Search

CN-122017549-A - Breaker state evolution prediction method based on digital twin model

CN122017549ACN 122017549 ACN122017549 ACN 122017549ACN-122017549-A

Abstract

The invention discloses a breaker state evolution prediction method based on a digital twin model, which relates to the technical field of digital twin and comprises the steps of collecting multi-source operation data, processing and generating an operation condition state data set; the method comprises the steps of constructing a digital twin model of the circuit breaker, carrying out state calibration and generating a state vector, establishing a two-way mapping of a state domain and an energy domain, executing forward mapping to generate the energy state vector, constructing a working condition evolution scheduler, extracting working condition semantics, determining channel activation sequence and activation intensity, constructing a multiphase flow type transition channel and jointly driving, outputting an energy state prediction result, reflecting and generating a future state evolution result, and outputting a state trend and a health prediction value. According to the invention, by constructing the digital twin model of the circuit breaker and introducing the state energy bidirectional mapping and multiphase flow type state transition mechanism, the precise modeling and the prospective prediction of the state evolution process of the circuit breaker under the complex working condition are realized.

Inventors

  • LI SONGWEI
  • ZHU BO

Assignees

  • 哈尔滨理工大学

Dates

Publication Date
20260512
Application Date
20260409

Claims (8)

  1. 1. The method for predicting the state evolution of the circuit breaker based on the digital twin model is characterized by comprising the following steps of: the method comprises the steps of collecting multi-source operation data in the operation process of the circuit breaker, and processing the multi-source operation data to form an operation condition state data set; constructing a digital twin model of the circuit breaker, performing state calibration on the digital twin model of the circuit breaker based on an operation condition state data set, and defining a state vector for representing the health state of the circuit breaker in the digital twin model of the circuit breaker; based on the current state vector and historical state change information in the operating condition state dataset, establishing a bidirectional mapping relation between a breaker state domain and an energy domain, and performing state-energy domain mapping processing on the current state vector to generate an energy state vector; Constructing a working condition evolution scheduler, respectively generating working condition semantic information and a working condition influence track through the working condition evolution scheduler based on the current operation working condition and the historical operation working condition sequence in the operation working condition state data set, and determining a channel activation sequence and a channel activation strength for controlling state transition; Constructing a multiphase flow type state transition channel on the basis of the energy state vector, wherein the multiphase flow type state transition channel comprises a progressive phase channel, a abrupt phase channel and a recovery phase channel, and the multiphase flow type state transition channel is driven in a combined way according to the channel activation sequence and the channel activation intensity to generate an energy state prediction result; And performing energy domain-to-state domain reflection processing on the energy state prediction result based on the bidirectional mapping relation to obtain a state evolution prediction result of the breaker in future multi-time steps, and outputting a state change trend and a comprehensive health state prediction result of the breaker.
  2. 2. The method for predicting the evolution of the state of a circuit breaker based on a digital twin model according to claim 1, wherein the multi-source operation data comprises circuit breaker opening and closing time data, opening and closing stroke data, opening and closing operation speed data, contact rebound characteristic data, arc duration data, opening and closing current data, contact resistance data, contact temperature rise data, circuit breaker shell temperature data, environment temperature data, load current level data, operation frequency data and overhaul maintenance record data.
  3. 3. The method for predicting the evolution of the state of a circuit breaker based on a digital twin model according to claim 1, wherein the processing of the multi-source operation data comprises the time synchronization, the anomaly rejection and the standardization of the multi-source operation data.
  4. 4. The method for predicting the evolution of the state of a circuit breaker based on a digital twin model according to claim 1, wherein the defining of the state vector for characterizing the state of health of the circuit breaker in the digital twin model of the circuit breaker comprises: dividing fields of multi-source operation data in an operation working condition state data set, setting opening current data, arc duration time data, load current level data, environment temperature data and operation frequency data as working condition input quantity fields, setting opening and closing time data, opening and closing stroke data, opening and closing operation speed data, contact rebound characteristic data, contact resistance data, contact temperature rise data and breaker shell temperature data as observation fields, setting overhaul and maintenance record data as maintenance event fields, and time sequencing all the fields according to a unified sampling period; the method comprises the steps of constructing a circuit breaker digital twin model, wherein the circuit breaker digital twin model consists of a state layer, a working condition input layer, an observation layer and a parameter set, the state layer comprises a state vector consisting of a contact wear state quantity, an operating mechanism fatigue state quantity and an insulation and thermal aging state quantity, the working condition input layer receives a working condition input quantity field of each sampling period, the observation layer receives an observation quantity field of the same sampling period, and the parameter set records a state update coefficient and an observation mapping coefficient; In each sampling period, calling a working condition input layer and a state layer, mapping the state vector of the previous sampling period with the working condition input quantity field of the current sampling period according to a state updating rule in a parameter set to generate a state vector of the current sampling period, and calling an observation layer to map the state vector of the current sampling period according to an observation mapping rule to generate an observation layer output; Taking an observed quantity field of the same sampling period in the operating condition state data set as a reference, performing error calculation on the output of the observation layer, resetting the initial value of a corresponding component of the state layer in the sampling period triggered by the maintenance event field, and iteratively adjusting a state update coefficient and an observation mapping coefficient in the parameter set until the error meets a preset threshold; and after the error meets a preset threshold, fixing the parameter set, outputting a state vector of the current sampling period and storing a state vector sequence.
