An excitation system working condition evaluation method and device based on simulation modeling and a medium

Abstract

The application discloses an excitation system working condition evaluation method and device based on simulation modeling, and a medium, relates to the twin simulation technical field, and comprises the following steps: collecting real-time operation data of the excitation system, unifying time benchmarks, and marking state sources to obtain a real-time state frame; grouping the real-time state frame, performing frequency measurement, correcting a sampling area, and switching data sources while retaining sampling credibility and state source marks to obtain standardized working condition state quantities; dividing nodes of regulating calculation states, limit auxiliary ring states, pulse readback states, and cabinet flow response states in the standardized working condition state quantities, configuring state transmission relationships and execution feedback relationships, and constructing a working condition causal twin simulation model. The application constructs pulse flow normal fingerprints corresponding to working condition boundary labels, realizes stable correlation expression of the pulse readback states and the cabinet flow response states under specific operation working conditions, and enables the twin simulation to obtain a standardized reference benchmark for deviation identification.

Description

Technical Field

[0001] This invention relates to the field of twin simulation technology, and in particular to a method, equipment and medium for evaluating the operating conditions of an excitation system based on simulation modeling. Background Technology

[0002] With the increase in generator capacity and the strengthening of power dynamic regulation requirements, excitation systems are gradually evolving from single regulation and control to multi-state collaborative sensing, online modeling, and operating condition assessment. Within the framework of computer-aided design, simulation modeling, and simulation verification technologies, twin simulation is widely used for equipment state mapping, operation process reproduction, and model consistency analysis. For excitation systems, technologies such as real-time sampling, clock synchronization, signal grouping, state modeling, and simulation deduction are gradually being combined, providing a data and model foundation for the dynamic relationship analysis between the excitation regulation process, the limiting auxiliary loop action, pulse execution feedback, and cabinet current response.

[0003] Existing excitation system condition assessments typically focus on judging operating parameters exceeding limits, matching fault characteristics, or comparing single simulation results. They lack a unified model for the chain-like transmission relationship between the adjustment calculation state, the limiting auxiliary loop state, the pulse readback state, and the cabinet current response state. Especially when there are time delays, condition boundary constraints, and feedback coupling between the pulse readback and the cabinet current response, existing methods struggle to accurately distinguish between normal operating condition changes and execution feedback deviations. They also struggle to further determine the risk contribution sources of deviations in the state transmission relationship and the execution feedback relationship, thus limiting the interpretability of the condition assessment results and the model's self-verification capability. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a simulation modeling-based method for evaluating the operating conditions of an excitation system to solve the problem that it is difficult to conduct causal evaluation of the deviation between pulse readback and cabinet current response in the excitation system under operating condition constraints.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, this invention provides a simulation modeling-based method for evaluating the operating conditions of an excitation system, comprising: collecting real-time operating data of the excitation system and unifying the time base and marking the state source to obtain a real-time state frame; performing grouped frequency measurement, sampling area correction, and data source switching on the real-time state frame, while retaining the sampling reliability and state source marking, to obtain standardized operating condition state quantities; dividing the adjustment calculation state, constraint auxiliary loop state, pulse readback state, and cabinet current response state in the standardized operating condition state quantities into nodes, configuring state transmission relationships and execution feedback relationships, and constructing a causal twin simulation model of the operating condition; and based on the causal twin simulation model of the operating condition, evaluating the adjustment calculation state and constraint auxiliary loop state... Operating condition boundary labels are generated, and a pulse flow normal fingerprint is constructed based on the stable correspondence between pulse readback state segments and cabinet flow response state segments under the operating condition boundary labels. The pulse readback state and cabinet flow response state in the current standardized operating condition state quantities are aligned with the matched pulse flow normal fingerprint under constrained timing to form a pulse flow deviation state. The operating condition is deduced based on the pulse flow deviation state-driven causal twin simulation model, and the risk contribution of the pulse flow deviation state in the state transmission relationship and execution feedback relationship is calculated to generate the operating condition risk contribution result. The current operating condition of the excitation system is simulated and verified, and the triggering cause, deviation source and simulation deduction basis are associated to generate an excitation system operating condition simulation evaluation report.

[0007] As a preferred embodiment of the excitation system condition evaluation method based on simulation modeling described in this invention, the steps for obtaining the real-time state frame are as follows: Real-time operating data of the excitation system is collected according to the preset framing cycle, and the local sampling time, status source marker and channel type marker are added to obtain the original sampling record; Based on the original sampling records, clock offset correction and clock drift correction are performed on the local sampling time using a unified clock. Sampling confidence is generated based on the clock correction results, state source marker, and channel type marker. Sampling records carrying sampling confidence, state source marker, and channel type marker are merged according to a preset framing period to obtain a real-time state frame.

[0008] As a preferred embodiment of the excitation system operating condition evaluation method based on simulation modeling described in this invention, the steps for obtaining the standardized operating condition state quantities are as follows: The excitation signal is grouped based on the channel type marker carried by each sampling record in the real-time status frame, and real-time frequency measurement and sampling area correction are performed on the grouped excitation signal. At the same time, the sampling reliability and status source marker in the real-time status frame are inherited to obtain the grouped correction status quantity. Based on the grouped correction state variables, the reliability of the original sampling area, the correction sampling area, and the short-term holding sampling area is compared, and the data source is switched according to the reliability comparison results to obtain the standardized operating condition state variables.

[0009] As a preferred embodiment of the excitation system condition evaluation method based on simulation modeling described in this invention, the steps for constructing the causal twin simulation model of the operating condition are as follows: The regulation calculation state, the limiting auxiliary loop state, the pulse readback state, and the cabinet flow response state are extracted from the standardized operating condition state variables and encapsulated into corresponding nodes to obtain a node encapsulation set. Based on the node encapsulation set, the adjustment calculation state node is mapped to the adjustment output port, the constraint auxiliary loop state node is mapped to the boundary constraint port, the pulse readback state node is mapped to the execution readback port, and the cabinet flow response state node is mapped to the response feedback port. According to the transmission rules from the adjustment output port to the boundary constraint port, from the adjustment output port to the execution readback port, from the boundary constraint port to the execution readback port, from the execution readback port to the response feedback port, from the response feedback port to the adjustment output port, and from the response feedback port to the boundary constraint port, a semantic contract table is generated. Candidate node relationships are generated based on the semantic contract table, and the causal interaction strength of each candidate node relationship is calculated. Candidate node relationships whose causal interaction strength meets the preset relationship threshold are retained as valid node relationships. Based on the valid node relationships, the state transmission relationship and execution feedback relationship between the adjustment calculation state node, the constraint auxiliary loop state node, the pulse readback state node and the cabinet flow response state node are configured and adjusted. The node connection direction and simulation call order are determined, and the working condition causal twin simulation model is constructed.

[0010] As a preferred embodiment of the excitation system condition evaluation method based on simulation modeling described in this invention, the steps for constructing the pulse current normal fingerprint are as follows: Based on the working condition causal twin simulation model, the execution feedback relationship between the pulse readback state node and the cabinet flow response state node is read, and the pulse readback state segment and cabinet flow response state segment within the same event window are extracted using the effective change time of the pulse readback state as the event starting point to obtain the pulse flow event segment. Read the adjustment calculation state and the constraint auxiliary loop state within the same event window from the pulse event fragment, and combine the output change range of the adjustment calculation state and the boundary action range of the constraint auxiliary loop state into the operating condition boundary label to generate pulse event samples; Historical pulse flow event samples were selected based on the following criteria: sampling confidence met a preset confidence threshold; pulse readback state change amplitude did not exceed a preset pulse mutation threshold; cabinet flow response state change amplitude did not exceed a preset cabinet flow fluctuation threshold; and neither pulse readback state segment nor cabinet flow response state segment was missing. The historical pulse flow event samples were grouped according to the working condition boundary label. Stable corresponding segments were determined based on the stable correspondence between pulse readback state segments and cabinet flow response state segments within the same group. The stable corresponding segments were aligned according to the relative framing position after the event start point. The median values ​​of pulse readback state values ​​and cabinet flow response state values ​​at the same relative framing position were extracted to generate standard pulse readback state sequences and standard cabinet flow response state sequences, thus constructing a normal pulse flow fingerprint.

