CN-122021105-A - Full-working-condition simulation modeling method and system for hydroelectric generating set

CN122021105ACN 122021105 ACN122021105 ACN 122021105ACN-122021105-A

Abstract

The invention discloses a full-working-condition simulation modeling method and system for a hydroelectric generating set, and relates to the technical field of hydroelectric generating sets, wherein the method comprises the steps of establishing a first twin model and a full-working-condition label matrix of the generating set by fusing multiple characteristics; the method comprises the steps of analyzing registration errors through a multi-scale grid operator, performing cross-scale reconstruction to generate a second twin model of the unit, respectively utilizing an all-condition label matrix to perform inversion accident tracing to construct a tracing graph, combining a fault label matrix to perform fault simulation factorization to establish a factorial graph, cooperatively fusing the all-condition inversion tracing graph and a fidelity abnormal factorial graph, and optimizing to generate a third twin model of the unit with higher fidelity. The technical problems of insufficient fidelity, weak full-working-condition coverage and dynamic inversion capability and low fault simulation and abnormal factorization accuracy of the digital twin model of the hydroelectric generating set in the prior art are solved, and the technical effects of constructing the digital twin model of the hydroelectric generating set covered by the full working conditions and improving the fidelity, the interpretation and the cross-scale consistency are achieved.

Inventors

CONG BINBIN
CHEN LIWEI
MEI XIAOLONG
LIU QINGWANG
ZHANG KANG
LIU YINZHAN
ZHOU ZHENG

Assignees

大唐河北发电有限公司王快发电分公司

Dates

Publication Date: 20260512
Application Date: 20251217

Claims (10)

