Self-diagnosis method of substation measurement system state terminal

Abstract

The invention relates to a self-diagnosis method for a state terminal of a substation measurement system, which comprises the following steps of: comparing signal values at different sampling nodes of signals from the same physical source in each measurement channel through the state terminal, and judging that the channels have physical chain degradation or poor contact; performing perturbation modulation on a measurement sampling process based on a clock signal controlled by a crystal oscillator in the state terminal, and if response deviation abnormity is detected, judging that a sampling time base has drift, time sequence disorder or cache backlog so as to identify a time base health state of a sampling chain in the terminal; in the operation process of the transformer substation, response differences generated by the excitation in different measurement channels are collected, and if response delay, amplitude distortion or jitter starting point deviation is found, it is judged that the anti-interference capacity of the corresponding channel is degraded; and according to the electric energy change rhythm sensed by the terminal in real time, adaptively adjusting the sensitivity threshold value of the jump response, and if frequent misjudgment or slow response of the jump response is detected, marking that the behavior sensitivity of the terminal sampling system is abnormal.

Description

TECHNICAL FIELD

[0001] The present application relates to a self-diagnosis method of a substation terminal, in particular to a self-diagnosis method of a substation measurement system state terminal. BACKGROUND

[0002] The prior art such as Chinese patent CN110011414A intelligent substation sampling loop fault judgment method based on homologous data comparison, although its method has certain practicability in improving the diagnosis efficiency of intelligent substation sampling loop, but still has obvious defects and disadvantages, especially in dealing with complex interference environment, system multi-dimensional degradation mode recognition and diagnosis closed loop, the adaptability is poor, and it is difficult to meet the high reliability, high robustness and high self-adaptability requirements of intelligent self-diagnosis system of future digital substation. Its main problems can be summarized as follows: first, the fault judgment of this patent depends on the amplitude and phase angle comparison model between homologous data, which is essentially a static consistency detection mechanism based on redundant comparison. When facing such as AD double channel, AB double set or cross interval configuration, if there is synchronization deviation, noise disturbance or sampling delay between channels, false alarm or missed alarm is easy to occur. For example, this method requires that the amplitude of two channels must be greater than 0.1Xn, and the phase angle difference satisfies the fixed range (within ±180°), which is applicable in ideal state, but in actual substation environment, external electromagnetic interference, communication jitter and multi-path reflection will cause signal phase angle transient fluctuation, thereby affecting the comparison effectiveness. In addition, this model adopts threshold judgment method, which cannot dynamically adapt to the fluctuation characteristics under different running stages or load conditions, lacks self-adaptive threshold adjustment mechanism, resulting in fixed sensitivity and high false alarm rate. Secondly, this method does not involve the behavior modeling and trend judgment of multi-dimensional state evolution of the measurement system. In other words, this patent focuses on the comparison result at the current time, and fails to establish a logical fluctuation tracking mechanism facing time sequence. For example, a certain AD channel deviates from the reference value for a long time, but the deviation changes slowly and does not trigger the transient threshold, so the original method may not be able to identify this abnormal trend for a long time, and finally lacks the basis for prediction when failure occurs.

[0003] Thirdly, the method does not consider the influence of dynamic interference factors of the physical environment where the measurement system is located on the judgment result. All its judgment processes assume that the sampling chain is in an ideal working condition, but in actual engineering, the cabinet, space, electromagnetic coupling strength and temperature and humidity fluctuations of the channel may cause measurement abnormalities. Fourthly, the invention fails to provide a comprehensive judgment strategy for multi-channel structure degradation or logic level failure. Its core logic is point-to-point channel level comparison, without the ability to establish correlation modeling and matrix analysis across channels. For example, when multiple channels fail simultaneously due to shared motherboard aging or sampling chip failure, the method cannot identify such common cause structure degradation, but simply reports channel comparison mismatch. Fifthly, the diagnosis strategy of the original method is essentially a one-time comparison and static threshold alarm, lacking self-feedback optimization capability, and not embodying the logic drift identification and response strategy adjustment mechanism between modules. Especially in the case of unstable or drifting multi-module output, the invention does not provide any degradation strategy or reasoning module freezing mechanism, which poses the risk of long-term failure of module judgment but still participating in the decision chain.

[0004] In summary, the existing technology of homologous data comparison method is suitable for basic sampling circuit consistency judgment, but cannot meet the needs of future intelligent substation in multi-channel structure degradation identification, logic drift diagnosis, environmental interference adaptation and state self-evolution learning. SUMMARY

[0005] The purpose of the present application is to provide a self-diagnosis method for substation measurement system state terminal, thereby solving some of the problems and deficiencies pointed out in the background art.

[0006] The technical scheme adopted by the present application to solve the above technical problems is as follows: a self-diagnosis method for a substation measurement system state terminal, comprising: comparing signal values at different sampling nodes of a plurality of physical paths, including sensors, amplifiers, analog-to-digital converters, buffers and communication modules, of signals from the same physical source in each measurement channel through the state terminal, and if trend mismatch, amplitude mutation or phase disorder are detected between the signals in each path, it is determined that the channel has physical chain degradation or poor contact;

[0007] Based on the clock signal controlled by the internal crystal oscillator of the state terminal, the measurement sampling process is perturbed and modulated, and the response of the measurement value to the perturbation is monitored. If the response deviation is abnormal, it is determined that the sampling time base has drift, timing disorder or buffer accumulation, to identify the time base health status of the internal sampling chain of the terminal. During the operation of the substation, the electromagnetic interference generated by circuit breakers, transformers or other external devices is used as a natural excitation signal, and the response difference of the excitation in different measurement channels is collected. If the response delay, amplitude distortion or jitter starting point offset is found, it is determined that the anti-interference ability of the corresponding channel has degraded;

[0008] According to the terminal real-time sensing of the electric energy change rhythm, the sensitivity threshold of the jump response is adaptively adjusted, and if the jump response is frequently misjudged or sluggish, the terminal sampling system is marked as abnormal in behavior sensitivity.

[0009] Further, when the signal value comparison is performed, the original reference signal from the primary side of the mutual inductor is introduced as a comparison reference to eliminate the gain error influence introduced by the analog link transmission; when it is detected that the trends of multiple measurement channels are simultaneously mismatched and the phase disorder is consistent, the terminal determines that it is a common node degradation according to the channel topology.

[0010] Further, the perturbation modulation includes disturbance segment tests implemented at different amplitude levels, and the load jitter sensitivity of the sampling chain is determined by extracting the nonlinear response threshold point position; when it is identified that there is a backlog trend in the sampling buffer, the terminal triggers the buffer refresh and flow control mechanism to restore the data transmission rhythm and prevent false triggering of abnormalities.

[0011] Further, the response difference is based on the amplitude offset and the signal start jitter noise frequency change of the response, which is used to identify the filter performance degradation in the channel; if multiple channels produce asynchronous responses to the same interference event, the terminal compares the response start time difference to detect the difference in the common mode interference resistance of each measurement module.

[0012] Further, the terminal performs multi-event clustering analysis on the jump misjudgment records, and if a certain type of misjudgment repeatedly occurs in consecutive different time periods, it is marked as a structural anomaly; when the sensitivity threshold of the jump response is adaptively adjusted, the current load fluctuation level of the measurement channel is used as an adjustment factor.

