An electrical equipment fault prediction method and system based on digital twinning

Abstract

The present application relates to the technical field of electrical equipment fault prediction, in particular to an electrical equipment fault prediction method and system based on digital twinning; the method comprises the following steps: obtaining equipment basic data, acquisition node configuration data, excitation configuration data and segmentation rule data of a target electrical equipment, constructing a healthy digital twin, establishing local constraint parameters, generating an encoded excitation sequence and a benchmark echo constraint set; collecting echo response data to form an echo response sequence, performing piecewise corresponding alignment and local correction search to obtain a minimum correction set and determine a sensitive constraint axis; generating a single-step misplacement compression sequence based on the sensitive constraint axis, obtaining compression echo response data and generating a compression minimum correction set; comparing the compression minimum correction set with the minimum correction set to determine a fault critical preposition state and output a predicted fault result. The present application realizes preposition identification of electrical equipment implicit local deterioration state and accurate determination of fault critical preposition state.

Description

Technical Field

[0001] This invention relates to the field of electrical equipment fault prediction technology, and specifically to a method and system for electrical equipment fault prediction based on digital twins. Background Technology

[0002] With the increasing complexity of power systems, industrial power distribution systems, and electrical control systems, the operating state of electrical equipment is no longer limited to changes in single parameters such as steady-state voltage, current, and temperature rise. Instead, it is simultaneously affected by structural connection status, local insulation status, excitation transmission path, response recovery process, and external operating disturbances. To improve the predictability and accuracy of fault prediction, related technologies have begun to combine digital twins, response acquisition, and constraint analysis to model and determine the response behavior of target electrical equipment at different operating stages. This enables the identification and handling of state changes before the formation of explicit faults. Therefore, how to effectively characterize the response relationship of electrical equipment without deviating from actual operating conditions and to complete fault prediction accordingly has become an important research direction in the field of electrical equipment testing and fault prediction technology. In application scenarios where electrical equipment operates for long periods, frequently switches between operating conditions, and experiences slow changes in local structural states, there are still problems such as the difficulty in forming stable precursor responses to latent local degradation and the difficulty in identifying critical instability boundaries in advance. Specifically, on the one hand, the response data generated by the target electrical equipment under natural operating conditions is easily affected by load fluctuations, environmental changes, differences in acquisition locations, and changes in local propagation paths. Even if there are changes in local contact states, local insulation paths, or local response lags within the equipment, the external acquisition results may still appear as discrete, short-term, or non-continuous states, making it difficult to form repeatable, locatable, and identifiable precursor evidence. On the other hand, some target electrical equipment can still maintain a normal surface response at the current operating point, but its internal constraint relationship is already close to the instability boundary. Under subsequent small excitation offsets, superposition of local disturbances, or switching between adjacent operating conditions, the response structure may suddenly change, causing the equipment to quickly transition from a non-obvious abnormal state to an obvious fault state, making it difficult for conventional single-condition detection results to identify such precursor risks in a timely manner. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for predicting electrical equipment faults based on digital twins, so as to solve the problems mentioned in the background art, such as the difficulty in forming stable precursor responses for latent local degradation and the difficulty in identifying critical instability boundaries in advance.

[0004] To achieve the above objectives, the technical solution of the present invention is: a method for predicting electrical equipment faults based on digital twins, comprising: S1. Obtain the basic equipment data, acquisition node configuration data, excitation configuration data and segmentation rule data of the target electrical equipment. Based on the basic equipment data, acquisition node configuration data and excitation configuration data, construct a health digital twin and establish local constraint parameters in the health digital twin to generate coded excitation sequences and reference echo constraint sets. Among them, the health digital twin is a digital twin reference body that characterizes the excitation-response benchmark relationship under the health state of the target electrical equipment; the local constraint parameters are the constraint parameter items set in the health digital twin corresponding to the local response units; and the benchmark echo constraint set is the benchmark constraint set set corresponding to the coded excitation sequence. S2. Input the coded excitation sequence into the target electrical equipment, acquire the echo response data at the acquisition node, form the echo response sequence according to the segmentation rule data, align the echo response sequence with the reference echo constraint set segment by segment to obtain the alignment result, and perform local correction search on the local constraint parameters associated with the alignment result based on the health digital twin to obtain the minimum correction set and determine the sensitive constraint axis. Among them, the minimum correction set is the set of the fewest correction terms that make the echo response sequence meet the requirements of the reference echo constraint set; the sensitive constraint axis is the constraint direction in the minimum correction set that corresponds to the main mismatch direction of the current echo response. S3. Based on the sensitive constraint axis, the coded excitation sequence is processed by single-axis misalignment to generate a single-step misaligned pressure test sequence. The single-step misaligned pressure test sequence is input to the target electrical equipment. Pressure test echo response data is acquired at the acquisition node. The pressure test echo response sequence is generated according to the segmentation rule data. The pressure test echo response sequence is aligned segment by segment with the pressure test reference echo constraint set to obtain the pressure test alignment result. Based on the health digital twin, the local constraint parameters associated with the pressure test alignment result are locally corrected and searched to generate the minimum correction set of the pressure test. Among them, the single-step misaligned stress test sequence is the stress test excitation sequence formed after the encoded excitation sequence is misaligned once along the sensitive constraint axis; the minimum correction set of stress test is the minimum set of correction terms determined for the stress test echo response sequence; S4. Compare the minimum correction set of the pressure test with the minimum correction set to determine the critical pre-fault state of the target electrical equipment, and output the predicted fault result based on the minimum correction set of the pressure test. Among them, the critical pre-fault state is the pre-fault state in which the target electrical equipment has not yet entered the manifest fault state and no longer meets the health state requirements under single-step misalignment pressure test conditions.

[0005] Preferably, in step S1, the acquisition nodes are data acquisition nodes set at various local response positions of the target electrical equipment; the health digital twin performs digital twin mapping on various local response positions of the target electrical equipment based on the equipment's basic data, and the health digital twin is set with multiple local response units, which are response units in the health digital twin that correspond to different local response positions of the target electrical equipment; the local constraint parameters are set according to the local response units in groups, with each group of local constraint parameters corresponding to one local response unit, and different local response units corresponding to different groups of local constraint parameters, and each local response unit establishes a correspondence with the acquisition node; the excitation configuration data is used to limit the arrangement order and path correspondence of each coded excitation segment, and the coded excitation sequence is formed by arranging multiple coded excitation segments in the order determined by the excitation configuration data, and each coded excitation segment establishes a correspondence with a local response unit; the reference echo constraint set is set in segments according to the coded excitation sequence, and each reference echo constraint is set according to the local constraint parameters of the corresponding local response unit, and each reference echo constraint is associated with the corresponding coded excitation segment, local response unit, and acquisition node.

[0006] Preferably, in step S2, the echo response data is the node response data collected by the acquisition node under the action of the coded excitation sequence of the target electrical equipment; the echo response sequence is a sequential response sequence formed by arranging the echo response data into segments according to segmentation rule data; the specific method for forming the echo response sequence according to segmentation rule data is as follows: the echo response data is divided into boundaries according to the segment order corresponding to the coded excitation sequence to form multiple echo segments, and the echo response sequence is formed according to the segment order corresponding to each echo segment; wherein, the echo segment includes a propagating echo segment, a reflected echo segment, and a recovered echo segment; the echo... The specific method for aligning the response sequence with the reference echo constraint set segment by segment is as follows: The reference echo constraint set is divided into multiple constraint segments according to the segment order of the coded excitation sequence, and each echo segment and each constraint segment are matched at the segment level according to the same segment order; the alignment result is a set of segment-level correspondence results formed by the correspondence between each echo segment and each constraint segment, and each correspondence item in the alignment result includes an echo segment identifier, a constraint segment identifier, a path identifier, and a node identifier; the path identifier in the alignment result is used to characterize the coded excitation path to which the corresponding item belongs, and the node identifier is used to characterize the acquisition node to which the corresponding item belongs.

[0007] Preferably, in step S2, the local constraint parameters associated with the alignment result are local constraint parameters determined based on the path identifier and node identifier in the alignment result; the local correction search is a constraint correction search operation performed item by item on the local constraint parameters, used to determine the correction terms corresponding to the constraint mismatch of the current echo response sequence relative to the benchmark echo constraint set; the specific method for performing local correction search on the local constraint parameters associated with the alignment result based on the health digital twin is as follows: candidate correction terms are generated for the local constraint parameters associated with the alignment result, and correction term combinations are constructed based on the candidate correction terms; constraint matching judgment is performed on each correction term combination to determine the correction term combination that satisfies the matching requirement between the echo response sequence and the benchmark echo constraint set; the minimum correction set is the local correction search... The final determined minimum set of correction terms is used to characterize the constraint correction distribution corresponding to the current echo response sequence meeting the requirements of the reference echo constraint set. The minimum set of correction terms includes correction position, correction type, and correction amount. The minimum set of correction terms is determined by having the fewest correction terms as the first determining condition. If the number of correction terms is the same, the minimum constraint matching error is determined by having the smallest constraint matching error as the second determining condition. The correction position corresponds to the local response unit, the correction type corresponds to the local constraint parameter category, and the correction amount corresponds to the offset of the local constraint parameter. The number of correction terms is the number of correction terms included in the minimum set of correction terms, and the constraint matching error is the error amount of the remaining constraint deviation between the echo response sequence and the reference echo constraint set. The multiple correction terms in the minimum set of correction terms are arranged according to the magnitude of the correction amount to form a correction term sequence.