1. The full-working-condition simulation modeling method for the hydroelectric generating set is characterized by comprising the following steps of: carrying out multi-characteristic fusion modeling according to a hydroelectric generating set, obtaining a first twin model of the generating set, and synchronously constructing an all-condition label matrix of the hydroelectric generating set; introducing a multi-scale grid operator to perform scale transformation registration error analysis on the first twin model of the unit, and obtaining a modeling registration error graph network; performing trans-scale characteristic reconstruction on the first twin model of the unit according to the modeling registration error graph network to generate a second twin model of the unit; Performing full-working-condition inversion accident tracing on the second twin model of the unit according to the full-working-condition label matrix, and establishing a full-working-condition inversion tracing graph network; Performing fault simulation fidelity anomaly factorization on a second twin model of the unit according to the unit fault label matrix of the hydroelectric unit, and establishing a fidelity anomaly factorization graph network; and carrying out double-network collaborative fusion optimization on the second twin model of the unit according to the full-working-condition inversion tracing graph and the fidelity anomaly factorial graph to generate a third twin model of the unit.
2. The method for full-condition simulation modeling of a hydroelectric generating set according to claim 1, wherein the method for multi-characteristic fusion modeling of the hydroelectric generating set to obtain a first twin model of the generating set comprises the following steps: According to the hydroelectric generating set, multi-characteristic collection and arrangement are carried out, and a set multi-characteristic vector is constructed, wherein the set multi-characteristic vector comprises a set mechanical characteristic vector, a set electromagnetic characteristic vector and a set thermal characteristic vector; Carrying out layered characteristic twin modeling according to the multi-characteristic vector of the unit to generate a unit layered characteristic model; And constructing a full-quantity characteristic coupling map according to the multi-characteristic vector of the unit, guiding the layered characteristic model of the unit to perform multi-characteristic fusion according to the full-quantity characteristic coupling map, and generating a first twin model of the unit.
3. The full-condition simulation modeling method of a hydroelectric generating set according to claim 1, wherein the step of introducing a multi-scale grid operator to conduct scale transformation registration error analysis on a first twin model of the set to obtain a modeling registration error graph comprises the steps of: Activating the multi-scale grid operator, wherein the multi-scale grid operator comprises a first-scale grid operator, a second-scale grid operator and a third-scale grid operator; Carrying out registration error analysis on a first twin model of the unit according to the first scale grid operator to obtain a first registration error graph network; carrying out registration error analysis on the first twin model of the unit according to the second scale grid operator to obtain a second registration error graph network; Carrying out registration error analysis on the first twin model of the unit according to the third scale grid operator to obtain a third registration error graph network; And integrating the first registration error graph network, the second registration error graph network and the third registration error graph network to generate the modeling registration error graph network.
4. A method for full-condition simulation modeling of a hydroelectric generating set according to claim 3, wherein performing registration error analysis on the first twin model of the set according to the first scale grid operator to obtain a first registration error graph network comprises: dividing a first twin model of the unit according to the first scale grid operator to obtain a first unit grid model; carrying out modeling characteristic identification of each grid according to the first unit grid model to obtain modeling characteristic vectors of each grid; Acquiring actual measurement characteristics of each grid of the hydroelectric generating set according to the first generating set grid model to obtain actual measurement characteristic vectors of each grid; Carrying out grid-by-grid registration error analysis on the grid modeling characteristic vectors according to the grid actual measurement characteristic vectors to obtain grid registration error vectors; And carrying out self-adaptive association mapping on the first unit grid model according to the grid registration error vectors to generate the first registration error graph net.
5. The full-working-condition simulation modeling method of the hydroelectric generating set according to claim 1, wherein full-working-condition inversion accident tracing is performed on the second twin model of the hydroelectric generating set according to the full-working-condition tag matrix, and a full-working-condition inversion tracing graph is established, comprising: Traversing the full-working-condition label matrix to extract a first working-condition label; Acquiring working condition logs of the hydroelectric generating set based on the first working condition label to obtain a first working condition log sequence; Performing mechanical response inversion accident tracing on the second twin model of the unit according to the first working condition log sequence to obtain a first working condition mechanical response inversion tracing space; performing electromagnetic response inversion accident tracing on the second twin model of the unit according to the first working condition log sequence to obtain a first working condition electromagnetic response inversion tracing space; Performing thermal response inversion accident tracing on the second twin model of the unit according to the first working condition log sequence to obtain a first working condition thermal response inversion tracing space; And carrying out graph structure representation according to the first working condition mechanical response inversion tracing space, the first working condition electromagnetic response inversion tracing space and the first working condition thermal response inversion tracing space to obtain a first working condition inversion tracing graph.
6. The full-condition simulation modeling method of a hydroelectric generating set according to claim 5, wherein performing mechanical response inversion accident tracing on the second twin model of the set according to the first condition log sequence to obtain a first condition mechanical response inversion tracing space comprises: Extracting an Mth working condition log sample according to the first working condition log sequence, and carrying out characteristic identification on the Mth working condition log sample to obtain an Mth working condition control sample and an Mth working condition mechanical response sample, wherein M is a positive integer; performing simulation control on the second twin model of the unit according to the M working condition control sample to obtain M mechanical response simulation data and M simulation process data; Performing deviation recognition on the M mechanical response simulation data according to the M working condition mechanical response sample to obtain an M mechanical response inversion accident factor; performing intermediate trigger factor tracing on the M mechanical response inversion accident factor according to the M simulation process data to obtain an M intermediate factor tracing result; Performing basic trigger factor tracing on the M intermediate factor tracing result according to the M simulation process data to obtain an M basic factor tracing result; And carrying out Boolean logic association according to the M-th mechanical response inversion accident factor, the M-th intermediate factor tracing result and the M-th basic factor tracing result to generate an M-th mechanical response inversion tracing path, and adding the M-th mechanical response inversion tracing path into the first working condition mechanical response inversion tracing space.
7. The full-condition simulation modeling method of the hydroelectric generating set according to claim 1, wherein the fault simulation fidelity anomaly factorization is carried out on the second twin model of the hydroelectric generating set according to the generating set fault tag matrix of the hydroelectric generating set, and a fidelity anomaly factorization graph network is established, comprising: Performing fault log acquisition on the hydroelectric generating set according to the generating set fault tag matrix to obtain each tag fault log set; performing multi-mode fault semantic deconstructment according to the label fault log sets to obtain label fault scene mode sets and label fault symptom mode sets; Performing scene-driven fault dynamic evolution on the second twin model of the unit according to the label fault scene mode sets to obtain label simulated fault response sets; Performing fidelity exception analysis on the label simulated fault response sets according to the label fault symptom modal sets to obtain label fault simulated fidelity exception characteristics; And carrying out causal chain tracing on the fidelity abnormal characteristics of each label fault simulation according to the data set of each label fault simulation process to generate the fidelity abnormal factorial graph network.
8. The method for simulating and modeling the full working condition of the hydroelectric generating set according to claim 2, wherein constructing the full characteristic coupling map according to the multi-characteristic vector of the hydroelectric generating set comprises the following steps: carrying out multi-parameter correlation analysis on the multi-characteristic vector of the unit to construct unit characteristic association analysis distribution; Constructing graph nodes according to the multi-characteristic vector of the unit, and constructing a full-quantity characteristic node graph; and performing side association compensation on the full-quantity characteristic node diagram according to the unit characteristic association analysis distribution to generate the full-quantity characteristic coupling map.
9. The full-condition simulation modeling method of the hydroelectric generating set according to claim 2, wherein the multi-characteristic collection and arrangement is carried out according to the hydroelectric generating set, and a multi-characteristic vector of the generating set is constructed, comprising: Acquiring multiple characteristics according to the hydroelectric generating set to obtain a set characteristic database; And performing multi-characteristic classification cleaning according to the unit characteristic database to generate the unit multi-characteristic vector.
10. A hydroelectric generating set all-condition simulation modeling system, wherein the system is configured to implement the hydroelectric generating set all-condition simulation modeling method according to any one of claims 1 to 9, the system comprising: the fusion modeling unit is used for carrying out multi-characteristic fusion modeling according to the hydroelectric generating set to obtain a first twin model of the generating set, and synchronously constructing an all-working-condition label matrix of the hydroelectric generating set; The error analysis unit is used for introducing a multi-scale grid operator to conduct scale transformation registration error analysis on the first twin model of the unit, and obtaining a modeling registration error graph; The trans-scale reconstruction unit is used for performing trans-scale characteristic reconstruction on the first twin model of the unit according to the modeling registration error graph network to generate a second twin model of the unit; the accident tracing unit is used for carrying out full-working condition inversion accident tracing on the second twin model of the unit according to the full-working condition label matrix, and establishing a full-working condition inversion tracing graph network; The anomaly factorization unit is used for carrying out fault simulation fidelity anomaly factorization on the second twin model of the unit according to the unit fault tag matrix of the hydroelectric unit, and establishing a fidelity anomaly factorization graph network; And the collaborative optimization unit is used for carrying out double-network collaborative fusion optimization on the second twin model of the unit according to the full-working-condition inversion tracing graph and the fidelity anomaly factorization graph to generate a third twin model of the unit.

