Method and platform for determining the operating envelope of a nuclear power system

Abstract

The present specification provides a method and platform for determining the operating boundary of a nuclear power system. First, according to the current operating condition scene type of the target nuclear power system, the target boundary parameter that needs to be focused on is determined by constructing and based on the operating dependency matrix; then, by constructing a target function that is adapted to the target nuclear power system according to the current operating condition scene type of the target nuclear power system and the operating indication parameter of the target nuclear power system; at the same time, according to the function type of the target function and the parameter number of the target boundary parameter, a target boundary optimization model that matches is determined; then, by jointly using the target boundary optimization model and the preset nuclear power system operation fitting model to explore and optimize the target boundary parameter of the target nuclear power system, the target operating boundary is determined. Thus, under the premise of ensuring the overall operation safety of the target nuclear power system without stopping the reactor, the safety margin of the operating parameter can be fully tapped and utilized, and the target operating boundary with better effect can be accurately determined.

Description

Technical Field

[0001] This manual belongs to the field of nuclear power system operation and control technology, and in particular relates to the methods and platforms for determining the operating boundaries of nuclear power systems. Background Technology

[0002] With the development and promotion of nuclear technology, more and more nuclear power systems or devices (such as nuclear power generation equipment) are being applied to industrial production.

[0003] Existing methods for operating and controlling nuclear power systems are generally conservative, primarily focusing on the system's operational safety requirements. They typically use the "worst-case" principle to set relatively strict operational boundaries for the system's variables. However, these boundaries are often overly conservative, with significant safety margins. While controlling the nuclear power system according to these boundaries can effectively ensure its safety and stability, the reactor's output power and other operational performance are often less than ideal.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This specification provides a method and platform for determining the operating boundary of a nuclear power system. It can be precisely adapted to the specific operating scenario of the target nuclear power system. Under the premise of ensuring the overall safe operation of the target nuclear power system without reactor failure, it fully explores and utilizes the safety margin of operating parameters to accurately determine the target operating boundary with good effect and high application value.

[0006] This specification provides a method for determining the operating boundaries of a nuclear power system, including: Obtain the attribute information of the target nuclear power system and determine the current operating scenario type of the target nuclear power system; Using a pre-defined nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, an operation dependency matrix matching the current operating scenario type of the target nuclear power system is constructed through simulation testing; wherein, the operation dependency matrix includes at least the influence variable that triggers the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variable; Based on the current operating scenario type of the target nuclear power system, influence variables that meet the safety level requirements are selected from the operation dependency matrix as target key variables; and based on the target key variables, the corresponding target boundary parameters are determined. Based on the current operating scenario type of the target nuclear power system and its operating indication parameters, a matching objective function and constraints are constructed; wherein, the objective function includes: a single objective function or a multi-objective function; Based on the function type of the objective function and the number of parameters of the objective boundary parameters, a matching objective boundary optimization model is determined from the preset optimization models; wherein, the preset optimization model includes at least: a preset first optimization model and a preset second optimization model; By combining the target boundary optimization model and the preset nuclear power system operation fitting model, and based on the attribute information, objective function, and constraints of the target nuclear power system, exploratory optimization processing is performed on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements, which are then used as the target operating boundaries of the target nuclear power system.

[0007] In one embodiment, after determining the target boundary parameter values ​​that meet the requirements as the target operating boundary of the target nuclear power system, the method further includes: Based on the target operating boundary, update the target operating rules and corresponding parameter values ​​in the target technical specifications of the target nuclear power system to obtain the updated target operating rules and updated target technical specifications; Control the operation of the target nuclear power system according to the updated target operating rules and the updated target technical specifications.

[0008] In one embodiment, the working condition scenario type includes: a working condition type and a combination of sub-scenario types under the working condition type; The operating condition types include: normal operating condition, expected operating condition, and accident operating condition; Correspondingly, the sub-scenario types under normal operating conditions include: steady-state operation sub-scenario, load adjustment transient sub-scenario, and normal shutdown / startup sub-scenario; The sub-scenario types under the expected operating conditions include: load transient event sub-scenario, secondary loop heat dissipation increase sub-scenario, and secondary loop heat dissipation decrease sub-scenario. The sub-scenario types under the accident conditions include: primary loop cooling and flow abnormality sub-scenario, and reactive abnormal event sub-scenario.

[0009] In one embodiment, the step of using a preset nuclear power system operation fitting model to construct an operation dependency matrix matching the current operating scenario type of the target nuclear power system through simulation testing, based on the attribute information of the target nuclear power system and the current operating scenario type, includes: Using the preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, an evolution simulation test of the operating state of the target nuclear power system is conducted to obtain corresponding simulation test data. Based on the simulation test data and the causal relationship rules that match the current operating scenario type, a correlation analysis is performed on the changing relationship of the operating variables of the target nuclear power system and the triggering conditions of the reactor protection system of the target nuclear power system. Based on the correlation analysis results, the influencing variables that trigger the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, as well as the safety level of the influencing variables, are determined. Based on the influencing variables and their safety levels, an operational dependency matrix is ​​constructed that matches the current operating scenario type of the target nuclear power system.

[0010] In one embodiment, based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, a matching objective function and constraints are constructed, including: Based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, corresponding optimization target items are determined; wherein, the optimization target items include at least: system power output item; Based on the optimization objective, construct a matching objective function; Based on the safety parameter data of the shutdown operation of the non-triggered reaction push protection system, corresponding constraints are constructed.

[0011] In one embodiment, determining a matching target boundary optimization model from a preset optimization model based on the function type of the objective function and the number of parameters of the target boundary parameters includes: When the objective function is a single objective function and the number of parameters of the objective boundary parameters is less than or equal to a specified number, the preset first optimization model is determined as the matching objective boundary optimization model. When the objective function is a multi-objective function and / or the number of parameters of the objective boundary parameters is greater than a specified number, the preset second optimization model is determined as the matching objective boundary optimization model.

[0012] In one embodiment, when the target boundary optimization model is a preset first optimization model, the joint use of the target boundary optimization model and a preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system, the objective function, and the constraints, to perform exploratory optimization processing on the target boundary parameters during the operation of the target nuclear power system, in order to determine the target boundary parameter values ​​that meet the requirements, includes: Using a pre-defined nuclear power system operation fitting model, multiple initial sample point data are randomly generated based on the attribute information of the target nuclear power system, the target boundary parameters, and the current operating scenario type. Using the target boundary optimization model, corresponding running status labels are determined based on the multiple initial sample point data; and first initial sample data is constructed based on the initial sample point data and the corresponding running status labels. Using the target boundary optimization model, based on the first initial sample data, combined with the objective function and constraints, optimization is performed to determine the initial first operating boundary; Obtain and determine the matching target exploration strategy based on the physical evolution characteristics of the current working scenario type. Using the target boundary optimization model, based on the target exploration strategy and the initial first running boundary, multiple rounds of exploration and optimization processing are performed on the neighborhood of the running boundary to determine the target boundary parameter values ​​that meet the requirements.

[0013] In one embodiment, when the target boundary optimization model is a preset second optimization model, the joint use of the target boundary optimization model and a preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system, the objective function, and the constraints, to perform exploratory optimization processing on the target boundary parameters during the operation of the target nuclear power system, in order to determine the target boundary parameter values ​​that meet the requirements, includes: Using a pre-defined nuclear power system operation fitting model, multiple initial sample point data are randomly generated based on the attribute information of the target nuclear power system, the target boundary parameters, and the current operating scenario type. Using the target boundary optimization model, the corresponding non-stop heap probability value is predicted based on the multiple initial sample point data; and a second initial sample data is constructed based on the initial sample point data and the corresponding non-stop heap probability value. The target boundary optimization model is used to optimize and solve the problem based on the second initial sample data, combined with the physical evolution characteristics, objective function and constraints of the current working condition scenario, to determine the initial second operating boundary. Obtain and optimize the model prediction variance of the target boundary with respect to the initial second running boundary to determine the initial exploration range; Calculate and construct the initial acquisition function based on the objective function difference at the initial second running boundary and the shutdown probability value; Based on the initial acquisition function and the initial exploration range, the target boundary optimization model is used to perform multiple rounds of exploration and optimization processing on the operating boundary to determine the target boundary parameter values ​​that meet the requirements.

[0014] This specification also provides a platform for determining the operating boundaries of a nuclear power system, including: The acquisition module is used to acquire the attribute information of the target nuclear power system and determine the current operating scenario type of the target nuclear power system; The first construction module is used to utilize a preset nuclear power system operation fitting model to construct an operation dependency matrix that matches the current operating scenario type of the target nuclear power system through simulation testing, based on the attribute information of the target nuclear power system and the current operating scenario type; wherein, the operation dependency matrix includes at least the influence variable that triggers the start of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variable; The filtering module is used to filter out the impact variables that meet the safety level requirements from the operation dependency matrix as target key variables based on the current operating scenario type of the target nuclear power system; and to determine the corresponding target boundary parameters based on the target key variables. The second construction module is used to construct a matching objective function and constraints based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system; wherein, the objective function includes: a single objective function or a multi-objective function; The first determining module is used to determine a matching target boundary optimization model from a preset optimization model based on the function type of the objective function and the number of parameters of the target boundary parameters; wherein the preset optimization model includes at least: a preset first optimization model and a preset second optimization model; The second determining module is used to jointly use the target boundary optimization model and the preset nuclear power system operation fitting model to perform exploratory optimization processing on the target boundary parameters of the target nuclear power system during operation based on the attribute information, objective function, and constraints of the target nuclear power system, so as to determine the target boundary parameter values ​​that meet the requirements, which are used as the target operating boundaries of the target nuclear power system.

[0015] This specification also provides a computer program product comprising a computer program that, when executed by a processor, implements the steps of a method for determining the operating boundaries of the nuclear power system.