[0008] Preferably, in step S2, the sensitive constraint axis is a single constraint direction in the minimum correction set corresponding to the main mismatch direction of the current echo response. It is used to identify the dominant mismatch direction of the current echo response sequence and to perform sorting and extraction on each correction item in the minimum correction set. When sorting each correction item in the minimum correction set, the sensitive constraint axis is sorted according to the magnitude of the correction amount of each correction item, and the local constraint parameter corresponding to the first correction item in the sort is taken. The sensitive constraint axis is then determined together with the coded excitation segment associated with the local constraint parameter. Only one sensitive constraint axis is determined, and it is set separately from the constraint directions corresponding to the other correction items in the minimum correction set. It uniquely indicates the main mismatch direction of the current echo response sequence. The coded excitation segment corresponding to the sensitive constraint axis is the unique target coded excitation segment for single-axis misalignment processing.

[0009] Preferably, in S3, the single-axis misalignment processing is a single misalignment processing method performed on the unique target coded excitation segment in the coded excitation sequence, used to generate a single-step misalignment stress test sequence. The specific method for performing single-axis misalignment processing on the coded excitation sequence based on the sensitive constraint axis is as follows: determine the unique target coded excitation segment corresponding to the sensitive constraint axis, keep the remaining coded excitation segments in the coded excitation sequence unchanged, perform a single misalignment adjustment on the unique target coded excitation segment, and combine the misaligned unique target coded excitation segment with the remaining coded excitation segments in the original segment order to form a single-step misalignment stress test sequence. The single-step misalignment stress test sequence is a stress test excitation sequence formed after the coded excitation sequence performs one misalignment processing along the sensitive constraint axis, used to obtain the stress test echo response data of the target electrical equipment under stress test conditions. The single-step misalignment stress test sequence includes the target coded excitation segment after single-axis misalignment processing and the remaining coded excitation segments that have not undergone misalignment processing. The segment order of the coded excitation sequence remains consistent before and after the single-axis misalignment processing.

[0010] Preferably, in step S3, the pressure test echo response data is the node response data collected by the acquisition node under the action of the single-step misaligned pressure test sequence of the target electrical equipment, used to form the pressure test echo response sequence; the pressure test echo response sequence is a sequential response sequence formed by arranging the pressure test echo response data into segments according to segmentation rule data, used as the pressure test input sequence for local correction search, and the pressure test echo response sequence includes a pressure test propagation echo segment, a pressure test reflection echo segment, and a pressure test recovery echo segment; the specific method for forming the pressure test echo response sequence according to the segmentation rule data is as follows: the pressure test echo response data is divided into boundaries according to the segment order corresponding to the single-step misaligned pressure test sequence to form multiple pressure test echo segments, and each pressure test echo segment is processed... Each sound segment is assigned a corresponding segment identifier, and the segments are arranged according to the segment order corresponding to the segment identifiers to form a stress test echo response sequence. The stress test reference echo constraint set is a set of stress test constraints generated by the health digital twin based on the single-step misaligned stress test sequence. The stress test reference echo constraint set is segmented according to the single-step misaligned stress test sequence. The specific method for aligning the stress test echo response sequence with the stress test reference echo constraint set segment by segment is as follows: the stress test reference echo constraint set is divided into multiple stress test constraint segments according to the segment order of the single-step misaligned stress test sequence, and each stress test echo segment and each stress test constraint segment are matched at the segment level according to the same segment order. The stress test alignment result is a set of segment-level correspondence results formed by the correspondence between each stress test echo segment and each stress test constraint segment.

[0011] Preferably, in step S3, the minimum correction set for stress testing is the set of the fewest correction terms determined for the stress test echo response sequence, used to characterize the correction distribution corresponding to the stress test echo response sequence meeting the constraints of the health digital twin under single-step misalignment stress test conditions. The minimum correction set for stress testing includes correction position, correction type, and correction amount. The structure of the correction terms in the minimum correction set for stress testing is consistent with that of the minimum correction set, and the minimum number of correction terms is used as the first determining condition. When the number of correction terms is the same, the minimum constraint matching error is used as the second determining condition. The specific method for generating the minimum correction set for stress testing by performing local correction search on the local constraint parameters associated with the stress test alignment results based on the health digital twin is as follows: the corresponding local constraint parameters are determined according to the path identifier and node identifier in the stress test alignment results. Stress test candidate correction terms are generated for the corresponding local constraint parameters, and stress test correction term combinations are constructed based on the stress test candidate correction terms. Constraint matching judgment is performed on each stress test correction term combination to determine the stress test correction term combination that meets the matching requirements of the stress test echo response sequence and the stress test reference echo constraint set. The minimum correction set for stress testing is determined from the stress test correction term combinations that meet the requirements.

[0012] Preferably, in step S4, the critical pre-fault state of the target electrical equipment is a pre-fault state in which the target electrical equipment no longer meets the health requirements before a manifest fault occurs. The critical pre-fault state is determined based on the comparison result between the minimum correction set and the minimum correction set. When the number of correction items, the number of correction locations, and the number of correction types in the minimum correction set are all greater than the corresponding statistical results of the minimum correction set, the target electrical equipment is determined to be in the critical pre-fault state. The predicted fault result includes the predicted fault location and the predicted fault category. The predicted fault location is determined based on the aggregation result of the correction amounts corresponding to each correction location in the minimum correction set, and the target electrical equipment location corresponding to the correction location with the largest aggregated correction amount is taken as the predicted fault location. The predicted fault category is determined based on the aggregation result of the correction amounts corresponding to each correction type in the minimum correction set, and the correction type with the largest aggregated correction amount is taken as the predicted fault category. The predicted fault result is output after it is determined that the target electrical equipment is in the critical pre-fault state.

[0013] On the other hand, the present invention provides an electrical equipment fault prediction system based on digital twins, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the aforementioned electrical equipment fault prediction method based on digital twins.

[0014] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects: 1. In this invention, based on the synergistic effect of the health digital twin, the coded excitation sequence and the reference echo constraint set, the local contact changes, local insulation changes and local response anomalies that are not easily revealed by the target electrical equipment under natural operating conditions can be transformed into echo response results that can be obtained in segments, aligned segment by segment and compared uniformly. 2. In this invention, by performing a local correction search on the echo response sequence to form a minimum correction set, and performing a single-step misalignment pressure test along the sensitive constraint axis, it is possible to determine whether the target electrical equipment has entered the critical pre-fault state, and further realize the output of predicting the fault location and predicting the fault type. Attached Figure Description

[0015] Figure 1 This is a flowchart of one embodiment of the present invention. Detailed Implementation

[0016] Example 1, as Figure 1 As shown, the present invention proposes a method for predicting electrical equipment faults based on digital twins, and its specific implementation steps are as follows: S1. Obtain the basic equipment data, acquisition node configuration data, excitation configuration data and segmentation rule data of the target electrical equipment. Based on the basic equipment data, acquisition node configuration data and excitation configuration data, construct a health digital twin and establish local constraint parameters in the health digital twin to generate coded excitation sequences and reference echo constraint sets. Among them, the health digital twin is a digital twin reference body that characterizes the excitation-response benchmark relationship under the health state of the target electrical equipment; the local constraint parameters are the constraint parameter items set in the health digital twin corresponding to the local response units; and the benchmark echo constraint set is the benchmark constraint set set corresponding to the coded excitation sequence. S2. Input the coded excitation sequence into the target electrical equipment, acquire the echo response data at the acquisition node, form the echo response sequence according to the segmentation rule data, align the echo response sequence with the reference echo constraint set segment by segment to obtain the alignment result, and perform local correction search on the local constraint parameters associated with the alignment result based on the health digital twin to obtain the minimum correction set and determine the sensitive constraint axis. Among them, the minimum correction set is the set of the fewest correction terms that make the echo response sequence meet the requirements of the reference echo constraint set; the sensitive constraint axis is the constraint direction in the minimum correction set that corresponds to the main mismatch direction of the current echo response. S3. Based on the sensitive constraint axis, the coded excitation sequence is processed by single-axis misalignment to generate a single-step misaligned pressure test sequence. The single-step misaligned pressure test sequence is input to the target electrical equipment. Pressure test echo response data is acquired at the acquisition node. The pressure test echo response sequence is generated according to the segmentation rule data. The pressure test echo response sequence is aligned segment by segment with the pressure test reference echo constraint set to obtain the pressure test alignment result. Based on the health digital twin, the local constraint parameters associated with the pressure test alignment result are locally corrected and searched to generate the minimum correction set of the pressure test. Among them, the single-step misaligned stress test sequence is the stress test excitation sequence formed after the encoded excitation sequence is misaligned once along the sensitive constraint axis; the minimum correction set of stress test is the minimum set of correction terms determined for the stress test echo response sequence; S4. Compare the minimum correction set of the pressure test with the minimum correction set to determine the critical pre-fault state of the target electrical equipment, and output the predicted fault result based on the minimum correction set of the pressure test. Among them, the critical pre-fault state is the pre-fault state in which the target electrical equipment has not yet entered the manifest fault state and no longer meets the health state requirements under single-step misalignment pressure test conditions.