[0013] Further, when the measurement channel is determined to have an overlapping result of anti-interference degradation and physical link degradation, the terminal increases the fault weight of the channel and includes it in the repair recommendation list; the terminal establishes a state consistency record table for each diagnostic module, and when the output result of a certain module continuously fluctuates beyond a threshold, it is determined that the logic stability is decreased, which is marked as potential reasoning drift.

[0014] The measurement terminal first performs a composite judgment on the running state of each measurement channel; if a certain measurement channel is simultaneously determined to have anti-interference ability degradation and physical link degradation, the channel is regarded as a composite degradation channel, and the corresponding fault weight is increased and included in the priority repair recommendation list; the terminal in the present application continuously monitors and integrates the deviation analysis of the historical output state of each diagnostic module, and defines the module logic stability index in the form of the following original function:

[0015]

[0016] Wherein:

[0017] Φ(t) is the logic stability index at time t, the larger the value, the more intense the module output fluctuation; α s is the output result of the diagnosis module at time point s; is the output sliding mean value in the window period, indicating the stable reference of the diagnosis result; θ s is the signal offset intensity of the local channel at time s; β is a response enhancement factor, used to adjust the amplification effect of signal fluctuation on output judgment; is introduced to avoid false judgment caused by short-time signal jitter; Δ is the integral window length, controlling the time coverage of the analysis;

[0018] The logic stability index Φ(t) combines the deviation rate (derivative) of the historical output and the exponential suppression of the external signal disturbance in the calculation process, so as to distinguish the non-stable fluctuation caused by the body logic drift and the input abnormality; if the index continuously exceeds the dynamic threshold Φ th , the module will be marked as a potential reasoning drift module, and enter the subsequent freezing, weight reduction, or strategy rollback process;

[0019] The derivation process of the above Φ(t) formula includes:

[0020] Let the output of the diagnosis module at any time s be α s , and the average output in a certain sliding time window (length Δ) be The difference between the two reflects the instantaneous deviation; in order to pay more attention to rapid fluctuations rather than absolute deviation, the present application takes the derivative of the difference with respect to time, that is, introduces:

[0021]

[0022] But in order to suppress the non-real jump of the output caused by short-term disturbance of the channel (such as electromagnetic interference, load mutation), the present application introduces an exponential damping term under the derivative term, which is:

[0023]

[0024] Where θ s represents the signal disturbance intensity at time s, and β is an adjustable enhancement factor; when the disturbance is strong, θ s is large, the exponential term tends to zero, so that the derivative is weakened; while the disturbance is weak, the exponential term tends to 1, the derivative has the greatest influence, and has the effect of environmental shielding; the above two are combined to construct the ratio:

[0025]

[0026] This expression is a disturbance adjustment version of the overall deviation rate, and then its derivative is taken (to characterize the deviation acceleration rate):

[0027]

[0028] Since the judgment logic stability essentially concerns the strength of the deviation rate, not the positive or negative direction, the absolute value is taken, and the integral is taken in the sliding time interval to obtain the logic stability function, which integrates three levels of dynamic diagnosis:

[0029] 1. The short-term deviation trend (derivative) reflects the non-stationarity of the output;

[0030] 2. The exponential suppression term suppresses the amplification of false signals during high disturbance periods;

[0031] 3. The integral window sliding accumulation provides system memory and realizes trend stability tracking;

[0032] Finally, Φ(t) is used as a stability characterization index of the diagnostic module's logic behavior at time t, and once it exceeds the system set threshold Φ th , it can be marked as potential reasoning drift. This judgment process has mathematical interpretability, engineering executability, and logical scalability, and can be used for subsequent degradation control, strategy adjustment, or module freezing.

[0033] Further, the terminal introduces the ambient temperature, humidity, and electromagnetic interference level corresponding to the physical location of the measurement channel as an adjustment factor to correct the fault weight while increasing the fault weight; each channel in the maintenance recommendation list is also marked with its last maintenance period and cumulative operating hours to assist in determining whether to trigger preventive maintenance.

[0034] Further, when multiple channels are simultaneously included in the maintenance recommendation list, the terminal determines whether it is a structural degradation of the common motherboard, sampling chip, or regional module through the cross-correlation channel fault probability matrix; the fault weight is used to generate the list and adjust the priority ranking and sampling frequency of the channel in data upload.

[0035] Further, the state consistency record table records the amplitude of the output fluctuation of the diagnostic module and the condition hit count of the key nodes in the judgment path; when the diagnostic module is marked as potential reasoning drift, the terminal verifies the sampling accuracy and stability of the dependent input signal source channel to determine whether the drift is caused by external data link abnormalities.

[0036] Further, when the same module is marked as reasoning drift in two consecutive self-diagnoses, the terminal enters a logic freezing mode, suspends the module from participating in the final state judgment, and uses the results as a reference index; the terminal scores the drift index of all modules marked as potential reasoning drift and determines whether it is in a deteriorating state to trigger software version rollback or repair.

[0037] The beneficial effects of the present application: by fusing the physical link state, anti-interference ability, sampling buffer behavior and behavior sensitivity, etc. Multi-layer state characteristics, realize the composite judgment of the health state of the measurement channel, effectively avoid the misjudgment of single index, improve the accuracy and integrity of fault identification. Introduce the temperature, humidity, electromagnetic interference level and other environmental factors of the channel as dynamic adjustment parameters, used to correct the diagnosis output and fault weight, with stronger environmental robustness, avoid misjudgment or missed detection caused by extreme environmental fluctuations. By constructing the state consistency record table of the diagnosis module, extracting its output fluctuation behavior, judgment path hit characteristics and other dynamic characteristics, combined with the logic stability index to quantitatively analyze the reasoning stability, from the source to guarantee the reliability and credibility of long-term operation of the module logic.

[0038] In the case of using multi-channel joint maintenance list, the fault probability matrix between channels is automatically constructed, the structural degradation problem of shared motherboards, sampling chips or regional modules is identified, and the positioning ability of the system to hardware common cause failure is significantly improved. The diagnosis module marked as reasoning drift is set to freeze mechanism, and the drift deterioration state is monitored through time trend, once it is determined that the drift is continuous, version rollback or module repair can be triggered, forming a closed loop of logic reliability self-recovery. Fault weight is not only used for maintenance sorting, but also drives data upload priority adjustment and sampling frequency adaptive adjustment, realizes the priority sampling and high frequency upload of key channels, and improves the dynamic response ability to high risk signal chain. The system records and uses the maintenance period, cumulative operating hours and misjudgment clustering results to support preventive maintenance and structural analysis decision, reduce operation and maintenance cost, and avoid over-repair or delayed repair. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 The flowchart of the substation measurement system state terminal self-diagnosis method of the present application is simplified.

[0040] Figure 2 The functional relationship diagram of the substation measurement terminal multi-dimensional self-diagnosis of the present application.

[0041] Figure 3 The self-diagnosis and priority control flowchart of the substation measurement terminal of the present application.

[0042] Figure 4 The simplified schematic diagram of the substation measurement terminal multi-channel self-diagnosis of the present application embodiment 1.