[0011] As a preferred embodiment of the excitation system condition evaluation method based on simulation modeling described in this invention, the steps for forming the pulse current deviation state are as follows: Based on the pulse flow normal fingerprint, the pulse readback status and cabinet flow response status in the current standardized operating condition status are extracted, and the pulse flow normal fingerprint is matched according to the corresponding operating condition boundary label to obtain the pulse flow segment to be verified. The pulse flow segment to be verified is aligned with the matched normal pulse flow fingerprint under constrained timing, and the pulse flow deviation of the pulse flow segment to be verified relative to the normal pulse flow fingerprint is calculated. The deviation levels of pulse readback status and cabinet flow response status are marked based on the pulse flow deviation, and the pulse flow deviation, constrained timing alignment results and operating condition boundary labels are written into the execution feedback relationship to form the pulse flow deviation status.

[0012] As a preferred embodiment of the excitation system condition assessment method based on simulation modeling described in this invention, the steps for generating the condition risk contribution result are as follows: Based on the pulse deviation state, the time, direction and intensity of the deviation are determined and encapsulated as a deviation injection event. The deviation injection event is written into the execution feedback relationship in the causal twin simulation model of the working condition to obtain the deviation-driven deduction conditions. Based on the deviation-driven inference condition-driven causal twin simulation model, factual and counterfactual inferences are performed. The risk contribution is calculated based on the differences between the factual and counterfactual inference trajectories in the state transmission relationship and execution feedback relationship, and the working condition risk contribution result is generated.

[0013] As a preferred embodiment of the excitation system operating condition evaluation method based on simulation modeling described in this invention, the steps for generating the excitation system operating condition simulation evaluation report are as follows: Extract the corresponding state transmission relationship, execution feedback relationship, factual deduction trajectory and counterfactual deduction trajectory from the working condition risk contribution results, determine the execution feedback relationship corresponding to the risk contribution amount as the triggering relationship, determine the pulse flow deviation state associated with the triggering relationship as the triggering cause, and determine the deviation position between the pulse readback state and the cabinet flow response state in the pulse flow deviation state as the deviation source, and obtain the simulation verification association results; Based on the simulation verification correlation results, the factual inference trajectory, counterfactual inference trajectory, standardized operating condition state variables, triggering causes and deviation sources are correlated and compared within the same time window to generate an excitation system operating condition simulation evaluation report.

[0014] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the excitation system condition evaluation method based on simulation modeling as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the excitation system condition evaluation method based on simulation modeling as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: By constructing a pulse flow normal fingerprint corresponding to the operating condition boundary label, a stable correlation expression between the pulse readback state and the cabinet flow response state under specific operating conditions is realized, enabling twin simulation to obtain a standardized reference benchmark for deviation identification; it can reflect the response characteristics under different regulation outputs and limiting auxiliary loop conditions, so that the formation of pulse flow deviation state has clear operating condition constraints and timing basis, and provides a traceable data foundation for subsequent risk contribution calculation, simulation verification, and generation of excitation system operating condition simulation evaluation report, thereby improving the interpretability, stability, and verifiability of the excitation system operating condition evaluation results. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a simulation-based modeling method for evaluating the operating conditions of an excitation system.

[0019] Figure 2 The flowchart for constructing a causal twin simulation model for operating conditions.

[0020] Figure 3 This is a flowchart showing the normal and deviated states of pulse flow fingerprints.

[0021] Figure 4 This is a flowchart for generating a simulation evaluation report of the excitation system's operating conditions.

[0022] Figure 5 This is a data chart comparing the factual projection trajectory and the counterfactual projection trajectory within the same time window.

[0023] Figure 6 Heatmap of risk contribution to state transit relationships and execution feedback relationships. Detailed Implementation

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0027] Reference Figures 1-6 As one embodiment of the present invention, this embodiment provides a method for evaluating the operating conditions of an excitation system based on simulation modeling, comprising the following steps: S1. Collect real-time operating data of the excitation system and unify the time base and mark the state source to obtain the real-time state frame. Perform grouped frequency measurement, sampling area correction and data source switching on the real-time state frame, while retaining the sampling reliability and state source mark to obtain the standardized operating condition state quantity.

[0028] S1.1: Collect real-time operating data of the excitation system according to the preset framing cycle, and add local sampling time, status source mark and channel type mark to obtain the original sampling record; It should be noted that the real-time operating data of the excitation system includes the adjustment calculation status, the limiting auxiliary loop status, the pulse readback status, and the cabinet current response status; The adjustment calculation state is derived from the real-time calculation results of the excitation regulator on the terminal voltage, excitation current, setpoint and adjustment parameters. Its function is to characterize the current voltage regulation control output of the excitation system and the basis for subsequent pulse control. The state of the auxiliary loop is derived from the calculation results of the limiting protection links such as under-excitation limit, over-excitation limit, V / F limit and stator overcurrent limit. Its function is to characterize whether the excitation system is close to or enters the restricted operating boundary. The pulse readback status originates from the readback detection result of the actual output trigger pulse by the communication pulse module, and its function is to characterize whether the pulse control command is executed accurately. The cabinet current response status is derived from the current sampling results of the rectifier cabinet or power cabinet, and its function is to characterize the current distribution of each cabinet and the current sharing response between cabinets after the pulse is executed; The status source marker is used to indicate which of the following statuses the original sampled value comes from: the adjustment calculation status, the limiting auxiliary loop status, the pulse readback status, and the cabinet flow response status; The channel type label is used to indicate the data type of the original sampled value, such as voltage, current, angle, switching, or computational quantity.

[0029] Specifically, according to the preset framing period as the acquisition trigger interval, when each framing period arrives, the current original sampled values ​​of the adjustment calculation status, the limiting auxiliary loop status, the pulse readback status, and the cabinet flow response status are read. The reading time is recorded as the local sampling time, the status source mark is recorded according to the location where the original sampled value is generated, and the channel type mark is recorded according to the data type of the original sampled value. The original sampled value, the local sampling time, the status source mark, and the channel type mark are bound and saved to obtain the original sampling record.

[0030] It should be noted that the framing period is set based on the excitation regulation control calculation period and the pulse readback update period. The example value range is from one millisecond to five milliseconds. The value is determined by the shortest effective update interval that can cover the regulation calculation state, the limiting auxiliary loop state, the pulse readback state, and the cabinet current response state.

[0031] S1.2: Based on the original sampling records, clock offset correction and clock drift correction are performed on the local sampling time using a unified clock. Sampling confidence is generated according to the clock correction results, state source marker and channel type marker. Sampling records carrying sampling confidence, state source marker and channel type marker are merged according to the preset framing period to obtain the real-time state frame.