  5. 5. A method of predicting state evolution of a circuit breaker based on a digital twin model as recited in claim 1 wherein said generating an energy state vector comprises: Extracting a state vector corresponding to a current sampling period and historical state change information aligned with time from an operation condition state data set, generating a historical state change sequence based on the historical state change information, and extracting on-off current data, arc duration time data, load current level data, environment temperature data and operation frequency data in the same time range as an operation condition driving sequence; Establishing a bidirectional mapping relation between a breaker state domain and an energy domain, wherein the bidirectional mapping relation comprises a forward mapping relation from the state domain to the energy domain and a reverse mapping relation from the energy domain to the state domain, and the bidirectional mapping relation comprises the following steps: The forward mapping relation is formed by jointly determining components, component orders and component weights of the energy state vector by the state vector and the working condition driving sequence; The reverse mapping relation rebuilds the corresponding state vector and corrects the state evolution hysteresis feature according to the component composition, the component sequence and the component weight by combining the corresponding working condition driving sequence and the historical state change sequence; When a forward mapping relation is established, carrying out path identification processing on the state vector based on a historical state change sequence, generating a path identification for distinguishing a degraded path from a recovery path, and taking the path identification as an input item of the forward mapping relation; Performing state-to-energy domain mapping processing on a state vector of a current sampling period based on a forward mapping relation, outputting an energy state vector according to component constitution and component order of the energy state vector, and binding and storing the energy state vector and the path identifier; And converting the energy state vector into a reflection result of the state vector based on the reverse mapping relation, carrying out consistency check on the reflection result and the state vector of the current sampling period, fixing the bidirectional mapping relation after the consistency check passes, and outputting the energy state vector.
  6. 6. The method of predicting state evolution of a circuit breaker based on a digital twin model of claim 1, wherein said determining the channel activation order and the channel activation strength for controlling state transitions comprises: Extracting a current operation condition and a historical operation condition sequence from the operation condition state data set, completing time alignment and uniform sampling, and forming a condition input set corresponding to a current sampling period; The working condition evolution scheduler is constructed and consists of a semantic deconstructing layer, an influence track planning layer and a layout execution layer, wherein the semantic deconstructing layer receives a working condition input set and outputs working condition semantic information and a path identifier, the influence track planning layer outputs a working condition influence track and a layout time window set based on a historical operation working condition sequence, and the layout execution layer outputs a channel activation sequence and channel activation strength based on the working condition semantic information and the working condition influence track; carrying out semantic analysis on the current operation working condition through the semantic deconstructing layer, generating working condition semantic information consisting of driving semantic information, modulating semantic information and disturbance semantic information, and generating a recovery path identifier and a mutation path identifier for the corresponding working condition section according to the overhaul maintenance record and the short-circuit breaking record; generating a short-period impact track, a long-period evolution track and a recovery sensitive interval based on a historical operation condition sequence and a path identifier through the influence track planning layer, and generating a corresponding arrangement time window set and a priority sequence; And determining a channel activation sequence, channel activation strength and activation duration for controlling state transition according to the working condition semantic information, the working condition influence track, the arrangement time window set and the priority sequence by the arrangement execution layer, and applying conflict inhibition and mutual exclusion constraint in a time window where channels overlap.
  7. 7. The method for predicting state evolution of a circuit breaker based on a digital twin model of claim 1, wherein generating the energy state prediction result comprises: constructing a multiphase flow type state transition channel on the basis of an energy state vector, wherein the multiphase flow type state transition channel comprises a progressive phase channel, a abrupt phase channel and a recovery phase channel, establishing a channel energy account book, a sliding formation window and a channel memory unit, and recording a channel processing sequence, channel processing intensity and channel history residual errors; Driving the progressive communication channel to execute sectional accumulation and temperature fatigue coupling window processing according to the channel activation sequence and the channel activation intensity, and applying monotonic envelope and speed upper limit constraint to the energy state vector to generate a first intermediate energy state vector; Identifying impact fragments triggered by short circuit break and arc duration in the operation condition state data set, selecting a mutation processing template according to a path identifier, driving a mutation phase channel to avoid repeated counting and executing dynamic saturation cutting in a refractory period window, and generating a second intermediate energy state vector; distinguishing two types of recovery units, namely lubrication recovery and component replacement recovery, according to overhaul and maintenance record data in the operating condition state data set, driving a recovery phase channel to set a maximum recovery limit and a gradual time window, and carrying out recovery processing on the second intermediate energy state vector to generate a third intermediate energy state vector; performing conflict inhibition and mutual exclusion constraint on overlapping time windows of three types of channels in a sliding formation window, starting a channel residual error buffer zone of a channel memory unit to forward undigested residual errors, and keeping combination according to a channel activation sequence and a channel activation intensity completion sequence to obtain energy state prediction result candidates; and (3) carrying out energy consistency check on the channel energy account book, and confirming an energy state prediction result after the preset threshold value is met.
  8. 8. The method for predicting the state evolution of a circuit breaker based on a digital twin model according to claim 1, wherein the outputting the state change trend and the comprehensive health state prediction result of the circuit breaker comprises: Acquiring an energy state prediction result, and calling a mapping relation from an energy domain to a state domain in a bidirectional mapping relation between a breaker state domain and the energy domain as a basis for the energy domain-to-state domain reflection processing; According to the mapping relation from the energy domain to the state domain, performing component-by-component mapping processing on the energy state prediction result according to the component constitution, the component sequence and the component weight of the energy state vector, converting each energy component into a corresponding state variation quantity, and generating a prediction state vector; According to the definition structure of state vectors in the digital twin model of the circuit breaker, carrying out state decomposition on the predicted state vectors to obtain contact wear state quantity, operating mechanism fatigue state quantity and insulation and thermal aging state quantity, and applying value range constraint and change rate constraint on each state quantity to form constrained predicted state vectors; re-inputting the constrained predicted state vector into a mapping relation from a breaker state domain to an energy domain to obtain a corresponding energy re-mapping result, and comparing the energy re-mapping result with the energy state predicted result in a consistency way, and confirming that the predicted state vector is an effective result when the consistency meets a preset condition; And taking the confirmed effective predicted state vector as the current state vector of the next time step, sequentially inputting a state domain and energy domain bidirectional mapping process, a working condition evolution scheduling process and a multiphase flow state transition process flow, recursively obtaining a state evolution prediction result of the future multi-time step of the circuit breaker, and outputting a state change trend and a comprehensive health state prediction result of the circuit breaker.