[0043] Figure 5 The substation multi-channel adaptive diagnosis and logic stability processing flowchart of the present application embodiment 2. DETAILED DESCRIPTION

[0044] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0045] In combination with the accompanying Figure 1 The substation measurement system state terminal self-diagnosis method of the present application performs real-time monitoring and verification on the multi-stage signal transmission path inside the measurement channel. The state terminal establishes a complete sampling link mapping for the signal source in each measurement channel (such as the analog output from the same voltage transformer or current transformer). Starting from the source point, the signal successively passes through the sensor primary interface, the analog signal amplification module, the analog-to-digital converter (ADC), the cache processing module, and the final communication interface module to form a complete data flow path. The state terminal selects multiple representative signal sampling nodes at the above-mentioned key path points and performs parallel comparison and analysis of the values, trends, and phase characteristics of the signal source reflected on multiple nodes at the same time or within a synchronous time window. The signal value comparison is not limited to the absolute difference of the instantaneous amplitude, but also includes the change trend (such as the rising edge slope, change direction) of the signal at each node in the time dimension, the consistency of the frequency spectrum structure, and the phase offset. If any of the following abnormal conditions is detected within a certain time interval: 1) there is a significant trend mismatch in some path signals, i.e., some nodes show an upward trend while another road shows a downward trend; 2) the amplitude difference at the same time point between paths exceeds the system's preset dynamic tolerance limit, resulting in amplitude mutation; 3) the phase alignment relationship of the signal suddenly deviates or nonlinearly drifts, the system considers that the physical link integrity of the measurement channel has been damaged according to the comprehensive judgment logic, further infers that there are hardware-level risk hidden dangers such as physical link degradation (such as increased contact resistance, unstable amplifier gain) or poor contact (such as loose sensor connector, virtual solder joint), finally generates a state abnormality flag, and lists the channel in the priority diagnosis target list. At the same time, the abnormal event is reported to the remote monitoring platform for system maintenance and scheduling, thereby realizing a high-reliability measurement chain health self-diagnosis mechanism that does not require external excitation signals and only relies on system self-link observation.

[0046] By constructing the terminal internal clock modulation and electromagnetic interference response double mechanism, the time base health and channel anti-interference ability of the terminal internal sampling chain are jointly judged. Firstly, the system relies on the stable clock signal provided by the state terminal internal crystal oscillator to implement periodic perturbation modulation on the measurement sampling process. The modulation method makes small compression or stretching effect of the sampling time base by making small amplitude dynamic disturbance adjustment of the sampling clock period without affecting the stable operation of the main clock. Then the terminal monitors the response characteristics of the measurement values collected by each channel under the micro-perturbation time base. If the sampling chain is normal, the signal changes before and after the perturbation should show high consistency. If the response value jitter is enhanced, the waveform is distorted, the expected change rhythm is not met or nonlinear deviation occurs under the disturbance excitation, it means that there is sampling time base offset or system sampling buffer backlog, sampling scheduling disorder in the channel. The system determines that the time base stability of the terminal internal sampling chain is decreased, and the corresponding state is marked. Further, in the normal operation process of the transformer substation, the state terminal uses the inherent existing device behavior signal in the system, especially the electromagnetic interference caused by the operation of circuit breaker, transformer switching or other high-power equipment, as the natural excitation source. Such interference signal has high energy, strong transient characteristics and does not need to inject additional test signal. The state terminal collects the transient response difference of such electromagnetic disturbance in different measurement channels, and compares its performance in amplitude fluctuation, response delay, pulse starting consistency and other dimensions. If the response of a channel to the same excitation event is significantly delayed, attenuated, waveform unstable or shows starting offset, the system determines that the anti-interference ability of the channel has degraded. The degradation is caused by factors such as shielding layer aging, common mode interference coupling enhancement, signal conditioning module failure. The terminal lists the channel in the anti-interference abnormal record based on the above self-diagnosis judgment result, and triggers parameter adaptive adjustment or maintenance reminder if necessary, so as to realize the online self-diagnosis mechanism of dynamically identifying the terminal internal sampling stability and channel anti-interference ability in the running process without shutdown test.

[0047] Based on the continuous sensing ability of the terminal to the real-time change characteristics of electric energy, a dynamic adaptation mechanism of jump response behavior is constructed for evaluating the sensitivity and stability of the terminal sampling system in the fast response scene, wherein the electric energy change rhythm includes real-time analysis of the dynamic curve formed by voltage, current, active power, reactive power and their derivatives, the terminal continuously tracks the typical electric energy jump events such as load switching, bus voltage adjustment and power side fluctuation of the monitored loop, extracts the change amplitude, duration, rising / falling edge slope and occurrence frequency of these jumps within a certain time window, and constructs a change rhythm model under the current operating condition according to the above, which does not depend on external input and is completely generated by local perception calculation of the terminal. Based on the model, the terminal adjusts the sensitivity threshold set in its jump response mechanism adaptively, so that the system relaxes the threshold value of jump determination when the power grid is in a high dynamic fluctuation state to avoid false positives, and tightens the threshold value to improve sensitivity when the power grid runs smoothly to enhance the burst recognition ability. When the terminal detects frequent misjudgment (such as normal load switching is judged as abnormal disturbance) or sluggish response (such as not identifying the electric energy mutation in time after its occurrence) in its actual operation, the system determines that the current sampling system has abnormal behavior sensitivity according to the corresponding deviation between the jump event record and the response behavior, that is, the behavior judgment logic of the system is no longer accurately adapted to the current load dynamic environment. The abnormality is caused by internal response rule solidification, sampling frequency drift, algorithm threshold aging or channel delay jitter. After marking such behavior abnormalities, the terminal can further trigger the recalibration mechanism or generate prompt information uploaded to the upper platform for operation and maintenance analysis, thereby realizing a sampling response behavior level self-diagnosis mechanism based on the self-perception and adjustment ability of the dynamic operation background of the power grid.

[0048] The drawings are combined Figure 2In the process of comparing the signal values in each measurement channel, a high-reliability original reference signal and a topology-related identification mechanism are introduced to improve the accuracy and systematicness of fault identification. When the state terminal compares the voltage, current and other analog signals obtained from the same physical source in each measurement channel, it not only compares the sampling results at nodes such as sensors, amplifiers, analog-to-digital converters, buffers and communication modules in each path, but also introduces an original reference signal from the primary side of the transformer as a comparison reference. This original reference signal is accessed through a short path or bypass to the terminal's dedicated reference channel and does not participate in conditioning and amplification. It has the pure characteristics of not being processed by link gain and is the source of the actual signal. In this way, errors caused by amplifier gain drift, analog signal attenuation, connection impedance changes and other non-measurement factors are eliminated. By normalizing and correcting the differences in amplitude and trend between the sampling values of each channel and the original reference value, the accuracy of measurement channel degradation determination is improved. Further, during system operation, when the terminal detects that the sampling results of multiple channels have trend mismatches, i.e., multiple channels show opposite or mutually exclusive rising or falling trends in the same time period for the same power change, and the phase information of the sampling results of each channel shows consistent overall drift or alignment damage, the system triggers a cross-channel consistency diagnosis mechanism. Based on the terminal's preset topology information, including the distribution of physical transformers relied on by each measurement channel, the sharing relationship of conditioning modules and bus wiring structure, etc., the system analyzes whether there is a common hardware-dependent path for these abnormal channels. If these channels share the same transformer, the same power supply path or common amplification / sampling modules in the topology, the system determines that they are common node degradation, i.e., the problem is not in the individual channels themselves, but in the structural nodes that these channels rely on. Then, through local marking and sending fault types to the dispatch center, early detection, accurate positioning and maintenance priority sorting of structural degradation are achieved, thereby constructing a high-precision self-diagnosis method that combines original reference correction and cross-channel topology sensing capability.