[0016] Based on the method and platform for determining the operating boundary of a nuclear power system provided in this specification, the attribute information of the target nuclear power system is first obtained, and the current operating scenario type of the target nuclear power system is determined. Then, using a pre-set nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, an operation dependency matrix matching the current operating scenario type of the target nuclear power system is constructed through simulation testing. This operation dependency matrix includes at least the influencing variables that trigger the startup of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of these influencing variables. Based on the current operating scenario type of the target nuclear power system, influencing variables with acceptable safety levels are selected from the operation dependency matrix as target key variables. And based on the target key variables, the corresponding target boundary is determined. Boundary parameters; based on the current operating scenario type of the target nuclear power system and its operating indication parameters, a matching objective function and constraints are constructed; the objective function may be a single objective function or a multi-objective function; then, based on the function type of the objective function and the number of parameters of the target boundary parameters, a matching target boundary optimization model is determined from the preset optimization models; the preset optimization models include at least a preset first optimization model and a preset second optimization model; using the target boundary optimization model and the preset nuclear power system operation fitting model in combination, based on the attribute information of the target nuclear power system, the objective function, and the constraints, exploratory optimization processing is performed on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements, which serve as the target operating boundaries of the target nuclear power system. In this way, by first constructing and matching the operational dependency matrix based on the current operating scenario type of the target nuclear power system, the target boundary parameters that need to be focused on based on the current operating scenario type can be accurately selected. Then, by distinguishing different situations based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, a highly targeted objective function that considers one or more optimization objectives can be constructed. At the same time, by determining the function type of the objective function and the number of parameters of the determined target boundary parameters, a matching and effective target boundary optimization model can be accurately determined from the preset optimization models. Finally, by jointly using the target boundary optimization model and the preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system, the objective function, and the constraints, the target boundary parameters during the operation of the target nuclear power system are explored and optimized, and the most suitable target operating boundary for the target nuclear power system based on the current operating scenario type is accurately determined.This allows for precise adaptation to the specific operating scenarios of the target nuclear power system. Under the premise of ensuring the overall safe operation of the target nuclear power system without reactor failure, the safety margin of the operating parameters can be fully explored and utilized to accurately determine the target operating boundary with better effect and higher application value. Subsequently, by controlling the operation of the target nuclear power system according to the target operating boundary, the power output of the target nuclear power system can be effectively improved, and relatively better operating results can be obtained. Attached Figure Description

[0017] To more clearly illustrate the embodiments of this specification, the accompanying drawings used in the embodiments will be briefly introduced below. The drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating a method for determining the operating boundary of a nuclear power system according to one embodiment of this specification. Figure 2 This is a schematic diagram of one embodiment of the method for determining the operating boundary of a nuclear power system provided in the embodiments of this specification, applied in a scenario example. Figure 3 This is a schematic diagram of one embodiment of the method for determining the operating boundary of a nuclear power system provided in the embodiments of this specification, applied in a scenario example. Figure 4 This is a schematic diagram of one embodiment of the method for determining the operating boundary of a nuclear power system provided in the embodiments of this specification, applied in a scenario example. Figure 5 This is a schematic diagram of one embodiment of the method for determining the operating boundary of a nuclear power system provided in the embodiments of this specification, applied in a scenario example. Figure 6 This is a schematic diagram of the structural composition of a server provided in one embodiment of this specification; Figure 7 This is a schematic diagram of the structural composition of a platform for determining the operating boundary of a nuclear power system, provided in one embodiment of this specification. Figure 8 This is a schematic diagram of one embodiment of the method for determining the operating boundary of a nuclear power system provided in the embodiments of this specification, applied in a scenario example. Figure 9 This is a schematic diagram of one embodiment of the method for determining the operating boundary of a nuclear power system provided in the embodiments of this specification, applied in a scenario example. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0020] It should be noted that the information and data related to users involved in the embodiments of this specification are all information and data authorized by the user or fully authorized by the relevant parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with relevant laws, regulations, and standards, and necessary confidentiality measures have been taken. They do not violate public order and good morals, and corresponding operation entry points are provided for users or relevant parties to choose to authorize or refuse.

[0021] It should also be noted that in the embodiments of this specification, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0022] See Figure 1 As shown in the embodiments of this specification, a method for determining the operating boundary of a nuclear power system is provided. In specific implementation, this method may include the following: S101: Obtain the attribute information of the target nuclear power system and determine the current operating scenario type of the target nuclear power system; S102: Using a preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, through simulation testing, construct an operation dependency matrix that matches the current operating scenario type of the target nuclear power system; wherein, the operation dependency matrix includes at least the influence variable that triggers the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variable; S103: Based on the current operating scenario type of the target nuclear power system, select the influencing variables that meet the safety level requirements from the operation dependency matrix as target key variables; and determine the corresponding target boundary parameters based on the target key variables; S104: Based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, construct a matching objective function and constraints; wherein, the objective function includes: a single objective function or a multi-objective function; S105: Based on the function type of the objective function and the number of parameters of the objective boundary parameters, determine a matching objective boundary optimization model from the preset optimization models; wherein, the preset optimization model includes at least: a preset first optimization model and a preset second optimization model; S106: By combining the target boundary optimization model and the preset nuclear power system operation fitting model, and based on the attribute information, objective function, and constraints of the target nuclear power system, the target boundary parameters during the operation of the target nuclear power system are explored and optimized to determine the target boundary parameter values ​​that meet the requirements, which are then used as the target operating boundary of the target nuclear power system.

[0023] Specifically, the aforementioned target nuclear power system can be understood as the nuclear power system currently under control and in operation. For example, the target nuclear power system may include the nuclear power plant reactor and primary coolant system.

[0024] The aforementioned target nuclear power system may also be equipped with a reactor protection system. Specifically, this reactor protection system is used to perform corresponding protective operations on the target nuclear power system when abnormalities occur during operation, in order to eliminate the abnormalities as much as possible and avoid reactor shutdown.

[0025] The aforementioned target operating boundary can be specifically understood as the limit value of the operating control parameters of the target nuclear power system that, while ensuring the safe operation of the target nuclear power system without reactor shutdown, can strive for the best possible optimization effect (e.g., obtaining higher reactor output power, and / or having a faster system adjustment response speed).

[0026] The aforementioned attribute information of the target nuclear power system can be understood as the basic information used to characterize the target nuclear power system, such as the engineering parameters, material types, and physical limit values ​​of each component in the target nuclear power system.

[0027] The aforementioned work scenario types can specifically include: work scenario types, and combinations of sub-scenario types under work scenario types.

[0028] Specifically, the aforementioned operating condition types can include: normal operating condition, anticipated operating condition, and accident operating condition. The normal operating condition can be understood as operating under normal load and with a high probability of not triggering the protection operation of the reactor protection system configured in the target nuclear power system. The anticipated operating condition can be understood as operating under a higher load than the normal operating condition, with a probability of triggering the protection operation of the reactor protection system, and based on the high probability that the protection operation of the reactor protection system can eliminate the safety hazard, it will not trigger a reactor shutdown of the target nuclear power system. The accident operating condition can be understood as having a high probability of triggering the protection operation of the reactor protection system, and based on the fact that the protection operation of the reactor protection system cannot completely eliminate the safety hazard, there is still a significant probability that it will further trigger a reactor shutdown of the target nuclear power system.

[0029] Furthermore, each type of working condition can be further subdivided into several corresponding sub-scenario types.

[0030] Specifically, the sub-scenario types under the above-mentioned normal operating conditions may include: steady-state operation sub-scenario, load regulation transient sub-scenario, normal shutdown / startup sub-scenario, etc.; the sub-scenario types under the above-mentioned expected operating conditions may include: load transient event sub-scenario, secondary loop heat dissipation increase sub-scenario, secondary loop heat dissipation decrease sub-scenario, etc.; the sub-scenario types under the above-mentioned accident operating conditions may include: primary loop cooling and flow abnormality sub-scenario, reactive abnormality event sub-scenario, etc.

[0031] Furthermore, for each sub-scenario type under each working condition, the physical evolution characteristics of the working condition scenario corresponding to each working condition scenario type can be determined through cluster learning based on historical data.

[0032] It should be noted that, considering the relatively complex and variable operation and control of nuclear power systems, this study breaks down the operation of nuclear power systems into specific sub-scenarios under specific operating conditions. This allows for more refined analysis and, based on the operational characteristics and safety hazards of different operating scenarios, the selection of key variables that have a major impact on the corresponding operating scenario type for targeted strong coupling analysis. This approach aims to closely approximate the actual operation of the nuclear power system under that specific operating scenario type, effectively explore the margin of the operating boundary, and obtain an operating boundary with relatively high application value.

[0033] The aforementioned pre-defined nuclear power system operation fitting model can be understood as a simulation program that supports the simulation of the evolution of the operating state of a nuclear power system. Specifically, the aforementioned pre-defined nuclear power system operation fitting model can, for example, be a nuclear power plant reactor and primary loop system operation simulation program developed based on the PCTRAN model.

[0034] The aforementioned first and second preset optimization models can be understood as algorithmic models based on different algorithmic approaches, suitable for exploring and optimizing the operational boundaries of nuclear power systems under different conditions to find the limit boundaries. Specifically, the aforementioned first preset optimization model can be a lightweight model suitable for handling low-dimensional and simple cases, while the aforementioned second preset optimization model can be a refined model suitable for handling high-dimensional and complex cases.

[0035] Specifically, the aforementioned preset first optimization model can be, for example, a boundary optimization model based on an improved support vector machine. This preset first optimization model is well-suited for low-dimensional, relatively simple nuclear power system operations. By combining active learning with uncertainty sampling, it can quickly approximate nonlinear boundaries, enabling efficient and accurate boundary exploration and optimization.

[0036] The aforementioned pre-defined second optimization model can be, for example, a boundary optimization model based on improved ensemble Bayesian learning. This pre-defined second optimization model is well-suited for high-dimensional, relatively complex nuclear power system operations. It characterizes and utilizes the prediction variance of multiple integrated neural networks to leverage the uncertainty in model predictions, guiding in-depth exploration of the operational boundary margin to achieve exploratory optimization of the operational boundary.

[0037] Specifically, corresponding sensors can be installed on each component and connecting pipeline of the target nuclear power system. Accordingly, in specific implementation, the current operating status parameters of the target nuclear power system can be collected through these sensors; then, based on the current operating status parameters, a preset operating status template is matched; and based on the matching result, the current operating condition scenario type of the target nuclear power system is determined.

[0038] In practice, a pre-defined nuclear power system operation fitting model can be used to obtain simulation test data on the target nuclear power system's operation under the current operating scenario, based on the target nuclear power system's attribute information and the current operating scenario type. Based on this simulation test data, and combined with causal rules matching the current operating scenario type, a correlation analysis is performed on the rate of change of various operating variables of the target nuclear power system and the triggering conditions of the reactor protection system. Based on the correlation analysis results, operating variables with correlation parameters greater than a pre-defined correlation threshold are selected as influencing variables. The safety level of these influencing variables is determined based on their correlation parameters. Finally, an Operational Dependency Matrix (ODM) matching the target nuclear power system is constructed based on the influencing variables and their safety levels.

[0039] The aforementioned safety levels can specifically include three levels: high, medium, and low. Higher-level operating variables are more likely to cause malfunctions in the target nuclear power system, thus posing safety hazards.

[0040] In practice, based on the current operating scenario type of the target nuclear power system and the mapping curve of the operating state change determined by historical data, the degree of drastic change in the operating state of the target nuclear power system under the current operating scenario type is determined.

[0041] Specifically, when the degree of change in the operating state exceeds the first threshold, it indicates that the target nuclear power system is experiencing drastic and rapid changes in its operating state based on the current operating scenario. Almost any change in any influencing variable could cause a significant instantaneous change in the overall operating state of the target nuclear power system. In this case, influencing variables at three levels—high, medium, and low—can be simultaneously selected from the operating dependency matrix as the target key variables that meet the requirements.

[0042] When the degree of change in the operating state is less than or equal to the first threshold and greater than the second threshold, it indicates that the target nuclear power system is experiencing relatively significant changes in its operating state based on the current operating scenario, and the rate of change is relatively normal. In this case, high-level and intermediate-level influencing variables can be selected from the operating dependency matrix as the target key variables that meet the requirements.

