Safety dispatching system and method based on intelligent operation agent of hydroelectric generating set

Abstract

The application discloses a kind of safe scheduling systems and methods based on water turbine unit wisdom operation intelligent agent, the system includes: operation perception module, state cognitive module, abnormal reasoning module, safety decision module, execution control module and intelligent optimization module;Operation perception module obtains multi-source operation data, state cognitive module carries out feature extraction and associated fusion processing, generates dynamic health state representation value;Abnormal reasoning module combines wisdom operation knowledge model to carry out abnormal state reasoning, generates disposal strategy;Safety decision module generates safety scheduling control scheme according to health state information, abnormal risk state information and safety operation constraint condition;Intelligent optimization module updates model parameters according to state feedback information.The application realizes water turbine unit dynamic health state representation, abnormal evolution reasoning and safety scheduling collaborative optimization, improves water turbine unit operation reliability and safety regulation ability.

Description

Technical Field

[0001] This invention relates to the field of hydropower unit scheduling technology, and in particular to a safety scheduling system and method based on a smart operation and maintenance agent for hydropower units. Background Technology

[0002] Hydropower units, as important adjustable power sources in the power system, undertake crucial tasks such as peak shaving, frequency regulation, and renewable energy integration. With the increasing proportion of renewable energy integration in the power system, the grid's operating status exhibits greater randomness and volatility, placing higher demands on the rapid adjustment capabilities, operational reliability, and safety control capabilities of hydropower units. Hydropower units typically operate in complex environments for extended periods, their operating status influenced by various factors, including internal mechanical, electrical, and hydraulic conditions, as well as external grid operating conditions. When hydropower units experience abnormal conditions, failure to promptly identify the evolving trends and implement appropriate scheduling measures can lead to performance degradation and even affect the safe and stable operation of the power grid. Therefore, constructing a safe scheduling method capable of real-time sensing of hydropower unit operating status, intelligently identifying abnormal risks, and combining equipment safety status for scheduling control is of great significance for improving the operational reliability of hydropower units and ensuring the safe operation of the power grid. However, existing hydropower scheduling methods still have the following shortcomings:

[0003] 1. The current methods for detecting the operating status of hydropower units mainly rely on a single operating parameter or a fixed threshold to determine anomalies. They do not fully explore the correlation and change characteristics between multiple sources of operating data, which leads to the failure to identify early abnormal states of the equipment in a timely manner.

[0004] 2. The methods for handling abnormalities in hydropower units are mainly based on matching and judging according to preset rules or historical failure cases. They lack the ability to autonomously reason about the correlation between multiple coupled abnormal states and the evolution process, resulting in low efficiency in abnormal response.

[0005] 3. The hydropower unit dispatch control method only adjusts power based on the grid dispatch target, without dynamically linking the unit's health status, abnormal risk trends, and equipment safety operation boundaries with the dispatch strategy, resulting in insufficient safe adjustment capability of the unit under complex operating conditions. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention provides a safety scheduling system and method based on a smart operation and maintenance intelligent agent for hydropower units.

[0007] On the one hand, the present invention provides a safety scheduling system based on a smart operation and maintenance intelligent agent for hydropower units, the system comprising: an operation perception module, a status cognition module, an anomaly reasoning module, a safety decision-making module, an execution control module, and an intelligent optimization module;

[0008] The operation perception module is used to acquire multi-source operation data during the operation of the hydropower unit and send the multi-source operation data to the status cognition module;

[0009] The state cognition module is used to perform feature extraction and correlation fusion processing on the multi-source operation data to generate health status information that characterizes the dynamic health status of the hydropower unit, and send the health status information to the anomaly inference module.

[0010] The anomaly reasoning module is used to perform anomaly reasoning based on the health status information and the intelligent operation and maintenance knowledge model, generate anomaly diagnosis results and corresponding handling strategies, and send the anomaly diagnosis results and corresponding handling strategies to the security decision module.

[0011] The safety decision module is used to generate a safety dispatch control scheme based on the anomaly diagnosis results, the handling strategy, and the safety operation constraints of the hydropower unit, combined with the power grid dispatch objectives, and send the safety dispatch control scheme to the execution control module.

[0012] The execution control module is used to control the hydropower unit to perform operation adjustment operations according to the safety dispatch control scheme, and send the status feedback information during the execution process to the intelligent optimization module;

[0013] The intelligent optimization module is used to update the intelligent operation and maintenance knowledge model and the safety scheduling decision model according to the status feedback information, and to feed back the updated model parameters to the status cognition module and the anomaly reasoning module.

[0014] Furthermore, the state cognition module includes a feature extraction unit, an association construction unit, and a health representation unit;

[0015] The feature extraction unit is used to extract time-series features from the multi-source operation data to obtain state features characterizing the change process of the hydropower unit's operating state.

[0016] The association construction unit is used to construct running state association relationships based on the changing relationships between different state characteristics;

[0017] The health characterization unit is used to generate dynamic health status characterization values ​​based on the operational status correlation and the state characteristic change process. The dynamic health status characterization values ​​are calculated using the following formula:

[0018]

[0019] in, This represents the dynamic health status indicator value of the hydropower unit; This represents the correlation score between different operational state characteristics at the current moment; This indicates the offset of the current operating state characteristics relative to the baseline operating state characteristics; This represents the baseline characteristic value corresponding to normal operating conditions; This indicates the rate of change of operational status characteristics over time; Indicates the duration of the state change.

[0020] Furthermore, the anomaly reasoning module includes an anomaly association construction unit, an anomaly evolution reasoning unit, and a handling strategy generation unit;

[0021] The abnormal association construction unit is used to establish the association relationship between different operating state characteristics based on the state change characteristics in the health status information, and to construct the abnormal state association model of the hydropower unit based on the association relationship.

[0022] The abnormal evolution reasoning unit is used to obtain the state evolution path corresponding to the current abnormal state according to the abnormal state association model, and to deduce the development trend of the abnormal state along the state evolution path in combination with real-time running status information, so as to obtain the scope of abnormal influence and future state change results.