[0049] In the detection of the time base health of the terminal sampling chain, a composite mechanism based on perturbation segmentation test and buffer dynamic identification is adopted. First, the terminal implements perturbation modulation on the sampling clock signal based on the internal crystal oscillator. The modulation is not limited to a single perturbation form, but adopts a multi-level perturbation segmentation method, that is, different amplitude levels of sampling time base perturbation are introduced in continuous sampling periods, for example, ±1%, ±3%, ±5% and other proportional period adjustments are applied in turn according to the microsecond level. Through this step-by-step perturbation, the response behavior of the system under slight, medium and edge states to the perturbation is excited. Then the terminal records the response of the sampling value under each perturbation level in real time, and identifies whether there is an obvious nonlinear turning point or abnormal mutation inflection point through the analysis of the signal change slope, response lag degree, signal amplitude offset and other indicators. The position of the threshold point represents the critical region of the system from stable response state to sensitive state transition. According to this, the sensitivity of the sampling chain to load disturbance is determined, so as to judge the robustness and time base pressure performance of the system. When the nonlinear response threshold point is at a lower perturbation level and the response changes sharply, it indicates that there is a potential hidden danger of high sensitivity and low fault tolerance in the sampling chain. Further, in order to judge whether the system data link exists response lag caused by flow blockage, the terminal continuously monitors the state of the sampling buffer area, records the dynamic difference value between the data in and out speeds, analyzes the buffer depth growth trend, and when it is identified that the buffer length continuously rises in a short time and exceeds the normal periodic fluctuation range, it is considered that there is a buffer accumulation trend. The system immediately triggers the self-recovery logic, including immediately refreshing the low-priority sampling data in the buffer, releasing the buffer space occupied by the non-critical diagnostic module, and dynamically adjusting the flow control rate of the communication interface. Through the above mechanism, the rhythm and continuity of data flow are restored, the false triggering, false diagnosis or result drift caused by sampling delay are avoided, and finally the joint self-diagnosis of the abnormal sensitivity of the sampling chain and the health status of the data channel is realized, so as to ensure that the state terminal can still maintain high reliability in the face of micro-perturbation disturbance and link pressure conditions.

[0050] Based on the fine observation of the electromagnetic interference response behavior of the measurement channel, the response characteristics are extracted from multiple dimensions for determining the channel internal filtering performance and anti-interference ability. Firstly, the terminal uses the naturally existing electromagnetic disturbance source in the operation of the substation as the excitation input, including the transient disturbance wave caused by the events such as circuit breaker action, transformer excitation inrush, bus switching, etc. The disturbance signal does not need to be externally injected, and has the characteristics of randomness and typicality. In the process of collecting the response of each measurement channel to the same disturbance event, the terminal not only analyzes the traditional response amplitude offset, i.e. the amplitude peak difference of each channel output signal in the initial disturbance stage, but also further extracts the jitter characteristics and high-frequency noise frequency change carried by the response signal in the initial stage, including the micro-fluctuation frequency of the rising edge of the response signal, the change of pulse envelope density and the change of instantaneous frequency domain energy distribution, etc. Such high-frequency response behavior is mainly determined by the internal analog filter, digital filter or hybrid filter structure of the channel. If the jitter enhancement of the response signal starting point, the frequency band widening or the rising of random noise components of a channel are continuously observed in multiple disturbance samples, the terminal determines that the filter performance of the channel has degraded due to filter cutoff frequency drift, device aging or high-frequency bypass failure, etc. On this basis, the terminal further compares the response starting time of each channel to the same disturbance event. If multiple channels should logically respond synchronously (such as originating from the same bus voltage disturbance), but the actual acquisition result shows that the response starting time exists obvious offset, i.e. there is a non-systematic delay in milliseconds or even sub-milliseconds, the terminal analyzes whether there is a difference in the anti-common mode interference ability between channels according to the measurement link topology and hardware structure, combined with the delay difference between channels. If the suppression ability of a channel to external common mode signal is obviously weaker than that of other channels, the channel has problems such as poor common mode filtering, decreased electrical isolation, or unbalanced reference ground, etc. Finally, the terminal establishes the diagnosis label of filtering performance degradation and anti-common mode ability deterioration based on the above response difference recognition result, and forms the health index update of the sub-channel for subsequent abnormal monitoring, parameter correction and maintenance plan reference, thereby constructing a multi-dimensional dynamic filtering state recognition self-diagnosis method which integrates amplitude response, noise frequency change and response time difference judgment.

[0051] By introducing multi-event clustering analysis and load-aware adjustment mechanism, the intelligent identification and sensitivity dynamic management of jump response misjudgment behavior are carried out. The terminal classifies and records all jump response behaviors occurring in the running process, especially pays attention to the events judged as misjudgment subsequently, and puts them into the misjudgment behavior log table, wherein each record contains the timestamp of misjudgment occurrence, response channel number, trigger type, waveform characteristics, system load state and other information. Then, the terminal carries out high-dimensional similarity clustering on all misjudgment events based on the event feature vector through the multi-event clustering algorithm based on time window sliding. If a misjudgment type repeatedly occurs in multiple non-continuous but adjacent or periodic time periods, that is, it presents time-related distribution pattern and characteristic stability, the terminal determines that this misjudgment type is not a sporadic judgment error, but a structural abnormal behavior caused by diagnostic logic or physical environment, and labels it as a structural abnormality in the system for subsequent priority debugging and rule correction. At the same time, the sensitivity threshold adjustment strategy for jump response is no longer fixed proportion method or static setting, but introduces the current load fluctuation level of the channel as a dynamic adjustment factor, that is, by continuously calculating the current, voltage power change standard deviation, average change rate and power factor fluctuation rate of the channel in a certain time window, a transient load dynamic level model is constructed. If the current channel is in a volatile or high-frequency switching working condition, the sensitivity threshold of jump judgment is relaxed to avoid normal jump being misjudged as abnormal. If the load is stable, the system tightens the threshold to improve the identification ability, thereby realizing a terminal behavior level self-diagnosis method combining the abnormality identification mechanism based on historical misjudgment clustering and the jump judgment sensitivity self-adjustment mechanism driven by real-time load fluctuation, which enhances the adaptability and self-correction ability of the measurement system to the dynamic environment of power grid.