[0043] When the degree of change in the operating state is less than or equal to the second threshold, it indicates that the target nuclear power system, based on the current operating scenario, is not significantly changing in its operating state and the rate of change is relatively slow. In this case, only the high-level influencing variables can be selected from the operating dependency matrix as the target key variables that meet the requirements.

[0044] In practice, based on the key target variables, the adjustable parameters of the target nuclear power system corresponding to the key target variables can be determined by consulting the operating rules and technical specifications of the target nuclear power system. These parameters can then be used as the target boundary parameters of interest (e.g., reactor power setpoint, main pump speed, etc.). The determined target boundary parameters may include only one type of parameter or multiple types of parameters simultaneously.

[0045] Specifically, the aforementioned operating procedures and technical specifications may be technical documents designed in advance for the target nuclear power system to guide users (e.g., operators) in operating and controlling the target nuclear power system.

[0046] In practice, the required optimization objectives and potential safety hazards during system operation can be determined based on the current operating conditions and operating parameters of the target nuclear power system. Then, a corresponding objective function can be constructed based on the optimization objectives. At the same time, corresponding constraints can be constructed based on the potential safety hazards during system operation.

[0047] Specifically, the objective function mentioned above can be a single-objective function or a multi-objective function. Specifically, the single-objective function may include only one optimization objective; the multi-objective function may include at least two optimization objectives.

[0048] The aforementioned operating indication parameters can be specifically input by the user of the nuclear power system operating control target, and are parameter data used to characterize the optimization target and operating requirements.

[0049] In practice, based on the function type of the objective function and the number of parameters of the objective boundary parameters, a preset optimization model matching the current operating condition of the target nuclear power system can be determined from the preset optimization models as the objective boundary optimization model. Then, the objective boundary optimization model and the preset nuclear power system operation fitting model are used together. Based on the attribute information of the target nuclear power system, the objective function, and the constraints, the target boundary parameters during the operation of the target nuclear power system are explored and optimized. First, the initial boundary is determined. Then, based on the initial boundary, multiple rounds of exploration and optimization are carried out intelligently. Under the premise of ensuring the safe operation of the target nuclear power system based on the current operating scenario type according to the constraints, the margin of the operating boundary is explored and utilized as much as possible. At the same time, the objective function is continuously explored and optimized to find the limit value of the objective boundary parameters (i.e., the target boundary parameter value) that matches the current operating scenario type of the target nuclear power system and contains the relatively optimal solution that satisfies the optimization objective under the current operating scenario type. The corresponding target operating boundary is obtained.

[0050] Based on the above embodiments, firstly, by constructing and matching the operational dependency matrix according to the current operating scenario type of the target nuclear power system, the target boundary parameters of interest are determined. Then, by differentiating different situations based on the current operating scenario type and the operational indication parameters of the target nuclear power system, a suitable objective function is constructed. Simultaneously, by determining the function type of the objective function and the number of parameters of the determined target boundary parameters from a preset optimization model, a matching target boundary optimization model is selected. Finally, by jointly using the target boundary optimization model and the preset nuclear power system operation fitting model, and based on the attribute information of the target nuclear power system, the objective function, and the constraints, exploratory optimization processing of the target boundary parameters during the operation of the target nuclear power system is performed to determine the target operating boundary. This allows for precise adaptation to the specific operating scenario of the target nuclear power system. While ensuring the overall operational safety of the target nuclear power system without reactor shutdown, it fully explores and utilizes the safety margin of the operating parameters, accurately determining the target operating boundary with better effects and higher application value. Subsequently, by controlling the operation of the target nuclear power system according to this target operating boundary, the power output of the target nuclear power system can be effectively improved.

[0051] In some embodiments, after determining the target boundary parameter values ​​that meet the requirements as the target operating boundary of the target nuclear power system, the method may further include the following: S1: Based on the target operating boundary, update the target operating rules and corresponding parameter values ​​in the target technical specifications of the target nuclear power system to obtain the updated target operating rules and updated target technical specifications; S2: Control the operation of the target nuclear power system according to the updated target operating rules and the updated target technical specifications.

[0052] Among them, the updated target operation rules and updated target technical specifications may at least include parameters such as target operation boundaries based on the corresponding working condition scenario type.

[0053] Furthermore, the updated target operating rules and updated target technical specifications may also include recommended values ​​for relevant adjustable parameters that enable the target nuclear power system to achieve relatively optimal operating performance based on the corresponding operating scenario type.

[0054] Specifically, users can adjust the relevant adjustable parameters of the target nuclear power system according to the updated target operating rules and the updated target technical specifications to control the operation of the target nuclear power system.

[0055] In this way, on the one hand, while ensuring the safe operation of the target nuclear power system based on the relevant scenario type, users can have greater room for adjustment to flexibly adjust and control the target nuclear power system according to specific situations and needs, avoiding overly conservative adjustments; on the other hand, it can also support users to accurately control the target nuclear power system to obtain relatively good operating results and meet complex and diverse application requirements.

[0056] In some embodiments, the working condition scenario type includes: a working condition type and a combination of sub-scenario types under the working condition type; Specifically, the operating condition types may include: normal operating conditions, expected operating conditions, and accident operating conditions; Accordingly, the sub-scenario types under normal operating conditions may specifically include: steady-state operation sub-scenario, load adjustment transient sub-scenario, normal shutdown / startup sub-scenario, etc. The sub-scenario types under the expected operating conditions may specifically include: load transient event sub-scenario, secondary loop heat dissipation increase sub-scenario, secondary loop heat dissipation decrease sub-scenario, etc. The specific sub-scenario types under the accident conditions may include: primary loop cooling and flow abnormality sub-scenario, reactive abnormal event sub-scenario, etc.

[0057] Before implementation, a large amount of historical usage data on nuclear power systems can be acquired and used to determine the types of operating scenarios through clustering learning, along with the physical evolution characteristics of each operating scenario type. See Table 1 for details.

[0058] Table 1

[0059] Based on the above embodiments, the corresponding working condition scenario types can be accurately classified by clustering learning of historical data, and the physical evolution characteristics of the working condition scenarios corresponding to each working condition scenario type can be determined.

[0060] In some embodiments, see Figure 2 As shown, the above-mentioned operation dependency matrix, which uses a preset nuclear power system operation fitting model, constructs an operation dependency matrix that matches the current operating scenario type of the target nuclear power system through simulation testing based on the attribute information of the target nuclear power system and the current operating scenario type. In specific implementation, this may include the following: S2-1: Using the preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, conduct an evolution simulation test on the operating state of the target nuclear power system to obtain corresponding simulation test data. S2-2: Based on the simulation test data and the causal relationship rules that match the current operating scenario type, perform a correlation analysis on the relationship between the changes in the operating variables of the target nuclear power system and the triggering conditions of the reactor protection system of the target nuclear power system; S2-3: Based on the correlation analysis results, determine the influencing variables that trigger the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, as well as the safety level of the influencing variables; S2-4: Based on the aforementioned influencing variables and their safety levels, construct an operational dependency matrix that matches the current operating scenario type of the target nuclear power system.

[0061] Based on the above embodiments, by utilizing a preset nuclear power system operation fitting model, the operation dependency matrix that matches the current operating scenario type of the target nuclear power system can be accurately and efficiently determined.

[0062] In some embodiments, during implementation, historical data can be used to determine the physical evolution characteristics of each working scenario type through cluster learning. Simultaneously, based on the specific working scenario type and the corresponding operational dependency matrix, the target boundary parameters to be focused on can be determined. See Table 2 for details. Here, the boundary dimension can be understood as the number of parameters of the target boundary parameters.

[0063] Table 2

[0064] In some embodiments, the above-mentioned objective function and constraints are constructed based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system. In specific implementation, these may include the following: S1: Based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, determine the corresponding optimization target items; wherein, the optimization target items include at least: system power output items; S2: Construct a matching objective function based on the optimization objective term; S3: Based on the safety parameter data of the shutdown operation of the non-triggered reaction push protection system, construct the corresponding constraints.

[0065] The aforementioned safety parameter data can be determined based on the triggering conditions of the reactor protection system's shutdown operation.

[0066] When the objective function is a multi-objective function, the optimization objective terms included in the objective function may further include at least one of the following: running boundary margin interval term, control stability indication parameter term, system adjustment response speed term, etc.

[0067] Specifically, for example, in a simple case, the objective function can be constructed using only the system power output term as follows:

[0068] in, The reactor power setpoint for the target nuclear power system. The maximum value set for reactor power. The maximum value set for reactor power.

[0069] By optimizing the objective function, we can find the values ​​of the relevant adjustable parameters corresponding to the relatively maximum system power output under the current operating conditions, while ensuring the safe operation of the target nuclear power system.

[0070] For example, in complex cases, the objective function can be constructed by combining the system power output term, the operating boundary margin interval term, and the control stability indication parameter term, as follows:

[0071] in, The reactor power setpoint for the target nuclear power system. The setpoint for the main pump speed of the target nuclear power system. The primary circuit thermal regression temperature of the target nuclear power system. The maximum value set for the main pump speed. The minimum value set for the main pump speed. This is the normal temperature value for the primary circuit thermal deactivation at the current system output power (e.g., 350°C when the system outputs full power; this value may vary for different reactor types). This represents the maximum value of the primary circuit thermal annealing temperature. This represents the minimum thermal deactivation temperature of the primary loop. The values ​​of a1, a2, and a3 can be: the first weighting coefficient corresponding to the system power output term, the second weighting coefficient corresponding to the operating boundary margin interval term, and the third weighting coefficient corresponding to the control stability indication parameter term, respectively. Specifically, for example, the values ​​of a1, a2, and a3 can be 0.5, 0.3, and 0.2, respectively.

[0072] In the above objective function, the first function term (system power output term) mainly considers the reactor power output of the nuclear power system in order to maximize the system power output. Correspondingly, the first weight coefficient can be set to the maximum value. The second function term (operational boundary margin interval term) considers the operating range margin of the operated variable (adjustable parameter). Specifically, for example, it aims to make the set value of the main pump speed as large as possible to avoid the main pump speed set value being too low, which would lead to a decrease in mass flow rate and trigger a reactor shutdown. The third function term (control stability indication parameter term) considers the stability of the control process. Specifically, for example, it aims to keep the primary loop thermal decooling temperature as close as possible to the normal value at the current power, with the smaller the fluctuation, the better. Therefore, it is set as a negative term with an absolute value added.

[0073] By optimizing the objective function, we can find the values ​​of adjustable parameters that can optimally balance various requirements such as system power output, operating boundary margin range, and control stability, while ensuring the safe operation of the target nuclear power system.

[0074] Specifically, for example, based on the safety parameter data of the reactor shutdown operation of the protection system without triggering a reaction, and combined with the objective function, the corresponding constraints can be constructed in the following manner:

[0075] in, The target is the operating pressure of the nuclear power system. This is the primary circuit thermal deactivation temperature. For the primary loop mass flow rate, The rated primary loop mass flow rate, This represents the neutron fluence rate.