[0023] The disposal strategy generation unit is used to generate corresponding disposal strategies based on the scope of the anomaly, the future state change results, and the current operating state of the hydropower unit, and to update the anomaly state association model based on the state change after the disposal strategy is executed.

[0024] Furthermore, the security decision module includes a security constraint generation unit, a collaborative scheduling decision unit, and a dynamic verification unit;

[0025] The safety constraint generation unit is used to dynamically adjust the equipment safety operation constraints based on the health status information, abnormal risk status information, and the changing trend of the hydropower unit's operating status.

[0026] The collaborative scheduling decision unit is used to associate and match the power grid scheduling target with the hydropower unit operating capacity according to the dynamically adjusted safety operation constraints, and generate a collaborative scheduling control scheme that meets the equipment safety status and power grid regulation requirements.

[0027] The dynamic verification unit is used to perform real-time verification of the collaborative scheduling control scheme based on the status feedback information after the hydropower unit executes the scheduling control scheme, and to adjust the subsequent scheduling control scheme when the equipment status is detected to deviate from the safe operation constraints.

[0028] Furthermore, the intelligent optimization module includes a feedback analysis unit, a model update unit, and a knowledge synchronization unit;

[0029] The feedback analysis unit is used to obtain the running status information fed back by the execution control module, and to evaluate the abnormal diagnosis results and the execution effect of the safety scheduling control scheme.

[0030] The model update unit is used to update the abnormal correlations in the intelligent operation and maintenance knowledge model and the control parameters in the safety scheduling decision model according to the evaluation results.

[0031] The knowledge synchronization unit is used to synchronize the updated model parameters to the state cognition module and the anomaly reasoning module.

[0032] On the other hand, the present invention provides a safety scheduling method based on a smart operation and maintenance agent for hydropower units, applied to the above-mentioned system, the method comprising the following steps:

[0033] S1, acquire multi-source operation data during the operation of the hydropower unit, and send the multi-source operation data to the status cognition module;

[0034] S2, perform feature extraction and correlation fusion processing on the multi-source operation data to generate health status information characterizing the dynamic health status of the hydropower unit;

[0035] S3. Based on the health status information and the intelligent operation and maintenance knowledge model, perform abnormal state reasoning to generate abnormal diagnosis results and corresponding handling strategies.

[0036] S4. Based on the abnormal diagnosis results, the handling strategy, and the safety operation constraints of the hydropower unit, a safety dispatch control scheme is generated in conjunction with the power grid dispatch objectives.

[0037] S5, according to the safety dispatch control scheme, control the hydropower unit to perform operation adjustment operation, and obtain status feedback information during the execution process;

[0038] S6. Update the intelligent operation and maintenance knowledge model and the security scheduling decision model according to the status feedback information, and feed back the updated model parameters to the status cognition module and the anomaly reasoning module.

[0039] Furthermore, step S2 includes:

[0040] The multi-source operation data is subjected to time-series feature extraction to obtain state features that characterize the change process of the hydropower unit's operating state. The operation state correlation relationship is constructed according to the change relationship between different state features, and a dynamic health status characterization value is generated according to the operation state correlation relationship and the change process of state features.

[0041] The dynamic health status representation value is calculated using the following formula:

[0042]

[0043] in, This represents the dynamic health status indicator value of the hydropower unit; This represents the correlation score between different operational state characteristics at the current moment; This indicates the offset of the current operating state characteristics relative to the baseline operating state characteristics; This represents the baseline characteristic value corresponding to normal operating conditions; This indicates the rate of change of operational status characteristics over time; Indicates the duration of the state change.

[0044] Furthermore, step S3 includes:

[0045] Based on the status change characteristics in the health status information, establish the correlation between different operating status characteristics, construct an abnormal status correlation model based on the correlation, obtain the status evolution path corresponding to the current abnormal status based on the abnormal status correlation model, and combine real-time operating status information to deduce the development trend of the abnormal status along the status evolution path to obtain the scope of abnormal impact and future status change results. Based on the scope of abnormal impact, the future status change results, and the current operating status, generate a disposal strategy.

[0046] Furthermore, step S4 includes:

[0047] Based on the health status information, abnormal risk status information, and the changing trend of the hydropower unit's operating status, the safety operation constraints of the equipment are dynamically adjusted. Based on the dynamically adjusted safety operation constraints, the power grid dispatching objectives are correlated and matched with the operating capacity of the hydropower unit to generate a collaborative dispatching control scheme that meets the safety status of the equipment and the regulation needs of the power grid. The collaborative dispatching control scheme is then dynamically adjusted based on the status feedback information during the execution process.

[0048] Furthermore, step S6 includes:

[0049] The system acquires status feedback information after the hydropower unit executes the safety dispatch control scheme. Based on the status feedback information, it evaluates the anomaly reasoning results and the execution effect of the safety dispatch control scheme. Based on the evaluation results, it updates the anomaly correlation in the intelligent operation and maintenance knowledge model and the control parameters in the safety dispatch decision model. The updated model parameters are then applied to the subsequent hydropower unit operation status perception, anomaly state reasoning, and safety dispatch control processes.

[0050] Compared with existing technologies, the present invention has the following advantages: The present invention acquires multi-source operational data of hydropower units through an operational perception module, and performs feature extraction and correlation fusion processing by a state cognition module to generate dynamic health status representation values, thereby achieving comprehensive perception of the hydropower unit's operational status and improving early anomaly identification capabilities; An anomaly reasoning module constructs an anomaly state correlation model, infers anomaly development trends based on state evolution paths, and generates handling strategies, enabling autonomous reasoning and proactive handling in multi-factor coupled anomaly scenarios, thus improving anomaly handling efficiency; A safety decision module generates a collaborative scheduling control scheme based on health status information, anomaly risk status information, and safe operation constraints, and dynamically optimizes it in conjunction with state feedback, achieving collaborative control of equipment health status and scheduling strategies, thereby improving the safe regulation capability and operational stability of hydropower units in complex power grid environments. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 A structural diagram of a safety scheduling system based on a smart operation and maintenance agent for hydropower units, provided in an embodiment of the present invention;

[0053] Figure 2 A flowchart of a safety scheduling method based on a smart operation and maintenance agent for hydropower units, provided as an embodiment of the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely one unit of embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0055] As one embodiment of the present invention, such as Figure 1 As shown in the figure, this embodiment provides a safety scheduling system based on a smart operation and maintenance intelligent agent for hydropower units. The system includes: an operation perception module, a status cognition module, an anomaly reasoning module, a safety decision-making module, an execution control module, and an intelligent optimization module.