Description

Breaker state evolution prediction method based on digital twin model Technical Field The invention relates to the technical field of digital twinning, in particular to a breaker state evolution prediction method based on a digital twinning model. Background The circuit breaker is used as a key control and protection device in a power system, and the running state of the circuit breaker is directly related to the safety and reliability of a power grid. With the continuous expansion of the capacity of the power system and the increasing complexity of the operation working condition, the circuit breaker can be influenced by various effects such as current impact, mechanical abrasion, thermal aging, environmental factors and the like in the long-term operation process, and the internal state presents the characteristic of continuous evolution along with time. Therefore, how to effectively model and predict the running state of the circuit breaker has become an important research content in the fields of power equipment state monitoring and intelligent operation and maintenance. In the prior art, the state evaluation and prediction method of the circuit breaker is mostly based on a single operation index or a simple time sequence analysis model, and the estimation of the health state of the equipment is realized by carrying out statistics or trend fitting on parameters such as contact wear, operation times or temperature rise, and the like. However, the method generally regards the state change of the circuit breaker as a single degradation path, and is difficult to simultaneously describe the real state evolution process under the superposition of various behaviors such as long-term operation degradation, burst impact such as short-circuit break and state recovery caused by overhaul and maintenance. The prior art generally lacks of system modeling on an operation working condition sequence and an evolution sequence thereof, is difficult to reflect the influence of different working condition combinations, working condition duration and historical state paths on a state evolution result, and the state variables are processed through direct mapping or linear relation, so that the energy accumulation, dissipation and irreversible evolution characteristics cannot be effectively reflected, the prediction result is insensitive to complex working condition changes, the prediction stability and the foresight are insufficient, and reliable support is difficult to provide for the refined operation and maintenance decision of the circuit breaker. Therefore, how to provide a circuit breaker state evolution prediction method based on a digital twin model is a problem that needs to be solved by those skilled in the art. Disclosure of Invention The invention aims to provide a state evolution prediction method of a circuit breaker based on a digital twin model, which is characterized in that the state evolution prediction method of the circuit breaker is implemented by carrying out calculation processing and state modeling on multi-source operation data of the circuit breaker, constructing the digital twin model of the circuit breaker, introducing a bidirectional mapping mechanism between a state domain and an energy domain, working condition evolution scheduling and multiphase flow type state transition processing, and realizing prediction analysis on the state evolution process of the circuit breaker under complex operation conditions. According to the embodiment of the invention, the method for predicting the state evolution of the circuit breaker based on the digital twin model comprises the following steps: the method comprises the steps of collecting multi-source operation data in the operation process of the circuit breaker, and processing the multi-source operation data to form an operation condition state data set; constructing a digital twin model of the circuit breaker, performing state calibration on the digital twin model of the circuit breaker based on an operation condition state data set, and defining a state vector for representing the health state of the circuit breaker in the digital twin model of the circuit breaker; based on the current state vector and historical state change information in the operating condition state dataset, establishing a bidirectional mapping relation between a breaker state domain and an energy domain, and performing state-energy domain mapping processing on the current state vector to generate an energy state vector; Constructing a working condition evolution scheduler, respectively generating working condition semantic information and a working condition influence track through the working condition evolution scheduler based on the current operation working condition and the historical operation working condition sequence in the operation working condition state data set, and determining a channel activation sequence and a channel activation strength for controlling state transi