[0076] Based on the above embodiments, different situations can be distinguished according to the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, and a target function with high adaptability and good effect, as well as constraints, can be constructed to meet the diverse requirements under different operating scenario types.

[0077] In some embodiments, the above-mentioned determination of a matching target boundary optimization model from a preset optimization model based on the function type of the objective function and the number of parameters of the target boundary parameters may specifically include the following: When the objective function is a single objective function and the number of parameters (or parameter dimensions) of the objective boundary parameters is less than or equal to a specified number, the preset first optimization model is determined as the matching objective boundary optimization model. When the objective function is a multi-objective function and / or the number of parameters of the objective boundary parameters is greater than a specified number, the preset second optimization model is determined as the matching objective boundary optimization model.

[0078] Specifically, the specified quantity mentioned above can be 2. In actual implementation, depending on the specific application scenario and processing requirements, the specified quantity can also be set to other appropriate values ​​such as 3 or 4.

[0079] Based on the above embodiments, by jointly considering the function type of the objective function and the number of parameters of the objective boundary parameters, different situations can be effectively and precisely distinguished. In this way, a matching objective boundary optimization model can be determined and used to participate in subsequent exploration and optimization processes, so as to find the target boundary parameter values ​​that meet the requirements more efficiently and accurately.

[0080] In some embodiments, when the target boundary optimization model is a preset first optimization model, refer to Figure 3 As shown, the above-mentioned combined use of the target boundary optimization model and the preset nuclear power system operation fitting model, based on the attribute information, objective function, and constraints of the target nuclear power system, performs exploratory optimization processing on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements. In specific implementation, this may include the following: S3-1: Using a preset nuclear power system operation fitting model, multiple initial sample point data are randomly generated based on the attribute information of the target nuclear power system, the target boundary parameters, and the current operating scenario type. S3-2: Using the target boundary optimization model, determine the corresponding running status labels based on the multiple initial sample point data; and construct the first initial sample data based on the initial sample point data and the corresponding running status labels. S3-3: Using the target boundary optimization model, based on the first initial sample data, combined with the objective function and constraints, optimization is performed to determine the initial first running boundary; S3-4: Obtain and determine the matching target exploration strategy based on the physical evolution characteristics of the current working condition scenario type. S3-5: Using the target boundary optimization model, based on the target exploration strategy and the initial first running boundary, perform multiple rounds of exploration and optimization processing on the neighboring region of the running boundary to determine the target boundary parameter values ​​that meet the requirements.

[0081] Specifically, the preset first optimization model may include a boundary optimization model based on an improved Support Vector Machine (SVM).

[0082] The above-mentioned target boundary optimization model, based on the target exploration strategy and the initial first running boundary, performs multi-round exploration optimization processing on the neighborhood of the running boundary. In specific implementation, this can include: performing the current round of exploration optimization processing on the neighborhood of the running boundary in the following manner: S1: Determine the adjacent area of ​​the first running boundary of the previous round; S2: Based on the target exploration strategy, randomly collect multiple neighboring sample point data from the neighboring region of the first running boundary of the previous round to construct the candidate point set for the current round; S3: Calculate the Euclidean distance between each candidate point data in the current round's candidate point set and the first sample dataset of the previous round; and based on the Euclidean distance, select the first sample data point of the current round from the candidate point set of the current round. S4: Using the target boundary optimization model, determine the corresponding running status label (e.g., stopped heap or not stopped heap) based on the first sample data point of the current round; and construct the first sample data of the current round based on the sample point data of the current round and the corresponding running status label. S5: Using the target boundary optimization model, based on the first sample data of the current round, combined with the objective function and constraints, optimization is performed to determine the first running boundary of the current round; S6: Using the nuclear power system operation fitting model, calculate the failure rate of the current wheel based on the first operating boundary of the current wheel; and check whether the failure rate of the current wheel and the failure rate of the previous wheel are less than the preset failure rate threshold. S7: When it is determined that the failure rate of the current round is less than the preset failure rate threshold, and the failure rate of the previous round is less than the preset failure rate threshold, the exploration optimization process ends; and the target boundary parameter value that meets the requirements is determined according to the first running boundary of the current round.

[0083] The aforementioned target exploration strategy can be determined based on the physical evolution characteristics of the current operating scenario. Specifically, the region adjacent to the first operating boundary of the previous round can be understood as the range of parameter values ​​near the first operating boundary of the previous round.

[0084] When the failure rate of the current round is determined to be greater than or equal to the preset failure rate threshold, and / or the failure rate of the previous round is greater than or equal to the preset failure rate threshold, the next round of exploration and optimization processing is triggered.

[0085] In practice, the failure probability of the first operating boundary in the current round can be determined based on the classifier in the target boundary optimization model; and the failure rate of the current round can be determined based on the failure probability of the first operating boundary in the current round. The failure rate of the current round is used to evaluate the coverage and completeness of the operating boundaries identified by the current round.

[0086] In practice, when both the failure rate of the current round and the failure rate of the previous round are less than a preset failure rate threshold (e.g., 0.01), that is, when the failure rates of two consecutive rounds are less than the preset failure rate threshold, it can be determined that the multi-round exploratory optimization process has converged, and thus the exploratory optimization process can be terminated. Based on the first running boundary of the current round, the target boundary parameter values ​​that meet the requirements are determined. Otherwise, the next round of exploratory optimization process needs to be repeated until the failure rate of the current round is less than the preset failure rate threshold, and the failure rate of the previous round is also less than the preset failure rate threshold.

[0087] Based on the above embodiments, by using a preset first optimization model to perform multiple rounds of exploratory optimization, it can be well applied to low-dimensional and simple cases, and can efficiently determine the target operating boundary that meets the requirements.

[0088] In some embodiments, when the target boundary optimization model is a preset second optimization model, refer to Figure 4 As shown, the above-mentioned combined use of the target boundary optimization model and the preset nuclear power system operation fitting model, based on the attribute information, objective function, and constraints of the target nuclear power system, performs exploratory optimization processing on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements. In specific implementation, this may include the following: S4-1: Using a preset nuclear power system operation fitting model, multiple initial sample point data are randomly generated based on the attribute information of the target nuclear power system, the target boundary parameters, and the current operating scenario type. S4-2: Using the target boundary optimization model, predict the corresponding non-stop heap probability value based on the multiple initial sample point data; and construct the second initial sample data based on the initial sample point data and the corresponding non-stop heap probability value. S4-3: Using the target boundary optimization model, based on the second initial sample data, combined with the physical evolution characteristics, objective function, and constraints of the current working scenario, the initial second operating boundary is determined by optimizing the solution. S4-4: Obtain and optimize the model prediction variance of the target boundary with respect to the initial second running boundary to determine the initial exploration range; S4-5: Calculate and construct the initial acquisition function based on the objective function difference based on the initial second running boundary and the shutdown probability value; S4-6: Based on the initial acquisition function and the initial exploration range, the target boundary optimization model is used to perform multiple rounds of exploration and optimization processing on the running boundary to determine the target boundary parameter values ​​that meet the requirements.

[0089] Specifically, the pre-defined second optimization model may include a boundary optimization model based on improved ensemble Bayesian learning. Correspondingly, the pre-defined second optimization model includes at least multiple ensemble network sub-models (e.g., neural networks). These multiple ensemble network sub-models can be combined as a classifier for the pre-defined second optimization model; each network sub-model can be different, so that the non-refueling probability of a nuclear power system can be predicted from multiple different perspectives based on the multiple network sub-models.

[0090] Specifically, the aforementioned pre-defined second optimization model may include five integrated network sub-models. Each network sub-model may contain four fully connected layers, each layer may contain 64 neurons, and the ReLU activation function is used between layers. The number of neurons in each layer can be dynamically and adaptively adjusted according to the number of parameters of the target boundary parameters.

[0091] Based on the initial acquisition function and initial exploration range, the above-mentioned target boundary optimization model is used to perform multi-round exploration and optimization processing on the operating boundary to determine the target boundary parameter values ​​that meet the requirements. In specific implementation, it may include: performing the current round of exploration and optimization processing on the operating boundary in the following manner: S1: Obtain the second running boundary, the collection function, and the exploration range of the previous round; S2: Using the collection function from the previous round, and based on the exploration range of the previous round, determine the second sample point data for the current round; S3: Using the target boundary optimization model, predict the corresponding non-stop heap probability value based on the second sample point data of the current round; and construct the second sample data of the current round based on the second sample point data of the current round and the corresponding non-stop heap probability value. S4: Using the target boundary optimization model, based on the second sample data of the current wheel, combined with the physical evolution characteristics, objective function, and constraints of the current working condition scenario, the second running boundary of the current wheel is determined by optimizing the solution. S5: Obtain and optimize the model prediction variance of the second running boundary of the current round based on the target boundary, and check whether the preset exploration termination condition is met; S6: When it is determined that the preset exploration end conditions are met, the target boundary parameter values ​​that meet the requirements are determined according to the second running boundary of the current round.

[0092] In practice, the failure rate of the current round can be calculated; and it can be checked whether the failure rate of the current round is less than a preset failure rate threshold, and whether the difference between the failure rate of the current round and the failure rate of the previous round is less than a preset difference threshold. When it is determined that the failure rate of the current round is less than the preset failure rate threshold, and the difference between the failure rate of the current round and the failure rate of the previous round is less than the preset difference threshold, the preset exploration termination condition is determined to be met. When it is determined that the failure rate of the current round is greater than or equal to the preset failure rate threshold, and / or, the difference between the failure rate of the current round and the failure rate of the previous round is greater than or equal to the preset difference threshold, the preset exploration termination condition is determined to be not met.

[0093] In practice, when it is determined that the preset exploration termination condition is not met, the exploration range of the current round can be determined first by the model prediction variance of the target boundary optimization model about the second operating boundary of the current round; then, the acquisition function of the current round can be constructed by calculating and based on the function difference of the objective function based on the second operating boundary of the current round and the shutdown probability value; based on the acquisition function of the current round and the exploration range of the current round, the next round of exploration optimization processing about the operating boundary can be performed using the target boundary optimization model.

[0094] In practice, the upper limit of constraints (U) and the lower limit of constraints (L) can be determined based on the model prediction variance of the second running boundary of the current round output by the classifier of the target boundary optimization model; then, the exploration range of the current round can be determined based on the upper limit of constraints and the lower limit of constraints.

[0095] Specifically, for example, the upper and lower limits of the constraints can be determined using the following formula:

[0096] Where U is the upper limit of the constraint and L is the lower limit of the constraint. Let x be the model prediction variance, x be the parameter value of the relevant adjustable parameter on the second running boundary of the current round, and A be the baseline reference value.

[0097] Specifically, the baseline reference value can be calculated and adaptively adjusted based on the difference between the failure rate of the current round and the failure rate of the previous round. For example, if the difference between the failure rate of the current round and the failure rate of the previous round is a large value, the baseline reference value can be increased to allow for exploration and optimization processing in the next round within a relatively large numerical range; if the difference between the failure rate of the current round and the failure rate of the previous round is a small value, the baseline reference value can be decreased to allow for exploration and optimization processing in the next round within a relatively small numerical range.