[0032] Specifically, based on the original sampling records, the midpoint of the framing period is used as the corresponding time of the unified clock. Local sampling times are read one by one, and the time difference between the local sampling time and the midpoint of the framing period is calculated. When the time difference corresponding to three consecutive framing periods under the same state source marker does not exceed a preset offset threshold, the time difference is used as a clock offset to shift and correct the local sampling time. When the change in time difference for three consecutive framing periods under the same state source marker exceeds a preset drift threshold and the direction of change is consistent, the local sampling time is drift compensated and corrected according to the direction of time difference change. After correction, if the corrected sampling time is within the center range of the current framing period, it is marked as high sampling time fit; if the corrected sampling time is within the current framing period but outside the center range of the current framing period, it is marked as medium sampling time fit. If the time exceeds the current framing period or the clock correction exceeds the preset correction limit, it is marked as low sampling time fit. Based on the preset mapping relationship between the state source marker and the channel type marker, the source clarity and data type matching are determined. High sampling confidence is generated when the sampling time fit is high, the source is clear, and the data type matches; medium sampling confidence is generated when the sampling time fit is medium, the source is clear, and the data type matches; otherwise, low sampling confidence is generated. Sampling records that have completed clock correction and carry sampling confidence, state source marker, and channel type marker are merged according to the framing period. If multiple sampling records with the same state exist within the same framing period, the sampling record with the highest sampling confidence is retained. If the sampling confidence is the same, the sampling record with the smallest absolute time difference between the corrected sampling time and the midpoint of the framing period is retained to obtain the real-time state frame.

[0033] It should be noted that the offset threshold is set based on the allowable fixed time difference between the local sampling time and the midpoint of the framing period under the same state source mark. The example value range is one-tenth to one-fifth of the framing period. The value is determined based on the alignment of the adjustment calculation state, the limiting auxiliary loop state, the pulse readback state, and the cabinet flow response state within the same real-time state frame without affecting the adjustment calculation state, the limiting auxiliary loop state, the pulse readback state, and the cabinet flow response state. The drift threshold is set based on the allowable variation of the time difference between local sampling moments between consecutive frame cycles. The example value range is from one-twentieth to one-tenth of the frame cycle. The value is determined by the ability to distinguish between normal sampling jitter and persistent clock drift. The center range of the framing period is set based on the degree of fit between the real-time intra-frame sampling record and the preset midpoint of the framing period. The example value range is one-fifth of the period before and after the midpoint of the framing period. The value is determined to ensure that the merged sampling record can represent the state of the current framing period. The upper limit of correction is set based on the maximum allowable correction magnitude of clock offset correction and clock drift correction. The example value range is one-third to one-half of the framing period. The basis for the value is to avoid over-correction that would cause the sampled records to be incorrectly merged into adjacent real-time status frames.

[0034] S1.3: Based on the channel type marker carried by each sampling record in the real-time status frame, the excitation signal is grouped, and real-time frequency measurement and sampling area correction are performed on the grouped excitation signal. At the same time, the sampling reliability and status source marker in the real-time status frame are inherited to obtain the grouped correction status quantity. Specifically, based on the real-time status frame, the channel type marker carried by each sampling record is read, and sampling records with the same channel type marker are grouped into the same excitation signal group. For excitation signal groups with AC variation characteristics, the real-time period is determined by the interval between adjacent zero-crossing points in the same direction. The effective sampling area within the current framing period is re-determined according to the real-time period. Sampling records that deviate from the effective sampling area are adjusted to adjacent effective sampling positions. At the same time, the original sampling reliability and status source marker of the sampling record are retained to obtain the group correction status quantity.

[0035] It should be noted that the effective sampling area refers to the sampling time range that can represent the current change state of the excitation signal, determined within the current framing period based on the real-time period of the excitation signal grouping.

[0036] S1.4: Based on the grouped correction state variables, the reliability of the original sampling area, the correction sampling area, and the short-term holding sampling area is compared, and the data source is switched according to the reliability comparison results to obtain the standardized operating condition state variables.

[0037] Specifically, based on the grouped correction state variables, the sampling records that have not been corrected by the sampling area are taken as the original sampling area, the sampling records that have completed real-time frequency measurement and sampling area correction are taken as the correction sampling area, and the sampling records that were not marked as low sampling confidence in the previous framing period and have complete state source marking and consistent channel type marking are continued to form a short-term holding sampling area. The sampling confidence, state source marking completeness and channel type marking consistency of the original sampling area, correction sampling area and short-term holding sampling area are compared respectively. The sampling area with the highest sampling confidence and complete state source marking and consistent channel type marking is selected as the current data source. When the sampling confidence is the same, the correction sampling area that has completed the sampling area correction is selected. The corresponding state values ​​are organized according to the selected current data source to obtain the standardized operating condition state variables.

[0038] S2. Node division is performed on the adjustment calculation state, constraint auxiliary loop state, pulse readback state, and cabinet flow response state in the standardized operating condition state variables. State transmission relationships and execution feedback relationships are configured to construct a causal twin simulation model of the operating condition.

[0039] S2.1: Extract the adjustment calculation state, the limiting auxiliary loop state, the pulse readback state, and the cabinet flow response state from the standardized operating condition state variables, and encapsulate them into corresponding nodes to obtain the node encapsulation set; Specifically, the states are categorized from the standardized operating condition state quantities according to the state source marker and channel type marker. The state values, sampling reliability, and state source markers corresponding to the adjustment calculation state are selected and encapsulated as adjustment calculation state nodes. The state values, sampling reliability, and state source markers corresponding to the restriction auxiliary loop state are selected and encapsulated as restriction auxiliary loop state nodes. The state values, sampling reliability, and state source markers corresponding to the pulse readback state are selected and encapsulated as pulse readback state nodes. The state values, sampling reliability, and state source markers corresponding to the cabinet flow response state are selected and encapsulated as cabinet flow response state nodes. The adjustment calculation state nodes, restriction auxiliary loop state nodes, pulse readback state nodes, and cabinet flow response state nodes are saved in the same frame order to obtain a node encapsulation set.

[0040] S2.2: Based on the node encapsulation set, the adjustment calculation state node is mapped to the adjustment output port, the constraint auxiliary loop state node is mapped to the boundary constraint port, the pulse readback state node is mapped to the execution readback port, and the cabinet flow response state node is mapped to the response feedback port. According to the transmission rules from the adjustment output port to the boundary constraint port, from the adjustment output port to the execution readback port, from the boundary constraint port to the execution readback port, from the execution readback port to the response feedback port, from the response feedback port to the adjustment output port, and from the response feedback port to the boundary constraint port, a semantic contract table is generated. Specifically, based on the node encapsulation set, the status source flag and channel type flag carried by the adjustment calculation status node, the limiting auxiliary loop status node, the pulse readback status node and the cabinet flow response status node are read, and the correspondence between the status type represented by the status source flag and the data type represented by the channel type flag is checked. When the status source flag carried by the adjustment calculation status node corresponds to the adjustment calculation status and the channel type flag corresponds to the calculation quantity, the adjustment calculation status node is confirmed to be valid, and the node name, status value, sampling confidence level and framing order of the adjustment calculation status node are registered to the adjustment output port. When the state source flag carried by the restricted auxiliary loop state node corresponds to the restricted auxiliary loop state and the channel type flag corresponds to the computational quantity or switching quantity, the restricted auxiliary loop state node is confirmed to be valid, and the node name, state value, sampling confidence level and framing order of the restricted auxiliary loop state node are registered to the boundary constraint port. When the status source flag carried by the pulse readback status node corresponds to the pulse readback status and the channel type flag corresponds to the switch quantity or angle quantity, the pulse readback status node is confirmed to be valid, and the node name, status value, sampling confidence level and framing order of the pulse readback status node are registered to the execution readback port. When the status source marker of the cabinet current response status node corresponds to the cabinet current response status and the channel type marker corresponds to the current quantity, the cabinet current response status node is confirmed to be valid, and the node name, status value, sampling confidence level and framing order of the cabinet current response status node are registered to the response feedback port. After port registration is completed, the semantic contract table is generated by recording the port role, transmission direction, transmission content, and feedback content according to the link relationship between the regulating output port and the boundary constraint port for transmitting the working condition boundary, the regulating output port and the execution readback port for transmitting the execution basis, the boundary constraint port and the execution readback port for transmitting the execution condition after constraint, the execution readback port and the response feedback port for transmitting the pulse execution result, and the response feedback port and the regulating output port and the boundary constraint port for feeding back the cabinet flow change result.