[0056] The operation perception module is used to acquire multi-source operation data during the operation of the hydropower unit and send the multi-source operation data to the status cognition module;

[0057] The state cognition module is used to perform feature extraction and correlation fusion processing on the multi-source operation data to generate health status information that characterizes the dynamic health status of the hydropower unit, and send the health status information to the anomaly inference module.

[0058] The anomaly reasoning module is used to perform anomaly reasoning based on the health status information and the intelligent operation and maintenance knowledge model, generate anomaly diagnosis results and corresponding handling strategies, and send the anomaly diagnosis results and corresponding handling strategies to the security decision module.

[0059] The safety decision module is used to generate a safety dispatch control scheme based on the anomaly diagnosis results, the handling strategy, and the safety operation constraints of the hydropower unit, combined with the power grid dispatch objectives, and send the safety dispatch control scheme to the execution control module.

[0060] The execution control module is used to control the hydropower unit to perform operation adjustment operations according to the safety dispatch control scheme, and send the status feedback information during the execution process to the intelligent optimization module;

[0061] The intelligent optimization module is used to update the intelligent operation and maintenance knowledge model and the safety scheduling decision model according to the status feedback information, and to feed back the updated model parameters to the status cognition module and the anomaly reasoning module.

[0062] Specifically, the operation sensing module is used to acquire multi-source operation data during the operation of the hydropower unit. The operation sensing module establishes data connections with the existing monitoring and control systems of the hydropower unit, as well as the power grid dispatch data interface, and continuously acquires multi-source operation data reflecting the current operating conditions and equipment status of the hydropower unit according to a preset operation status evaluation cycle. For situations where different data sources have inconsistent sampling times and cycles, the multi-source operation data is time-aligned according to a unified time benchmark. Data belonging to the same operation status evaluation cycle is divided into corresponding data processing windows, and abnormal and missing sampled values ​​are preprocessed to form a continuous and time-corresponding multi-source operation data sequence. After completing data processing, the operation sensing module sends the resulting multi-source operation data to the status cognition module as input data for representing the dynamic health status of the hydropower unit.

[0063] By performing unified time-series processing on multi-source operating data from different sources of hydropower units, different operating status information can be mapped to the same operating conditions and evaluation cycle, reducing the impact of data sampling time differences on state correlation analysis, and providing a unified data foundation for the subsequent construction of correlation relationships between different operating status characteristics and dynamic health status characterization.

[0064] Furthermore, the state cognition module includes a feature extraction unit, an association construction unit, and a health representation unit;

[0065] The feature extraction unit is used to extract time-series features from the multi-source operation data to obtain state features characterizing the change process of the hydropower unit's operating state.

[0066] The association construction unit is used to construct running state association relationships based on the changing relationships between different state characteristics;

[0067] The health characterization unit is used to generate dynamic health status characterization values ​​based on the operational status correlation and the state characteristic change process. The dynamic health status characterization values ​​are calculated using the following formula:

[0068]

[0069] in, This represents the dynamic health status indicator value of the hydropower unit; This represents the correlation score between different operational state characteristics at the current moment; This indicates the offset of the current operating state characteristics relative to the baseline operating state characteristics; This represents the baseline characteristic value corresponding to normal operating conditions; This indicates the rate of change of operational status characteristics over time; Indicates the duration of the state change.

[0070] Specifically, the state cognition module includes a feature extraction unit, an association construction unit, and a health characterization unit. The feature extraction unit extracts temporal features from the multi-source operational data sent by the operation perception module according to the continuous operation state evaluation window, and establishes a corresponding baseline operation state based on the current operating condition of the hydropower unit and historical data of normal operation, obtaining state features characterizing the process of changes in the hydropower unit's operation state. The association construction unit constructs operational state associations based on the synchronous changes of different state features within the continuous evaluation period, and obtains association scores between different operational state features at the current moment. The health characterization unit calculates dynamic health state characterization values ​​based on the operational state associations and the process of changes in state features, and then sends the hydropower unit's health state information to the anomaly inference module based on the dynamic health state characterization values.

[0071] By simultaneously introducing the correlation scores between different operating state characteristics, the degree of deviation of the current operating state from the normal operating state, and the continuous accumulation process of state changes, the dynamic health status characterization value not only reflects whether a single operating parameter is abnormal, but also whether the original correlation between different operating state characteristics is destroyed and whether the abnormal state continues to develop. This overcomes the problem that a single monitoring parameter and a fixed threshold are difficult to reflect the overall dynamic health status of the hydropower unit, and is conducive to identifying the early abnormal state corresponding to the changes in the correlation between multiple parameters before a single parameter reaches the abnormal threshold.

[0072] Furthermore, the anomaly reasoning module includes an anomaly association construction unit, an anomaly evolution reasoning unit, and a handling strategy generation unit;

[0073] The abnormal association construction unit is used to establish the association relationship between different operating state characteristics based on the state change characteristics in the health status information, and to construct the abnormal state association model of the hydropower unit based on the association relationship.

[0074] The abnormal evolution reasoning unit is used to obtain the state evolution path corresponding to the current abnormal state according to the abnormal state association model, and to deduce the development trend of the abnormal state along the state evolution path in combination with real-time running status information, so as to obtain the scope of abnormal influence and future state change results.

[0075] The disposal strategy generation unit is used to generate corresponding disposal strategies based on the scope of the anomaly, the future state change results, and the current operating state of the hydropower unit, and to update the anomaly state association model based on the state change after the disposal strategy is executed.