[0098] In practice, the penalty term can be calculated and constructed based on the function difference of the objective function based on the second operating boundary of the current cycle. At the same time, the corresponding guidance coefficient can be determined by calculating and determining the shutdown probability value of the second operating boundary of the current cycle using the objective boundary optimization model. Then, the acquisition function of the current cycle can be constructed based on the above penalty term and guidance coefficient.

[0099] In this way, the current round's data collection function, combined with the current round's exploration range, can be used to guide the model to regress into the feasible region for the next round of exploration and optimization.

[0100] Based on the above embodiments, by using a preset second optimization model as the target boundary optimization model for multiple rounds of exploratory optimization processing, it can be well applied to high-dimensional and complex situations and can efficiently determine the target operating boundary that meets the requirements.

[0101] In some embodiments, after determining the target boundary parameter values ​​that meet the requirements, a corresponding target operating boundary range map can be constructed based on the aforementioned target boundary parameter values. See details in [link to relevant documentation]. Figure 5 As shown in the diagram, based on the target operating boundary range diagram, the feasible and infeasible regions of the relevant adjustable parameters of the target nuclear power system under the current operating scenario type, while satisfying the relevant optimization objectives, can be clearly delineated.

[0102] See Figure 5 It can be seen that the target operating boundary (corresponding to the predicted boundary) determined by the method for determining the operating boundary of the nuclear power system provided in the embodiments of this specification is obviously closer to the real boundary than the operating boundary (corresponding to the unoptimized boundary) determined by the existing method. Compared with the operating boundary determined by the existing method, it can obviously more fully explore and utilize the safety margin of the operating parameters.

[0103] In some embodiments, while controlling the operation of the target nuclear power system according to the updated target operating rules and the updated target technical specifications, the method further includes: monitoring the operating scenario type of the target nuclear power system; when a change in the operating scenario type of the target nuclear power system is detected, determining the changed operating scenario type of the target nuclear power system; and redetermining the target operating boundary of the target nuclear power system based on the changed operating scenario type.

[0104] As can be seen from the above, the method for determining the operating boundary of a nuclear power system provided in this specification first obtains the attribute information of the target nuclear power system and determines the current operating scenario type of the target nuclear power system; then, using a preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, a simulation test is conducted to construct an operation dependency matrix that matches the current operating scenario type of the target nuclear power system; wherein, the operation dependency matrix includes at least the influence variables that trigger the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variables; based on the current operating scenario type of the target nuclear power system, influence variables with the required safety level are selected from the operation dependency matrix as target key variables; and based on the target key variables, the corresponding target... Boundary parameters; based on the current operating scenario type of the target nuclear power system and its operating indication parameters, a matching objective function and constraints are constructed; the objective function may be a single objective function or a multi-objective function; then, based on the function type of the objective function and the number of parameters of the target boundary parameters, a matching target boundary optimization model is determined from the preset optimization models; the preset optimization models include at least a preset first optimization model and a preset second optimization model; using the target boundary optimization model and the preset nuclear power system operation fitting model in combination, based on the attribute information of the target nuclear power system, the objective function, and the constraints, exploratory optimization processing is performed on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements, which serve as the target operating boundaries of the target nuclear power system.

[0105] First, based on the current operating scenario type of the target nuclear power system, a matching operational dependency matrix is ​​constructed to determine the target boundary parameters of interest. Then, based on the current operating scenario type and the operating indication parameters of the target nuclear power system, different situations are differentiated, and a suitable objective function is constructed. Simultaneously, based on the function type of the objective function and the number of parameters of the determined target boundary parameters, a matching target boundary optimization model is determined from a pre-set optimization model. Finally, by jointly using the target boundary optimization model and the pre-set nuclear power system operation fitting model, based on the attribute information of the target nuclear power system, the objective function, and the constraints, exploratory optimization processing of the target boundary parameters during the operation of the target nuclear power system is performed to determine the target operating boundary. This allows for precise adaptation to the specific operating scenario of the target nuclear power system. While ensuring the overall safe operation of the target nuclear power system without reactor shutdown, the safety margin of the operating parameters is fully explored and utilized to accurately determine the target operating boundary with better effects and higher application value. Subsequently, by controlling the operation of the target nuclear power system according to this target operating boundary, the power output of the target nuclear power system can be effectively improved.

[0106] This specification provides an embodiment of a server, see below. Figure 6 As shown. The server includes a network communication port 601, a processor 602, and a memory 603. These structures are connected by internal cables so that they can perform specific data interaction.

[0107] Specifically, the network communication port 601 can be used to obtain the attribute information of the target nuclear power system and determine the current operating scenario type of the target nuclear power system.

[0108] The processor 602 can specifically be used to utilize a preset nuclear power system operation fitting model to construct an operation dependency matrix matching the current operating scenario type of the target nuclear power system through simulation testing, based on the attribute information of the target nuclear power system and the current operating scenario type. The operation dependency matrix includes at least the influencing variables that trigger the startup of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of these influencing variables. Based on the current operating scenario type of the target nuclear power system, influencing variables with acceptable safety levels are selected from the operation dependency matrix as target key variables. Based on the target key variables, corresponding target boundary parameters are determined. Based on the current operating scenario type of the target nuclear power system and the target... The operation indication parameters of the nuclear power system are used to construct a matching objective function and constraints. The objective function may be a single-objective function or a multi-objective function. Based on the function type of the objective function and the number of parameters of the target boundary parameters, a matching target boundary optimization model is determined from a preset optimization model. The preset optimization model includes at least a preset first optimization model and a preset second optimization model. Using the target boundary optimization model and the preset nuclear power system operation fitting model in combination, and based on the attribute information of the target nuclear power system, the objective function, and the constraints, exploratory optimization processing is performed on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements, which are then used as the target operation boundary of the target nuclear power system.

[0109] The memory 603 can be used to store the corresponding instruction program and related intermediate data.

[0110] Based on the above method, the relevant structural performance of the server can be effectively utilized to improve the data processing speed of electronic devices and efficiently realize the data processing for determining the operating boundary of the nuclear power system.

[0111] In this embodiment, the network communication port 601 can be a virtual port bound to different communication protocols, thereby enabling the sending or receiving of different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. Furthermore, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM or CDMA; it can also be a Wi-Fi chip; or it can be a Bluetooth chip.

[0112] In this embodiment, the processor 602 can be implemented in any suitable manner. For example, the processor can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc. This specification is not limiting.

[0113] In this embodiment, the memory 603 may include multiple layers. In a digital system, anything that can store binary data can be a memory. In an integrated circuit, a circuit with storage function but no physical form is also called a memory, such as RAM, FIFO, etc. In a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0114] This specification also provides a computer-readable storage medium for a method of determining the operating boundary of a nuclear power system, wherein the computer-readable storage medium stores computer program instructions that, when executed, perform the following: acquiring attribute information of a target nuclear power system and determining the current operating scenario type of the target nuclear power system; using a preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, constructing an operation dependency matrix matching the current operating scenario type of the target nuclear power system through simulation testing; wherein the operation dependency matrix includes at least the influence variables that trigger the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variables; and selecting influence variables with acceptable safety levels from the operation dependency matrix as target key variables based on the current operating scenario type of the target nuclear power system. Variables; and based on the key target variables, determine the corresponding target boundary parameters; based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, construct a matching objective function and constraints; wherein, the objective function includes: a single objective function or a multi-objective function; based on the function type of the objective function and the number of parameters of the target boundary parameters, determine a matching target boundary optimization model from the preset optimization models; wherein, the preset optimization models include at least: a preset first optimization model and a preset second optimization model; using the target boundary optimization model and the preset nuclear power system operation fitting model in combination, based on the attribute information of the target nuclear power system, the objective function, and the constraints, perform exploratory optimization processing on the target boundary parameters during the operation of the target nuclear power system, so as to determine the target boundary parameter values ​​that meet the requirements, as the target operating boundary of the target nuclear power system.

[0115] In this embodiment, the storage medium includes, but is not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), Cache, Hard Disk Drive (HDD), or Memory Card. The memory can be used to store computer program instructions. The network communication unit can be an interface configured according to standards specified in the communication protocol for network connection communication.

[0116] In this embodiment, the specific functions and effects implemented by the program instructions stored in the computer-readable storage medium can be explained in comparison with other embodiments, and will not be repeated here.

[0117] This specification also provides a computer program product, comprising at least a computer program, which, when executed by a processor, implements the following method steps: acquiring attribute information of a target nuclear power system and determining the current operating scenario type of the target nuclear power system; using a preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, constructing an operation dependency matrix matching the current operating scenario type of the target nuclear power system through simulation testing; wherein, the operation dependency matrix at least includes an influence variable that triggers the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variable; selecting influence variables with acceptable safety levels from the operation dependency matrix as target key variables based on the current operating scenario type of the target nuclear power system; and determining the safety level of the reactor protection system based on the target key variables. The corresponding target boundary parameters are determined; based on the current operating scenario type of the target nuclear power system and its operating indication parameters, a matching objective function and constraints are constructed; wherein the objective function includes a single objective function or a multi-objective function; based on the function type of the objective function and the number of parameters of the target boundary parameters, a matching target boundary optimization model is determined from a preset optimization model; wherein the preset optimization model includes at least a preset first optimization model and a preset second optimization model; using the target boundary optimization model and the preset nuclear power system operation fitting model in combination, based on the attribute information of the target nuclear power system, the objective function, and the constraints, exploratory optimization processing is performed on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements, which are used as the target operating boundaries of the target nuclear power system.

[0118] See Figure 7As shown in the embodiments of this specification, a platform for determining the operating boundary of a nuclear power system is also provided. This device may specifically include the following structural modules: The acquisition module 701 can be used to acquire the attribute information of the target nuclear power system and determine the current operating scenario type of the target nuclear power system; The first construction module 702 can be used to utilize a preset nuclear power system operation fitting model to construct an operation dependency matrix that matches the current operating scenario type of the target nuclear power system through simulation testing, based on the attribute information of the target nuclear power system and the current operating scenario type; wherein, the operation dependency matrix includes at least the influence variable that triggers the start of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variable. The screening module 703 can be used to select impact variables that meet the safety level requirements from the operation dependency matrix as target key variables based on the current operating scenario type of the target nuclear power system; and determine the corresponding target boundary parameters based on the target key variables. The second construction module 704 can be specifically used to construct a matching objective function and constraints based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system; wherein, the objective function includes: a single objective function or a multi-objective function; The first determining module 705 is specifically used to determine a matching target boundary optimization model from a preset optimization model based on the function type of the objective function and the number of parameters of the target boundary parameters; wherein the preset optimization model includes at least: a preset first optimization model and a preset second optimization model; The second determining module 706 can be used to jointly use the target boundary optimization model and the preset nuclear power system operation fitting model to perform exploratory optimization processing on the target boundary parameters of the target nuclear power system during operation based on the attribute information, objective function, and constraints of the target nuclear power system, so as to determine the target boundary parameter values ​​that meet the requirements as the target operation boundary of the target nuclear power system.