[0041] S2.3: Generate candidate node relationships based on the semantic contract table, calculate the causal interaction strength of each candidate node relationship, retain candidate node relationships whose causal interaction strength meets the preset relationship threshold as valid node relationships, configure and adjust the state transmission relationship and execution feedback relationship between the calculation state node, the constraint auxiliary loop state node, the pulse readback state node and the cabinet flow response state node based on the valid node relationship, determine the node connection direction and simulation call order, and construct the working condition causal twin simulation model.

[0042] Specifically, based on the semantic contract table, the starting port, target port, transmission direction, and allowable lag range of each record are read. Nodes registered at the starting port are paired with nodes registered at the target port to obtain candidate node relationships. The starting node state value, target node state value, and sampling confidence are read from the continuous framed sequence corresponding to the candidate node relationship. Sampling records corresponding to low sampling confidence are removed. Within the current calculation window, the changes in the starting node state value and the target node state value are normalized according to their corresponding value spans, so that the data corresponding to different channel type labels are transformed into dimensionless changes. Within the allowable lag range defined by the semantic contract table, the matching degree of the starting node state value change preceding the target node state value change is calculated using lag-normalized cross-correlation, and the maximum matching degree is taken as the causal interaction strength of the candidate node relationship. When the causal interaction strength meets the preset relationship threshold, the candidate node relationship is retained as a valid node relationship. Based on the transmission direction corresponding to the effective node relationship, the adjustment output port to the boundary constraint port, the adjustment output port to the execution readback port, and the boundary constraint port to the execution readback port are configured as state transmission relationships, and the execution readback port to the response feedback port, the response feedback port to the adjustment output port, and the response feedback port to the boundary constraint port are configured as execution feedback relationships. The forward connection directions of the adjustment calculation state node, the constraint auxiliary loop state node, the pulse readback state node, and the cabinet flow response state node are determined according to the state transfer relationship, and the execution feedback relationship is registered as the feedback correction connection direction for the next framing cycle. The simulation call order of the current framing cycle is determined according to the order of the adjustment calculation state node, the constraint auxiliary loop state node, the pulse readback state node, and the cabinet flow response state node, and the feedback call order of the next framing cycle is determined according to the order in which the cabinet flow response state node feeds back to the adjustment calculation state node and the constraint auxiliary loop state node. Each valid node relationship is registered as a state update rule. The dimensionless state change of the starting node within the allowable lag range is used as the forward input of the target node. The correction magnitude of the forward input on the standardized state value of the target node is determined according to the causal interaction strength of the corresponding valid node relationship. The current standardized state value of the target node is updated according to the correction magnitude to obtain the standardized state value of the target node in the next frame cycle. Among them, the adjustment calculation state node is updated according to the execution feedback relationship from the response feedback port to the adjustment output port; the constraint auxiliary loop state node is updated according to the state transmission relationship from the adjustment output port to the boundary constraint port and the execution feedback relationship from the response feedback port to the boundary constraint port; the pulse readback state node is updated according to the state transmission relationship from the adjustment output port to the execution readback port and the boundary constraint port to the execution readback port; and the cabinet flow response state node is updated according to the execution feedback relationship from the execution readback port to the response feedback port. When the updated standardized state value exceeds the standardized value range of the corresponding state, the updated standardized state value is corrected to the boundary of the standardized value range of the corresponding state, and a causal twin simulation model of the working condition is constructed.

[0043] It should be noted that the allowable hysteresis range is set based on the actual response sequence between the adjustment calculation state, the constraint auxiliary loop state, the pulse readback state, and the cabinet flow response state. The example value range is one to five framing cycles, and the value is based on the maximum normal propagation delay between the generation of the coverage adjustment output, the constraint boundary constraint, the pulse readback update, and the cabinet flow response change. The relationship threshold is set based on the distribution of causal interaction intensity of candidate node relationships under historical normal operating conditions. The example value range is 0.6 to 0.8. The basis for the value is to retain candidate node relationships that are stable and reflect the change of the state of the starting node leading the change of the state of the target node, and to remove accidental correlation relationships.

[0044] The expression for generating candidate node relationships is: ; In the formula, Represents the set of candidate node relationships; Indicates the starting node; Indicates the target node; Indicates starting node Point to target node The node relationships; Indicates the starting node The starting port for registration; Represents the target node The target port to be registered; The semantic contract table is represented by a starting node. Point to target node The allowable lag range; Represents a semantic contract table.

[0045] It should be noted that the current calculation window refers to the continuous frame interval in the continuous frame sequence corresponding to the candidate node relationship, where neither the starting node nor the target node is marked as having low sampling confidence. When the length of the continuous frame interval exceeds the preset window length, the current calculation window is truncated according to the preset window length. The preset window length is set based on the allowable hysteresis range and the normal response duration. The example value is the upper limit of the allowable hysteresis range plus five to twenty frame cycles.

[0046] The expression for calculating the normalized state change is: ; ; In the formula, Indicates the starting node In the State values ​​under each framed sequence; Indicates the frame sequence number; Indicates the first The previous frame order of the current frame order; Represents the target node In the State values ​​under each framed sequence; and Representing the starting node The maximum and minimum state values ​​within the current calculation window are extracted from the current calculation window according to the continuous frame sequence corresponding to the candidate node relationship. and Representing the target node respectively The maximum and minimum state values ​​within the current calculation window; Indicates the starting node In the The dimensionless state change after normalization of the value span under the framed sequence; Represents the target node In the The dimensionless state change after normalization of the value span under each frame sequence; all are dimensionless change quantities. When the corresponding value span is zero, the causal interaction strength of the candidate node relationship is recorded as zero.

[0047] The expression for calculating the causal interaction strength of candidate node relationships is: ; In the formula, Indicates starting node Point to target node The strength of causal interaction; Indicates the starting node State value change leading the target node The number of frames with lag in state value changes; Indicates the starting node Point to target node And the number of delayed frames is At that time, the starting node and target node None of them were marked as a set of framed sequences with low sampling confidence; Indicates the starting node In the Dimensionless state changes in a framed sequence.

[0048] The expression for determining valid node relationships is: ; In the formula, Represents the set of valid node relationships; This indicates the preset relationship threshold.

[0049] S3. Based on the causal twin simulation model of the working condition, the working condition boundary label is formed by adjusting the calculation state and restricting the auxiliary loop state. Based on the stable correspondence between the pulse readback state segment and the cabinet flow response state segment under the working condition boundary label, the pulse flow normal fingerprint is constructed. The pulse readback state and cabinet flow response state in the current standardized working condition state quantity are aligned with the matched pulse flow normal fingerprint in a constrained time sequence to form the pulse flow deviation state.

[0050] S3.1: Based on the working condition causal twin simulation model, read the execution feedback relationship between the pulse readback state node and the cabinet flow response state node, and use the effective change time of the pulse readback state as the event starting point to extract the pulse readback state segment and the cabinet flow response state segment within the same event window to obtain the pulse flow event segment. Specifically, based on the causal twin simulation model of the operating condition, the execution feedback relationship between the pulse readback state node and the cabinet flow response state node is read, and the feedback connection direction and allowable hysteresis range are obtained from the execution feedback relationship. The pulse readback state is checked in the framed order in the standardized operating condition state variables. When the state value of the pulse readback state changes in the adjacent framed period and the corresponding sampling confidence is not low sampling confidence, the framed time of the change is determined as the valid change time. The valid change time is used as the event start point, and the event window end point is determined according to the allowable hysteresis range in the execution feedback relationship. The pulse readback state segment is intercepted between the event start point and the event window end point, and the cabinet flow response state segment is intercepted in the same event window. The pulse readback state segment and the cabinet flow response state segment are bound together to obtain the pulse flow event segment.