[0052] Combining the drawings Figure 3 In the process of collecting and judging the running state of multiple measurement channels, first, a composite degradation identification logic is established for channel health. When a measurement channel is simultaneously judged to have degradation of anti-interference ability and physical link degradation, the terminal classifies the channel as a composite degradation channel and increases its fault weight, so that the channel is in a higher priority position in fault handling and maintenance planning. At the same time, the channel will be included in the maintenance recommendation list for subsequent system to call when dispatching maintenance tasks. In order to enhance the stability identification ability of the terminal internal diagnosis logic module, a logic stability monitoring method based on historical output behavior is further proposed. This method constructs the logic stability index Φ(t) of the diagnosis module through continuity analysis and nonlinear function integration, which is defined as follows:

[0053]

[0054] Where Φ(t) represents the logic stability index at time t, reflecting the stability of the module output; a larger value indicates more severe output fluctuations; α s This represents the raw output of the diagnostic module at time point s. Let θ be the average output within a sliding window centered at s, representing the average steady state of the module during that time period; s β represents the signal offset strength of the local channel at time s, which indicates the severity of the external disturbance; β is the response enhancement factor, used to adjust the degree of influence of the external disturbance on the logic judgment. This is an exponential damping suppression term used to suppress amplified abnormal signals under high-disturbance environments and prevent short-term jitter from falsely triggering system judgments; Δ is the integral window length, controlling the time range for historical analysis of the control logic state. In actual operation, the output of the diagnostic module may jitter due to unstable external inputs, power grid disturbances, sampling errors, or fluctuations in algorithm logic. The derivation process of this exponent is as follows: First, construct the deviation term between the instantaneous output of the module and its average value. This term characterizes the output instability of the module. To enhance the dynamic sensitivity of this deviation term, its time derivative is introduced. Furthermore, considering that short-term disturbances in practical applications should not cause significant changes in judgment behavior, the disturbance intensity θ is... s Introducing the exponential decay function As the denominator, it achieves environmental weighted suppression of the deviation rate, forming a perturbation correction expression. The derivative of this expression is used to characterize the deviation acceleration rate, and its absolute value is then integrated within a sliding time window to finally obtain the logic stability index Φ(t). This function has three technical features: First, it identifies non-steady states by characterizing the instantaneous change trend of the module output through the derivative; second, it suppresses pseudo-fluctuations caused by high-frequency interference through the perturbation index decay factor, thereby improving robustness; third, it introduces the cumulative perception of historical states through the integration operation, realizing the ability to perform delayed analysis of logic behavior trends. Therefore, when the terminal detects that the Φ(t) index of a certain diagnostic module is continuously higher than the dynamic threshold Φ set by the system for multiple cycles, it can detect non-steady states. th When the output behavior of a module is determined to be unstable, it is marked as a potential inference drift module. This module will be assigned an untrusted label in the system and will enter processes such as freezing, downgrading, or rollback of control strategies to prevent its output from continuing to participate in measurement decisions or causing larger system misjudgments. This constructs a logical stability assessment mechanism with mathematical interpretability, dynamic adaptability, and engineering feasibility, enhancing the state terminal's ability to monitor and adjust the reliability of its own judgment chain.

[0055] In the process of increasing the fault weight, the environmental factors of the physical location of the channel are considered, including temperature, humidity, and electromagnetic interference level and other external conditions as adjustment factors to correct the fault weight. Specifically, the terminal obtains the temperature, humidity and electromagnetic interference intensity value of the area where the channel is located by real-time monitoring of the running state of each measurement channel and combining the feedback data of the environmental perception module. Changes in these environmental factors will directly affect the working stability and accuracy of the measurement channel. In particular, in high temperature, high humidity or strong electromagnetic interference environment, the probability of measurement error and equipment failure will increase, therefore by introducing these factors as adjustment factors, the fault weight can be dynamically adjusted according to environmental changes. The specific adjustment method is that the terminal corrects the fault weight by a preset weighting model according to the temperature, humidity and electromagnetic interference level data collected by the environmental sensor, which quantifies the relationship between environmental factors and fault weight to generate real-time updated fault weight values. In this way, the terminal not only evaluates the maintenance priority of the channel according to its own fault condition, but also adjusts the priority according to the influence of environmental changes, so as to realize more accurate and efficient fault diagnosis and maintenance decision. At the same time, in the maintenance suggestion list, the terminal will mark the last maintenance cycle and the cumulative running hours of each channel, which will be used as auxiliary judgment basis to evaluate whether preventive maintenance should be triggered. When the cumulative running hours of the channel exceed a certain threshold, or the running period since the last maintenance is relatively long, the terminal will make a judgment of preventive maintenance. If the channel fault weight is high at this time, the system will preferentially recommend the channel for inspection or maintenance, thereby effectively reducing the risk of failure caused by equipment aging or overuse, and ensuring the stability and reliability of the measurement system of the substation.

[0056] When multiple channels are included in the maintenance recommendation list at the same time, the terminal uses the interrelated channel failure probability matrix to analyze the failure relationship of these channels in depth, so as to judge whether the failure is caused by common hardware modules or structural problems. The terminal first establishes a channel failure probability matrix according to the failure type, failure frequency, time period of each channel, etc. By analyzing the failure occurrence correlation between channels, it is judged whether multiple channel failures are concentrated in the same hardware component or regional module (for example, shared motherboard, commonly used sampling chip or measurement circuit located in the same regional module). If the failure of multiple channels shows significant time correlation and these channels have commonality in hardware structure, the terminal determines that it is a structural degradation problem at the hardware level. Through this analysis, the terminal can further optimize the formulation of the maintenance recommendation list, prioritize the handling of structural failure problems, and avoid repeated checking of each channel failure, thereby improving maintenance efficiency and saving maintenance resources. In addition, the terminal uses the generated failure weight to dynamically sort the maintenance recommendation list, and the channels with higher failure weight will be prioritized in the maintenance list to ensure that important failures and potential risks are handled in a timely manner. At the same time, the failure weight is also used to adjust the priority sorting and sampling frequency of each channel in the data upload process. When the failure weight of a certain channel is high, this channel will be given a higher sampling frequency in the subsequent data collection process, so as to more accurately monitor the state change of the channel, and the sampling data of this channel will be uploaded to the monitoring system in priority, so as to perform more urgent monitoring and intervention; while for the channel with lower failure weight, the sampling frequency can be appropriately reduced, and the priority of data upload can be adjusted according to its importance, so as to ensure the normal operation of the system while reasonably allocating resources, optimizing data transmission efficiency and network load.

[0057] By establishing a state consistency record table, the output fluctuation amplitude of the diagnostic module is continuously monitored and recorded, and the condition hit frequency of the key nodes in the diagnostic path is counted to quantify the stability of the module under different operating states. Specifically, the state consistency record table will record the output fluctuation of each run of the diagnostic module in detail, including the amplitude change of the output signal and the condition hit frequency of the key judgment nodes, where each judgment node represents a branch point in the decision path, and the condition hit frequency of the node reflects whether the path is frequently activated, which in turn affects the accuracy and consistency of the diagnostic results. By monitoring and analyzing these data, the terminal can effectively identify whether the diagnostic module has a phenomenon of excessive output fluctuation or frequent jump in the continuous working process, so as to judge whether the module has a potential reasoning drift. If the output stability of the diagnostic module decreases and deviates from the predetermined logic path, the system will mark it as a potential reasoning drift module and trigger further fault detection and repair process. In this process, the terminal further verifies the input signal source channel relied on by the module, detects its sampling accuracy and stability, and judges whether the reasoning drift is caused by external data link abnormality. The terminal compares the collected input signal with the preset normal waveform, analyzes whether the sampling accuracy is drifted, whether the signal waveform is affected by external interference or whether there are problems such as packet loss, delay, etc. in the data link, and if it is confirmed that the input signal is abnormal and causes the reasoning drift, the system will perform data link diagnosis and repair process to ensure the stability and reliability of the measurement system.