[0119] In some embodiments, after determining the target boundary parameter values ​​that meet the requirements as the target operating boundary of the target nuclear power system, the device may further be used to: update the target operating rules and corresponding parameter values ​​in the target technical specifications of the target nuclear power system according to the target operating boundary, to obtain updated target operating rules and updated target technical specifications; and control the operation of the target nuclear power system according to the updated target operating rules and updated target technical specifications.

[0120] In some embodiments, the working condition scenario type may specifically include: a working condition type and a combination of sub-scenario types under the working condition type; Specifically, the operating condition types may include: normal operating conditions, expected operating conditions, and accident operating conditions; Accordingly, the sub-scenario types under normal operating conditions may specifically include: steady-state operation sub-scenario, load adjustment transient sub-scenario, normal shutdown / startup sub-scenario, etc. The sub-scenario types under the expected operating conditions may specifically include: load transient event sub-scenario, secondary loop heat dissipation increase sub-scenario, secondary loop heat dissipation decrease sub-scenario, etc. The specific sub-scenario types under the accident conditions may include: primary loop cooling and flow abnormality sub-scenario, reactive abnormal event sub-scenario, etc.

[0121] In some embodiments, when the first construction module 702 is specifically implemented, it can utilize a preset nuclear power system operation fitting model to construct an operation dependency matrix matching the current operating scenario type of the target nuclear power system through simulation testing, based on the attribute information of the target nuclear power system and the current operating scenario type: Using the preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, perform evolution simulation testing on the operating state of the target nuclear power system to obtain corresponding simulation test data; based on the simulation test data and causal relationship rules matching the current operating scenario type, perform correlation analysis on the changing relationship of the operating variables of the target nuclear power system and the triggering conditions of the reactor protection system of the target nuclear power system; based on the correlation analysis results, determine the influencing variable that triggers the start of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influencing variable; based on the influencing variable and the safety level of the influencing variable, construct an operation dependency matrix matching the current operating scenario type of the target nuclear power system.

[0122] In some embodiments, when the second construction module 704 is specifically implemented, it can construct a matching objective function and constraints based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system in the following manner: Based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, determine the corresponding optimization objective item; wherein the optimization objective item includes at least: a system power output item; construct a matching objective function based on the optimization objective item; and construct corresponding constraints based on the safety parameter data of the shutdown operation of the non-triggered reaction push protection system.

[0123] In some embodiments, when the first determining module 705 is specifically implemented, it can determine a matching target boundary optimization model from a preset optimization model according to the function type of the objective function and the number of parameters of the target boundary parameters in the following manner: when the function type of the objective function is a single objective function and the number of parameters of the target boundary parameters is less than or equal to a specified number, the preset first optimization model is determined as the matching target boundary optimization model; when the function type of the objective function is a multi-objective function and / or the number of parameters of the target boundary parameters is greater than a specified number, the preset second optimization model is determined as the matching target boundary optimization model.

[0124] In some embodiments, when the target boundary optimization model is a preset first optimization model, the second determining module 706 can, in specific implementation, combine the target boundary optimization model and the preset nuclear power system operation fitting model in the following manner: based on the attribute information, objective function, and constraints of the target nuclear power system, perform exploratory optimization processing on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements: using the preset nuclear power system operation fitting model to randomly generate multiple initial sample point data based on the attribute information, target boundary parameters, and current operating scenario type of the target nuclear power system; using the target boundary optimization... The model determines the corresponding operating state labels based on the multiple initial sample point data; and constructs the first initial sample data based on the initial sample point data and the corresponding operating state labels; using the target boundary optimization model, based on the first initial sample data and combined with the objective function and constraints, it performs optimization to determine the initial first operating boundary; it acquires and determines the matching target exploration strategy based on the physical evolution characteristics of the current operating scenario type; and using the target boundary optimization model, based on the target exploration strategy and the initial first operating boundary, it performs multiple rounds of exploration and optimization processing on the neighborhood of the operating boundary to determine the target boundary parameter values ​​that meet the requirements.

[0125] In some embodiments, when the target boundary optimization model is a preset second optimization model, the second determining module 706, when specifically implemented, can use the target boundary optimization model and the preset nuclear power system operation fitting model in combination as follows: based on the attribute information, objective function, and constraints of the target nuclear power system, exploratory optimization processing of the target boundary parameters during the operation of the target nuclear power system is performed to determine the target boundary parameter values ​​that meet the requirements: using the preset nuclear power system operation fitting model to randomly generate multiple initial sample point data based on the attribute information, target boundary parameters, and current operating scenario type of the target nuclear power system; using the target boundary optimization model to predict the corresponding non-stop reactor probability value based on the multiple initial sample point data; and Based on the initial sample point data and the corresponding non-stop probability value, a second initial sample data is constructed. Using the target boundary optimization model, based on the second initial sample data, combined with the physical evolution characteristics, objective function, and constraints of the current operating scenario, an initial second operating boundary is determined. The initial exploration range is determined by obtaining and calculating the model prediction variance of the target boundary optimization model regarding the initial second operating boundary. An initial acquisition function is constructed based on the function difference of the objective function based on the initial second operating boundary and the stop probability value. Based on the initial acquisition function and the initial exploration range, the target boundary optimization model is used to perform multiple rounds of exploration and optimization processing regarding the operating boundary to determine the target boundary parameter values ​​that meet the requirements.

[0126] It should be noted that the units, devices, or modules described in the above embodiments can be implemented by computer chips or physical entities, or by products with certain functions. For ease of description, the above devices are described by dividing them into various modules according to their functions. Of course, in implementing this specification, the functions of each module can be implemented in one or more software and / or hardware, or the module that implements the same function can be implemented by a combination of multiple sub-modules or sub-units, etc. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection between the devices or units shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0127] As can be seen from the above, the platform for determining the operating boundary of a nuclear power system provided in the embodiments of this specification can be precisely adapted to the specific operating scenario of the target nuclear power system. Under the premise of ensuring the overall safe operation of the target nuclear power system without reactor failure, it can fully explore and utilize the safety margin of the operating parameters to accurately determine the target operating boundary with good effect and high application value. Subsequently, by controlling the operation of the target nuclear power system according to the target operating boundary, the power output of the target nuclear power system can be effectively improved.

[0128] In a specific scenario example, the method for determining the operating boundary of a nuclear power system provided in this specification can be applied to identify and optimize the operating boundary of a nuclear power plant. The specific implementation process may include the following:

[0129] In this scenario example, the identification and optimization of the operational boundary of a nuclear power plant is essentially an optimization problem under complex constraints. It possesses the following characteristics: multi-parameter nature: this is reflected in the operator's need to coordinate and optimize multiple variables such as reactor power setpoints and main pump speed; nonlinearity and coupling: nuclear power plants exhibit extremely strong physical nonlinearity during transient processes, and there are complex mutual coupling relationships between various parameters, resulting in nonlinear geometric characteristics of their operational boundary; the operational boundary problem of nuclear power plants may also have multiple objectives: optimization objectives often involve maximizing power output (economy), stability of parameter operating ranges (safety), and rapid response capability of power regulation (flexibility), among other dimensions.

[0130] However, the current identification of the safety boundary of nuclear power plants mainly relies on the deterministic safety analysis (DSA) method, that is, through the iterative process of value taking-simulation-verification, based on envelope analysis and reverse calculation simulation, starting from the extreme accident consequences of "just violating the safety criteria", the parameter safety range of the system under the current operating conditions is deduced. There is still no method for identifying and optimizing the operable boundary of nuclear power plants. Although the existing DSA method has played an important role in ensuring the safe operation of nuclear power plants, it still has the following significant defects when facing increasingly complex operating requirements and performance mining: (1) The boundary setting is too conservative and the resource mining is insufficient. Traditional methods are mostly based on the "worst-case" principle for envelope analysis, which leads to the parameter limits in the operating procedures and technical specifications being designed too conservatively. This excessive safety margin squeezes the actual operating space of nuclear power plants, limits the energy output potential, and causes the system to fail to achieve optimal economy in many scenarios where it could operate safely; (2) There is a lack of strong coupling analysis between parameters. Traditional analysis methods often adopt the isolated thinking of "single variable fluctuation", that is, analyzing the boundary of a single parameter while fixing other parameters. This method fails to fully consider the combined effect of multiple parameters in multidimensional space and the nonlinear coupling effect, making it difficult to identify complex operating boundaries under the coordinated changes of multiple variables, resulting in identification loopholes or redundancy; (3) It has low computational efficiency and relies on expert engineering experience. The iterative process of "value taking-simulation-verification" is highly dependent on engineering experience for point placement, and the sampling is highly blind. In order to obtain a more accurate boundary of a complex system, a large-scale operation simulation program call is required, which is computationally expensive and time-consuming, making it difficult to achieve rapid boundary identification; (4) It lacks an adaptive sampling and objective evaluation system. The traditional sampling method has insufficient sampling density in the boundary region, resulting in limited model recognition accuracy of the boundary. In the attempt to introduce machine learning, there is often a lack of efficient sampling strategies for "boundary exploration". At the same time, there is a lack of a unified algorithm evaluation system that takes into account both convergence speed and recognition accuracy, making it difficult to quantify and evaluate the deviation between the recognition result and the "real boundary"; (5) The construction of the recognition platform is lagging behind, and the support of automated tools is insufficient. Due to the lack of strong support from the automated platform, the formulation of operating procedures is limited, resulting in the inability to fully release the operating potential under specific tasks.

[0131] To address the aforementioned problems and their root causes, this scenario example proposes a method for identifying and optimizing the operational boundaries of nuclear power plants. (See reference...) Figure 8 As shown, it may include the following:

[0132] First, given a library of operating scenarios, define the operating scenarios that the simulation program can simulate (e.g., the types of operating scenarios). Then, establish an operational dependency matrix based on the simulation program (e.g., a pre-defined fitting model for nuclear power system operation) and the operating scenario library.

[0133] Second, select key variables (e.g., target key variables) based on the runtime dependency matrix, and determine the dimensions of the output boundary (e.g., the number of parameters of the target boundary parameters) based on the dimensions of the key parameters.

[0134] Third, determine the optimization objective of the operable boundary (e.g., construct the objective function), such as maximizing the power demand under safe conditions as the optimization objective.

[0135] Fourth, machine learning algorithms (e.g., target boundary optimization models) are used to identify the boundary, explore the operational margin, and design algorithms based on support vector machines (e.g., the pre-defined first optimization model) and ensemble learning Bayesian optimization (e.g., the pre-defined second optimization model).

[0136] Fifth, it outputs the operable boundaries under the current working conditions and visualizes the operable boundaries.