[0051] S3.2: Read the adjustment calculation state and the constraint auxiliary loop state within the same event window from the pulse event segment, and combine the output change range of the adjustment calculation state and the boundary action range of the constraint auxiliary loop state into a working condition boundary label to generate a pulse event sample; Specifically, based on the event start point and event window end point recorded in the pulse flow event segment, the adjustment calculation state and constraint auxiliary loop state within the same event window are read from the standardized operating condition state variables. The initial state value, ending state value, and change direction of the adjustment calculation state between the event start point and the event window end point are collected into the output change interval. The constraint boundary, action state, and range of action of the constraint auxiliary loop state within the same event window are collected into the boundary action interval. The output change interval and the boundary action interval are bound together according to the same event window to form the operating condition boundary label. The operating condition boundary label is bound to the pulse flow event segment and saved to generate a pulse flow event sample.

[0052] S3.3: Based on the pulse flow event samples, historical pulse flow event samples are selected that meet the preset confidence threshold, the pulse readback state change amplitude does not exceed the preset pulse mutation threshold, the cabinet flow response state change amplitude does not exceed the preset cabinet flow fluctuation threshold, and neither the pulse readback state segment nor the cabinet flow response state segment is missing. The historical pulse flow event samples are grouped according to the working condition boundary label. Stable corresponding segments are determined based on the stable correspondence between the pulse readback state segment and the cabinet flow response state segment within the same group. The stable corresponding segments are aligned according to the relative framing position after the event start point. The median values ​​of the pulse readback state value and the cabinet flow response state value at the same relative framing position are extracted to generate a standard pulse readback state sequence and a standard cabinet flow response state sequence, thus constructing a normal pulse flow fingerprint.

[0053] Specifically, based on the operating condition boundary labels in the pulse flow event samples, sample records are selected from historical pulse flow event samples that meet the preset confidence threshold, whose pulse readback state change amplitude does not exceed the preset pulse mutation threshold, whose cabinet flow response state change amplitude does not exceed the preset cabinet flow fluctuation threshold, and whose pulse readback state segments and cabinet flow response state segments are not missing. Historical pulse flow event samples with consistent operating condition boundary labels are grouped into the same group. Within the same group, the pulse readback state segments and cabinet flow response state segments in each historical pulse flow event sample are compared in frame order, retaining those with consistent pulse readback state change directions. Segments whose cabinet flow response states change in the same direction, whose sequential relationship conforms to the state transit relationship, and whose response interval is within the allowable lag range are identified as stable corresponding segments. Taking the effective change moment of the pulse readback state as a unified starting point, the stable corresponding segments are converted into a relative framed position sequence after the event starting point. The median values ​​of the pulse readback state value and the cabinet flow response state value at the same relative framed position are extracted to generate a standard pulse readback state sequence and a standard cabinet flow response state sequence. The standard pulse readback state sequence, the standard cabinet flow response state sequence, and the corresponding operating condition boundary labels are bound and saved to construct a pulse flow normal fingerprint.

[0054] It should be noted that the confidence threshold is set based on the normal distribution of sampling confidence in historical pulse flow event samples and the data reliability required to construct normal pulse flow fingerprints. The example value range is above medium sampling confidence or quantization confidence not less than 0.7. The basis for the value is to remove low sampling confidence samples and retain historical pulse flow event samples that can stably represent the correspondence between pulse readback state and cabinet flow response state. The pulse mutation threshold is set based on the distribution of adjacent frame variation amplitudes of historical normal pulse readback state segments, and the cabinet flow fluctuation threshold is set based on the distribution of adjacent frame variation amplitudes of historical normal cabinet flow response state segments. Example values ​​can be the sum of the mean of the corresponding historical normal variation amplitude and three times the standard deviation, respectively. The value is determined by removing historical pulse flow event samples that have mutations, missing values, or significantly deviated from the normal response range. The standard pulse readback state sequence and the standard cabinet flow response state sequence are used as standard references for constrained timing alignment and pulse flow deviation calculation of subsequent pulse flow segments to be verified. The median value is used to reduce the impact of random fluctuations of individual historical pulse flow event samples on the normal fingerprint of pulse flow.

[0055] S3.4: Based on the pulse flow normal fingerprint, the pulse readback state and cabinet flow response state in the current standardized operating condition state quantity are extracted as events, and the pulse flow normal fingerprint is matched according to the corresponding operating condition boundary label to obtain the pulse flow segment to be verified. Specifically, based on the pulse flow normal fingerprint, the pulse readback state and cabinet flow response state are read from the current standardized operating condition state variables. The event starting point is determined according to the effective change time of the pulse readback state, and the current pulse readback state segment and the current cabinet flow response state segment are extracted according to the event window corresponding to the pulse flow normal fingerprint. The adjustment calculation state and the constraint auxiliary loop state within the same event window are read from the current standardized operating condition state variables to form the current operating condition boundary label. The current operating condition boundary label is matched with the operating condition boundary label in the constructed pulse flow normal fingerprint. The pulse flow normal fingerprint with the same operating condition boundary label is selected as the verification benchmark. The current pulse readback state segment, the current cabinet flow response state segment and the matched pulse flow normal fingerprint are bound and saved to obtain the pulse flow segment to be verified.

[0056] S3.5: Perform constrained timing alignment between the pulse flow segment to be verified and the matched normal pulse flow fingerprint, and calculate the pulse flow deviation of the pulse flow segment to be verified relative to the normal pulse flow fingerprint. Specifically, the pulse readback state sequence and cabinet flow response state sequence in the pulse flow segment to be verified are compared with the pulse readback state sequence and cabinet flow response state sequence in the matched normal pulse flow fingerprint. The allowable hysteresis range corresponding to the working condition boundary label is read, and the first sampling position of the pulse flow segment to be verified and the first sampling position of the normal pulse flow fingerprint are used as the alignment start point, and the last sampling position of the pulse flow segment to be verified and the last sampling position of the normal pulse flow fingerprint are used as the alignment end point. According to the rule of monotonically advancing sampling position, continuously advancing adjacent alignment points, and the alignment offset not exceeding the allowable hysteresis range, a set of constrained timing alignment paths is determined. On each constrained timing alignment path in the set of constrained timing alignment paths, the joint difference between the pulse readback state and the cabinet flow response state is calculated point by point, and the average value of the joint difference is used as the path cost. The constrained timing alignment path with the minimum path cost is selected, and the minimum path cost is bounded to obtain the pulse flow deviation of the pulse flow segment to be verified relative to the normal pulse flow fingerprint.

[0057] The deterministic expression for the constrained timing alignment path set is: ; In the formula, Represents the set of allowed constrained timing alignment paths; Indicates the allowable lag range; This represents a constrained timing alignment path; This indicates the first sampling position of the pulse segment to be verified; This indicates the first sampling position of a normal pulse flow fingerprint; This indicates the sampling position of the pulse segment to be verified corresponding to the last alignment point of the constrained timing alignment path; This indicates the pulse flow normal fingerprint sampling position corresponding to the last alignment point of the constrained timing alignment path; This represents the total number of sampling points for the pulse segment to be verified; This represents the total number of sampling points in a normal pulse flow fingerprint. Indicates the sequential number of alignment points in the constrained timing alignment path; Indicates the first The adjacent alignment points after the first alignment point are numbered sequentially; This indicates the total number of alignment points contained in the constrained timing alignment path; Indicates the pulse segment to be checked that is related to the first... The sampling positions corresponding to each alignment point; This indicates that the pulse flow in the fingerprint is similar to the first... The sampling positions corresponding to each alignment point; Indicates the pulse segment to be checked that is related to the first... The sampling positions corresponding to each alignment point; This indicates that the pulse flow in the fingerprint is similar to the first... The sampling positions corresponding to each alignment point; Indicates the lower bound of the allowable hysteresis range; This indicates the upper bound of the allowed lag range.