[0017] In this embodiment S1, the acquisition nodes are data acquisition nodes set at each local response position of the target electrical equipment; the health digital twin performs digital twin mapping on each local response position of the target electrical equipment based on the equipment basic data, and the health digital twin is set with multiple local response units, which are response units in the health digital twin that correspond to different local response positions of the target electrical equipment; the local constraint parameters are set according to the local response units in groups, each group of local constraint parameters corresponds to one local response unit, different local response units correspond to different groups of local constraint parameters, and each local response unit establishes a correspondence with the acquisition node; the excitation configuration data is used to limit the arrangement order and path correspondence of each coded excitation segment, the coded excitation sequence is formed by arranging multiple coded excitation segments in the order determined by the excitation configuration data, and each coded excitation segment establishes a correspondence with the local response unit; the reference echo constraint set is set in segments according to the coded excitation sequence, each reference echo constraint is set according to the local constraint parameters of the corresponding local response unit, and each reference echo constraint is associated with the corresponding coded excitation segment, local response unit and acquisition node respectively.

[0018] In this embodiment S1, the target electrical equipment can be a switchgear, circuit breaker, transformer, cable terminal, relay protection device, motor control unit, or other electrical equipment with excitation transmission and response acquisition capabilities. The equipment basic data describes the basic structure and basic state of the target electrical equipment, specifically including the compositional relationships of equipment components, conductive connection relationships, insulation isolation relationships, local response location distribution relationships, rated operating parameters, and basic configuration relationships related to the excitation transmission path. The acquisition node configuration data describes the location of the acquisition nodes, node numbers, the correspondence between nodes and local response locations, and the nodes under different excitation paths. The acquisition covers the coverage relationship; the excitation configuration data describes the arrangement order of the coded excitation segments, the connection relationship between segments, the correspondence of excitation paths, and the scope of action of the excitation segments; the segmentation rule data describes the segment boundary division basis, the segment sequence formation basis, and the identification rules of different echo segments in the echo response data; the above types of data can be obtained from equipment design data, factory configuration data, installation and commissioning data, online configuration tables, or operation and maintenance files, and are uniformly read, organized, and standardized by the control terminal, edge computing unit, or detection host to form the basic data set required for subsequent construction of the health digital twin, generation of coded excitation sequences, and establishment of the benchmark echo constraint set.

[0019] In this embodiment S1, the health digital twin is established around the excitation response benchmark relationship of the target electrical equipment in a healthy state. Its establishment process is based on the local response positions determined by the equipment's basic data, supported by the node coverage relationship determined by the acquisition node configuration data, and constrained by the excitation segment sequence and path relationship determined by the excitation configuration data. This forms a local mapping structure corresponding to the target electrical equipment in the digital space. Each local response position is a positional unit in the target electrical equipment that has independent response significance to the coded excitation input, echo propagation, echo reflection, and echo recovery. These positional units can be divided according to the equipment's internal structural boundaries, connection boundaries, insulation boundaries, path turning boundaries, or detection interest boundaries. After completing the local response location division, multiple local response units are further set in the health digital twin. Each local response unit corresponds to the digital response representation object of the local response location in the health state, and is used to carry the reference response relationship, constraint relationship and node association relationship of the location under the action of different coded excitation segments. The actual setting position of the acquisition node in the target electrical equipment is mapped to the local response unit in the health digital twin, so that the echo response data acquired by the acquisition node on the physical equipment side can find the corresponding digital response unit in the health digital twin, thereby ensuring that the subsequent echo response sequence and the segment-by-segment correspondence alignment between the reference echo constraint set have a clear object basis and position basis.

[0020] In this embodiment S1, the local constraint parameters are a set of parameter items established around each local response unit within the healthy digital twin. Their purpose is to limit the local response constraints that each local response unit should satisfy in response to the encoded stimulus input under healthy conditions. The local constraint parameters are set in groups according to local response units, with each group corresponding to one local response unit. Different local response units correspond to different groups of local constraint parameters, thus forming a parameter grouping structure with local response units as the basic organizational unit. Each group of local constraint parameters can respectively correspond to propagation constraints, reflection constraints, recovery constraints, and their local constraints related to node acquisition relationships and path transmission relationships. The constraints are defined as follows: multiple parameters within the same group jointly limit the local echo behavior of the corresponding local response unit in a healthy state; different groups limit the echo behavior at different local response locations. Since each local response unit has established a correspondence with the acquisition node, the local constraint parameters are not only related to the corresponding local response unit, but also indirectly related to the acquisition node through the local response unit. This allows for accurate location of the parameter group range to be searched based on the alignment results of the echo response sequence during subsequent local correction searches, without having to indiscriminately process all parameters in the healthy digital twin, thus ensuring that the correction search object has clear boundaries.

[0021] In this embodiment S1, the coded excitation sequence is formed by arranging multiple coded excitation segments in an order determined by the excitation configuration data. Each coded excitation segment is a basic excitation unit in the coded excitation sequence, used to apply distinctive excitation inputs to the target electrical equipment at different segment positions. The excitation configuration data limits the arrangement order and path correspondence of the coded excitation segments, so that each coded excitation segment forms a predetermined configuration in terms of input order, action path, and target range. Each coded excitation segment establishes a correspondence with a local response unit, which can be based on the passage position, action position, and acquisition coverage position of the excitation path in the target electrical equipment. This is to determine the different coded excitation segments so that they can form distinguishable response excitation effects for different local response units. After the coded excitation sequence is generated, the subsequently collected echo response data can be divided and associated at the segment level according to the order and path logic of the coded excitation segments, thereby providing a serialized input basis for forming the echo response sequence, establishing segment-level correspondence, and performing local correction search. In this embodiment, the organization of the coded excitation sequence emphasizes segment-level distinguishability, sequential traceability, and path mapping to ensure that the echo response data can be traced back to the corresponding coded excitation segment and the corresponding local response unit in subsequent processing.

[0022] In this embodiment S1, the reference echo constraint set is a set of reference constraints further formed based on the health digital twin and excitation configuration data after the coded excitation sequence is generated. This reference constraint set is set in segments according to the coded excitation sequence, with each reference echo constraint corresponding to a coded excitation segment, and is established based on the local response unit and its local constraint parameters associated with the coded excitation segment. Since each local response unit has established a correspondence with the acquisition node, each reference echo constraint is simultaneously associated with the corresponding coded excitation segment, local response unit, and acquisition node, thereby forming a three-in-one set of segment-level reference constraints, local response positions, and node acquisition positions. A unified constraint structure is established. Under this structure, the coded excitation segment determines the segment order and path range of the excitation input, the local response unit determines the local response constraints that should be satisfied under healthy conditions, and the acquisition node determines the physical acquisition location of subsequent response sampling. The three together determine the boundary and content of a single reference echo constraint. Multiple reference echo constraints are arranged according to the segment order of the coded excitation sequence to form a reference echo constraint set. This constraint set is used in subsequent steps to perform segment-by-segment alignment with the echo response sequence and serves as a reference object for local correction search, so as to ensure that the formation of the subsequent minimum correction set is based on the unified constraints of the healthy state.

[0023] In this embodiment S2, the echo response data is the node response data collected by the acquisition node under the action of the coded excitation sequence of the target electrical equipment; the echo response sequence is a sequential response sequence formed by arranging the echo response data into segments according to segmentation rule data; the specific method for forming the echo response sequence according to the segmentation rule data is as follows: the echo response data is divided into boundaries according to the segment order corresponding to the coded excitation sequence to form multiple echo segments, and the echo response sequence is formed according to the segment order corresponding to each echo segment; wherein, the echo segment includes the propagating echo segment, the reflected echo segment, and the recovered echo segment; the echo... The specific method for aligning the response sequence with the reference echo constraint set segment by segment is as follows: The reference echo constraint set is divided into multiple constraint segments according to the segment order of the coded excitation sequence, and each echo segment and each constraint segment are matched at the segment level according to the same segment order; the alignment result is a set of segment-level correspondence results formed by the correspondence between each echo segment and each constraint segment, and each correspondence item in the alignment result includes an echo segment identifier, a constraint segment identifier, a path identifier, and a node identifier; the path identifier in the alignment result is used to characterize the coded excitation path to which the corresponding item belongs, and the node identifier is used to characterize the acquisition node to which the corresponding item belongs.

[0024] In this embodiment S2, after inputting the coded excitation sequence to the target electrical equipment, each acquisition node collects the node response generated by the target electrical equipment under the excitation according to the acquisition position corresponding to the acquisition node configuration data. The obtained data constitutes echo response data. The echo response data takes the acquisition node as the acquisition position, the segment order of the coded excitation sequence as the time organization basis, and the internal excitation propagation, local reflection, and recovery regression process of the target electrical equipment as the response source. Therefore, the node response data collected by each acquisition node has a common segment order under the action of the same coded excitation segment, and has distinguishable segment-level boundaries under the action of different coded excitation segments. When forming the echo response sequence according to the segmentation rule data, the arrangement order of each coded excitation segment in the coded excitation sequence is read first, and the echo response data is arranged according to the corresponding... The segment sequence is divided into boundaries so that the node responses formed during the same coded excitation segment are grouped into the same echo segment. Following the pre-defined segment boundary identification rules, inter-segment transition rules, and segment order arrangement rules in the segmentation rule data, each echo segment is processed at the segment level to form an echo response sequence with a clear chronological order. The echo segments include propagating echo segments, reflected echo segments, and recovering echo segments. The propagating echo segment corresponds to the path propagation response interval after the coded excitation input; the reflected echo segment corresponds to the return response interval caused by local structures in the propagation path; and the recovering echo segment corresponds to the interval of the node response regression process after the excitation ends. All three types of echo segments are uniformly included in the echo response sequence according to the segment order of the coded excitation segment to ensure a one-to-one correspondence can be established at the same segment level during subsequent alignment processing.