[0058] The logic freezing mechanism is combined with the drift index scoring mechanism to enhance the adaptive repair capability of the terminal to reasoning drift. When the terminal performs self-diagnosis, if a diagnostic module is marked as reasoning drift in two consecutive self-diagnosis processes, that is, there is obvious inconsistency or instability between the results output by the module in two independent time windows and the expected results, the terminal enters the logic freezing mode, suspends the module from participating in the subsequent final state judgment, at this time, the diagnostic result of the module will no longer directly affect the decision output of the system, but is temporarily marked as a reference index for comparison and analysis in the subsequent diagnostic process. In the logic freezing mode, the terminal continuously monitors the state of the module and cross- verifies the output results of the module with the results of other modules, so that when the module returns to normal, it can participate in system decision again; at the same time, the terminal scores the drift index of all modules marked as potential reasoning drift. The calculation of the drift index is based on the output stability of the module and the deviation degree from the normal operation mode, including weighted evaluation of the output fluctuation amplitude, output change frequency and consistency with historical output, to obtain a quantitative drift score. The drift index is used to reflect the stability level of the module and the severity of reasoning drift. On this basis, the terminal also judges the time trend of the module drift. If the drift index of the module shows a continuous upward trend or the drift fluctuation continuously expands in a continuous time period, the system determines that the module enters the deterioration state, and triggers the corresponding repair mechanism, including starting software version rollback or system patch update, to restore the normal logic judgment function of the module. Through this mechanism, the terminal can actively isolate abnormal modules when detecting reasoning drift, avoid their negative impact on measurement results, and flexibly adjust the repair strategy through drift index and time trend analysis, to ensure that the system always maintains efficient and stable operation state, thereby improving the intelligent diagnosis and maintenance capability of the substation measurement system.

[0059] Embodiment 1

[0060] In combination with the accompanying drawings Figure 4 In this embodiment, the measurement channels in a certain substation are subjected to self-diagnosis. In this process, the measurement terminal compares the signals of multiple channels with the original reference signal on the primary side of the transformer. It is assumed that the substation measurement system has four measurement channels (channels A, B, C and D), which are responsible for measuring the voltage and current signals at different positions in the substation, and the primary side of the transformer provides a standard reference signal. It is assumed that channels A and B are located in different areas, and channels C and D share the same measurement motherboard during the measurement process.

[0061] When performing signal value comparison, the terminal compares the signals of these measurement channels one by one according to the reference signal collected from the primary side of the transformer. Specifically, the reference signal (set as an alternating current signal) has a known amplitude and frequency, and the terminal calculates the amplitude, frequency and phase of each channel through the collected signal and compares them with the reference signal. Suppose that at a certain time, the amplitude of the voltage signal of channel A is 240V, while the amplitude of the reference signal is 230V, and the amplitude of the current signal of channel B is 10A, while the amplitude of the reference signal is 9.8A. In this example, there is a certain difference in the amplitudes of channel A and channel B, which is the deviation caused by the analog link gain. In order to eliminate this gain error, the terminal compares the reference signal with the signals of these channels, finds the difference between them, and then adjusts the measurement results to compensate for this error, thereby improving the accuracy of the system.

[0062] However, if the trend of multiple channels is mismatched and the phase is consistent, the terminal will further analyze the signals and determine whether there is a deeper hardware problem according to the channel topology relationship. For example, if the signals of channel C and channel D not only have amplitude mismatch, but also have obvious phase shift at the same time point, and these two channels share the same motherboard, the system will determine that the common node is degraded, and mark these two channels as potential fault channels. At this time, the terminal will generate a maintenance suggestion and give priority to checking the hardware of these two channels, because they have a common hardware problem, such as a shared power supply, sampling chip or motherboard failure, which causes their measurement results to have consistency problems.

[0063] The following data are given: the amplitude of the reference signal is 230V, and the frequency is 50Hz, while the measurement results of channels A, B, C and D are as follows:

[0064] Channel A: amplitude 240V, frequency 50.1Hz, phase error 0.2°

[0065] Channel B: amplitude 245V, frequency 50.0Hz, phase error 0.5°

[0066] Channel C: amplitude 238V, frequency 50.1Hz, phase error 1.2°

[0067] Channel D: amplitude 241V, frequency 50.1Hz, phase error 1.3°

[0068] When comparing these signals, the terminal finds that the amplitude of channels A, B, and C has a small deviation, but the amplitude of channel D is significantly higher, and the phase error is also larger. After further analysis, the terminal finds that the phase error of channel C and channel D has a highly consistent trend over time, and they share the same motherboard and power supply, so it is speculated that this is a hardware problem caused by power supply interference or common mode interference. Therefore, the terminal determines channels C and D as common node degradation and adds them to the maintenance recommendation list, which is given priority for repair.

[0069] There are multiple measurement channels in the system, including voltage and current signal sampling channels. Each channel is closely connected to the power equipment of the substation, performs real-time data collection, and uploads to the remote monitoring system through the measurement terminal.

[0070] First, the terminal performs a series of perturbation modulation tests when collecting signals. This test evaluates the load jitter sensitivity of each channel by introducing perturbations of different amplitude levels on the sampling time base of each measurement channel. Specifically, the current signal of channel A fluctuates around 10A under normal conditions, while the voltage signal of channel B fluctuates around 230V. The terminal performs perturbation modulation on the sampling time base at different amplitude levels such as ±1%, ±3%, and ±5% to generate perturbations of different intensities, and records the response of each channel under perturbation. For example, if the amplitude of the sampling signal of channel A remains stable under ±3% perturbation but shows a large response fluctuation under ±5% perturbation, then the channel has a higher sensitivity to larger perturbations and may experience load jitter problems. Through this test, the terminal can determine which channels produce jitter when the load changes, thereby identifying high-risk channels that affect measurement accuracy in advance.

[0071] Next, when the terminal identifies that the sampling buffer of a channel has a backlog trend, the system triggers the buffer refresh and flow control mechanism. For example, during the sampling process of channel A, due to the high-frequency data collection and transmission load of the system, it is found that the buffer area has been continuously growing for a period of time, and the buffer has not been released in time, causing data backlog. Through real-time monitoring of the buffer, the terminal finds that its growth rate exceeds the normal fluctuation range, and the system performs buffer refresh operation to release invalid data and adjust the sampling frequency to reduce the risk of buffer backlog. Through this mechanism, the system can ensure the continuity and accuracy of data transmission, and prevent false triggering of exceptions or data loss caused by buffer backlog.

[0072] During this process, the terminal also analyzes the response differences in the collected signals, especially through amplitude offset and initial jitter frequency changes to detect filter performance degradation in the channel. Set in a certain period, the response of channel A and channel B has a significant amplitude offset, and the signal of channel A has a significant initial jitter when the same interference event occurs, and its noise frequency is higher than that of channel B. This indicates that the filter of channel A has performance degradation problem, which fails to effectively suppress the influence of high-frequency noise or other interference sources. By analyzing these signal differences, the terminal can identify which channels have filter faults and mark them as targets for further inspection and repair.

