A stress monitoring method, device, electronic device and storage medium

Abstract

Embodiments of the present application disclose a stress monitoring method and device, electronic equipment and a storage medium. The method comprises: constructing a stress prediction model according to historical operation data, initializing at least two hyperparameters of the stress prediction model according to a chaotic mapping rule to obtain a hyperparameter combination; optimizing the hyperparameter combination according to a fitness function and a preset grey wolf algorithm to determine an optimal hyperparameter combination in the hyperparameter combination; and taking the optimal hyperparameter combination as an optimal result of the stress prediction model, so that the optimal result of the stress prediction model is packaged into a data packet to monitor stress. According to the embodiments of the present application, the hyperparameters of the stress prediction model are initialized according to the chaotic mapping rule, which can improve the diversity of hyperparameter optimization. The hyperparameter combination is optimized according to the fitness function and the preset grey wolf algorithm to monitor stress, which can improve the real-time performance and accuracy of stress monitoring and ensure safe and stable operation of a unit.

Description

Technical Field

[0001] This invention relates to the field of stress monitoring technology for thermal power units, and in particular to a stress monitoring method, device, electronic equipment, and storage medium. Background Technology

[0002] With the introduction of the "dual carbon" target, the construction and promotion of new energy sources such as wind power and photovoltaics have been vigorously advanced. The role of thermal power units has gradually shifted towards providing a safety net and flexible power source, leading to more frequent participation of thermal power units in grid peak shaving. During rapid start-up and shutdown and load shifting conditions, key components of thermal power units, such as the turbine rotor, will experience significant temperature gradient changes. A large temperature difference exists between the inner and outer surfaces of the turbine rotor, resulting in rotor stress levels significantly higher than the design conditions, causing substantial lifespan loss. Simultaneously, during generator operation, due to the high-speed rotation of the turbine rotor, existing contact-based stress monitoring methods are insufficient for online real-time monitoring of turbine rotor surface stress. Therefore, a real-time turbine rotor stress monitoring method is urgently needed to ensure the safe and stable operation of the turbine.

[0003] Currently, in engineering, methods such as stress simplification algorithms based on unsteady-state heat conduction equations, temperature probe methods based on field data, and inertial link methods based on control theory are generally used to calculate and monitor the stress of steam turbine rotors. However, these methods suffer from problems such as poor real-time performance and poor calculation accuracy, and cannot guarantee the safe and stable operation of steam turbines. Summary of the Invention

[0004] In view of this, the present invention provides a stress monitoring method, device, electronic device and storage medium, which can ensure the diversity in the hyperparameter combination optimization process, improve the real-time performance and accuracy of stress monitoring, and ensure the safe and stable operation of the unit and the service life of the unit.

[0005] According to one aspect of the present invention, an embodiment of the present invention provides a stress monitoring method, the method comprising: A stress prediction model is constructed based on historical operating data; The stress prediction model is initialized according to the chaotic mapping rule to obtain at least one set of hyperparameter combinations. The optimal hyperparameter combination among the at least one set of hyperparameter combinations is determined by optimizing the at least one set of hyperparameter combinations based on the pre-constructed fitness function and the preset gray wolf algorithm. The optimal combination of hyperparameters is used as the optimal result of the stress prediction model, so that the optimal result of the stress prediction model is packaged into a data packet for stress monitoring.

[0006] According to another aspect of the present invention, embodiments of the present invention also provide a stress monitoring device, the device comprising: The model building module is used to build stress prediction models based on historical operational data. An initialization module is used to initialize at least two hyperparameters of the stress prediction model according to the chaotic mapping rule to obtain at least one set of hyperparameter combinations. The optimization module is used to optimize the at least one set of hyperparameter combinations according to the pre-constructed fitness function and the preset gray wolf algorithm, and determine the optimal hyperparameter combination among the at least one set of hyperparameter combinations; The monitoring module is used to take the optimal combination of hyperparameters as the optimal result of the stress prediction model, so that the optimal result of the stress prediction model is encapsulated into a data packet for stress monitoring.