[0076] Specifically, the anomaly reasoning module includes an anomaly association construction unit, an anomaly evolution reasoning unit, and a disposal strategy generation unit. The anomaly association construction unit receives health status information output by the state cognition module and constructs an anomaly state association model based on the characteristics of anomaly changes and the correlation between corresponding operating states in the intelligent operation and maintenance knowledge model. This allows different anomaly states to form a continuously reasoning-enabled association structure according to their operational correlations and evolutionary order. The anomaly evolution reasoning unit obtains the corresponding state evolution path from the anomaly state association model based on the characteristics of the currently occurring anomaly changes. It compares the real-time operating state with the state change direction corresponding to each candidate state evolution path, excluding state evolution paths inconsistent with real-time operating state changes, and continuing to reason forward for state evolution paths consistent with real-time operating state changes to obtain the potential impact range of the current anomaly and the future state change results. The disposal strategy generation unit identifies disposal locations from the state evolution path that can block the continued propagation of the anomaly based on the anomaly impact range, future state change results, and the current operating state of the hydropower unit. It then forms corresponding disposal strategies based on these disposal locations and sends the anomaly diagnosis results and disposal strategies to the safety decision module.

[0077] By forming an abnormal state association model based on state feature correlations and dynamically filtering state evolution paths according to real-time operating status, the abnormal reasoning process is transformed from static matching of historical fault cases to dynamic evolution reasoning for the current multi-factor coupled state. Furthermore, by selecting the handling position that can block the propagation of abnormality based on subsequent abnormal states that have not yet actually occurred, a handling strategy adapted to the current operating state can be generated before the abnormality continues to expand, thereby reducing the dependence on fixed rules and human experience and improving the timeliness of abnormal reasoning and fault handling in multi-factor coupled abnormal scenarios.

[0078] Furthermore, the security decision module includes a security constraint generation unit, a collaborative scheduling decision unit, and a dynamic verification unit;

[0079] The safety constraint generation unit is used to dynamically adjust the equipment safety operation constraints based on the health status information, abnormal risk status information, and the changing trend of the hydropower unit's operating status.

[0080] The collaborative scheduling decision unit is used to associate and match the power grid scheduling target with the hydropower unit operating capacity according to the dynamically adjusted safety operation constraints, and generate a collaborative scheduling control scheme that meets the equipment safety status and power grid regulation requirements.

[0081] The dynamic verification unit is used to perform real-time verification of the collaborative scheduling control scheme based on the status feedback information after the hydropower unit executes the scheduling control scheme, and to adjust the subsequent scheduling control scheme when the equipment status is detected to deviate from the safe operation constraints.

[0082] Specifically, the security decision module includes a security constraint generation unit, a collaborative scheduling decision unit, and a dynamic verification unit. The safety constraint generation unit receives the anomaly diagnosis results and handling strategies sent by the anomaly reasoning module, and combines them with the health status information obtained by the state cognition module and the changing trends of the hydropower unit's operating status to dynamically adjust the original safe operation boundaries of the hydropower unit, forming safe operation constraints that match the current equipment health status and anomaly risks. When the dynamic health status characterization value decreases or the anomaly status shows a trend of continuing to propagate to the critical operating status, the current allowable operating adjustment range of the hydropower unit is narrowed. When the dynamic health status characterization value recovers and the anomaly risk decreases, the corresponding operating adjustment range is gradually restored based on continuous state feedback. Within the dynamically adjusted safe operation constraints, the collaborative scheduling decision unit associates and matches the power grid scheduling target with the current safe operating capacity of the hydropower unit. When the power grid scheduling target is within the current safe adjustment capacity range, a collaborative scheduling control scheme is formed according to the power grid scheduling target. When the power grid scheduling target exceeds the current safe adjustment capacity range, the actual adjustment range is corrected according to the dynamic safe operation constraints. The dynamic verification unit continuously receives state feedback information during the execution of the scheduling control scheme. When it detects that the equipment status deviates from the current safe operation constraints, the subsequent scheduling control scheme is readjusted.

[0083] This embodiment dynamically adjusts safe operation constraints based on health status information, abnormal risk status information, and operational status change trends. This allows the equipment health status and abnormal risks of the hydropower unit to directly participate in the generation process of the dispatching plan, rather than merely serving as evaluation information after dispatching. Furthermore, by associating and matching the grid dispatching objectives with the unit's current safe operating capacity, the unit proactively reduces its adjustment range when its health status declines and gradually restores its adjustment capacity when its health status recovers. This achieves dynamic coordination between equipment safety status and grid adjustment needs, improving the safe adjustment capability of hydropower units in complex grid environments.

[0084] Specifically, the execution control module receives the safety scheduling control scheme formed by the safety decision module, generates corresponding operation adjustment instructions based on the safety scheduling control scheme, and sends the operation adjustment instructions to the hydropower unit control system, so that the hydropower unit performs the corresponding operation adjustment operation according to the safety scheduling control scheme. During the operation adjustment process, the execution control module continuously acquires the actual operating status of the hydropower unit according to a preset control cycle, and compares the actual operating status with the target operating status corresponding to the safety scheduling control scheme and the current safety operating constraints. When the actual operating status meets the safety operating constraints, the subsequent operation adjustment continues to be executed. When the actual operating status approaches or exceeds the safety operating constraints, the corresponding operation adjustment stops and status feedback information is generated. The status feedback information is sent to the intelligent optimization module, and also serves as the feedback basis for the safety decision module to dynamically verify the subsequent scheduling control scheme.

[0085] This embodiment continuously collects the actual operating status during the execution of the safety dispatch control scheme and performs process verification in combination with dynamic safety operation constraints. This transforms the dispatch scheme from a one-time control issuance to a closed-loop control process with operational feedback. This reduces the equipment operation risk caused by continuing to execute the original dispatch scheme when there is a deviation between the model prediction state and the actual response of the unit, and improves the actual execution reliability of the safety dispatch control scheme.

[0086] Furthermore, the intelligent optimization module includes a feedback analysis unit, a model update unit, and a knowledge synchronization unit;

[0087] The feedback analysis unit is used to obtain the running status information fed back by the execution control module, and to evaluate the abnormal diagnosis results and the execution effect of the safety scheduling control scheme.