[0025] In this embodiment S2, when aligning the echo response sequence with the reference echo constraint set segment by segment, the segment order of the coded excitation sequence is used as the unified correspondence benchmark. The reference echo constraint set is divided into multiple constraint segments according to the same segment order as the coded excitation sequence. Then, each echo segment in the echo response sequence is matched with each constraint segment in the reference echo constraint set according to the same segment order. The above segment-level correspondence does not perform an overall comparison of the entire data segment, but rather establishes a mapping item by item between the propagation, reflection, and recovery response relationships of the echo segment and the segment-level constraint relationships in the corresponding constraint segment under the same segment order, thereby forming the alignment result. The alignment result consists of multiple segment-level correspondence items, and each correspondence item includes at least an echo segment identifier, a constraint segment identifier, and a path identifier. The system includes an echo segment identifier, a constraint segment identifier, a path identifier, and a node identifier. The echo segment identifier indicates the echo segment sequence and category to which the corresponding item belongs. The constraint segment identifier indicates the constraint segment sequence. The path identifier represents the encoding stimulus path to which the corresponding item belongs. The node identifier represents the acquisition node to which the corresponding item belongs. Since the path identifier and node identifier together define the response source path and response acquisition location, the alignment result not only represents the segment-level correspondence between echo segments and constraint segments but also serves as an index for local constraint parameters in subsequent local correction searches. This allows the local correction search to be performed directly on the parameter items related to the current segment-level correspondence without extending to the entire parameter range in the health digital twin.

[0026] In this embodiment S2, the local constraint parameters associated with the alignment result are determined based on the path identifier and node identifier in the alignment result; the local correction search is a constraint correction search operation performed item by item on the local constraint parameters, used to determine the correction terms corresponding to the constraint mismatch of the current echo response sequence relative to the reference echo constraint set; the specific method for performing local correction search on the local constraint parameters associated with the alignment result based on the health digital twin is as follows: candidate correction terms are generated for the local constraint parameters associated with the alignment result, and correction term combinations are constructed based on the candidate correction terms. Constraint matching judgment is performed on each correction term combination to determine the correction term combination that meets the matching requirements between the echo response sequence and the reference echo constraint set; the minimum correction set is the set after the local correction search. A defined minimum set of correction terms is used to characterize the constraint correction distribution corresponding to the current echo response sequence meeting the requirements of the reference echo constraint set. The minimum set of correction terms includes correction position, correction type, and correction amount. The minimum set of correction terms is determined by having the fewest correction terms as the first determining condition. If the number of correction terms is the same, the minimum constraint matching error is determined by having the smallest constraint matching error as the second determining condition. The correction position corresponds to the local response unit, the correction type corresponds to the local constraint parameter category, and the correction amount corresponds to the offset of the local constraint parameter. The number of correction terms is the number of correction terms included in the minimum set of correction terms, and the constraint matching error is the error amount of the remaining constraint deviation between the echo response sequence and the reference echo constraint set. Multiple correction terms in the minimum set of correction terms are arranged in order of correction amount to form a correction term sequence.

[0027] In this embodiment S2, the local constraint parameters associated with the alignment result are determined based on the path identifier and node identifier in the alignment result. Specifically, firstly, the path identifier of each corresponding item in the alignment result is read to determine the coded excitation path to which the corresponding item belongs. Then, the node identifier of the corresponding item is read to determine the acquisition coverage position of the coded excitation path in the target electrical equipment. The determined coded excitation path and acquisition coverage position are compared with the mapping relationship of local response units in the health digital twin. Local constraint parameters related to the current corresponding item are extracted from the local response units corresponding to the coded excitation path and acquisition coverage position, thereby forming a set of local constraint parameters associated with the alignment result. The association is a mapping relationship. The mapping association means that the path identifiers and node identifiers in the alignment result establish a deterministic correspondence with the local constraint parameters through local response units, rather than directly splicing or logically inferring the echo response data and local constraint parameters. Therefore, different corresponding items in the same alignment result can be associated with different sets of local constraint parameters. The corresponding items in different alignment results correspond to different local response units and different sets of local constraint parameters when the path identifiers or node identifiers are different. The set of local constraint parameters only covers the segment-level mapping range corresponding to the current echo response sequence and does not cover all local constraint parameters in the healthy digital twin, thereby limiting the subsequent local correction search to the local constraint range pointed to by the current alignment result.

[0028] In this embodiment S2, the local correction search is a constraint correction search operation performed item by item on the local constraint parameters associated with the alignment result. Specifically, it includes three consecutive processing stages: candidate correction item generation, correction item combination construction, and constraint matching determination. During candidate correction item generation, a corresponding candidate correction item is established for each associated local constraint parameter. Each candidate correction item records at least the correction position, correction type, and correction amount. The correction position is determined by the local response unit to which the local constraint parameter belongs, the correction type is determined by the constraint category to which the local constraint parameter belongs, and the correction amount is determined by the offset of the local constraint parameter relative to the health state constraint requirement. After the candidate correction item generation is completed, the same path is followed... Multiple candidate correction items are organized using path identifiers, adjacent node identifiers, or the attribution relationship of the same local response unit to form multiple correction item combinations. These correction item combinations are not arbitrarily aggregated, but rather established around the segment-level correspondence between the current echo response sequence and the reference echo constraint set. Subsequently, constraint matching determination is performed on each correction item combination. The constraint matching determination is based on whether the remaining constraint deviation between the echo response sequence and the reference echo constraint set after the intervention of the correction item combination meets the matching requirements. Correction item combinations that meet the matching requirements are retained, while those that do not meet the matching requirements are eliminated. This yields the range of candidate correction item combinations that meet the current alignment result constraint requirements.

[0029] In this embodiment S2, the minimum correction set is determined within the range of candidate correction term combinations that satisfy the constraint matching requirements. The minimum correction set is the minimum set of correction terms determined after local correction search, which represents the constraint correction distribution corresponding to the current echo response sequence satisfying the baseline echo constraint set requirements. The minimum correction set includes correction position, correction type, and correction amount, where the correction position corresponds to the local response unit, the correction type corresponds to the local constraint parameter category, and the correction amount corresponds to the offset of the local constraint parameter. The determination of the minimum correction set is performed under two conditions in sequence: the first condition is that the number of correction terms is the minimum, where the number of correction terms is the number of correction terms included in the minimum correction set; the second condition is that the constraint matching error is the minimum, where the constraint matching error is the difference between the echo response sequence and the baseline echo constraint. The error of the residual constraint deviation between sets is specifically processed by first selecting the combination with the fewest correction items from all combinations of correction items that meet the constraint matching requirements, and then selecting the combination with the smallest constraint matching error from the combinations with the same number of correction items. This combination is then determined as the minimum correction set. After the minimum correction set is determined, multiple correction items are sorted according to the magnitude of the correction amount to form a correction item sequence. The magnitude of the correction amount is specifically limited to sorting them from largest to smallest by the absolute value of the correction amount of each correction item. If the absolute values ​​of the correction amounts are the same, they are sorted from largest to smallest by the decrease in constraint matching error caused by the corresponding correction item. If the decrease in constraint matching error is still the same, they are sorted from front to back by the segment order of the corresponding correction item's associated coded excitation segment in the coded excitation sequence.

[0030] In this embodiment S2, the sensitive constraint axis is a single constraint direction in the minimum correction set corresponding to the main mismatch direction of the current echo response. It is used to identify the dominant mismatch direction of the current echo response sequence and to perform sorting and extraction on each correction item in the minimum correction set. When sorting each correction item in the minimum correction set, the sensitive constraint axis is sorted according to the magnitude of the correction amount of each correction item, and the local constraint parameter corresponding to the first correction item in the sort is taken. The sensitive constraint axis is then determined together with the coded excitation segment associated with the local constraint parameter. Only one sensitive constraint axis is determined, and it is set separately from the constraint directions corresponding to the other correction items in the minimum correction set. It uniquely indicates the main mismatch direction of the current echo response sequence. The coded excitation segment corresponding to the sensitive constraint axis is the unique target coded excitation segment for single-axis misalignment processing.

[0031] In this embodiment S2, the sensitive constraint axis is a single constraint direction extracted from the minimum correction set. It corresponds to the main mismatch direction in the current echo response sequence. The extraction basis of the sensitive constraint axis is the sequence of correction items in the minimum correction set, rather than the parallel values ​​of all correction directions. Specifically, after the minimum correction set has formed a sequence of correction items according to the magnitude of the correction amount, the first correction item in the sequence is first identified. The first correction item is the correction item that is first under the constraint of the magnitude of the correction amount. The local constraint parameter corresponding to this correction item is determined as the current main mismatch parameter. Then, the associated coded excitation segment of this local constraint parameter in the healthy digital twin is read, and the sensitive constraint axis is determined by the constraint direction represented by the local constraint parameter and the segment-level positional relationship of the associated coded excitation segment. Thus, the sensitive constraint axis contains both the constraint direction attribute and the coded excitation segment positioning attribute. The sensitive constraint axis is not a general representation of all correction item directions, but a single direction object determined by a single dominant correction item.