[0007] According to another aspect of the present invention, embodiments of the present invention also provide an electronic device, the terminal device comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the stress monitoring method according to any embodiment of the present invention.

[0008] According to another aspect of the present invention, embodiments of the present invention also provide a computer-readable storage medium storing computer instructions for causing a processor to execute and implement the stress monitoring method described in any embodiment of the present invention.

[0009] In this embodiment of the invention, a stress prediction model is constructed using historical operating data, and at least two hyperparameters of the stress prediction model are initialized according to chaotic mapping rules, which can improve the diversity of the hyperparameter combination optimization process. The at least one set of hyperparameter combinations is optimized by a pre-constructed fitness function and a preset gray wolf algorithm to obtain the optimal hyperparameter combination, thereby monitoring stress. This can improve the real-time performance and accuracy of stress monitoring, realize online real-time monitoring of the stress of high-speed rotating turbine rotors, and provide guidance for setting parameters such as the unit's temperature change rate and power change rate, ensuring the safe and stable operation of the unit and extending the unit's lifespan.

[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating a stress monitoring method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a key component of a steam turbine rotor according to an embodiment of the present invention; Figure 3 A flowchart illustrating yet another stress monitoring method provided in an embodiment of the present invention; Figure 4 This is a schematic diagram comparing chaotic initialization and random initialization according to an embodiment of the present invention; Figure 5 This is a schematic diagram of a crossover mutation operation provided in an embodiment of the present invention; Figure 6 A schematic diagram showing the comparison between a prior art and an improved Grey Wolf algorithm according to an embodiment of the present invention; Figure 7 This is a flowchart illustrating another stress monitoring method provided in an embodiment of the present invention; Figure 8 This is a structural block diagram of a stress monitoring device provided in an embodiment of the present invention; Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0013] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0014] It should be noted that the terms "first," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0015] In one embodiment, Figure 1 This is a flowchart illustrating a stress monitoring method according to an embodiment of the present invention. This embodiment is applicable to stress monitoring of key components of a thermal power unit's steam turbine. The method can be executed by a stress monitoring device, which can be implemented in hardware and / or software and can be configured in an electronic device.

[0016] like Figure 1 As shown, the specific steps include: S110. Construct a stress prediction model based on historical operating data.

[0017] Among them, historical operating data refers to the relevant historical operating data of key parts of the turbine rotor, which includes, but is not limited to, main steam pressure, main steam temperature, metal temperature, speed, temperature change rate, power change rate, and rotor volume average temperature.

[0018] In this embodiment, relevant historical operating data corresponding to key components of the turbine rotor can be obtained from the distributed control system based on the specific application scenario and the requirements for constructing the stress prediction model. A stress prediction model, along with its corresponding constraints, can then be constructed based on this historical operating data. Of course, the historical operating data corresponding to key components of different units may be the same for some and different for others; the appropriate stress prediction model will be constructed according to the requirements.

[0019] S120. Initialize at least two hyperparameters of the stress prediction model according to the chaotic mapping rule to obtain at least one set of hyperparameter combinations.

[0020] Among them, the chaotic mapping rule can be understood as generating a chaotic sequence through a certain iterative method, which has the characteristics of randomness, ergodicity and sensitivity to initial values.

[0021] In this embodiment, the hyperparameter combination includes at least two hyperparameters, which can be the kernel width (g) and the penalty factor (c), or other hyperparameters; this embodiment does not impose any restrictions. This embodiment can obtain at least one set of hyperparameter combinations by initializing at least two hyperparameters of the stress prediction model. Of course, each set of hyperparameter combinations must include at least two hyperparameters: the kernel width (g) and the penalty factor (c).

[0022] In this embodiment, at least two hyperparameters corresponding to a set of stress prediction models can be randomly generated. A Logistic mapping is then used to map the randomly generated hyperparameters to N sets of chaotic hyperparameter combinations. The components corresponding to the N sets of chaotic hyperparameter combinations are then mapped to the value ranges of the variables corresponding to the at least two hyperparameters, respectively, to obtain at least one set of hyperparameter combinations. By employing chaotic initialization in this embodiment, the individuals in the initial population can be discretely distributed within the set solution space, which can very effectively improve the diversity of hyperparameters.