[0073] During the measurement process, jump misjudgment may also occur, which is usually caused by power grid mutations or external interference. Set in a high load switching event, multiple channels respond at the same time, but channel C and channel D have frequent misjudgments, and the judgment results are inconsistent with the actual changes. The terminal will perform multi-event clustering analysis on these misjudgments and classify them as the same type of misjudgment event. If a certain type of misjudgment occurs repeatedly in multiple time periods, the terminal will mark it as a structural anomaly and perform further analysis. For example, set channel C and channel D show the same misjudgment pattern in the same power jump event, and they have appeared in multiple different time periods, it can be inferred that they share a hardware or logic problem, which causes misjudgment on the same interference event, and the terminal system will mark it as a structural anomaly and suggest to check the related hardware or logic module first.

[0074] In addition, in order to ensure the accuracy and timeliness of the response, the terminal will dynamically adjust the threshold when adjusting the sensitivity threshold of the jump response, combined with the current load fluctuation level of the measurement channel. For example, set the load fluctuation of channel A and channel B is small, the system tightens their jump threshold, and improves the sensitivity to small amplitude jump; if the load fluctuation of channel C and channel D is large, the system will appropriately relax the threshold to avoid misjudging the jump event when the load mutates. Set in a load switching event, the current of channel A suddenly jumps from 10A to 15A, the system adjusts the sensitivity threshold according to the load fluctuation to ensure accurate response to current change, and will not misjudge as equipment failure or interference event due to mutation.

[0075] Embodiment 2:

[0076] Combined with the Figure 5 Based on embodiment 1, the fault diagnosis and repair process of multiple measurement channels improves the stability and reliability of the system by combining adaptive fault weight and logic stability monitoring mechanism. In the substation, four measurement channels A, B, C, and D are responsible for monitoring the signals of different devices in the substation, such as transformers, voltage and current transformers, etc. In order to improve the signal acquisition accuracy and fault diagnosis capability, the terminal uses self-diagnosis method.

[0077] First, the terminal performs signal value comparison, and introduces the original reference signal from the primary side of the transformer as the comparison reference during signal value comparison to eliminate the gain error in analog link transmission. The amplitude of the current signal collected by channel A is set to 50 A, and the amplitude of the reference signal is 49.8 A. Although the difference between the two amplitudes is small, the system can detect the existence of gain error by comparing the reference signal with the signal of channel A. In order to eliminate this error, the terminal adjusts the measurement value to be consistent with the original reference signal. Next, the terminal performs more complex signal comparison to check the trend mismatch and phase disorder of other channels. For example, at a certain time, the signal amplitudes of channel C and channel D are 55 A and 54 A respectively, while the reference signal amplitude is 53.5 A, and the phase synchronization of the two channels is disordered. The terminal determines that there is a common node degradation in channels C and D according to the channel topology, and marks them as fault channels and includes them in the priority repair recommendation list.

[0078] After detecting the above faults, the terminal increases the fault weight of channels C and D based on the fault weight increasing mechanism, and prioritizes the repair of channels C and D. These fault weights will not only affect the generation of the repair list, but also affect the priority and sampling frequency of data upload. The fault weight of channels C and D is high, so the terminal will prioritize sampling the data of these channels and uploading them to the remote monitoring platform during subsequent data transmission. This ensures that the fault channels can be repaired and monitored in a timely manner.

[0079] Next, the terminal continuously monitors the output state of each diagnostic module and performs integral deviation analysis, and determines whether the diagnostic module has reasoning drift by defining a module logic stability index. Assuming that at a certain time, the output result of channel B is 50.2 A, and the output of channel A is 50 A, and the output fluctuation gradually increases, the terminal analyzes it by the following formula:

[0080]

[0081] wherein, α s is the output result of channel B at time point s, is the sliding mean value of the output in the window period, θ s is the signal offset strength, β is the response enhancement factor, and Δ is the integral window length. Assuming that the output fluctuation of channel B is large within the time window Δ = 5 seconds, the calculation gives:

[0082]

[0083] After calculation, if the obtained Φ(t) value exceeds the set threshold Φ th= 0.1, the system determines that channel B has potential reasoning drift and marks it as a logically unstable module. At this time, the terminal freezes the module, suspends its participation in the final state judgment, and takes its output result as a reference index for subsequent analysis and repair.

[0084] Further, the terminal also scores the drift index of the module marked as potential reasoning drift, and judges whether the module is in a deteriorating state in combination with the time trend. For example, the terminal calculates the drift index of channel B. If it is found that the drift index of the channel gradually increases in a continuous period of time, it means that the reasoning drift of the module continues to intensify, and the system triggers the software version rollback or module repair process to restore the stability of the module.

[0085] In subsequent diagnosis, the drift index of channel B rises from 0.12 to 0.15 over time, and the drift continues to intensify, so the system determines that the channel enters a deteriorating state and triggers a rollback operation to restore the previous stable software version, thereby avoiding its impact on subsequent data.

[0086] The substation is equipped with multiple measurement channels (channels A to F) responsible for real-time measurement of current, voltage, and frequency parameters of key devices such as transformers, capacitors, and current transformers within the station. Initially, the system identifies that channels B and E have both anti-interference ability degradation and physical chain transmission deterioration during the inspection period after the first load switching. After preliminary determination, they are identified as composite degradation channels, and the fault weight is significantly increased and enters the maintenance recommendation list. To prevent misjudgment, the state terminal further introduces three external factors such as the ambient temperature, humidity, and electromagnetic interference level of the physical location of the channel as dynamic adjustment parameters of the fault weight. Taking channel E as an example, the area where the channel is located is close to the high-voltage disconnector, and the on-site ambient temperature is 45°C, the humidity is 80%, and the electromagnetic interference level is level 4 (full score 5). After comprehensive analysis, the system determines that the fault of channel E is caused by external environmental extreme influence, so the original weight is corrected downward. The ambient temperature and humidity of the area where channel B is located are moderate, and the interference level is only level 1, so the fault weight remains unchanged to ensure that resources are concentrated on high-certainty abnormal channels.

[0087] After the maintenance recommendation list is generated, the terminal further evaluates whether preventive maintenance strategies need to be triggered in combination with the last maintenance time and the cumulative operating hours of the channel. The system queries the historical database and finds that channel B has accumulated more than 15,000 hours since the last maintenance, and the last maintenance has been 22 months, which exceeds the regular 12-month maintenance cycle, so it is marked as a high-priority preventive maintenance object and is pushed to the operation and maintenance scheduling system, suggesting that replacement or precision calibration should be prioritized in the next maintenance window.

[0088] As channels C and D also gradually entered the maintenance recommendation list, the terminal began to trigger the structural degradation identification mechanism, analyzed the commonality between multiple channels included in the list by constructing the mutual correlation channel failure probability matrix. The system found that channels C, D and E shared the same sampling mainboard and digital-to-analog conversion chip, and in the past three inspection cycles, the failure score curves of the three channels had high similarity, the delay time, amplitude offset and noise characteristic trend were basically synchronous, so the system marked the three channels as a structural degradation suspected group, and recommended to check whether the shared hardware had problems such as aging, capacitor drift, chip misadjustment. At the same time, the fault weight is also used to dynamically adjust the priority and sampling frequency of data upload in real time. Taking channels B, C and D as an example, because they are confirmed as high-risk channels, their sampling frequency is increased from 1 Hz to 10 Hz, and the data upload delay is controlled within 100 ms, to ensure that the data basis of the system in the judgment process is more accurate and timely.