Description

Full-working-condition simulation modeling method and system for hydroelectric generating set Technical Field The invention relates to the technical field related to hydroelectric generating sets, in particular to a full-working-condition simulation modeling method and system for a hydroelectric generating set. Background The hydroelectric generating set is used as core equipment for hydropower energy conversion, the operation safety, stability and high efficiency of the hydroelectric generating set are directly related to the reliability and economy of an electric power system, along with the expansion of the scale of the hydroelectric generating set and the improvement of the intelligent level, the operation environment of the hydroelectric generating set is more and more complex, the working condition changes frequently, the fault modes are various, and the requirements of the modern hydropower system on fine management and intelligent operation and maintenance are difficult to meet by the traditional modeling analysis method based on single physical characteristics or local working conditions. The current hydropower modeling is mostly dependent on a mechanism model or a data driving model, often has a simplification error under complex working conditions and multi-fault coupling, and is poor in performance under data scarcity or abnormal working conditions, and lacks uniform description of full-working-condition cross-scale dynamic characteristics, so that the model has limitations in working condition conversion, fault tracing and abnormal analysis. Therefore, in the related technology of the present stage, the technical problems of insufficient fidelity of the digital twin model of the hydroelectric generating set, weak full-working-condition coverage and dynamic inversion capability and low fault simulation and abnormal factorization accuracy exist. Disclosure of Invention The application solves the technical problems of insufficient fidelity, weak full-working-condition coverage and dynamic inversion capability and low fault simulation and abnormal factorization accuracy of the digital twin model of the hydroelectric generating set in the prior art by providing the full-working-condition simulation modeling method and the full-working-condition simulation modeling system of the hydroelectric generating set, and achieves the technical effects of constructing the digital twin model of the full-working-condition coverage hydroelectric generating set and improving the fidelity, the interpretation and the trans-scale consistency. The application provides a full-working-condition simulation modeling method of a hydroelectric generating set, which comprises the steps of carrying out multi-characteristic fusion modeling according to the hydroelectric generating set, obtaining a first twin model of the generating set, synchronously constructing a full-working-condition label matrix of the hydroelectric generating set, introducing a multi-scale grid operator to carry out scale transformation registration error analysis on the first twin model of the generating set, obtaining a modeling registration error graph, carrying out trans-scale characteristic reconstruction on the first twin model of the generating set according to the modeling registration error graph, generating a second twin model of the generating set, carrying out full-working-condition inversion accident tracing on the second twin model of the generating set according to the full-working-condition label matrix, establishing a full-working-condition inversion graph, carrying out fault simulation fidelity anomaly factorial on the second twin model of the generating set according to the generating set fault simulation fidelity anomaly factorial graph, carrying out double-network collaborative optimization on the second twin model of the generating set according to the full-working-condition inversion graph and the fidelity anomaly factorial graph, and generating a third twin model of the generating set. In a possible implementation mode, the all-condition simulation modeling method of the hydroelectric generating set further comprises the following steps of collecting and sorting multiple characteristics according to the hydroelectric generating set, constructing a generating set multiple characteristic vector, wherein the generating set multiple characteristic vector comprises a generating set mechanical characteristic vector, a generating set electromagnetic characteristic vector and a generating set thermal characteristic vector, carrying out layered characteristic twin modeling according to the generating set multiple characteristic vector, constructing a full-quantity characteristic coupling map according to the generating set multiple characteristic vector, and guiding the generating set layered characteristic model to carry out multiple characteristic fusion according to the full-quantity characteristic coupling map to generate the generatin