[0058] It should be noted that before calculating the joint difference of the alignment points, the pulse readback state values ​​and cabinet flow response state values ​​in the pulse flow segment to be verified and the normal pulse flow fingerprint are respectively normalized in terms of their range. Specifically, the pulse readback state values ​​are normalized according to the maximum and minimum values ​​of the pulse readback state sequence in the matched normal pulse flow fingerprint, and the cabinet flow response state values ​​are normalized according to the maximum and minimum values ​​of the cabinet flow response state sequence in the matched normal pulse flow fingerprint, so that the pulse readback state values ​​and cabinet flow response state values ​​are converted into dimensionless state values.

[0059] The joint difference calculation expression for alignment points is: ; In the formula, Indicates the first pulse segment to be verified The sampling location is compared with the first normal fingerprint in pulse flow. Joint differences between sampling locations; Indicates the sampling position number in the pulse segment to be verified; Indicates the sampling position number in a normal pulse fingerprint; Indicates the pulse segment to be checked is in the 1st... The dimensionless state value after normalization of the pulse readback state value at each sampling position; Indicating normal pulse flow fingerprint in the first... The dimensionless state value after normalization of the pulse readback state value at each sampling position; Indicates the pulse segment to be checked is in the 1st... The dimensionless state value after normalization of the cabinet flow response state value at each sampling location; Indicating normal pulse flow fingerprint in the first... The dimensionless state value obtained by normalizing the cabinet flow response state value at each sampling location.

[0060] The expression for calculating the pulse deviation is: ; In the formula, This indicates the amount of pulse flow deviation of the pulse flow segment to be verified relative to the normal pulse flow fingerprint; This indicates the pulse flow relationship between the pulse readback state and the cabinet current response state; Represents the set of allowed constrained timing alignment paths. Select the constrained temporal alignment path with the minimum path cost; Represents a constrained timing alignment path The total number of alignment points included; Indicates the sampling location number and sampling location number The alignment points formed belong to the constrained timing alignment path. ; Represents a constrained timing alignment path The sum of the joint differences of all alignment points; This represents the normalized reference constant used to convert the minimum path cost into a bounded deviation of less than one.

[0061] It should be noted that, The range of values ​​is ;when When the value is close to zero, it indicates a high degree of consistency between the pulse flow segment to be verified and the normal pulse flow fingerprint. An increase indicates a greater deviation of the pulse flow segment to be verified from the normal pulse flow fingerprint.

[0062] S3.6: Based on the pulse flow deviation, mark the deviation level of the pulse readback state and cabinet flow response state, and write the pulse flow deviation, constrained timing alignment result and working condition boundary label into the execution feedback relationship to form the pulse flow deviation state.

[0063] Specifically, based on the pulse flow deviation, the pulse readback state and cabinet flow response state are marked as normal deviation, slight deviation, significant deviation, and severe deviation. When the pulse flow deviation does not reach the preset slight deviation threshold, it is marked as normal deviation; when the pulse flow deviation reaches the preset slight deviation threshold but does not reach the preset significant deviation threshold, it is marked as slight deviation; when the pulse flow deviation reaches the preset significant deviation threshold but does not reach the preset severe deviation threshold, it is marked as significant deviation; and when the pulse flow deviation reaches the preset severe deviation threshold, it is marked as severe deviation. The pulse flow deviation, deviation level, constrained timing alignment result, and operating condition boundary label are registered in the execution feedback relationship between the pulse readback state node and the cabinet flow response state node to form the pulse flow deviation state.

[0064] It should be noted that the slight deviation threshold is set based on the distribution of pulse deviation amount of historical normal pulse event samples relative to normal pulse fingerprints. The example value range is 0.15 to 0.25, and the value is determined to distinguish between normal fluctuations and recordable slight deviations. The significant deviation threshold is set based on the degree of influence of the pulse flow deviation on the stability of the correspondence between the pulse readback state and the cabinet flow response state. The example value range is 0.35 to 0.55, and the value is determined by identifying significant deviations that need to be entered into the working condition simulation. The severe deviation threshold is set based on the degree to which the pulse flow deviation affects the reliability of the execution feedback relationship. The example value range is 0.65 to 0.8. The value is determined by identifying a severe deviation between the pulse readback state and the cabinet flow response state that is clearly inconsistent with the normal pulse flow fingerprint.

[0065] S4. Based on the causal twin simulation model of the pulse current deviation state driving the operating condition, perform operating condition deduction, calculate the risk contribution of the pulse current deviation state in the state transmission relationship and execution feedback relationship, generate the operating condition risk contribution result, perform simulation verification of the current operating condition of the excitation system, and associate the triggering cause, deviation source and simulation deduction basis to generate an excitation system operating condition simulation evaluation report.

[0066] S4.1: Based on the pulse flow deviation state, determine the deviation action time, deviation action direction and deviation action intensity, encapsulate them as deviation injection events, write the deviation injection events into the execution feedback relationship in the working condition causal twin simulation model, and obtain the deviation-driven deduction conditions. Specifically, based on the pulse flow deviation status, the pulse flow deviation amount, deviation level, constrained timing alignment result, and operating condition boundary label are read, and the frame formation time when the pulse flow segment to be verified first deviates from the normal pulse flow fingerprint is located according to the constrained timing alignment result. The frame formation time when the first deviation occurs is determined as the deviation action time. The order and magnitude of the changes in the pulse readback state and the cabinet flow response state relative to the normal pulse flow fingerprint are compared at the same frame formation time. The correspondence between the cabinet flow response state and the pulse readback state, which is lagging, leading, larger, or smaller, is determined as the deviation action direction. The deviation action intensity is determined based on the pulse flow deviation amount and deviation level. The deviation action time, deviation action direction, deviation action intensity, and operating condition boundary label are bound as deviation injection events, and the deviation injection events are written into the execution feedback relationship between the pulse readback state node and the cabinet flow response state node in the operating condition causal twin simulation model to obtain the deviation-driven deduction conditions.

[0067] S4.2: Based on the deviation-driven deduction conditions, the causal twin simulation model of the working condition is used to perform factual deduction and counterfactual deduction. The risk contribution is calculated based on the differences between the factual deduction trajectory and the counterfactual deduction trajectory in the state transmission relationship and execution feedback relationship, and the working condition risk contribution result is generated.

[0068] Specifically, based on the deviation-driven deduction conditions, the deviation action time, deviation action direction, deviation action intensity, and operating condition boundary labels in the deviation injection event are read. Taking the deviation action time as the deduction starting point, the deviation action direction and deviation action intensity are written into the execution feedback relationship between the pulse readback state node and the cabinet flow response state node in the operating condition causal twin simulation model. According to the state update rules corresponding to the state transmission relationship and execution feedback relationship, the correction magnitude of the corresponding forward input to the standardized state value of the target node is determined by the causal interaction strength of the effective node relationship. The standardized state values ​​of the adjustment calculation state node, the limiting auxiliary loop state node, the pulse readback state node, and the cabinet flow response state node are corrected and updated in sequence to obtain the factual deduction trajectory that retains the pulse flow deviation state. Under the same inference starting point, the same working condition boundary label, the same initial standardized working condition state quantity, and the same simulation call order, the pulse flow deviation state is replaced with the normal associated state corresponding to the pulse flow normal fingerprint. According to the same state update rule, the pulse readback state sequence and cabinet flow response state sequence corresponding to the pulse flow normal fingerprint are used as the execution feedback input between the pulse readback state node and the cabinet flow response state node. The standardized state values ​​of the adjustment calculation state node, the constraint auxiliary loop state node, the pulse readback state node, and the cabinet flow response state node are corrected and updated again to obtain the counterfactual inference trajectory. State transit relationships and execution feedback relationships are read from the causal twin simulation model of the working condition, and these relationships are merged into a set of relationships to be calculated. Based on the set of relationships to be calculated, the relationships to be calculated are determined one by one. Within the inference frame range corresponding to each relationship to be calculated, the standardized state values ​​of the associated states of the relationship to be calculated are extracted in the factual inference trajectory and the counterfactual inference trajectory. The root mean square difference between the factual inference trajectory and the counterfactual inference trajectory is calculated, and the root mean square difference is converted into a bounded risk contribution. The risk contribution is collected according to the relationship number to be calculated, and the working condition risk contribution result is generated.