[0023] S130. Optimize at least one set of hyperparameter combinations based on the pre-constructed fitness function and the preset gray wolf algorithm, and determine the optimal hyperparameter combination among at least one set of hyperparameter combinations.

[0024] The fitness function can be understood as the root mean square error function between the actual stress value and the predicted stress value. The pre-defined Grey Wolf algorithm is an improved version, mainly reflected in the following aspects: initializing hyperparameters using chaotic mapping, performing crossover and mutation operations on hyperparameters, and dynamically updating hyperparameters using a weighted average based on the fitness scaling factor.

[0025] The optimal hyperparameter combination can be understood as the optimal hyperparameter combination obtained through the improved Grey Wolf algorithm and the iterative update of the fitness function. This optimal hyperparameter combination contains at least two hyperparameters: kernel width (g) and penalty factor (c).

[0026] In some embodiments, the pre-constructed fitness function can be expressed by the formula: ,in, Represented as the true stress value, This is expressed as a predicted stress value.

[0027] In this embodiment, based on a pre-constructed fitness function, three hyperparameter combinations in at least one set of hyperparameter combinations whose error between the true value and the predicted value of the first stress reaches a first preset error threshold can be identified. A cross-operation method is used to cross-mutate at least two hyperparameters in at least one set of hyperparameter combinations to obtain at least one set of first hyperparameter combinations after cross-mutation. Then, based on the pre-constructed fitness function, a second fitness value corresponding to at least one set of first hyperparameter combinations is determined. Based on the second fitness value, three sets of first hyperparameter combinations whose error between the true value and the predicted value of the second stress reaches a second preset error threshold are identified. By comparing the three sets of hyperparameter combinations whose error reaches the first preset error threshold with the three sets whose error reaches the second preset error threshold, three target hyperparameter combinations whose error reaches a third preset error threshold are selected as the current optimal three sets of hyperparameter combinations. Furthermore, a dynamic weighted average can be performed based on the levels corresponding to the current optimal three sets of hyperparameter combinations and a pre-configured fitness ratio coefficient to update the hyperparameter combinations. This iterative update continues until the optimal hyperparameter combinations are obtained.

[0028] In some embodiments, the dynamic weighted average based on the pre-configured fitness scaling coefficient to update the hyperparameter combination can be performed after the chaotic mapping initialization process, or after comparing the three sets of hyperparameter combinations before and after the crossover operation mutation. This embodiment does not impose any restrictions on this.

[0029] S140. The optimal combination of hyperparameters is used as the optimal result of the stress prediction model, so that the optimal result of the stress prediction model is packaged into a data packet for stress monitoring.

[0030] In this embodiment, at least one set of hyperparameter combinations is optimized using a pre-constructed fitness function and a preset gray wolf algorithm. The optimal hyperparameter combination is found through iterative updates, and this optimal combination is used as the optimal result of the stress prediction model. This optimal result is then packaged into a data packet to monitor the stress in key components of the turbine rotor. For example, to better understand the monitoring of stress in key components of the turbine rotor... Figure 2 This is a schematic diagram of a key component of a steam turbine rotor provided in an embodiment of the present invention.

[0031] The technical solution of this invention constructs a stress prediction model using historical operating data and initializes at least two hyperparameters of the stress prediction model according to chaotic mapping rules, which can improve the diversity of the hyperparameter combination optimization process. By optimizing at least one set of hyperparameter combinations using a pre-constructed fitness function and a preset gray wolf algorithm, the optimal hyperparameter combination is obtained, thereby monitoring stress. This improves the real-time performance and accuracy of stress monitoring, realizes online real-time monitoring of the stress of high-speed rotating turbine rotors, and provides guidance for setting parameters such as the unit's temperature change rate and power change rate, ensuring the safe and stable operation of the unit and extending its service life.

[0032] In one embodiment, Figure 3 This is a flowchart of another stress monitoring method provided in an embodiment of the present invention. Based on the above embodiments, this embodiment further refines the process by initializing at least two hyperparameters of the stress prediction model according to the chaotic mapping rule to obtain at least one set of hyperparameter combinations, and by optimizing at least one set of hyperparameter combinations according to a pre-constructed fitness function and a preset gray wolf algorithm to determine the optimal hyperparameter combination among at least one set of hyperparameter combinations.