[0032] In this embodiment S2, only one sensitive constraint axis is determined and is set separately from the constraint directions corresponding to the other correction terms in the minimum correction set. Specifically, although the other correction terms in the minimum correction set also reflect the local constraint deviation between the echo response sequence and the reference echo constraint set, their sorting position in the correction term sequence is not at the first position, so they are not used as the dominant mismatch direction of the current echo response sequence. The uniqueness of the sensitive constraint axis is guaranteed by the unique determination result of the first correction term in the correction term sequence. Once the first correction term is determined, its corresponding local constraint parameters and associated coded excitation segments are fixed together, and the sensitive constraint axis is fixed accordingly. The uniqueness is reflected in the fact that there is only one sensitive constraint axis that can be used as the dominant mismatch direction within the same minimum correction set, and this sensitive constraint axis exists in parallel with the other correction directions but is not selected in parallel, so that the dominant direction, secondary direction and auxiliary direction among the multiple mismatch directions in the current echo response sequence form a clear distinction relationship.

[0033] In this embodiment S2, after the sensitive constraint axis is extracted from the minimum correction set, it continues to serve as the basis for subsequent single-axis misalignment processing. Specifically, the coded excitation segment corresponding to the sensitive constraint axis is determined as the unique target coded excitation segment for single-axis misalignment processing. The remaining coded excitation segments maintain their original segment order and configuration in subsequent single-axis misalignment processing. The difference between the unique target coded excitation segment and the remaining coded excitation segments is that the unique target coded excitation segment directly corresponds to the main mismatch direction in the current echo response sequence, while the remaining coded excitation segments correspond only to non-dominant mismatch directions or local response relationships that have not been extracted as sensitive directions. Therefore, when performing subsequent single-axis misalignment processing, only the coded excitation segment corresponding to the sensitive constraint axis is subjected to a single misalignment adjustment, and a single-step misalignment stress test sequence is formed accordingly, so that the single-axis misalignment processing object, the sensitive constraint axis, and the dominant mismatch direction are consistent.

[0034] In this embodiment S3, single-axis misalignment processing is a single-time misalignment processing method performed on the unique target coded excitation segment in the coded excitation sequence, used to generate a single-step misalignment stress test sequence. The specific method for performing single-axis misalignment processing on the coded excitation sequence based on the sensitive constraint axis is as follows: determine the unique target coded excitation segment corresponding to the sensitive constraint axis, keep the remaining coded excitation segments in the coded excitation sequence unchanged, perform a single misalignment adjustment on the unique target coded excitation segment, and combine the misaligned unique target coded excitation segment with the remaining coded excitation segments in the original segment order to form a single-step misalignment stress test sequence. The single-step misalignment stress test sequence is a stress test excitation sequence formed after the coded excitation sequence performs one misalignment processing along the sensitive constraint axis, used to obtain the stress test echo response data of the target electrical equipment under stress test conditions. The single-step misalignment stress test sequence includes the target coded excitation segment after single-axis misalignment processing and the remaining coded excitation segments that have not undergone misalignment processing. The segment order of the coded excitation sequence remains consistent before and after the single-axis misalignment processing.

[0035] In this embodiment S3, the sensitive constraint axis is no longer regenerated, nor is it repeatedly screened for all correction directions. Instead, it directly adopts the dominant mismatch direction result already determined within the minimum correction set. The first correction item in the minimum correction set corresponds to a local constraint parameter. This local constraint parameter is associated with an coded excitation segment in the healthy digital twin. This coded excitation segment is the unique target coded excitation segment corresponding to the sensitive constraint axis. The unique target coded excitation segment is a segment-level excitation unit in the coded excitation sequence that directly corresponds to the current major mismatch direction, distinguished from other coded excitation segments that only participate in excitation transmission but are not extracted as the main mismatch direction carrying segment. Therefore, in single-axis misalignment processing, "single axis" refers to a single constraint direction defined along the sensitive constraint axis. Instead of performing segment-level processing, synchronous processing along multiple constraint directions, or overall reconstruction of the entire coded stimulus sequence, the "target" in the single-axis misalignment processing refers to processing only the unique target coded stimulus segment, without performing misalignment adjustments on other coded stimulus segments. The "single" in the single-axis misalignment processing refers to performing a misalignment adjustment only once on the unique target coded stimulus segment within a stress test processing cycle, without recursive misalignment, repeated iterative misalignment, or performing multiple offset processing on multiple coded stimulus segments in the same processing cycle. This ensures a single correspondence between the sensitive constraint axis, the unique target coded stimulus segment, and the single misalignment adjustment, and enables subsequent single-step misalignment stress test sequences to fully inherit the main mismatch direction information corresponding to the minimum correction set.

[0036] In this embodiment S3, when performing single-axis misalignment processing on the coded stimulus sequence based on the sensitive constraint axis, the unique target coded stimulus segment associated with the sensitive constraint axis is first read and identified from the original segment order structure of the coded stimulus sequence. Then, the segment order relationship, inter-segment connection relationship, and non-target segment configuration relationship of the remaining coded stimulus segments in the coded stimulus sequence remain unchanged. Only the unique target coded stimulus segment is subjected to a single misalignment adjustment. This single misalignment adjustment acts on the segment-level configuration position of the unique target coded stimulus segment, causing the target coded stimulus segment to be misaligned within a limited range in the original coded stimulus sequence, without changing the arrangement order and inter-connection of the remaining coded stimulus segments in the coded stimulus sequence. After completing a single misalignment adjustment, the unique target coded excitation segment after misalignment adjustment is recombined with the remaining coded excitation segments that have not undergone misalignment processing according to their original segment order to form a single-step misaligned stress test sequence. The recombination does not involve recalculating a new coded segment order or reordering all coded excitation segments. Instead, it involves placing the unique target coded excitation segment after the single misalignment adjustment back into the original segment order structure while retaining the original segment order framework. This ensures that the single-step misaligned stress test sequence and the coded excitation sequence remain isomorphic in the overall segment order structure, with a single misalignment difference only at the segment level position corresponding to the unique target coded excitation segment. This allows subsequent stress test echo response data to have a comparable segment level basis with normal echo response data.

[0037] In this embodiment S3, the single-step misaligned stress test sequence is composed of a unique target coded excitation segment after single-axis misalignment processing and the remaining coded excitation segments that have not undergone misalignment processing. The unique target coded excitation segment after single-axis misalignment processing carries the stress test disturbance information corresponding to the current major mismatch direction. The remaining coded excitation segments that have not undergone misalignment processing maintain the original segment-level configuration in the coded excitation sequence to preserve the original excitation background except for the major mismatch direction. The single-step misaligned stress test sequence is still a segment-level derived sequence of the coded excitation sequence, rather than an independent test sequence completely separated from the coded excitation sequence. Its sequence composition structure is consistent with the coded excitation sequence, the only difference being the target coded excitation segment. Since the single misalignment adjustment has been completed, after inputting the single-step misalignment stress test sequence to the target electrical equipment, the target electrical equipment will produce a segment-level response change in the local response direction corresponding to the unique target coded excitation segment that is different from the normal coded excitation condition. However, in the local response directions corresponding to the other coded excitation segments, the excitation basis remains consistent with the normal coded excitation condition. This allows the single-step misalignment stress test sequence to retain the overall segment-level framework of the normal excitation sequence while applying a single-step limited segment-level disturbance in the dominant misalignment direction. As a result, the stress test echo response data subsequently acquired by the acquisition node has stress test condition characteristics that are isomorphic to the coded excitation sequence, correspond to the sensitive constraint axis, and are consistent with the single misalignment processing.

[0038] In this embodiment S3, the segment order of the coded stimulus sequence remains consistent before and after the single-axis misalignment process. This consistency means that the order of each coded stimulus segment in the coded stimulus sequence is not reordered due to a single misalignment adjustment, nor is the relative position of the remaining coded stimulus segments in the original sequence changed due to the misalignment of the unique target coded stimulus segment. The single-axis misalignment process only changes the segment-level processing state of the unique target coded stimulus segment and does not change the segment order structure of the entire sequence. Therefore, when the single-step misaligned stress test sequence is subsequently used to form the stress test echo response sequence, the boundary division and segment order arrangement can still be performed according to the same segmentation rule data as the coded stimulus sequence. Furthermore, when the stress test reference echo constraint set is subsequently generated, the segment-level correspondence logic continues to be used, enabling the stress test echo response sequence and the stress test reference echo constraint set to perform segment-by-segment correspondence alignment. The consistency of the segment order also limits the processing boundary of the single-axis misalignment process, that is, the single-axis misalignment process only performs a single misalignment adjustment on the unique target coded stimulus segment.