[0088] The model update unit is used to update the abnormal correlations in the intelligent operation and maintenance knowledge model and the control parameters in the safety scheduling decision model according to the evaluation results.

[0089] The knowledge synchronization unit is used to synchronize the updated model parameters to the state cognition module and the anomaly reasoning module.

[0090] Specifically, the intelligent optimization module includes a feedback analysis unit, a model update unit, and a knowledge synchronization unit. The feedback analysis unit acquires state feedback information sent by the execution control module, compares the future state change results generated by the anomaly inference module with the actual state change process after execution, and compares the target operating state corresponding to the safety scheduling control scheme with the actual operating state of the hydropower unit to form anomaly inference results and an evaluation result of the scheduling control effect. The model update unit adjusts the anomaly correlations in the intelligent operation and maintenance knowledge model and the control parameters in the safety scheduling decision model based on the evaluation results. When the predicted state evolution path is consistent with the actual state evolution process, it maintains or enhances the correlation between corresponding anomaly states; when the predicted state evolution path is inconsistent with the actual state change, it corrects the corresponding anomaly state correlations and adjusts the corresponding control parameters based on the difference between the scheduling control scheme and the actual equipment response. The knowledge synchronization unit sends the updated model parameters to the state cognition module and the anomaly inference module for subsequent hydropower unit health state cognition and anomaly state inference.

[0091] This embodiment utilizes the actual state feedback after anomaly handling and safe scheduling to continuously update the intelligent operation and maintenance knowledge model and the safe scheduling decision model. This allows the correlation between abnormal states and the safe scheduling control parameters to be corrected as the actual operating state of the hydropower unit changes, avoiding the mismatch between the intelligent agent model and the actual equipment state caused by long-term use of fixed knowledge relationships and fixed control parameters.

[0092] As one embodiment of the present invention, such as Figure 2 As shown, this embodiment provides a safety scheduling method based on a smart operation and maintenance agent for hydropower units, applied to the aforementioned system. The method includes the following steps:

[0093] S1, acquire multi-source operation data during the operation of the hydropower unit, and send the multi-source operation data to the status cognition module;

[0094] S2, perform feature extraction and correlation fusion processing on the multi-source operation data to generate health status information characterizing the dynamic health status of the hydropower unit;

[0095] S3. Based on the health status information and the intelligent operation and maintenance knowledge model, perform abnormal state reasoning to generate abnormal diagnosis results and corresponding handling strategies.

[0096] S4. Based on the abnormal diagnosis results, the handling strategy, and the safety operation constraints of the hydropower unit, a safety dispatch control scheme is generated in conjunction with the power grid dispatch objectives.

[0097] S5, according to the safety dispatch control scheme, control the hydropower unit to perform operation adjustment operation, and obtain status feedback information during the execution process;

[0098] S6. Update the intelligent operation and maintenance knowledge model and the security scheduling decision model according to the status feedback information, and feed back the updated model parameters to the status cognition module and the anomaly reasoning module.

[0099] Specifically, the operation perception module continuously acquires multi-source operation data during the operation of the hydropower unit according to the preset operation status evaluation cycle. Data from different sampling cycles and different sampling times are time-aligned, and data processing windows are divided with a unified time benchmark. Abnormal and missing sampling values ​​within the data processing windows are processed to form a continuous multi-source operation data sequence. The processed multi-source operation data is then sent to the status cognition module.

[0100] Furthermore, step S2 includes:

[0101] The multi-source operation data is subjected to time-series feature extraction to obtain state features that characterize the change process of the hydropower unit's operating state. The operation state correlation relationship is constructed according to the change relationship between different state features, and a dynamic health status characterization value is generated according to the operation state correlation relationship and the change process of state features.

[0102] Specifically, time-series features are extracted from multi-source operational data according to the continuous operational status evaluation window to obtain... The system identifies operational status characteristics to characterize the changes in the operating status of hydropower units, and obtains the corresponding normal operating baseline status from the historical normal operating data of the hydropower units based on the current operating conditions. The system constructs operational status correlation relationships based on the changing relationships between different operational status characteristics, and calculates the dynamic health status characterization value of the hydropower units based on the operational status correlation score, the degree of deviation of the current operational status characteristics from the baseline operational status characteristics, and the continuous change process of the operational status over time.

[0103] The dynamic health status representation value is calculated using the following formula:

[0104]

[0105] in, This represents the dynamic health status indicator value of the hydropower unit; This represents the correlation score between different operational state characteristics at the current moment; This indicates the offset of the current operating state characteristics relative to the baseline operating state characteristics; This represents the baseline characteristic value corresponding to normal operating conditions; This indicates the rate of change of operational status characteristics over time; Indicates the duration of the state change.

[0106] (1) Operational status related score Calculation

[0107] The operational status correlation score is calculated based on the degree of correlation change between current operational status characteristics:

[0108]

[0109] in, Indicates the first Each operational state characteristic at the current moment eigenvalues; Indicates the first Each operational state characteristic at the current moment eigenvalues; Indicates the first The historical average value of each operating status feature within the corresponding normal operating time window; Indicates the first The historical average value of each operating status feature within the corresponding normal operating time window; Indicates the first The standard deviation of the variation of each operating status characteristic within the normal operating time window; Indicates the first The standard deviation of the variation of each operating status characteristic within the normal operating time window.

[0110] Specifically, firstly, the historical average value and standard deviation of each operating state feature are obtained from the historical normal operation data corresponding to the current operating conditions of the hydropower unit. Then, the current deviation of any two operating state features from their respective historical average values ​​is calculated. The product of the two deviations is compared with the product of the corresponding standard deviations to obtain the correlation status of the corresponding state feature pairs. Finally, all state feature pairs are averaged to obtain the operating state correlation score at the current moment. .

[0111] (2) State offset Calculation

[0112] The state offset is calculated according to the following formula:

[0113]

[0114] in, Indicates the first The normal operating baseline value corresponding to each operating status characteristic.