[0033] like Figure 3 As shown, the stress monitoring method in this embodiment may specifically include the following steps: S310. Construct a stress prediction model based on historical operating data.

[0034] S320. Randomly generate at least two hyperparameters corresponding to a set of stress prediction models; wherein the hyperparameters include at least the kernel function width (g) and the penalty factor (c).

[0035] In this embodiment, after constructing a pressure monitoring and prediction model using historical operating data of key parts of the turbine rotor, a set of two or more hyperparameters corresponding to the stress monitoring test model are randomly generated. The randomly generated hyperparameters include at least the kernel function width (g) and the penalty factor (c).

[0036] S330. Use Logistic mapping to map a randomly generated set of hyperparameters to generate N sets of chaotic hyperparameter combinations.

[0037] In this embodiment, a randomly generated set of hyperparameters can be mapped to N sets of chaotic hyperparameter combinations using a Logistic mapping. The Logistic mapping is expressed by the following formula: ,in, It is a chaotic variable with a value range of (0,1). The value range of is [0,4]. When When the value is 4, the population is in a completely chaotic state. In this embodiment, to facilitate a better understanding of Logistic mapping initialization, the individuals of the initial population can be discretely distributed within a set solution space, which can very effectively improve diversity. Figure 4 This is a schematic diagram comparing chaotic initialization and random initialization in an embodiment of the present invention.

[0038] S340. Transpose the component carriers corresponding to the N sets of chaotic hyperparameter combinations onto the preset value ranges corresponding to at least two hyperparameters respectively, to obtain at least one set of hyperparameter combinations; wherein each set of hyperparameter combinations contains at least the kernel function width (g) and the penalty factor (c).

[0039] The preset value range refers to the value range corresponding to the hyperparameter. This value range can be set based on experience or manually through experiments, and this embodiment does not impose any restrictions. For example, the kernel function width (g) has a corresponding preset value range, which can be set to a maximum value of 2 and a minimum value of 0.001; the penalty factor (c) also has a corresponding preset value range, which can be set to a maximum value of 500 and a minimum value of 2.

[0040] In this embodiment, after mapping a set of hyperparameters to generate N sets of chaotic hyperparameter combinations, the component carriers corresponding to the N sets of chaotic hyperparameter combinations can be mapped to the preset value ranges corresponding to at least two hyperparameters, resulting in at least one set of hyperparameter combinations. Each set of hyperparameter combinations includes at least a kernel function width (g) and a penalty factor (c). For example, if the penalty factor (c) is 0.5, then it is necessary to set ( ) * 0.5, where, The maximum value of the penalty factor (c) Since the minimum value of the penalty factor (c) is given, this method can be applied to map all values ​​to the corresponding preset range of kernel function width (g) and penalty factor (c).

[0041] S350. Based on the pre-constructed fitness function, determine at least three sets of hyperparameter combinations in which the error between the true value of the first stress and the predicted value of the first stress reaches a first preset error threshold.

[0042] Here, the first true stress value refers to the true stress value before the cross-operation variation, and the first predicted stress value refers to the predicted stress value before the cross-operation variation. The first preset error threshold refers to the error threshold between the first true stress value and the first predicted stress value. This threshold can be set manually or through other methods.

[0043] In this embodiment, based on a pre-constructed fitness function, three hyperparameter combinations in at least one set of hyperparameter combinations whose error between the true value of the first stress and the predicted value of the first stress reaches a first preset error threshold can be determined. Specifically, at least one set of hyperparameter combinations can be input into the pre-constructed fitness function to determine the first fitness value corresponding to each of the at least one set of hyperparameter combinations, and the three hyperparameter combinations whose error between the true value of the first stress and the predicted value of the first stress reaches the first preset error threshold can be determined from the first fitness value. This can be understood as finding the three hyperparameter combinations with the smallest error between the true value of stress and the predicted value of stress.