[0137] In this scenario example, regarding the construction of the operational dependency matrix for nuclear power operating conditions based on a simulation program, the specific implementation involves establishing a scenario library as shown in Table 1 using the simulation program. This library covers the entire range of scenarios for nuclear power plants, from steady-state operation to complex transients and accident conditions. Each scenario defines specific physical evolution characteristics, laying the foundational data environment for the targeted extraction of key variables and the construction of optimization functions in subsequent steps.

[0138] After determining the operating scenarios, a preliminary operating scenario evolution simulation was conducted using a nuclear power plant reactor and primary loop system operation simulation program developed based on the PCTRAN model. This constructed a "scenario-causal relationship-operational safety priority" mapping model, namely the Operational Dependency Matrix (ODM). Input: Operating scenario type. Related items: Nuclear physics causal chains calculated based on the self-developed operation simulation program. For example: loss of feedwater flow → decrease in steam generator cooling capacity → increase in primary loop average temperature → triggering reactor shutdown. This causal chain combines physical principles with the program's simulation trend analysis. Output: Operational safety rating (high / medium / low) for each parameter. The logic of the operational safety rating is the correlation between the rate of change of a parameter after deviating from its rated value and the triggering of the Reactor Protection System (RPS) setting. If the rate of deviation of a variable is strongly correlated with the threat level of protection actions, a higher operational safety rating is assigned to it.

[0139] In this scenario example, regarding the selection of key variables based on the runtime dependency matrix, this step selects the adjustable parameters of the runtime simulation program corresponding to the key parameters from high to low based on the runtime security rating of the runtime dependency matrix constructed earlier. The dimensions of the output boundary are then determined based on the dimensions of the key parameters.

[0140] Implementation Example (Taking Loss of Feedwater Flow (LOFW) as an Example): Key variables rated "High": primary coolant average temperature, etc. These variables directly determine whether protection actions are triggered. Variables rated "Medium": coolant flow rate, primary loop pressure, steam flow rate, etc. Adjustable parameters in the simulation program include reactor power setpoint, main pump speed setpoint, pressure and opening degree / power of pressurizers (spray valve / relief valve / safety valve / heater), etc. Therefore, the key variables selected are: reactor power setpoint, main pump speed setpoint, etc. See Table 2 for details.

[0141] In this scenario example, regarding the establishment of the operational optimization objective function, in practice, the optimization objective is set according to operational requirements to identify the maximum operating range while ensuring safety. Single-objective optimization: For example, maximizing the reactor power output of a nuclear power plant to reduce the frequency of power loss. If multiple objectives are considered, such as simultaneously considering the reactor power output, the operating range margin of the manipulated variables, and the stability of the control process, a comprehensive objective function can be constructed. Constraints: The system is constrained by setting the reactor protection system to avoid triggering, and the system state is fed back through a simulation program.

[0142] In this scenario example, regarding boundary identification and exploration optimization based on uncertainty optimization, this step utilizes machine learning algorithms in conjunction with a lightweight operational simulation program to achieve rapid identification and accurate discovery of the operational boundaries of the nuclear power plant. This includes an adaptive boundary identification algorithm based on SVM and a boundary optimization algorithm based on ensemble learning Bayesian optimization.

[0143] (1) SVM-based adaptive boundary recognition algorithm (suitable for low-dimensional / single-target scenarios)

[0144] For scenarios with low-dimensional key operational variables (e.g., ≤ 2 dimensions) and a single optimization objective, a Gaussian kernel-based Support Vector Machine (SVM) active learning algorithm is adopted. Its core is to intelligently select highly informative sample points to efficiently approximate the nonlinear true boundary of the feasible region. The specific steps are as follows: 1) Initial modeling: Randomly collect N initial samples (e.g., N=10), obtain their running status labels (stopped or not stopped) by running a simulation program, and train the initial SVM boundary classifier.

[0145] 2) Uncertainty Sampling: Search near the decision boundary generated by the current SVM. This region has the highest uncertainty; select new points within this region for simulation verification.

[0146] 3) Support Vector Exploration and Diversity Optimization: Randomly generate a set of candidate points near the current support vectors. Calculate the Euclidean distance between the candidate points and the collected sample set, and select the point with the farthest distance as the sampling point. This strategy aims to ensure the comprehensiveness of boundary exploration and avoid getting trapped in local optima. Convergence Judgment: Introduce the failure probability (i.e., the probability of stopping) index to evaluate the coverage of the current identification boundary. If the relative increase of the failure probability in two consecutive rounds is less than a set threshold (e.g., 0.01), the algorithm converges and outputs the final boundary; otherwise, update the model and return to step 2.

[0147] (2) Boundary optimization algorithm based on ensemble learning Bayesian optimization (suitable for high-dimensional / multi-objective scenarios)

[0148] For complex scenarios involving multi-parameter coupling (e.g., ≥3 dimensions) and multi-objective optimization, an ensemble learning Bayesian optimization algorithm is employed. This algorithm utilizes the variance of the ensemble neural network to characterize the uncertainty of the model's predictions, thereby guiding the in-depth mining of boundary margins.

[0149] 1) Ensemble Classifier Construction: An ensemble model consisting of 5 neural networks (NNs) is constructed. Each NN contains 4 fully connected layers, and each layer contains 64 neurons (the number of neurons is dynamically adjusted according to the input dimension). The ReLU activation function is used between layers. Variational inference is used for training, and 10 sets of initial operational variable combinations are collected for pre-training.

[0150] 2) Constraint Region and Acquisition Function Optimization: Dynamic Search Interval Limitation: Establishing upper and lower constraints U and L based on uncertainty.

[0151] in, This represents the variance of the probability values ​​of the outputs of multiple neural networks that do not stop the heap. The baseline value A is usually set to 0.5, but can be adjusted according to the degree of exploration preference for infeasible regions (such as heap stopping regions). For example, increasing the degree of exploration of infeasible regions can moderately reduce the baseline value of A.

[0152] 3) Acquisition Function Design: Design an acquisition function based on the difference in the objective function. Calculate the objective function increment between the candidate sampling point and the current optimal solution. Select the next sampling point by maximizing the acquisition function. If the increment is greater than 0 and the probability of continuous heaping is ≥0.5, assign a high weight; otherwise, if the probability is <0.5, return a negative value as a penalty to guide the algorithm back to the feasible region.

[0153] 4) Model Iteration and Convergence: Obtain new sample labeling by running the simulation program and update the ensemble model parameters. Simultaneously calculate the failure probability; if it is less than a threshold (e.g., 0.01) for M consecutive rounds, stop the iteration.

[0154] (3) Decision-making logic of the algorithm's applicable scenarios

[0155] Among them, the SVM active learning algorithm focuses on single-objective, low-dimensional scenarios. Its technical advantage lies in its fast decision-making speed and low computational cost for simple nonlinear boundaries.

[0156] Ensemble learning Bayesian optimization algorithm: focuses on multi-objective, high-dimensional scenarios. It solves the problem of small sample size through a collective voting mechanism of neural networks and achieves robust recognition of complex boundaries by utilizing variance information.

[0157] It should be noted that in this scenario example, the SVM-based adaptive boundary recognition algorithm used is from "global search" to "precise outlining". The original SVM is a general classification algorithm, while this method achieves efficient mining of nuclear power operation boundaries by introducing an active learning mechanism.

[0158] Specifically, on the one hand, this method addresses the shortcomings of traditional heuristic sampling algorithms based on "uncertainty measures": they typically require pre-collecting a large number of samples for training, but in nuclear power simulations, random sampling across the entire space is highly inefficient and costly. The improvement lies in the algorithm's approach: instead of reading all data at once, it actively seeks points with the most information near the "decision boundary" generated by the current SVM after initial modeling with a small sample, performing simulation verification. Physically, this concentrates computational resources on the critical zone between "safety" and "shutdown," achieving the highest accuracy in boundary identification with the fewest simulations (low computational cost). On the other hand, it improves the logic by incorporating support vector exploration with "diversity constraints": it introduces Euclidean distance calculation in boundary point selection, forcing the algorithm to select new points furthest from the already collected sample points. The innovative value lies in solving the problem of the original SVM easily getting trapped in "local boundary regions" under complex nonlinear boundaries, ensuring that the algorithm can detect isolated and extreme operational risk points that may exist in the multidimensional space of the nuclear power system.

[0159] The ensemble learning Bayesian optimization algorithm used: From "point estimation" to "uncertainty quantification". For high-dimensional multi-objective scenarios, this method combines deep neural networks (NN) with Bayesian optimization to solve the robustness problem under complex coupled parameters.

[0160] Specifically, on the one hand, this can be achieved based on the "collective intelligent voting" operational safety evaluation mechanism: an ensemble model consisting of five neural networks is constructed, using the variance of collective predictions to dynamically characterize the "uncertainty" of the system state. Its innovative value lies in the fact that the original algorithm often gives a binary conclusion of reactor shutdown. This method quantifies the model's "confidence" in a specific operating condition through variance. For areas with high uncertainty, the algorithm automatically strengthens its exploration to prevent misjudgments caused by model bias, meeting the stringent safety margin requirements of the nuclear power industry. On the other hand, the logic is improved through a custom acquisition function with "physical constraint penalties": in the Bayesian optimized acquisition function, the "reactor shutdown probability" is explicitly used as a penalty term (a negative value is assigned if the reactor shutdown probability is < 0.5). Its physical significance lies in directly integrating the physical hard constraint of the nuclear power plant's "reactor protection system setting value" into the algorithm's mathematical optimization logic. Compared to the original algorithm, it can more effectively guide the system to explore its operational potential while always remaining within a safe feasible region.

[0161] In this scenario example, regarding the output and application of the operational boundary, in specific implementation, the multi-dimensional operational boundary under the current operating conditions is output. This result can be directly used to: optimize operating procedures and parameter limits in technical specifications, reducing unnecessary conservative redundancy; and to explore system operating margins, reduce unplanned reactor shutdown frequency, and improve the economy and flexibility of nuclear power plants.

[0162] For details regarding platform architecture and interface design in this scenario example, please refer to [link / reference]. Figure 9 As shown, the platform can consist of five functional modules, achieving closed-loop feedback through data interfaces. Among them, Figure 9 The final output boundary map can be found in [reference]. Figure 5 As shown.

[0163] User input module: Receives the specified scenario conditions to be analyzed (such as "partial loss of water supply flow").

[0164] Key variable extraction module: Based on the runtime dependency matrix (ODM), identify key parameters (such as reactor power setpoint).

[0165] Algorithm library module: Includes SVM (low-dimensional / single-objective) and ensemble learning Bayesian optimization (high-dimensional / multi-objective) algorithms.

[0166] The simulation program module serves as a simulation verification base, driving algorithm iteration by providing feedback on system status (whether a shutdown protection signal is triggered).

[0167] The platform output module enables boundary visualization, comparison, and feasible region division. This includes the "physical true boundary" (used only to verify the accuracy of the boundary generated by the method described in this invention; in actual engineering situations, there is often no such thing as a true boundary), the initial procedure boundary determined according to the current technical specifications or operating procedures of the nuclear power plant, and the boundary generated using the operating boundary identification and optimization method described in this invention. The operating conditions are distinguished by color and feature points. Green areas and green feasible region points represent operating states where the system is within safety margins and has not triggered an emergency shutdown, while red areas and red infeasible region points represent restricted operating conditions where system parameters have reached protection system settings and triggered an emergency shutdown. For details, please refer to [link to relevant documentation]. Figure 5 As shown.