[0115] Specifically, based on the operating conditions of the current head and current operating load, the normal operating baseline values ​​corresponding to each operating state characteristic are obtained from the historical normal operating data. Calculate the absolute difference between the current state characteristic value and the normal operating baseline value, and then calculate... The average of the absolute differences of the state features is used to obtain the overall offset of the current operating state relative to the normal operating state. .

[0116] (3) Benchmark eigenvalues Calculation

[0117] The benchmark feature value is calculated according to the following formula:

[0118]

[0119] in, Indicates the number of uptime windows. Indicates the first The status characteristic values ​​corresponding to each normal operating time window.

[0120] Specifically, data is selected from historical normal operating data corresponding to the current operating conditions of the hydropower units. For each consecutive or intermittently set normal operation time window, obtain the corresponding status characteristic value for each normal operation time window. and calculate The average value of the state feature values ​​of each normal operating time window is used as the benchmark feature corresponding to the current operating condition. This reduces the impact of random fluctuations during a single normal operating moment on the baseline state.

[0121] (4) Rate of state change Calculation

[0122] The rate of state change is calculated using the state characteristic changes at adjacent sampling times:

[0123]

[0124] in, Indicates the current sampling time The corresponding running state characteristic value, This represents the operating state feature value corresponding to the previous sampling period. This indicates the sampling time interval between adjacent sampling moments.

[0125] Specifically, during the continuous operation status evaluation process, the operation status feature values ​​corresponding to the current evaluation time and the previous evaluation time are obtained. The difference between the operation status feature values ​​at the two evaluation times is divided by the sampling time interval to obtain the rate of change of the current operation status feature over time.

[0126] (5) Duration of state change Acquisition

[0127] The duration of the state change This indicates the duration during which the current operating state characteristic continuously deviates from the normal operating state. When the operating state characteristic first exhibits a continuous deviation from the normal fluctuation range, the corresponding start time is recorded, and this is continuously assessed during subsequent operating state evaluations to determine whether the operating state has recovered to the normal fluctuation range. If the operating state has not recovered, the time difference between the current evaluation time and the aforementioned start time is taken as the duration of the state change. Once the operating status returns to the normal fluctuation range, the start time will be recorded again.

[0128] The dynamic health status characterization value is obtained from the above parameters. When the correlation score between different operational state characteristics decreases, Changes occur; when the overall operating state of the hydropower unit gradually deviates from the normal operating state. Increase; when the state deviates from continuous development, the integral term accumulates continuously with the state change process, thus... It can simultaneously reflect the correlation changes between multiple operating states, the degree of deviation of the current state, and the degree of continuous evolution of anomalies.

[0129] This embodiment uses the correlation scores between different operating state characteristics, the offset of the operating state relative to the normal operating benchmark, and the cumulative change of the operating state over time to form a dynamic health status characterization value. This transforms the evaluation of the health status of hydropower units from a single-parameter static judgment to a dynamic characterization that combines the effects of multi-state correlation and time evolution. In particular, when individual operating parameters have not yet reached the fixed abnormal threshold, but the original correlation between parameters has changed or the status continues to deviate from the normal operating trend, the dynamic health status characterization value can reflect potential abnormalities in advance, thereby improving the ability to identify early abnormal states of hydropower units.

[0130] Furthermore, step S3 includes:

[0131] Based on the status change characteristics in the health status information, establish the correlation between different operating status characteristics, construct an abnormal status correlation model based on the correlation, obtain the status evolution path corresponding to the current abnormal status based on the abnormal status correlation model, and combine real-time operating status information to deduce the development trend of the abnormal status along the status evolution path to obtain the scope of abnormal impact and future status change results. Based on the scope of abnormal impact, the future status change results, and the current operating status, generate a disposal strategy.

[0132] Specifically, based on the status change characteristics in the health status information obtained in step S2, the operational status characteristics with abnormal changes are mapped to the corresponding operational status in the intelligent operation and maintenance knowledge model. An abnormal status association model is formed based on the correlation between each operational status and the historical evolution sequence of abnormal statuses. Taking the currently identified abnormal status as the starting point for reasoning, candidate status evolution paths connected to the current abnormal status are extracted from the abnormal status association model. Each candidate status evolution path is dynamically filtered according to the real-time operational status. Status evolution paths that are inconsistent with the current real-time status change direction are excluded, and status evolution paths that are consistent with the current real-time status change are further extrapolated to obtain the scope of the abnormal impact and the future status change results. Based on the scope of the abnormal impact and the future status change results, status nodes that can block the continued propagation of the abnormality are identified, and corresponding disposal strategies are generated in combination with the current operational status of the hydropower unit.

[0133] This embodiment dynamically filters the state evolution path in the abnormal state association model based on the real-time operating status, so that the abnormal diagnosis is no longer just based on fixed rules or historical cases for result matching, but can continuously deduce the future development process of the abnormal by combining the current multi-factor coupled operating status, and actively generate handling strategies for the direction of abnormal propagation, thereby improving the autonomous reasoning ability and fault handling efficiency in complex abnormal scenarios.

[0134] Furthermore, step S4 includes:

[0135] Based on the health status information, abnormal risk status information, and the changing trend of the hydropower unit's operating status, the safety operation constraints of the equipment are dynamically adjusted. Based on the dynamically adjusted safety operation constraints, the power grid dispatching objectives are correlated and matched with the operating capacity of the hydropower unit to generate a collaborative dispatching control scheme that meets the safety status of the equipment and the regulation needs of the power grid. The collaborative dispatching control scheme is then dynamically adjusted based on the status feedback information during the execution process.

[0136] Specifically, based on the health status information formed in step S2 and the abnormal impact range and future state change results formed in step S3, the current abnormal risk status of the hydropower unit is obtained, and the safe operation boundary of the equipment is dynamically adjusted according to the health status and abnormal risk status. When the dynamic health status characterization value decreases or the abnormal development trend points to the critical equipment status, the current allowable operation adjustment range of the hydropower unit is narrowed. When the health status recovers and the abnormal development trend weakens, the operation adjustment range is gradually restored according to continuous state feedback. Within the dynamically adjusted safe operation constraints, the grid dispatch target is associated and matched with the current safe operation capacity of the hydropower unit, and the adjustment target that meets the safe operation constraints is converted into a collaborative dispatch control scheme. During the execution of the collaborative dispatch control scheme, the safety constraints are continuously checked according to the latest state feedback information, and the subsequent dispatch control scheme is adjusted when the operation status deviates from the current safe operation constraints.