[0044] In one embodiment, based on a pre-constructed fitness function, three sets of hyperparameter combinations are determined where the error between the actual value of the first stress and the predicted value of the first stress reaches a first preset error threshold. These combinations include: Determine the true value of the first stress corresponding to each historical operating data; The first stress prediction value is obtained by inputting historical operating data into the stress prediction model; At least one set of hyperparameter combinations is input into a pre-constructed fitness function to determine the first fitness value corresponding to each of the at least one set of hyperparameter combinations. Based on the first fitness value, determine three sets of hyperparameter combinations where the error between the true value of the first stress and the predicted value of the first stress reaches the first preset error threshold.

[0045] In this embodiment, the first true stress value corresponding to the historical operating data of key parts of the turbine rotor is determined; the first predicted stress value is obtained by inputting the historical operating data into the stress prediction model; at least one set of hyperparameter combinations is input into a pre-constructed fitness function to determine the first fitness value corresponding to each set of hyperparameter combinations, and based on the first fitness value, three sets of hyperparameter combinations whose error between the first true stress value and the first predicted stress value reaches a first preset error threshold are determined.

[0046] S360. Use the crossover operation method to perform crossover mutation on at least two hyperparameters in at least one set of hyperparameter combinations to obtain at least one set of first hyperparameter combinations after crossover mutation.

[0047] The crossover operation method can be understood as performing a crossover operation on at least two hyperparameters, changing their positions to produce mutations. For example, before mutation, the hyperparameter kernel function width (g) and penalty factor (c) are 250 and 0.1, respectively; the crossover mutation operation can be understood as obtaining the value of penalty factor (c') through penalty factor (c) 0.1, and obtaining the value of kernel function width (g') through kernel function width (g) 250.

[0048] The first hyperparameter combination refers to the hyperparameter combination generated after crossover mutation, which includes at least two hyperparameters: kernel width (g) and penalty factor (c).

[0049] In this embodiment, a crossover operation method can be used to perform crossover mutation on at least two hyperparameters from one or more hyperparameter combinations to obtain multiple first hyperparameter combinations after crossover mutation. Each first hyperparameter combination contains at least two hyperparameters: kernel width (g) and penalty factor (c). Specifically, the upper and lower limit ranges of the values ​​corresponding to at least two hyperparameters in the hyperparameter combination can be determined. Based on the current values ​​of the at least two hyperparameters and their upper and lower limit ranges, the proportions corresponding to the upper and lower limit ranges can be determined. Based on the proportions, the upper and lower limit ranges, and the lower limit of the values ​​corresponding to the at least two hyperparameters, the at least two hyperparameters after crossover mutation can be determined, and the at least two hyperparameters after crossover mutation can be used as the first hyperparameter combination.

[0050] In this embodiment, to facilitate a better understanding of the cross-operation method, Figure 5 This is a schematic diagram of a crossover mutation operation provided in an embodiment of the present invention. Figure 5 In Represented as the dim hyperparameters before crossover mutation. Figure 5 In These are the mutated dim hyperparameters.

[0051] In one embodiment, a crossover operation method is used to perform crossover mutation on at least two hyperparameters in at least one set of hyperparameter combinations to obtain at least one set of first hyperparameter combinations after crossover mutation, including: Determine the upper and lower bounds of the values ​​for at least two hyperparameters in at least one set of hyperparameter combinations; Based on the current values ​​and upper and lower limits of at least two hyperparameters, determine the first ratio corresponding to the upper and lower limits of the values, and cross-change the first ratio to obtain the second ratio; Based on the second ratio, the upper and lower limits of the value range, and the lower limits of the values ​​corresponding to the at least two hyperparameters, determine at least two hyperparameters after crossover mutation, and use the at least two hyperparameters after crossover mutation as at least one set of first hyperparameter combinations.

[0052] The upper and lower limits of the values ​​can be understood as the upper and lower limits of the values ​​corresponding to the hyperparameters. Each hyperparameter has its own upper and lower limits, which can be set based on experience, experiments, or by the user. For example, the upper limit of the kernel width (g) is 2, and the lower limit is 0.001; the upper limit of the penalty factor (c) is 500, and the lower limit is 2.