[0089] In addition, in order to ensure the logical accuracy of the diagnosis module itself, the terminal continuously records the state consistency of the output of each module, including the fluctuation amplitude of the output value, the trend offset, and the hit frequency statistics of the key logic nodes in the internal judgment path. For example, the diagnosis module of channel D has continuously output inconsistent signal determination results in the past 24 hours, the system finds that two key nodes in its judgment path are frequently triggered, and the output result fluctuates sharply in a short time, significantly exceeding the system's set behavior stability threshold, so it is marked as potential reasoning drift. In order to exclude the influence of external factors, the terminal then checks the accuracy and stability of the input signal source (current transformer signal from channel D) it relies on. The check shows that there are short-time interference peaks in the signal source at multiple measurement points, and the sampling accuracy deviates from the normal range, so it is preliminarily determined that the root cause of the module's reasoning drift is the data chain anomaly. The system recommends to check the signal chain in stages, and includes the binding relationship between the data chain and the diagnosis module into the drift causality matrix.

[0090] If the diagnosis module is determined to be reasoning drift in two consecutive rounds of inspection, the terminal will start the logic freeze mode, suspend the module from participating in the final state judgment, and all output results will only be used for reference and will not participate in the logic decision-making process. During the freeze period, the system continuously scores the drift index of the module, and evaluates whether it is in a worsening state through time series analysis. Taking module M as an example, its drift index increased from 0.21 to 0.46 in 24 hours, and the fluctuation frequency increased, indicating that its drift trend is intensifying. The system therefore triggers the built-in recovery mechanism, rolls back to the previous stable software version, and starts self-checking, resetting its input processing logic and intermediate parameters. If the module's drift index decreases and remains stable after rolling back, the module will participate in system diagnosis again; otherwise, it will enter the degradation isolation strategy, and manual intervention is recommended to check its underlying logic or external dependencies.

[0091] The foregoing merely illustrates the principles of the application and application of its more prominent features. Those skilled in the art will appreciate that the application is not limited to the embodiments described and illustrated and that various modifications and improvements can be made thereto without departing from the spirit and scope of the application. The scope of the application is delimited by the appended claims and their equivalents.

Claims

1. A self-diagnostic method for the status terminal of a substation measurement system, characterized in that... include: The status terminal compares the signal values ​​of signals from the same physical source in each measurement channel at different sampling nodes through multiple physical paths, including sensors, amplifiers, analog-to-digital converters, buffers, and communication modules. If a trend mismatch, amplitude abrupt change, or phase disorder is detected in the signal values ​​between different paths, it is determined that the channel has physical chain degradation or poor contact. The measurement sampling process is perturbed by the clock signal controlled by the internal crystal oscillator of the state terminal, and the response of the measured value to the perturbation is monitored. If an abnormal response deviation is detected, it is determined that there is drift, timing disorder or buffer backlog in the sampling time base, so as to identify the time base health status of the internal sampling chain of the terminal. During the operation of a substation, electromagnetic interference generated by circuit breakers, transformers or other external equipment is used as a natural excitation signal. The response differences generated by this excitation in different measurement channels are collected. If response delay, amplitude distortion or jitter start offset is found, the anti-interference capability of the corresponding channel is determined to be degraded. Based on the real-time sensing of electrical energy changes by the terminal, the sensitivity threshold of the switching response is adaptively adjusted. If frequent misjudgments or sluggish responses are detected in the switching response, the terminal sampling system is marked as having abnormal behavioral sensitivity.

2. The self-diagnosis method for the status terminal of a substation measurement system according to claim 1, characterized in that... When comparing the signal values, the original reference signal from the primary side of the transformer is introduced as the comparison benchmark to eliminate the gain error introduced by the analog link transmission. When the trends of multiple measurement channels are simultaneously mismatched and the phases are disordered, the terminal determines that the common node is degraded based on the channel topology.

3. The self-diagnosis method for the status terminal of the substation measurement system according to claim 2, characterized in that... The perturbation modulation includes perturbation segmentation tests implemented at different amplitude levels, and the load jitter sensitivity of the sampling chain is determined by extracting the nonlinear response threshold point position; when a backlog trend is identified in the sampling buffer, the terminal triggers a buffer refresh and flow control mechanism to restore the data transmission rhythm and prevent accidental triggering of anomalies.

4. The self-diagnosis method for the status terminal of a substation measurement system according to claim 3, characterized in that... The response difference is based on amplitude offset and changes in the frequency of the initial jitter noise of the response signal, and is used to identify filter performance degradation in the channel; If multiple channels respond asynchronously to the same interference event, the terminal compares the time difference of the response start points to detect the differences in the common-mode interference immunity of each measurement module.

5. The self-diagnosis method for the status terminal of a substation measurement system according to claim 4, characterized in that... The terminal performs multi-event clustering analysis on the jump misjudgment records. If a certain type of misjudgment occurs repeatedly in different consecutive time periods, it is marked as a structural anomaly. When the sensitivity threshold of the jump response is adaptively adjusted, the current load fluctuation level of the measurement channel is used as the adjustment factor.

6. The self-diagnosis method for the status terminal of a substation measurement system according to claim 5, characterized in that... When the measurement channel is determined to have overlapping results of anti-interference degradation and physical chain degradation, the terminal increases the fault weight of the channel and includes it in the maintenance suggestion list; the terminal establishes a state consistency record table for each diagnostic module. When the output result of a certain module fluctuates continuously beyond the threshold, it is judged as a decrease in logic stability and marked as potential inference drift.

7. The self-diagnosis method for the status terminal of a substation measurement system according to claim 6, characterized in that... While increasing the fault weight, the terminal introduces the ambient temperature, humidity, and electromagnetic interference level corresponding to the physical location of the measurement channel as adjustment factors to correct the fault weight; each channel in the maintenance suggestion list is also marked with its most recent maintenance cycle and cumulative operating hours to help determine whether preventive maintenance is triggered.

8. The self-diagnostic method for the status terminal of a substation measurement system according to claim 7, characterized in that... When multiple channels are included in the maintenance recommendation list at the same time, the terminal uses the cross-correlation channel fault probability matrix to determine whether the structural degradation is due to a common motherboard, sampling chip, or regional module; the fault weights are used to generate the list and adjust the priority order and sampling frequency of the channels in data upload.

9. The self-diagnosis method for the status terminal of a substation measurement system according to claim 8, characterized in that... The state consistency record table records the amplitude of the output fluctuation of the diagnostic module and the number of condition hits of key nodes in the judgment path; Once the diagnostic module is flagged as a potential inference drift, the terminal verifies the sampling accuracy and stability of the dependent input signal source channels to determine whether the drift is caused by an external data link anomaly.

10. The self-diagnosis method for the status terminal of a substation measurement system according to claim 9, characterized in that... If the status terminal marks the same module as inference drift in two consecutive self-diagnoses, it enters the logic freeze mode, suspends the module from participating in the final status judgment, and uses the result as a reference indicator. The terminal scores all modules marked as potential inference drift with a drift index and judges whether they are in a deteriorating state in combination with the time trend, so as to trigger software version rollback or repair.

Images