[0039] In this embodiment S3, the pressure test echo response data is the node response data collected by the acquisition node under the action of the single-step misaligned pressure test sequence of the target electrical equipment, which is used to form the pressure test echo response sequence. The pressure test echo response sequence is a sequential response sequence formed by arranging the pressure test echo response data into segments according to the segmentation rule data, which is used as the pressure test input sequence for local correction search. The pressure test echo response sequence includes the pressure test propagation echo segment, the pressure test reflection echo segment, and the pressure test recovery echo segment. The specific method for forming the pressure test echo response sequence according to the segmentation rule data is as follows: the pressure test echo response data is divided into boundaries according to the segment order corresponding to the single-step misaligned pressure test sequence to form multiple pressure test echo segments. Each pressure test echo segment is assigned a corresponding segment identifier, and the corresponding segment identifier is used to form the signal. The segmented arrangement forms a stress test echo response sequence; the stress test reference echo constraint set is a set of stress test constraints generated by the health digital twin based on the single-step misaligned stress test sequence, and the stress test reference echo constraint set is segmented according to the single-step misaligned stress test sequence; the specific method for aligning the stress test echo response sequence and the stress test reference echo constraint set segment by segment is as follows: the stress test reference echo constraint set is divided into multiple stress test constraint segments according to the segmented order of the single-step misaligned stress test sequence, and each stress test echo segment and each stress test constraint segment are matched at the segment level according to the same segmented order; the stress test alignment result is a set of segment-level correspondence results formed by the correspondence between each stress test echo segment and each stress test constraint segment; each stress test echo segment in the stress test echo response sequence is matched one-to-one with each coded excitation segment in the single-step misaligned stress test sequence.

[0040] In this embodiment S3, the pressure test echo response data is the node response data collected by the acquisition nodes under the action of the single-step misaligned pressure test sequence. The acquisition nodes complete the data acquisition at each local response position of the target electrical equipment according to the acquisition node configuration data. The node response data acquired by each acquisition node is collected according to the corresponding local response position, the corresponding coded excitation segment, and the corresponding excitation path, thereby forming a node-level response data set under the pressure test state. The node-level response data set covers the propagation response during the action of the single-step misaligned pressure test sequence, the reflection response at the local structural position, and the recovery response after the excitation action ends. Moreover, the node response data corresponding to each acquisition node has a clear acquisition position identifier and segment-level attribution relationship, so that the pressure test echo response data has a unified foundation in terms of data source, path attribution, and segment sequence organization, providing original data support for the subsequent formation of the pressure test echo response sequence.

[0041] In this embodiment S3, the stress test echo response sequence is formed by performing segment-level division and sequential arrangement of the stress test echo response data according to the segmentation rule data. Specifically, the stress test echo response data is first read according to the coded excitation segment order in the single-step misaligned stress test sequence. Then, the starting boundary and ending boundary of each response interval are determined according to the boundary division rules in the segmentation rule data, so that the node responses formed during the action of the same coded excitation segment are classified into the same stress test echo segment. Subsequently, each stress test echo segment is assigned a corresponding segment identifier, and each stress test echo segment is arranged in sequence according to the segment order corresponding to the segment identifier, thereby forming a stress test echo response sequence. The stress test echo response sequence includes a stress test propagation echo segment, a stress test reflection echo segment, and a stress test recovery echo segment. The stress test propagation echo segment corresponds to the response interval formed by the transmission along the excitation path after the excitation input. The stress test reflection echo segment corresponds to the return response interval at the local response position. The stress test recovery echo segment corresponds to the interval where the node response return process is located. All stress test echo segments together constitute a stress test input sequence with a unified segment order structure.

[0042] In this embodiment S3, the load testing reference echo constraint set is generated by the health digital twin based on the single-step misaligned load testing sequence. Each load testing reference echo constraint is segmented according to the order of the coded excitation segments in the single-step misaligned load testing sequence. Each load testing reference echo constraint corresponds to a coded excitation segment, a local response unit, and a set of acquisition nodes associated with the local response unit. After reading the target coded excitation segment configuration and the configuration of the remaining coded excitation segments in the single-step misaligned load testing sequence, the health digital twin establishes the segment-level constraint relationship under the load testing state by combining the local constraint parameters corresponding to each local response unit. This forms multiple load testing constraint segments. The load testing constraint segments are arranged according to the order of the coded excitation segments and together constitute the load testing reference echo constraint set. The load testing reference echo constraint set carries the response constraints, path constraints, and node constraints that each segment should satisfy under the load testing state and is consistent with the load testing echo response sequence at the segment order level.

[0043] In this embodiment S3, when aligning the stress test echo response sequence with the stress test reference echo constraint set segment by segment, firstly, each stress test constraint segment in the stress test reference echo constraint set is identified according to the segment order corresponding to the single-step misaligned stress test sequence. Then, each stress test echo segment in the stress test echo response sequence is established with the stress test constraint segment of the same segment order to form a segment-level correspondence relationship, thereby forming a stress test alignment result. The stress test alignment result consists of multiple segment-level correspondence items. Each segment-level correspondence item includes at least a stress test echo segment identifier, a stress test constraint segment identifier, a path identifier, and a node identifier. The stress test echo segment identifier is used to record the position and category of the stress test echo segment to which the current correspondence item belongs. The stress test constraint segment identifier is used to record the position of the stress test constraint segment to which the current correspondence item belongs. The path identifier is used to record the encoding excitation path to which the current correspondence item belongs. The node identifier is used to record the acquisition node to which the current correspondence item belongs. All segment-level correspondence items together constitute the stress test alignment result set and serve as the direct basis for subsequently determining the local stress test constraint parameters and performing the local stress test correction search.

[0044] In this embodiment S3, the minimum correction set for stress testing is the set of the fewest correction terms determined for the stress test echo response sequence. It is used to characterize the correction distribution corresponding to the stress test echo response sequence meeting the constraints of the health digital twin under single-step misalignment stress test conditions. The minimum correction set for stress testing includes correction position, correction type, and correction amount. The structure of the correction terms in the minimum correction set for stress testing is consistent with that of the minimum correction set. The minimum number of correction terms is used as the first determining condition. When the number of correction terms is the same, the minimum constraint matching error is used as the second determining condition. The specific method for generating the minimum correction set for stress testing by performing local correction search on the local constraint parameters associated with the stress test alignment results based on the health digital twin is as follows: the corresponding local constraint parameters are determined according to the path identifier and node identifier in the stress test alignment results. Stress test candidate correction terms are generated for the corresponding local constraint parameters. Stress test correction term combinations are constructed based on the stress test candidate correction terms. Constraint matching judgment is performed on each stress test correction term combination to determine the stress test correction term combination that meets the matching requirements of the stress test echo response sequence and the stress test reference echo constraint set. The minimum correction set for stress testing is determined from the stress test correction term combinations that meet the requirements.

[0045] In this embodiment S3, the minimum correction set for stress testing is determined for the stress test echo response sequence. Its formation is based on the segment-level correspondence between the stress test alignment result and the stress test reference echo constraint set. Specifically, after completing the segment-level correspondence, each stress test echo segment in the stress test echo response sequence obtains its corresponding path identifier, node identifier, and stress test constraint segment identifier, and maps them to the corresponding local response unit and corresponding local constraint parameters within the health digital twin. This forms the range of local constraint parameters under stress testing conditions. After performing a local correction search within this parameter range, multiple sets satisfying the stress test requirements can be obtained. The combination of stress test correction terms required for constraint matching is selected from the above combination of stress test correction terms. The minimum set of stress test correction terms represents the correction distribution state corresponding to the stress test echo response sequence reaching the constraint requirements of a healthy digital twin under single-step misalignment stress test conditions. Each correction term has a clear correspondence in local response position, local constraint category and parameter offset magnitude. The entire set of correction distribution is established based on the segment-level response change under stress test conditions. Therefore, the minimum set of stress test correction terms has the structural characteristics of clear stress test path, clear node source, clear correction range and clear segment-level affiliation.

[0046] In this embodiment S3, the minimum correction set for stress testing includes correction location, correction type, and correction amount. Correction location indicates the local response unit to which the stress test constraint deviation falls; correction type indicates the category of local constraint parameter to which the stress test constraint deviation belongs; and correction amount indicates the offset of the corresponding local constraint parameter relative to the health state constraint requirement. The correction item structure in the minimum correction set for stress testing maintains the same organizational method as the minimum correction set, so that in subsequent comparison processing, quantity statistics, location aggregation, and type aggregation are performed according to the same field rules. The determination of the minimum correction set for stress testing adopts a two-level conditional selection process. The first determination condition is the number of correction items. The first condition is to minimize the number of correction items, where the number of correction items is the count result of the correction items in the stress test correction item combination. The second condition is to minimize the constraint matching error, where the constraint matching error is the error of the remaining constraint deviation between the stress test echo response sequence and the stress test reference echo constraint set. In the specific process, firstly, the number of correction items is filtered for all stress test correction item combinations that meet the stress test constraint matching requirements. Then, the constraint matching error is filtered for combinations with the same number of correction items. Finally, the stress test correction item combination that meets both of the above conditions is retained as the minimum stress test correction set. This makes the minimum stress test correction set have unified rules at the structural, screening, and statistical levels.