[0137] This embodiment dynamically changes the safety operation constraints of hydropower units based on real-time health status and abnormal risks, so that the equipment health status, abnormal risks and safety operation boundaries directly participate in the generation of dispatching schemes, and uses dynamic safety operation constraints to limit the grid dispatching objectives, so as to achieve coordinated matching between grid dispatching needs and actual equipment safety regulation capabilities, reduce the risk of hydropower units with declining health status undertaking excessive regulation tasks, and improve the unit safety regulation capabilities in complex grid environments.

[0138] Specifically, the safety dispatch control scheme formed in step S4 is converted into the corresponding operation adjustment command of the hydropower unit and sent to the hydropower unit control system for execution; the actual operating status of the hydropower unit is continuously acquired according to the preset control cycle, and the actual operating status is compared with the target operating status corresponding to the collaborative dispatch control scheme and the current safety operating constraints. When the actual operating status meets the safety operating constraints, the subsequent operation adjustment continues to be executed; when the actual operating status is close to or exceeds the safety operating constraints, the corresponding operation adjustment is stopped, and the current operating status forms a status feedback information and is sent to the safety decision module and the intelligent optimization module.

[0139] This embodiment continuously verifies safety constraints based on the actual operating status during the execution of the scheduling scheme, enabling the hydropower unit scheduling process to be adjusted in a timely manner according to the actual response of the equipment, reducing continuous erroneous adjustments caused by the deviation between the predicted operating status and the actual operating status, and improving the reliability of the safety scheduling control process.

[0140] Furthermore, step S6 includes:

[0141] The system acquires status feedback information after the hydropower unit executes the safety dispatch control scheme. Based on the status feedback information, it evaluates the anomaly reasoning results and the execution effect of the safety dispatch control scheme. Based on the evaluation results, it updates the anomaly correlation in the intelligent operation and maintenance knowledge model and the control parameters in the safety dispatch decision model. The updated model parameters are then applied to the subsequent hydropower unit operation status perception, anomaly state reasoning, and safety dispatch control processes.

[0142] Specifically, the state feedback information generated in step S5 is obtained, and the future state change results derived in step S3 are compared with the actual state change results of the hydropower unit. At the same time, the target state corresponding to the safety dispatch control scheme generated in step S4 is compared with the actual operating state. When the predicted abnormal state evolution path is consistent with the actual state change process, the correlation between the corresponding abnormal states in the intelligent operation and maintenance knowledge model is retained or enhanced. When the predicted state evolution path is inconsistent with the actual state change process, the corresponding abnormal state correlation is corrected. The control parameters in the safety dispatch decision model are corrected according to the actual operating response after the execution of the safety dispatch control scheme, and the updated model parameters are fed back to the state cognition module and the abnormal reasoning module for subsequent dynamic health state representation and abnormal state reasoning.

[0143] This embodiment continuously updates the intelligent operation and maintenance knowledge model and the safe scheduling decision model based on the differences between the anomaly evolution prediction results and scheduling control results and the actual operating state. This enables the state correlation, anomaly evolution relationship and scheduling control parameters to be continuously corrected as the long-term operating state of the hydropower unit changes, forming a closed-loop optimization mechanism for the intelligent operation and maintenance agent. This improves the adaptability of subsequent state cognition, anomaly reasoning and safe scheduling results to the actual operating state of the hydropower unit.

[0144] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A safety dispatching system based on a smart operation and maintenance intelligent agent for hydropower units, characterized in that: The system includes: a runtime perception module, a state recognition module, an anomaly reasoning module, a security decision-making module, an execution control module, and an intelligent optimization module; The operation perception module is used to acquire multi-source operation data during the operation of the hydropower unit and send the multi-source operation data to the status cognition module; The state cognition module is used to perform feature extraction and correlation fusion processing on the multi-source operation data to generate health status information that characterizes the dynamic health status of the hydropower unit, and send the health status information to the anomaly inference module. The anomaly reasoning module is used to perform anomaly reasoning based on the health status information and the intelligent operation and maintenance knowledge model, generate anomaly diagnosis results and corresponding handling strategies, and send the anomaly diagnosis results and corresponding handling strategies to the security decision module. The safety decision module is used to generate a safety dispatch control scheme based on the anomaly diagnosis results, the handling strategy, and the safety operation constraints of the hydropower unit, combined with the power grid dispatch objectives, and send the safety dispatch control scheme to the execution control module. The execution control module is used to control the hydropower unit to perform operation adjustment operations according to the safety dispatch control scheme, and send the status feedback information during the execution process to the intelligent optimization module; The intelligent optimization module is used to update the intelligent operation and maintenance knowledge model and the safety scheduling decision model according to the status feedback information, and to feed back the updated model parameters to the status cognition module and the anomaly reasoning module.

2. The safety dispatching system based on the intelligent operation and maintenance agent of hydropower units according to claim 1, characterized in that: The state cognition module includes a feature extraction unit, an association construction unit, and a health representation unit; The feature extraction unit is used to extract time-series features from the multi-source operation data to obtain state features characterizing the change process of the hydropower unit's operating state. The association construction unit is used to construct running state association relationships based on the changing relationships between different state characteristics; The health characterization unit is used to generate dynamic health status characterization values ​​based on the operational status correlation and the state characteristic change process. The dynamic health status characterization values ​​are calculated using the following formula: in, This represents the dynamic health status indicator value of the hydropower unit; This represents the correlation score between different operational state characteristics at the current moment; This indicates the offset of the current operating state characteristics relative to the baseline operating state characteristics; This represents the baseline characteristic value corresponding to normal operating conditions; This indicates the rate of change of operational status characteristics over time; Indicates the duration of the state change.