[0053] In this embodiment, by determining the upper and lower limit ranges of values ​​corresponding to at least two hyperparameters in at least one set of hyperparameter combinations, a first ratio corresponding to the upper and lower limit ranges of values ​​is determined based on the current values ​​of the at least two hyperparameters and their respective upper and lower limit ranges. The first ratio is then cross-variantly changed to obtain a second ratio. Based on the second ratio, the upper and lower limit ranges of values, and the lower limit of values ​​corresponding to the at least two hyperparameters, at least two hyperparameters after cross-variation are determined. These at least two cross-variant hyperparameters are then used as at least one set of first hyperparameter combinations. Specifically, the first ratio can be expressed by the formula: ,in, This represents the upper limit of the hyperparameter's value. This indicates the lower bound of the hyperparameter value. This represents the current value corresponding to the hyperparameter. According to the formula... , thus obtaining at least two hyperparameters after mutation.

[0054] S370. Determine the second fitness value corresponding to at least one set of first hyperparameter combinations based on the pre-constructed fitness function, and determine three sets of first hyperparameter combinations whose error between the second stress true value and the second stress predicted value reaches the second preset error threshold based on the second fitness value.

[0055] Here, the second fitness value refers to the fitness value corresponding to at least two hyperparameters in the first hyperparameter combination after the crossover and mutation operation. The second true stress value refers to the true stress value corresponding to the crossover and mutation operation. The second predicted stress value refers to the predicted stress value corresponding to the crossover and mutation operation. The second preset error threshold is the error threshold between the second true stress value and the second predicted stress value.

[0056] In this embodiment, based on a pre-constructed fitness function, a second fitness value corresponding to one or more sets of cross-mutated first hyperparameter combinations is determined, and based on the second fitness value, three sets of cross-mutated first hyperparameter combinations whose error between the second stress true value and the second stress predicted value reaches a second preset error threshold are determined.

[0057] S380. Compare the three sets of hyperparameter combinations whose errors reach the first preset error threshold with the three sets of first hyperparameter combinations whose errors reach the second preset error threshold, and select the three sets of target hyperparameter combinations whose errors reach the third preset error threshold as the current optimal three sets of hyperparameter combinations.

[0058] Among them, the three optimal hyperparameter combinations are the three hyperparameter combinations with the smallest error that are suspended between the three hyperparameter combinations whose error reaches the first preset error threshold and the three first hyperparameter combinations whose error reaches the second preset error threshold.

[0059] In this embodiment, three sets of hyperparameter combinations with errors reaching a first preset error threshold are compared with three sets of first hyperparameter combinations with errors reaching a second preset error threshold. From these comparisons, three target sets with errors reaching a third preset error threshold are selected as the current optimal three sets of hyperparameter combinations. This can be understood as follows: if the stress error of the three sets of hyperparameter combinations with errors reaching the first preset error threshold is less than the stress error of the three sets of first hyperparameter combinations with errors reaching the second preset error threshold, then the three sets of hyperparameter combinations with errors reaching the first preset error threshold are selected as the current optimal three sets of hyperparameter combinations; if the stress error of the three sets of hyperparameter combinations with errors reaching the first preset error threshold is greater than the stress error of the three sets of first hyperparameter combinations with errors reaching the second preset error threshold, then the three sets of first hyperparameter combinations with errors reaching the second preset error threshold are selected as the current optimal three sets of hyperparameter combinations.

[0060] S390. Based on the levels corresponding to the three optimal hyperparameter combinations and the pre-configured fitness ratio coefficients, a dynamic weighted average is performed to update the hyperparameter combinations. The steps of returning the pre-constructed fitness function to determine the three hyperparameter combinations whose error between the true value of the first stress and the predicted value of the first stress reaches the first preset error threshold are iteratively updated until the optimal hyperparameter combination is obtained.