[0069] It should be noted that the extrapolation frame range is set based on the deviation action time, the allowable hysteresis range, and the normal response duration corresponding to the relationship to be calculated. The example value range is five to twenty frame cycles after the deviation action time. The value is based on the complete process of the coverage pulse deviation state propagating along the state transmission relationship and the execution feedback relationship to form a stable difference.

[0070] The root mean square difference between the factual and counterfactual trajectories is calculated using the following expression: ; In the formula, Represents the relation to be calculated The root mean square difference between the corresponding factual deduction trajectory and the counterfactual deduction trajectory; This represents a relation to be computed within a set of relations to be computed; Represents the relation to be calculated The associated set of states; Represents the relation to be calculated An associated state; Represents the relation to be calculated The corresponding inference frame range; Indicates the range of inference frames The framing sequence number within the frame; Representing state In the trajectory of factual deduction, the first Standardized state values ​​under a framed sequence; Representing state In the counterfactual deduction trajectory, the first Standardized state values ​​under a framed sequence; Indicates the trajectory of factual deduction; Indicates the counterfactual deduction trajectory; Represents the relation to be calculated The number of associated states; Represents the relation to be calculated The corresponding number of simulation frames.

[0071] The expression for calculating the risk contribution corresponding to the relationship to be calculated is: ; In the formula, Represents the relation to be calculated The corresponding risk contribution.

[0072] S4.3: Extract the corresponding state transmission relationship, execution feedback relationship, factual deduction trajectory and counterfactual deduction trajectory from the working condition risk contribution results, determine the execution feedback relationship corresponding to the risk contribution amount as the triggering relationship, determine the pulse flow deviation state associated with the triggering relationship as the triggering cause, and determine the deviation position between the pulse readback state and the cabinet flow response state in the pulse flow deviation state as the deviation source, and obtain the simulation verification association results; Specifically, the system reads the relationships to be calculated, relationship types, risk contribution amounts, factual deduction trajectories, and counterfactual deduction trajectories one by one from the operational risk contribution results, and distinguishes between state transmission relationships and execution feedback relationships according to relationship type. It then finds the execution feedback relationship with the highest risk contribution amount and identifies it as the triggering relationship. Based on the triggering relationship, it reads the pulse flow deviation states already written into the execution feedback relationship and identifies them as the triggering cause. According to the constrained timing alignment results in the pulse flow deviation states, it locates the frame position where the pulse readback state and the cabinet flow response state first deviate, identifies this frame position as the deviation position, and identifies the corresponding pulse readback state and cabinet flow response state as the deviation source. Finally, it binds and saves the state transmission relationship, execution feedback relationship, triggering relationship, triggering cause, deviation source, factual deduction trajectory, and counterfactual deduction trajectory to obtain the simulation verification association results.

[0073] S4.4: Based on the simulation verification correlation results, the factual inference trajectory, counterfactual inference trajectory, standardized operating condition state variables, triggering causes and deviation sources are correlated and compared within the same time window to generate an excitation system operating condition simulation evaluation report.

[0074] Specifically, based on the simulation verification results, the triggering cause, deviation source, factual deduction trajectory, counterfactual deduction trajectory, and corresponding state transmission and execution feedback relationships are read. The frame position corresponding to the deviation source is used as the comparison starting point, and the deduction frame range corresponding to the working condition risk contribution result is used as the same time window. The adjustment calculation state, limiting auxiliary loop state, pulse readback state, and cabinet current response state within the same time window are extracted from the standardized working condition state quantities. The factual deduction trajectory, counterfactual deduction trajectory, and standardized working condition state quantities are aligned frame by frame according to the frame order. The deviation of the factual deduction trajectory from the standardized working condition state quantities is recorded, as well as the recovery of the counterfactual deduction trajectory from the standardized working condition state quantities. The triggering cause, deviation source, same time window, state transmission relationship, execution feedback relationship, factual deduction trajectory comparison result, counterfactual deduction trajectory comparison result, and working condition risk contribution result are organized into corresponding items to generate an excitation system working condition simulation evaluation report.

[0075] Figure 5 The display shows a comparison of the changes in standardized operating condition state quantities, factual inference trajectories, and counterfactual inference trajectories within the same time window. The horizontal axis represents the frame cycle number, and the vertical axis represents the standardized state value. The three black and white curves represent the standardized operating condition state quantities, factual inference trajectories, and counterfactual inference trajectories, respectively. The factual inference trajectory represents the operating condition inference result after retaining the pulse flow deviation state, while the counterfactual inference trajectory represents the inference result after replacing the pulse flow deviation state with the normal correlation state corresponding to the normal pulse flow fingerprint. The greater the difference between the two, the more significant the impact of the pulse flow deviation state on the current operating condition.

[0076] Figure 6 The figure shows the distribution of average risk contribution for each state transmission relationship and execution feedback relationship under different restricted auxiliary loop state action types. The horizontal axis represents the restricted auxiliary loop state action type, and the vertical axis represents the state transmission relationship and execution feedback relationship. The color scale on the right represents the average risk contribution. The closer the color is to yellow, the greater the average risk contribution. The closer the color is to dark, the smaller the average risk contribution. The risk contribution of "execution readback port to response feedback port" is the highest overall in the figure. This indicates that the pulse flow deviation state mainly affects the execution feedback relationship between the pulse readback state and the cabinet flow response state, and can be further used to locate the deviation propagation path and risk source.

[0077] This embodiment also provides a computer device applicable to the excitation system condition evaluation method based on simulation modeling, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the excitation system condition evaluation method based on simulation modeling as proposed in the above embodiment.

[0078] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0079] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the simulation-based modeling-based excitation system condition evaluation method proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0080] In summary, this invention achieves a stable correlation between pulse readback state and cabinet flow response state under specific operating conditions by constructing a pulse flow normal fingerprint corresponding to the operating condition boundary label. This enables twin simulation to obtain a standardized reference benchmark for deviation identification. Furthermore, it reflects the response characteristics under different regulation outputs and limiting auxiliary loop conditions, providing clear operating condition constraints and timing basis for the formation of pulse flow deviation states. It also provides a traceable data foundation for subsequent risk contribution calculation, simulation verification, and the generation of excitation system operating condition simulation evaluation reports, thereby improving the interpretability, stability, and verifiability of the excitation system operating condition evaluation results.

