percentile) as the amplitude dispersion threshold.

[0042] A new spectrum is constructed by combining the phase spectrum with the new amplitude spectrum. An inverse Fourier transform is then performed on this new spectrum to obtain a new risk-modulated current sequence. After adjusting the spectral amplitude, the new amplitude spectrum is recombine with the original phase spectrum to form a new spectral structure. An inverse Fourier transform is then performed to convert the frequency domain signal back to the time domain, thus obtaining the new risk-modulated current sequence. Because the randomly diffused frequency structure has been standardized in the frequency domain, the reconstructed current sequence retains the main trends while reducing unstable fluctuations caused by random shocks. The result is a more stable and structurally clearer current sequence.

[0043] The rotation phase of the risk-modulated current sequence is removed to obtain the second absolute biased current sequence. In the initial step, a rotation phase was introduced into the current signal to embed risk information. Therefore, after the spectral structure reconstruction is completed, this rotation phase needs to be removed to restore the signal to the form corresponding to the actual current physical quantity. By removing the rotation phase, the signal after risk enhancement processing can be converted back into an absolute biased current sequence, while retaining the stable structure formed during the spectral adjustment process. The effect is to obtain a current sequence that retains its physical meaning and has a more stable trend of change.

[0044] Calculate the average value of the absolute bias current in the second absolute bias current sequence, and use it as the comprehensive absolute bias current for the substation area.

[0045] After obtaining the second absolute bias current sequence, the average current value at each moment in the sequence is calculated to obtain the comprehensive absolute bias current for that period. Since the sequence has been processed by risk modulation and spectral structure reconstruction, the effects of random shocks and spectral diffusion have been weakened. Therefore, the obtained average value can more realistically reflect the overall level of the absolute bias current for that period, thereby improving the stability and risk characterization capability of the comprehensive absolute bias current index.

[0046] Step 4: For non-substation areas, set weighting factors based on their distance from substation areas, determine the comprehensive absolute bias current of each non-substation area based on the partition weighted interpolation algorithm, determine the risk situation of the current period based on the comprehensive absolute bias current, and derive the risk assessment map.

[0047] The weighting factors are shown in the following formula: In the formula, As a weighting factor, This represents the distance from a point in the non-substation area to the nearest subway line. The attenuation constant can be obtained through field testing, which is an existing technology. This can be achieved by measuring the current at different distances using probes placed in the field, thus obtaining the distance (…). The curve of current decay is then used with an exponential function. Perform fitting distance ( - Current decay curve, to determine value.

[0048] The method for determining the comprehensive absolute bias current of each non-substation area based on a partitioned weighted interpolation algorithm is a current technology. Specifically, for any point in a non-substation area, the distance between that point and each substation area (center) is calculated. If the distance is less than 5 kilometers, the corresponding substation area is referred to as the effective substation for that point. The formula for calculating the comprehensive absolute bias current of that point is as follows: in, Let be the total absolute bias current of the i-th effective substation. Let be the distance between this point and the i effective substation areas (centers). It is the power exponent of distance decay (usually taking a value of 1-3). This represents the maximum absolute bias current value across all substation areas. The exponential decay coefficient, The distance is 5 kilometers. This is the distance from the point to the nearest valid substation area (or 0 if no valid substation is available). Let be a very small value close to 0 to prevent the denominator from becoming meaningless, and let i be the index of the valid substation. This represents the number of effective substations at that point.

[0049] The comprehensive absolute bias current at each location is divided according to Table 1, and a risk assessment diagram is derived. Please see Figures 2-4 , Figure 2 This is a DC bias risk diagram for the early departure period of this invention. Figure 3 This is a DC bias risk diagram for peak periods in this invention. Figure 4 This is a DC bias risk map for off-peak periods of the present invention. The substation in the map is the substation area described in this example.

[0050] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0051] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0052] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0053] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that cannot be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways, characterized in that, The specific steps include: Step 1: Obtain the dynamic operation data of trains in the current time period and input it into the coupling model of subway and regional power grid. Simulate the absolute bias current of each substation area at each time in the current time period and determine the risk level of each absolute bias current. Step 2: For each substation area, based on the risk level distribution of its absolute bias current in the current time period, determine its risk type in the current time period. The risk types include same level distribution, adjacent level distribution and irregular level distribution. Step 3: For substation areas belonging to the same level distribution, the average absolute bias current at each time moment is taken as its comprehensive absolute bias current. For substation areas belonging to adjacent level distribution, the transition of its absolute bias current and the average absolute bias current at the current time period are analyzed to determine its comprehensive absolute bias current. For substation areas belonging to irregular level distribution, the spectrum phase reconstruction analysis of its absolute bias current at the current time period is performed to determine its comprehensive absolute bias current. Step 4: For non-substation areas, set weighting factors based on their distance from substation areas, determine the comprehensive absolute bias current of each non-substation area based on the partition weighted interpolation algorithm, determine the risk situation of the current period based on the comprehensive absolute bias current, and derive the risk assessment map.

2. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 1, characterized in that: The train dynamic operation data for the current time period includes the train's position and traction current at each moment of the current time period; The absolute bias current is the absolute value of the DC bias current.

3. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 2, characterized in that, The logic for determining the risk level of each absolute bias current is as follows: by using a preset threshold, the risk level of each absolute bias current is determined, which includes four risk levels: Level 1, Level 2, Level 3 and Level 4.

4. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 3, characterized in that, The logic for determining the risk type of the substation area in the current time period is as follows: For any substation area, if the absolute bias current at all times during the current period belongs to the same risk level, then the risk type of this substation area is defined as the same level distribution; otherwise, the following logical judgment is made. The number of times the absolute bias current falls into each risk level in the substation area during the current time period is counted. A preset threshold is set, and the risk level with a number of times exceeding the threshold is called the primary risk level. If there is only one primary risk level in the substation area, and the risk level to which the absolute bias current belongs at other times is adjacent to the primary risk level, then the risk type of this substation area is defined as adjacent level distribution. If there are two or more primary risk levels in the substation area, and the primary risk levels are continuously distributed, then the risk type of this substation area is also defined as adjacent level distribution. For any substation area, if its risk type is neither distributed at the same level nor at adjacent levels, then its risk type is defined as an irregular level distribution.

5. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 4, characterized in that, The logic for determining the transition status in the current time period is as follows: For any substation area with adjacent risk levels, the logic for determining its comprehensive absolute bias current is as follows: For any moment starting from the second moment in the current time period, if the risk level of the absolute bias current changes relative to the previous moment, it is recorded as a transition; otherwise, it is recorded as a translation. If the risk level of the absolute bias current at this moment is greater than the risk level at the previous moment, the transition is called a positive transition; otherwise, the transition is called a negative transition.

6. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 5, characterized in that, The logic for analyzing the transition of the absolute bias current in the current time period is as follows: The number of transitions and translations occurring in the substation area are counted. The sum of the number of transitions and translations is called the total number of counts. The number of transitions divided by the total number of counts is called the transition density. A transition amplification factor is preset for each risk level. The average absolute bias current of the substation area at each moment in the current period is calculated, and the risk level to which the average absolute bias current belongs is determined as the target risk level. The transition amplification factor corresponding to the target risk level is extracted as the target transition amplification factor. The product of the transition density and the target transition amplification factor is calculated to obtain the transition adjustment ratio.

7. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 6, characterized in that, For substation areas belonging to adjacent grade distributions, the logic for determining their comprehensive absolute bias current is as follows: If the number of positive transitions is greater than the number of negative transitions, add 1 to the transition adjustment ratio and multiply by the average absolute bias current to obtain the comprehensive absolute bias current for the substation area; if the number of positive transitions is equal to the number of negative transitions, use the average absolute bias current as the comprehensive absolute bias current for the substation area; if the number of positive transitions is less than the number of negative transitions, subtract the transition adjustment ratio from 1 and multiply by the average absolute bias current to obtain the comprehensive absolute bias current for the substation area.

8. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 1, characterized in that, The logic for performing spectral phase reconstruction analysis on the absolute bias current in the current time period is as follows: For the absolute bias current of the substation area at each moment in the current period, the risk level corresponding to the assignment is embedded into the absolute bias current in the form of a rotated phase to obtain the risk modulation current. The risk modulation current at each moment is used to form a risk modulation current sequence. The risk modulation current sequence is subjected to a Fourier transform to obtain the amplitude spectrum and phase spectrum of the risk modulation current sequence. The frequency is clustered based on the phase in the phase spectrum to obtain several cluster sets. Each cluster set includes several frequencies under the same cluster. An amplitude adjustment rule is set, and the amplitude corresponding to the frequency of each cluster set is adjusted according to the amplitude adjustment rule to obtain a new amplitude spectrum. The phase spectrum and the new amplitude spectrum are combined to form a new spectrum. The new spectrum is subjected to an inverse Fourier transform to obtain a new risk modulation current sequence. The rotated phase of the rotated risk modulation current sequence is removed to obtain the second absolute bias current sequence.

9. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 8, characterized in that: The logic for setting amplitude adjustment rules is as follows: Calculate the variance of the amplitude corresponding to each frequency in each cluster set, which is called amplitude dispersion. Set an amplitude dispersion threshold. If the amplitude dispersion is greater than the amplitude dispersion threshold, retain the amplitude value of each frequency in the cluster set. If the amplitude dispersion is not greater than the amplitude dispersion threshold, calculate the average of the amplitude values ​​corresponding to all frequencies in the cluster set, which is called the comprehensive amplitude. Replace the amplitude value of each frequency in the cluster set with the comprehensive amplitude.

10. The dynamic risk assessment method for DC bias magnetization of transformers caused by stray currents in subways according to claim 8, characterized in that: For any substation area with an irregular risk level distribution, the logic for determining its comprehensive absolute bias current is as follows: Calculate the average value of the absolute bias current in the second absolute bias current sequence, and use it as the comprehensive absolute bias current for the substation area.

Images