[0061] In this embodiment, since each of the three sets of hyperparameters corresponds to a different stress error level, a dynamic weighted average can be performed based on the levels corresponding to the current optimal three sets of hyperparameters and the pre-configured fitness ratio coefficient to update the hyperparameters. The step of returning to the pre-constructed fitness function to determine the three sets of hyperparameters where the error between the first true stress value and the first predicted stress value reaches the first preset error threshold is iteratively updated until the optimal set of hyperparameters is obtained. The optimal set of hyperparameters includes at least two hyperparameters: kernel width (g) and penalty factor (c).

[0062] In one embodiment, a dynamic weighted average is performed based on the levels corresponding to the three currently optimal hyperparameter combinations and a pre-configured fitness scaling factor to update the hyperparameter combinations, including: Determine the level differences corresponding to the three optimal combinations of hyperparameters; The hyperparameter combination is updated by dynamically weighting the average based on the level differences and the pre-configured fitness ratio coefficient. The pre-configured fitness ratio coefficient is the ratio of the three optimal hyperparameter combinations to the sum of the three target hyperparameter combinations whose errors reach the third preset error threshold.

[0063] In this embodiment, determining the level differences corresponding to the three currently optimal hyperparameter combinations can be understood as sorting each hyperparameter in the three currently optimal hyperparameter combinations according to the stress error level to determine the level differences. The hyperparameter combinations are then updated by dynamically weighting the level differences and a pre-configured fitness ratio coefficient. In this embodiment, the pre-configured fitness ratio coefficient is the ratio of each of the three currently optimal hyperparameter combinations to the sum of the three target hyperparameter combinations whose errors reach a third preset error threshold. This can be expressed by the formula: ,in, and These represent the fitness function values ​​corresponding to the three optimal hyperparameter combinations, and are sorted by stress error magnitude. , It is a constant. , , They are respectively represented as and The fitness scaling factor has a larger scaling factor for smaller stress errors and a smaller scaling factor for larger stress errors. When updating hyperparameter combinations by dynamically weighting the scaling factor based on level differences and pre-configured fitness scaling factors, the formula is as follows:

[0064] In this embodiment, to facilitate a better understanding of the optimization effect of hyperparameter combinations, Figure 6 This is a schematic diagram comparing the effects of a prior art and an improved gray wolf algorithm according to an embodiment of the present invention, as shown in the figure. Figure 6 As shown, the improved Grey Wolf algorithm yields the optimal combination of hyperparameters.

[0065] S3100: The optimal combination of hyperparameters is used as the optimal result of the stress prediction model, so that the optimal result of the stress prediction model is packaged into a data packet for stress monitoring.

[0066] The above-described technical solution of this invention, by randomly generating at least two hyperparameters corresponding to a set of stress prediction models, using Logistic mapping to map the randomly generated set of hyperparameters to generate N sets of chaotic hyperparameter combinations, and then mapping the component carriers corresponding to the N sets of chaotic hyperparameter combinations to the preset value intervals corresponding to the at least two hyperparameters respectively, obtains at least one set of hyperparameter combinations, which can effectively improve diversity; by determining three sets of hyperparameter combinations in the at least one set of hyperparameter combinations whose error between the true value of the first stress and the predicted value of the first stress reaches a first preset error threshold through a pre-constructed fitness function, and using a crossover operation method to cross-mutate at least two hyperparameters in the at least one set of hyperparameter combinations to obtain at least one set of first hyperparameter combinations after cross-mutation, can increase the number of hyperparameter sets. The diversity of hyperparameter combinations avoids getting stuck in local optima during optimization. It uses a dynamic weighted average of the levels corresponding to the three currently optimal hyperparameter combinations and pre-configured fitness ratio coefficients to update the hyperparameter combinations. It returns to the pre-constructed fitness function to determine the three hyperparameter combinations whose error between the true first stress value and the predicted first stress value reaches a first preset error threshold. This iterative update continues until the optimal hyperparameter combination is obtained. This enables online real-time monitoring of the stress on the high-speed rotating turbine rotor, providing guidance for setting parameters such as the rate of change of unit start-up and shutdown temperature and the rate of change of power, ensuring the real-time safe and stable operation of the unit. It also provides stress monitoring data for calculating the turbine's life loss and assessing its remaining life, and provides data support for online calculation of turbine life loss during the later stages of turbine operation.