[0081] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for evaluating the operating conditions of an excitation system based on simulation modeling, characterized in that, include: Real-time operating data of the excitation system is collected and time reference is unified and state source is marked to obtain real-time state frames. Group frequency measurement, sampling area correction and data source switching are performed on the real-time state frames, while the sampling reliability and state source marking are retained to obtain standardized operating condition state quantities. The adjustment calculation state, the limiting auxiliary loop state, the pulse readback state, and the cabinet flow response state in the standardized operating condition state variables are divided into nodes, and the state transmission relationship and execution feedback relationship are configured to construct a causal twin simulation model of the operating condition. Based on the causal twin simulation model of the working condition, the working condition boundary label is formed by adjusting the calculation state and restricting the auxiliary loop state. Based on the stable correspondence between the pulse readback state segment and the cabinet flow response state segment under the working condition boundary label, the pulse flow normal fingerprint is constructed. The pulse readback state and cabinet flow response state in the current standardized working condition state quantity are aligned with the matched pulse flow normal fingerprint in a constrained time sequence to form the pulse flow deviation state. The operating conditions are simulated based on the causal twin simulation model driven by the pulse flow deviation state, and the risk contribution of the pulse flow deviation state in the state transmission relationship and execution feedback relationship is calculated. The operating condition risk contribution result is generated, and the current operating condition of the excitation system is simulated and verified. The triggering cause, deviation source and simulation simulation basis are associated to generate the excitation system operating condition simulation evaluation report.

2. The method for evaluating the operating conditions of an excitation system based on simulation modeling as described in claim 1, characterized in that, The steps to obtain the real-time status frame are as follows: The excitation system real-time operation data is collected according to the preset framing period, and the local sampling time, status source mark and channel type mark are added to obtain the original sampling record; Based on the original sampling records, clock offset correction and clock drift correction are performed on the local sampling time using a unified clock. Sampling confidence is generated based on the clock correction results, state source marker, and channel type marker. Sampling records carrying sampling confidence, state source marker, and channel type marker are merged according to a preset framing period to obtain a real-time state frame.

3. The method for evaluating the operating conditions of an excitation system based on simulation modeling as described in claim 1 or 2, characterized in that, The steps to obtain the standardized operating condition state quantities are as follows: The excitation signal is grouped based on the channel type marker carried by each sampling record in the real-time status frame, and real-time frequency measurement and sampling area correction are performed on the grouped excitation signal. At the same time, the sampling reliability and status source marker in the real-time status frame are inherited to obtain the grouped correction status quantity. Based on the grouped correction state variables, the reliability of the original sampling area, the correction sampling area, and the short-term holding sampling area is compared, and the data source is switched according to the reliability comparison results to obtain the standardized operating condition state variables.

4. The method for evaluating the operating conditions of an excitation system based on simulation modeling as described in claim 1, characterized in that, The steps for constructing the causal twin simulation model under the operating conditions are as follows: The regulation calculation state, the limiting auxiliary loop state, the pulse readback state, and the cabinet flow response state are extracted from the standardized operating condition state variables and encapsulated into corresponding nodes to obtain a node encapsulation set. Based on the node encapsulation set, the adjustment calculation state node is mapped to the adjustment output port, the constraint auxiliary loop state node is mapped to the boundary constraint port, the pulse readback state node is mapped to the execution readback port, and the cabinet flow response state node is mapped to the response feedback port. According to the transmission rules from the adjustment output port to the boundary constraint port, from the adjustment output port to the execution readback port, from the boundary constraint port to the execution readback port, from the execution readback port to the response feedback port, from the response feedback port to the adjustment output port, and from the response feedback port to the boundary constraint port, a semantic contract table is generated. Candidate node relationships are generated based on the semantic contract table, and the causal interaction strength of each candidate node relationship is calculated. Candidate node relationships whose causal interaction strength meets the preset relationship threshold are retained as valid node relationships. Based on the valid node relationships, the state transmission relationship and execution feedback relationship between the adjustment calculation state node, the constraint auxiliary loop state node, the pulse readback state node and the cabinet flow response state node are configured and adjusted. The node connection direction and simulation call order are determined, and the working condition causal twin simulation model is constructed.

5. The method for evaluating the operating conditions of an excitation system based on simulation modeling as described in claim 1 or 4, characterized in that, The steps for constructing a normal pulse flow fingerprint are as follows: Based on the working condition causal twin simulation model, the execution feedback relationship between the pulse readback state node and the cabinet flow response state node is read, and the pulse readback state segment and cabinet flow response state segment within the same event window are extracted using the effective change time of the pulse readback state as the event starting point to obtain the pulse flow event segment. Read the adjustment calculation state and the constraint auxiliary loop state within the same event window from the pulse event fragment, and combine the output change range of the adjustment calculation state and the boundary action range of the constraint auxiliary loop state into the operating condition boundary label to generate pulse event samples; Historical pulse flow event samples were selected based on the following criteria: sampling confidence met a preset confidence threshold; pulse readback state change amplitude did not exceed a preset pulse mutation threshold; cabinet flow response state change amplitude did not exceed a preset cabinet flow fluctuation threshold; and neither pulse readback state segment nor cabinet flow response state segment was missing. The historical pulse flow event samples were grouped according to the working condition boundary label. Stable corresponding segments were determined based on the stable correspondence between pulse readback state segments and cabinet flow response state segments within the same group. The stable corresponding segments were aligned according to the relative framing position after the event start point. The median values ​​of pulse readback state values ​​and cabinet flow response state values ​​at the same relative framing position were extracted to generate standard pulse readback state sequences and standard cabinet flow response state sequences, thus constructing a normal pulse flow fingerprint.

6. The method for evaluating the operating conditions of an excitation system based on simulation modeling as described in claim 5, characterized in that, The steps to form the pulse flow deviation state are as follows: Based on the pulse flow normal fingerprint, the pulse readback status and cabinet flow response status in the current standardized operating condition status are extracted, and the pulse flow normal fingerprint is matched according to the corresponding operating condition boundary label to obtain the pulse flow segment to be verified. The pulse flow segment to be verified is aligned with the matched normal pulse flow fingerprint under constrained timing, and the pulse flow deviation of the pulse flow segment to be verified relative to the normal pulse flow fingerprint is calculated. The deviation levels of pulse readback status and cabinet flow response status are marked based on the pulse flow deviation, and the pulse flow deviation, constrained timing alignment results and operating condition boundary labels are written into the execution feedback relationship to form the pulse flow deviation status.

7. The method for evaluating the operating conditions of an excitation system based on simulation modeling as described in claim 6, characterized in that, The steps for generating the risk contribution result for the operating condition are as follows: Based on the pulse deviation state, the time, direction and intensity of the deviation are determined and encapsulated as a deviation injection event. The deviation injection event is written into the execution feedback relationship in the causal twin simulation model of the working condition to obtain the deviation-driven deduction conditions. Based on the deviation-driven inference condition-driven causal twin simulation model, factual and counterfactual inferences are performed. The risk contribution is calculated based on the differences between the factual and counterfactual inference trajectories in the state transmission relationship and execution feedback relationship, and the working condition risk contribution result is generated.

8. The method for evaluating the operating conditions of an excitation system based on simulation modeling as described in claim 1 or 7, characterized in that, The steps for generating the excitation system operating condition simulation evaluation report are as follows: Extract the corresponding state transmission relationship, execution feedback relationship, factual deduction trajectory and counterfactual deduction trajectory from the working condition risk contribution results, determine the execution feedback relationship corresponding to the risk contribution amount as the triggering relationship, determine the pulse flow deviation state associated with the triggering relationship as the triggering cause, and determine the deviation position between the pulse readback state and the cabinet flow response state in the pulse flow deviation state as the deviation source, and obtain the simulation verification association results; Based on the simulation verification correlation results, the factual inference trajectory, counterfactual inference trajectory, standardized operating condition state variables, triggering causes and deviation sources are correlated and compared within the same time window to generate an excitation system operating condition simulation evaluation report.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the excitation system condition evaluation method based on simulation modeling as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the excitation system condition evaluation method based on simulation modeling as described in any one of claims 1 to 8.

Images