3. The safety dispatching system based on the intelligent operation and maintenance agent of hydropower units according to claim 2, characterized in that: The anomaly reasoning module includes an anomaly association construction unit, an anomaly evolution reasoning unit, and a handling strategy generation unit; The abnormal association construction unit is used to establish the association relationship between different operating state characteristics based on the state change characteristics in the health status information, and to construct the abnormal state association model of the hydropower unit based on the association relationship. The abnormal evolution reasoning unit is used to obtain the state evolution path corresponding to the current abnormal state according to the abnormal state association model, and to deduce the development trend of the abnormal state along the state evolution path in combination with real-time running status information, so as to obtain the scope of abnormal influence and future state change results. The disposal strategy generation unit is used to generate corresponding disposal strategies based on the scope of the anomaly, the future state change results, and the current operating state of the hydropower unit, and to update the anomaly state association model based on the state change after the disposal strategy is executed.

4. The safety dispatching system based on the intelligent operation and maintenance agent of hydropower units according to claim 3, characterized in that: The security decision module includes a security constraint generation unit, a collaborative scheduling decision unit, and a dynamic verification unit. The safety constraint generation unit is used to dynamically adjust the equipment safety operation constraints based on the health status information, abnormal risk status information, and the changing trend of the hydropower unit's operating status. The collaborative scheduling decision unit is used to associate and match the power grid scheduling target with the hydropower unit operating capacity according to the dynamically adjusted safety operation constraints, and generate a collaborative scheduling control scheme that meets the equipment safety status and power grid regulation requirements. The dynamic verification unit is used to perform real-time verification of the collaborative scheduling control scheme based on the status feedback information after the hydropower unit executes the scheduling control scheme, and to adjust the subsequent scheduling control scheme when the equipment status is detected to deviate from the safe operation constraints.

5. The safety dispatching system based on the intelligent operation and maintenance agent of hydropower units according to claim 4, characterized in that: The intelligent optimization module includes a feedback analysis unit, a model update unit, and a knowledge synchronization unit. The feedback analysis unit is used to obtain the running status information fed back by the execution control module, and to evaluate the abnormal diagnosis results and the execution effect of the safety scheduling control scheme. The model update unit is used to update the abnormal correlations in the intelligent operation and maintenance knowledge model and the control parameters in the safety scheduling decision model according to the evaluation results. The knowledge synchronization unit is used to synchronize the updated model parameters to the state cognition module and the anomaly reasoning module.

6. A safety scheduling method based on a smart operation and maintenance intelligent agent for hydropower units, applied to the system described in any one of claims 1-5, characterized in that: The method includes the following steps: S1, acquire multi-source operation data during the operation of the hydropower unit, and send the multi-source operation data to the status cognition module; S2, perform feature extraction and correlation fusion processing on the multi-source operation data to generate health status information characterizing the dynamic health status of the hydropower unit; S3. Based on the health status information and the intelligent operation and maintenance knowledge model, perform abnormal state reasoning to generate abnormal diagnosis results and corresponding handling strategies. S4. Based on the abnormal diagnosis results, the handling strategy, and the safety operation constraints of the hydropower unit, a safety dispatch control scheme is generated in conjunction with the power grid dispatch objectives. S5, according to the safety dispatch control scheme, control the hydropower unit to perform operation adjustment operation, and obtain status feedback information during the execution process; S6. Update the intelligent operation and maintenance knowledge model and the security scheduling decision model according to the status feedback information, and feed back the updated model parameters to the status cognition module and the anomaly reasoning module.

7. The safety scheduling method based on the intelligent operation and maintenance agent of hydropower units according to claim 6, characterized in that: Step S2 includes: The multi-source operation data is subjected to time-series feature extraction to obtain state features that characterize the change process of the hydropower unit's operating state. The operation state correlation relationship is constructed according to the change relationship between different state features, and a dynamic health status characterization value is generated according to the operation state correlation relationship and the change process of state features. The dynamic health status representation value is calculated using the following formula: in, This represents the dynamic health status indicator value of the hydropower unit; This represents the correlation score between different operational state characteristics at the current moment; This indicates the offset of the current operating state characteristics relative to the baseline operating state characteristics; This represents the baseline characteristic value corresponding to normal operating conditions; This indicates the rate of change of operational status characteristics over time; Indicates the duration of the state change.

8. The safety scheduling method based on the intelligent operation and maintenance agent of hydropower units according to claim 7, characterized in that: Step S3 includes: Based on the status change characteristics in the health status information, establish the correlation between different operating status characteristics, construct an abnormal status correlation model based on the correlation, obtain the status evolution path corresponding to the current abnormal status based on the abnormal status correlation model, and combine real-time operating status information to deduce the development trend of the abnormal status along the status evolution path to obtain the scope of abnormal impact and future status change results. Based on the scope of abnormal impact, the future status change results, and the current operating status, generate a disposal strategy.

9. The safety scheduling method based on the intelligent operation and maintenance agent of hydropower units according to claim 8, characterized in that: Step S4 includes: Based on the health status information, abnormal risk status information, and the changing trend of the hydropower unit's operating status, the safety operation constraints of the equipment are dynamically adjusted. Based on the dynamically adjusted safety operation constraints, the power grid dispatching objectives are correlated and matched with the operating capacity of the hydropower unit to generate a collaborative dispatching control scheme that meets the safety status of the equipment and the regulation needs of the power grid. The collaborative dispatching control scheme is then dynamically adjusted based on the status feedback information during the execution process.

10. The safety scheduling method based on the intelligent operation and maintenance agent of hydropower units according to claim 9, characterized in that: Step S6 includes: The system acquires status feedback information after the hydropower unit executes the safety dispatch control scheme. Based on the status feedback information, it evaluates the anomaly reasoning results and the execution effect of the safety dispatch control scheme. Based on the evaluation results, it updates the anomaly correlation in the intelligent operation and maintenance knowledge model and the control parameters in the safety dispatch decision model. The updated model parameters are then applied to the subsequent hydropower unit operation status perception, anomaly state reasoning, and safety dispatch control processes.

Images