[0168] Specifically, in terms of interface design, deep integration of algorithms and simulations is achieved through standardized encapsulation. The simulation program encapsulation interface features dynamic parameter injection, enabling real-time modification and feedback of thermal-hydraulic parameters. The algorithm logic encapsulation interface standardizes inputs (parameter names, search space) and outputs (boundary data point sets). The main control program mapping mechanism utilizes a built-in "scenario-variable name-value range" mapping in the configuration dictionary to achieve closed-loop scheduling.

[0169] The above scenario examples validate the method for determining the operating boundary of the nuclear power system provided in this manual: 1) Through a nuclear power plant reactor and primary loop system operation simulation program, the operating scenario of the nuclear power plant was simulated. 2) The key variable selection method based on the nuclear power plant operating dependency matrix (ODM) achieved the transformation from scenario conditions to core variables. 3) Through an adaptive boundary identification and optimization dual-strategy method, a differentiated algorithm path of SVM active learning and ensemble Bayesian optimization was adopted to address dimensional differences. 4) The framework and interface design of a workable boundary identification and optimization platform with a two-way closed-loop feedback mechanism: an integrated platform that integrates real-time algorithm-driven simulation and simulation result feedback to correct the algorithm.

[0170] While this specification provides the steps of operation for the methods described in the embodiments or flowcharts, more or fewer steps may be included based on conventional or non-inventive means. The order of steps listed in the embodiments is merely one possible order of execution among many steps and does not represent the only possible order. In actual device or client product execution, the methods shown in the embodiments or drawings may be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment, or even a distributed data processing environment). The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, product, or apparatus. Without further limitations, the presence of other identical or equivalent elements in a process, method, product, or apparatus that includes said elements is not excluded. The terms "first," "second," etc., are used to denote names and do not indicate any particular order.

[0171] Those skilled in the art will also know that, besides implementing the controller using purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the controller function as logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the devices within it used to implement various functions can also be considered structures within that hardware component. Alternatively, the devices used to implement various functions can be considered as both software modules implementing the method and structures within a hardware component.

[0172] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, classes, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer-readable storage media, including storage devices.

[0173] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this specification can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions of this specification can essentially be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments of this specification.

[0174] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. This specification can be used in numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable electronic devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0175] Although this specification has been described by way of examples, those skilled in the art will recognize that many variations and modifications are possible without departing from the spirit of this specification, and it is intended that the appended text include such variations and modifications without departing from the spirit of this specification.

Claims

1. A method for determining the operating boundary of a nuclear power system, characterized in that, include: Obtain the attribute information of the target nuclear power system and determine the current operating scenario type of the target nuclear power system; Using a pre-defined nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, an operation dependency matrix matching the current operating scenario type of the target nuclear power system is constructed through simulation testing; wherein, the operation dependency matrix includes at least the influence variable that triggers the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variable; Based on the current operating scenario type of the target nuclear power system, influence variables that meet the safety level requirements are selected from the operation dependency matrix as target key variables; and based on the target key variables, the corresponding target boundary parameters are determined. Based on the current operating scenario type of the target nuclear power system and its operating indication parameters, a matching objective function and constraints are constructed; wherein, the objective function includes: a single objective function or a multi-objective function; Based on the function type of the objective function and the number of parameters of the objective boundary parameters, a matching objective boundary optimization model is determined from the preset optimization models; wherein, the preset optimization model includes at least: a preset first optimization model and a preset second optimization model; By combining the target boundary optimization model and the preset nuclear power system operation fitting model, and based on the attribute information, objective function, and constraints of the target nuclear power system, exploratory optimization processing is performed on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements, which are then used as the target operating boundaries of the target nuclear power system.

2. The method according to claim 1, characterized in that, After determining the target boundary parameter values ​​that meet the requirements as the target operating boundary of the target nuclear power system, the method further includes: Based on the target operating boundary, update the target operating rules and corresponding parameter values ​​in the target technical specifications of the target nuclear power system to obtain the updated target operating rules and updated target technical specifications; Control the operation of the target nuclear power system according to the updated target operating rules and the updated target technical specifications.

3. The method according to claim 1, characterized in that, The working condition scenario types include: working condition types, and combinations of sub-scenario types under the working condition types; The operating condition types include: normal operating condition, expected operating condition, and accident operating condition; Correspondingly, the sub-scenario types under normal operating conditions include: steady-state operation sub-scenario, load adjustment transient sub-scenario, and normal shutdown / startup sub-scenario; The sub-scenario types under the expected operating conditions include: load transient event sub-scenario, secondary loop heat dissipation increase sub-scenario, and secondary loop heat dissipation decrease sub-scenario. The sub-scenario types under the accident conditions include: primary loop cooling and flow abnormality sub-scenario, and reactive abnormal event sub-scenario.

4. The method according to claim 3, characterized in that, The step of using a pre-set nuclear power system operation fitting model to construct an operation dependency matrix matching the current operating scenario type of the target nuclear power system through simulation testing, based on the attribute information of the target nuclear power system and the current operating scenario type, includes: Using the preset nuclear power system operation fitting model, based on the attribute information of the target nuclear power system and the current operating scenario type, an evolution simulation test of the operating state of the target nuclear power system is conducted to obtain corresponding simulation test data. Based on the simulation test data and the causal relationship rules that match the current operating scenario type, a correlation analysis is performed on the changing relationship of the operating variables of the target nuclear power system and the triggering conditions of the reactor protection system of the target nuclear power system. Based on the correlation analysis results, the influencing variables that trigger the start-up of the reactor protection system of the target nuclear power system based on the current operating scenario type, as well as the safety level of the influencing variables, are determined. Based on the influencing variables and their safety levels, an operational dependency matrix is ​​constructed that matches the current operating scenario type of the target nuclear power system.

5. The method according to claim 3, characterized in that, Based on the current operating scenario type of the target nuclear power system and its operating indication parameters, a matching objective function and constraints are constructed, including: Based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system, corresponding optimization target items are determined; wherein, the optimization target items include at least: system power output item; Based on the optimization objective, construct a matching objective function; Based on the safety parameter data of the shutdown operation of the non-triggered reaction push protection system, corresponding constraints are constructed.

6. The method according to claim 3, characterized in that, The step of determining a matching target boundary optimization model from a preset optimization model based on the function type of the objective function and the number of parameters of the target boundary parameters includes: When the objective function is a single objective function and the number of parameters of the objective boundary parameters is less than or equal to a specified number, the preset first optimization model is determined as the matching objective boundary optimization model. When the objective function is a multi-objective function and / or the number of parameters of the objective boundary parameters is greater than a specified number, the preset second optimization model is determined as the matching objective boundary optimization model.

7. The method according to claim 6, characterized in that, When the target boundary optimization model is a preset first optimization model, the joint use of the target boundary optimization model and the preset nuclear power system operation fitting model, based on the attribute information, objective function, and constraints of the target nuclear power system, performs exploratory optimization processing on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements, including: Using a pre-defined nuclear power system operation fitting model, multiple initial sample point data are randomly generated based on the attribute information of the target nuclear power system, the target boundary parameters, and the current operating scenario type. Using the target boundary optimization model, corresponding running status labels are determined based on the multiple initial sample point data; and first initial sample data is constructed based on the initial sample point data and the corresponding running status labels. Using the target boundary optimization model, based on the first initial sample data, combined with the objective function and constraints, optimization is performed to determine the initial first operating boundary; Obtain and determine the matching target exploration strategy based on the physical evolution characteristics of the current working scenario type. Using the target boundary optimization model, based on the target exploration strategy and the initial first running boundary, multiple rounds of exploration and optimization processing are performed on the neighborhood of the running boundary to determine the target boundary parameter values ​​that meet the requirements.

8. The method according to claim 6, characterized in that, When the target boundary optimization model is a preset second optimization model, the joint use of the target boundary optimization model and the preset nuclear power system operation fitting model, based on the attribute information, objective function, and constraints of the target nuclear power system, performs exploratory optimization processing on the target boundary parameters during the operation of the target nuclear power system to determine the target boundary parameter values ​​that meet the requirements, including: Using a pre-defined nuclear power system operation fitting model, multiple initial sample point data are randomly generated based on the attribute information of the target nuclear power system, the target boundary parameters, and the current operating scenario type. Using the target boundary optimization model, the corresponding non-stop heap probability value is predicted based on the multiple initial sample point data; and a second initial sample data is constructed based on the initial sample point data and the corresponding non-stop heap probability value. The target boundary optimization model is used to optimize and solve the problem based on the second initial sample data, combined with the physical evolution characteristics, objective function and constraints of the current working condition scenario, to determine the initial second operating boundary. Obtain and optimize the model prediction variance of the target boundary with respect to the initial second running boundary to determine the initial exploration range; Calculate and construct the initial acquisition function based on the objective function difference at the initial second running boundary and the shutdown probability value; Based on the initial acquisition function and the initial exploration range, the target boundary optimization model is used to perform multiple rounds of exploration and optimization processing on the operating boundary to determine the target boundary parameter values ​​that meet the requirements.

9. A platform for determining the operating boundary of a nuclear power system, characterized in that, include: The acquisition module is used to acquire the attribute information of the target nuclear power system and determine the current operating scenario type of the target nuclear power system; The first construction module is used to utilize a preset nuclear power system operation fitting model to construct an operation dependency matrix that matches the current operating scenario type of the target nuclear power system through simulation testing, based on the attribute information of the target nuclear power system and the current operating scenario type; wherein, the operation dependency matrix includes at least the influence variable that triggers the start of the reactor protection system of the target nuclear power system based on the current operating scenario type, and the safety level of the influence variable; The filtering module is used to filter out the impact variables that meet the safety level requirements from the operation dependency matrix as target key variables based on the current operating scenario type of the target nuclear power system; and to determine the corresponding target boundary parameters based on the target key variables. The second construction module is used to construct a matching objective function and constraints based on the current operating scenario type of the target nuclear power system and the operating indication parameters of the target nuclear power system; wherein, the objective function includes: a single objective function or a multi-objective function; The first determining module is used to determine a matching target boundary optimization model from a preset optimization model based on the function type of the objective function and the number of parameters of the target boundary parameters; wherein the preset optimization model includes at least: a preset first optimization model and a preset second optimization model; The second determining module is used to jointly use the target boundary optimization model and the preset nuclear power system operation fitting model to perform exploratory optimization processing on the target boundary parameters of the target nuclear power system during operation based on the attribute information, objective function, and constraints of the target nuclear power system, so as to determine the target boundary parameter values ​​that meet the requirements, which are used as the target operating boundaries of the target nuclear power system.

10. A computer program product, characterized in that, It includes a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1 to 8.

Images