[0047] In this embodiment S3, when performing local correction search on the local constraint parameters associated with the stress test alignment results based on the health digital twin, the corresponding local response unit is first determined according to the path identifier and node identifier in the stress test alignment results. Then, the local response unit extracts the local constraint parameters related to the current stress test item. Subsequently, stress test candidate correction items are generated for each local constraint parameter, and stress test candidate correction items with the same path affiliation, adjacent node affiliation, or the same local response unit affiliation are organized into stress test correction item combinations. After each stress test correction item combination is established, stress test constraint matching judgment is performed on each of them. The residual constraint deviation between the stress test echo response sequence and the stress test reference echo constraint set is used as the judgment object. All stress test correction item combinations that meet the stress test constraint matching requirements are included in the screening range. Then, the screening is completed according to two-level conditions: the number of correction items and the constraint matching error. The minimum stress test correction set is determined from the stress test correction item combinations that meet the screening conditions. The stress test candidate correction items, stress test correction item combinations, stress test constraint matching judgment and the minimum stress test correction set are sequentially connected to form a complete stress test local correction search chain, so that the minimum stress test correction set can completely record the correction position, correction type and correction amount distribution corresponding to each local response position under the stress test state.

[0048] In this embodiment S4, the specific method for comparing the minimum correction set of the pressure test with the minimum correction set is as follows: The number of correction items, correction locations, and correction types in both the minimum correction set and the minimum correction set are counted, and the corresponding statistical results are compared item by item. The critical pre-fault state of the target electrical equipment is the pre-fault state in which the target electrical equipment no longer meets the health requirements before a manifest fault occurs. The critical pre-fault state is determined based on the comparison results between the minimum correction set of the pressure test and the minimum correction set. When the number of correction items, correction locations, and correction types in the minimum correction set of the pressure test are all greater than the corresponding statistical results of the minimum correction set, the target electrical equipment is determined to be in the critical pre-fault state. The predicted fault result includes the predicted fault location and the predicted fault category. The predicted fault location is determined based on the aggregated result of the correction amounts corresponding to each correction location in the minimum correction set of the pressure test, and the target electrical equipment location corresponding to the correction location with the largest aggregated correction amount is taken as the predicted fault location. The predicted fault category is determined based on the aggregated result of the correction amounts corresponding to each correction type in the minimum correction set of the pressure test, and the correction type with the largest aggregated correction amount is taken as the predicted fault category. The predicted fault result is output after determining that the target electrical equipment is in the critical pre-fault state.

[0049] In this embodiment S4, when comparing the minimum correction set of the load test with the minimum correction set, firstly, all correction items in the minimum correction set of the load test are statistically merged, and then all correction items in the minimum correction set are statistically merged using the same caliber. The comparison uses the number of correction items, the number of correction locations, and the number of correction types as unified comparison dimensions. The number of correction items reflects the overall scale of the correction items in the current correction set, the number of correction locations reflects the range of local response locations covered by the current correction set, and the number of correction types reflects the range of local constraint categories involved in the current correction set. In specific processing, the number of correction items in the minimum correction set of the load test is first counted, and then duplicate correction locations are merged according to the correction location field to obtain the number of load test correction locations. Then, based on the correction type field, duplicate correction types are merged to obtain the number of pressure test correction types. The same method is used to count the number of correction items, correction locations, and correction types in the minimum correction set. After completing the above statistics, the corresponding statistical results of the minimum pressure test correction set and the minimum correction set are compared item by item in the order of the number of correction items, the number of correction locations, and the number of correction types. The same statistical caliber and the same field rules are used for item by item comparison to ensure that the comparison basis of the two types of correction sets is consistent in the dimensions of quantity, location, and type. This forms the comparison result set in S4. The comparison result set takes the statistical differences in the three comparison dimensions as its core content and is used to determine whether the target electrical equipment has entered the critical state before the fault.

[0050] In this embodiment S4, the critical pre-fault state of the target electrical equipment is determined based on the comparison result between the minimum correction set of the pressure test and the minimum correction set. The critical pre-fault state corresponds to the pre-fault state in which the target electrical equipment has already experienced a change in the health state constraint range before the formation of an obvious fault. During the determination, the number of correction items, the number of correction locations, and the number of correction types in the minimum correction set of the pressure test are compared one-to-one with the number of correction items, the number of correction locations, and the number of correction types in the minimum correction set, respectively. When the statistical results of the minimum correction set of the pressure test are higher than the corresponding statistical results of the minimum correction set, the target electrical equipment is determined to be in the critical pre-fault state. The increment in the number of correction items indicates that the scale of the constraints that need to be corrected under the stress test conditions has expanded. The increment in the number of correction locations indicates that the range of local response locations covered by the correction distribution under the stress test conditions has expanded. The increment in the number of correction types indicates that the range of local constraint categories that need to be corrected under the stress test conditions has expanded. The three statistical results together constitute the basis for determining the critical pre-fault state. After the determination is completed, the determination result is written into the state record item corresponding to the target electrical equipment, so that the critical pre-fault state forms a fixed correspondence with the current minimum correction set of the stress test, the current minimum correction set, and the current target electrical equipment, thereby establishing the state basis required for subsequent fault prediction output.

[0051] In this embodiment S4, the predicted fault result includes the predicted fault location and the predicted fault category. The predicted fault result is output after determining that the target electrical equipment is in a critical pre-fault state. The predicted fault location is formed based on the distribution of correction positions in the minimum correction set of the pressure test. Specifically, firstly, all correction items in the minimum correction set of the pressure test are classified according to the correction position field, and correction items belonging to the same correction position are grouped into the same position group. Then, the correction amounts in each position group are accumulated to obtain the aggregate correction amount corresponding to each correction position. After that, the magnitude of the aggregate correction amount corresponding to each correction position is compared, and the target electrical equipment location corresponding to the correction position with the largest aggregate correction amount is determined as the fault location. The predicted fault location and fault category are formed based on the distribution of correction types in the minimum correction set of the pressure test. Specifically, all correction items in the minimum correction set of the pressure test are first classified according to the correction type field. Correction items belonging to the same correction type are grouped into the same type group. Then, the correction amounts in each type group are accumulated to obtain the aggregate correction amount corresponding to each correction type. After that, the magnitude of the aggregate correction amount corresponding to each correction type is compared, and the correction type with the largest aggregate correction amount is determined as the predicted fault category. After the predicted fault location and predicted fault category are determined, they are written together with the fault critical pre-state into the fault prediction result record of the target electrical equipment to form the predicted fault result.

[0052] Example 2: The present invention proposes a digital twin-based electrical equipment fault prediction system, which is applied to the digital twin-based electrical equipment fault prediction method proposed in Example 1. It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the digital twin-based electrical equipment fault prediction method in Example 1.

[0053] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for predicting electrical equipment faults based on digital twins, characterized in that, Includes the following steps: S1. Obtain the basic equipment data, acquisition node configuration data, excitation configuration data and segmentation rule data of the target electrical equipment. Based on the basic equipment data, acquisition node configuration data and excitation configuration data, construct a health digital twin and establish local constraint parameters in the health digital twin to generate coded excitation sequences and reference echo constraint sets. Among them, the health digital twin is a digital twin reference body that characterizes the excitation-response benchmark relationship under the health state of the target electrical equipment; the local constraint parameters are the constraint parameter items set in the health digital twin corresponding to the local response units; and the benchmark echo constraint set is the benchmark constraint set set corresponding to the coded excitation sequence. S2. Input the coded excitation sequence into the target electrical equipment, acquire the echo response data at the acquisition node, form the echo response sequence according to the segmentation rule data, align the echo response sequence with the reference echo constraint set segment by segment to obtain the alignment result, and perform local correction search on the local constraint parameters associated with the alignment result based on the health digital twin to obtain the minimum correction set and determine the sensitive constraint axis. Among them, the minimum correction set is the set of the fewest correction terms that make the echo response sequence meet the requirements of the reference echo constraint set; the sensitive constraint axis is the constraint direction in the minimum correction set that corresponds to the main mismatch direction of the current echo response. S3. Based on the sensitive constraint axis, the coded excitation sequence is processed by single-axis misalignment to generate a single-step misaligned pressure test sequence. The single-step misaligned pressure test sequence is input to the target electrical equipment. Pressure test echo response data is acquired at the acquisition node. The pressure test echo response sequence is generated according to the segmentation rule data. The pressure test echo response sequence is aligned segment by segment with the pressure test reference echo constraint set to obtain the pressure test alignment result. Based on the health digital twin, the local constraint parameters associated with the pressure test alignment result are locally corrected and searched to generate the minimum correction set of the pressure test. Among them, the single-step misaligned stress test sequence is the stress test excitation sequence formed after the encoded excitation sequence is misaligned once along the sensitive constraint axis; the minimum correction set of stress test is the minimum set of correction terms determined for the stress test echo response sequence; S4. Compare the minimum correction set of the pressure test with the minimum correction set to determine the critical pre-fault state of the target electrical equipment, and output the predicted fault result based on the minimum correction set of the pressure test. Among them, the critical pre-fault state is the pre-fault state in which the target electrical equipment has not yet entered the manifest fault state and no longer meets the health state requirements under single-step misalignment pressure test conditions.

2. The method for predicting electrical equipment faults based on digital twins according to claim 1, characterized in that: In step S1, the acquisition nodes are data acquisition nodes set at various local response positions of the target electrical equipment; the health digital twin performs digital twin mapping on various local response positions of the target electrical equipment based on the equipment's basic data, and the health digital twin is set with multiple local response units, which are response units in the health digital twin that correspond to different local response positions of the target electrical equipment; the local constraint parameters are set according to the local response units in groups, with each group of local constraint parameters corresponding to one local response unit, and different local response units corresponding to different groups of local constraint parameters, and each local response unit establishes a correspondence with the acquisition nodes; the excitation configuration data is used to limit the arrangement order and path correspondence of each coded excitation segment, and the coded excitation sequence is formed by arranging multiple coded excitation segments in the order determined by the excitation configuration data, and each coded excitation segment establishes a correspondence with a local response unit; the reference echo constraint set is set in segments according to the coded excitation sequence, and each reference echo constraint is set according to the local constraint parameters of the corresponding local response unit, and each reference echo constraint is associated with the corresponding coded excitation segment, local response unit, and acquisition node.

3. The method for predicting electrical equipment faults based on digital twins according to claim 2, characterized in that: In S2, the echo response data is the node response data collected by the acquisition node under the action of the coded excitation sequence of the target electrical equipment; the echo response sequence is a sequential response sequence formed by arranging the echo response data into segments according to segmentation rule data; the specific method for forming the echo response sequence according to segmentation rule data is as follows: the echo response data is divided into boundaries according to the segment order corresponding to the coded excitation sequence to form multiple echo segments, and the echo response sequence is formed according to the segment order corresponding to each echo segment; wherein, the echo segment includes the propagating echo segment, the reflected echo segment, and the recovered echo segment; the echo response... The specific method for aligning the sequence with the reference echo constraint set segment by segment is as follows: the reference echo constraint set is divided into multiple constraint segments according to the segment order of the coded excitation sequence, and each echo segment and each constraint segment are matched at the segment level according to the same segment order; the alignment result is a set of segment-level correspondence results formed by the correspondence between each echo segment and each constraint segment, and each correspondence item in the alignment result includes an echo segment identifier, a constraint segment identifier, a path identifier, and a node identifier; the path identifier in the alignment result is used to characterize the coded excitation path to which the corresponding item belongs, and the node identifier is used to characterize the acquisition node to which the corresponding item belongs.

4. The method for predicting electrical equipment faults based on digital twins according to claim 3, characterized in that: In S2, the local constraint parameters associated with the alignment result are determined based on the path identifier and node identifier in the alignment result; the local correction search is a constraint correction search operation performed item by item on the local constraint parameters, used to determine the corresponding correction terms for the constraint mismatch of the current echo response sequence relative to the reference echo constraint set; the specific method for performing local correction search on the local constraint parameters associated with the alignment result based on the health digital twin is as follows: candidate correction terms are generated for the local constraint parameters associated with the alignment result, and correction term combinations are constructed based on the candidate correction terms. Constraint matching judgment is performed on each correction term combination to determine the correction term combination that meets the matching requirements between the echo response sequence and the reference echo constraint set; The minimum correction set is the set of the fewest correction terms determined after the local correction search, used to characterize the constraint correction distribution corresponding to the current echo response sequence when it meets the requirements of the benchmark echo constraint set. The minimum correction set includes correction position, correction type and correction amount. The minimum correction set is determined by having the minimum number of correction items as the first determining condition. When the number of correction items is the same, the minimum constraint matching error is determined as the second determining condition. The correction position corresponds to the local response unit, the correction type corresponds to the local constraint parameter category, and the correction amount corresponds to the offset of the local constraint parameter. Among them, the number of correction terms is the number of correction terms contained in the minimum correction set, and the constraint matching error is the error of the remaining constraint deviation between the echo response sequence and the reference echo constraint set. Multiple correction terms in the minimum correction set are arranged in order of correction amount to form a correction term sequence.

5. The method for predicting electrical equipment faults based on digital twins according to claim 4, characterized in that: In S2, the sensitive constraint axis is a single constraint direction in the minimum correction set corresponding to the main mismatch direction of the current echo response. It is used to identify the dominant mismatch direction of the current echo response sequence and to perform sorting and extraction on each correction item in the minimum correction set. When sorting the correction items in the minimum correction set, the sensitive constraint axis is sorted according to the magnitude of the correction amount of each correction item, and the local constraint parameter corresponding to the first correction item in the sort is taken. The sensitive constraint axis is then determined together with the coded excitation segment associated with the local constraint parameter. Only one sensitive constraint axis is determined, and it is set separately from the constraint directions corresponding to the other correction items in the minimum correction set. It uniquely indicates the main mismatch direction of the current echo response sequence. The coded excitation segment corresponding to the sensitive constraint axis is the unique target coded excitation segment for single-axis misalignment processing.

6. The method for predicting electrical equipment faults based on digital twins according to claim 5, characterized in that: In S3, the single-axis misalignment processing is a single-time misalignment processing method performed on the unique target coded excitation segment in the coded excitation sequence, used to generate a single-step misalignment stress test sequence. The specific method for performing single-axis misalignment processing on the coded excitation sequence based on the sensitive constraint axis is as follows: determine the unique target coded excitation segment corresponding to the sensitive constraint axis, keep the remaining coded excitation segments in the coded excitation sequence unchanged, perform a single misalignment adjustment on the unique target coded excitation segment, and combine the misaligned unique target coded excitation segment with the remaining coded excitation segments in the original segment order to form a single-step misalignment stress test sequence. The single-step misalignment stress test sequence is a stress test excitation sequence formed after the coded excitation sequence performs one misalignment processing along the sensitive constraint axis, used to obtain the stress test echo response data of the target electrical equipment under stress test conditions. The single-step misalignment stress test sequence includes the target coded excitation segment after single-axis misalignment processing and the remaining coded excitation segments that have not undergone misalignment processing. The segment order of the coded excitation sequence remains consistent before and after the single-axis misalignment processing.

7. The method for predicting electrical equipment faults based on digital twins according to claim 6, characterized in that: In S3, the pressure test echo response data is the node response data collected by the acquisition node under the action of the single-step misaligned pressure test sequence of the target electrical equipment, used to form the pressure test echo response sequence; the pressure test echo response sequence is a sequential response sequence formed by arranging the pressure test echo response data into segments according to segmentation rule data, used as the pressure test input sequence for local correction search, and the pressure test echo response sequence includes the pressure test propagation echo segment, the pressure test reflection echo segment, and the pressure test recovery echo segment; the specific method for forming the pressure test echo response sequence according to the segmentation rule data is as follows: the pressure test echo response data is divided into boundaries according to the segment order corresponding to the single-step misaligned pressure test sequence to form multiple pressure test echo segments, and each pressure test echo segment... Each segment is assigned a corresponding identifier, and the segments are arranged according to the segment order corresponding to the identifiers to form a stress test echo response sequence. The stress test reference echo constraint set is a set of stress test constraints generated by the health digital twin based on the single-step misaligned stress test sequence. The stress test reference echo constraint set is segmented according to the single-step misaligned stress test sequence. The specific method for aligning the stress test echo response sequence with the stress test reference echo constraint set segment by segment is as follows: the stress test reference echo constraint set is divided into multiple stress test constraint segments according to the segment order of the single-step misaligned stress test sequence, and each stress test echo segment and each stress test constraint segment are matched at the segment level according to the same segment order. The stress test alignment result is a set of segment-level correspondence results formed by the correspondence between each stress test echo segment and each stress test constraint segment.

8. The method for predicting electrical equipment faults based on digital twins according to claim 7, characterized in that: In S3, the minimum correction set of the stress test is the set of the minimum correction terms determined for the stress test echo response sequence, which is used to characterize the correction distribution corresponding to the stress test echo response sequence meeting the health digital twin constraint requirements under the single-step misaligned stress test condition. The minimum correction set for stress testing includes correction location, correction type and correction amount. The correction item structure in the minimum correction set for stress testing is consistent with the minimum correction set. The minimum number of correction items is used as the first determining condition. When the number of correction items is the same, the minimum constraint matching error is used as the second determining condition. The specific method for generating the minimum correction set of stress test based on the local constraint parameters associated with the stress test alignment results using a health digital twin is as follows: Local constraint parameters are determined based on the path identifier and node identifier in the stress test alignment results. Candidate correction items for each local constraint parameter are generated, and combinations of stress test correction items are constructed based on these candidate items. Constraint matching is performed on each combination of stress test correction items to determine the combination of stress test correction items that meets the matching requirements between the stress test echo response sequence and the stress test baseline echo constraint set. Finally, the minimum correction set of stress test is determined from the combinations of stress test correction items that meet the requirements.

9. The method for predicting electrical equipment faults based on digital twins according to claim 8, characterized in that: In step S4, the critical pre-fault state of the target electrical equipment is a state in which the target electrical equipment no longer meets the health requirements before a manifest fault occurs. The critical pre-fault state is determined based on the comparison result between the minimum correction set of the pressure test and the minimum correction set. When the number of correction items, the number of correction locations, and the number of correction types in the minimum correction set of the pressure test are all greater than the corresponding statistical results of the minimum correction set, the target electrical equipment is determined to be in the critical pre-fault state. The predicted fault result includes the predicted fault location and the predicted fault category. The predicted fault location is determined based on the aggregation result of the correction amount corresponding to each correction location in the minimum correction set of the pressure test, and the target electrical equipment location corresponding to the correction location with the largest aggregated correction amount is taken as the predicted fault location. The predicted fault category is determined based on the aggregation result of the correction amount corresponding to each correction type in the minimum correction set of the pressure test, and the correction type with the largest aggregated correction amount is taken as the predicted fault category. The predicted fault result is output after it is determined that the target electrical equipment is in the critical pre-fault state.

10. A fault prediction system for electrical equipment based on digital twins, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes a computer program to implement the electrical equipment fault prediction method based on digital twins as described in any one of claims 1-9.

Images