[0067] In one embodiment, to facilitate a better understanding of the stress monitoring method, Figure 7 This is a flowchart illustrating another stress monitoring method provided in this embodiment of the invention. In this embodiment, at least one set of hyperparameter combinations is optimized using a pre-constructed fitness function and a preset gray wolf algorithm to obtain the optimal hyperparameter combination, thereby monitoring stress. This embodiment uses kernel width (g) and penalty factor (c) as the two hyperparameters for illustration. This embodiment mainly improves three aspects (chaotic mapping initialization, crossover and mutation operations on kernel width (g) and penalty factor (c), and dynamic weighted average updates of kernel width (g) and penalty factor (c) using fitness scaling factors), aiming to improve the global exploration capability and local exploitation capability of the optimization process, ultimately resulting in higher prediction accuracy of the support vector machine prediction model.

[0068] This embodiment of the invention uses a combination of two hyperparameters, kernel width (g) and penalty factor (c), as an example for illustration. Figure 7 As shown, the stress monitoring method includes the following steps: S710. Based on the requirements for constructing the stress prediction model, obtain historical operating data of the turbine rotor from the distributed control system.

[0069] S720. Construct a stress prediction model based on historical operating data and build a corresponding fitness function.

[0070] S730. Determine N sets of chaotic hyperparameter combinations, the kernel function width (g) and the penalty factor (c) as two hyperparameters, and the maximum number of iterations.

[0071] S740. The kernel function width (g) and penalty factor (c) of the stress prediction model are initialized using chaotic mapping rules to obtain at least one set of hyperparameter combinations.

[0072] In this embodiment, chaos is a form of motion in a nonlinear dynamic system, characterized by randomness, ergodicity, and sensitivity to initial values. Using chaotic initialization allows the individuals of the initial population to be discretely distributed within a predetermined solution space, effectively improving population diversity. The specific steps are as follows: a1. Randomly generate two hyperparameters corresponding to a set of stress prediction models: kernel width (g) and penalty factor (c).

[0073] a2. Use Logistic mapping to generate N chaotic hyperparameter combinations, each combination containing two hyperparameters: kernel width (g) and penalty factor (c).

[0074] a3. Transpose the component carriers corresponding to the N sets of chaotic hyperparameter combinations onto the preset value ranges corresponding to the kernel function width (g) and the penalty factor (c) to obtain at least one set of hyperparameter combinations; wherein each set of hyperparameter combinations includes the kernel function width (g) and the penalty factor (c).

[0075] S750. Based on the pre-constructed fitness function, determine at least three sets of hyperparameter combinations in which the error between the true value of the first stress and the predicted value of the first stress reaches a first preset error threshold.

[0076] S760. Use the crossover operation method to perform crossover mutation on at least two hyperparameters in at least one set of hyperparameter combinations to obtain at least one set of first hyperparameter combinations after crossover mutation.

[0077] In this embodiment, to prevent the optimization process from getting stuck in local optima and failing to find the optimal kernel function width (g) and penalty factor (c), a crossover operation is used to change the kernel function width (g) and penalty factor (c), generating a variant kernel function width (g') and penalty factor (c'), increasing diversity and avoiding getting stuck in local optima.

[0078] S770. Calculate the second fitness value of at least one set of first hyperparameter combinations after mutation based on the pre-constructed fitness function, and determine three sets of first hyperparameter combinations whose error between the second stress true value and the second stress predicted value reaches a second preset error threshold based on the second fitness value.

[0079] S780. Compare the three sets of hyperparameter combinations whose errors reach the first preset error threshold with the three sets of first hyperparameter combinations whose errors reach the second preset error threshold, and select the three sets of target hyperparameter combinations whose errors reach the third preset error threshold as the current optimal three sets of hyperparameter combinations.

[0080] S790. Based on the current optimal three sets of hyperparameter combinations, the corresponding levels and the pre-configured fitness ratio coefficients are dynamically weighted and averaged to update the hyperparameter combinations.

[0081] S7100: Determine whether the current iteration count has reached the maximum iteration count. If yes, execute S7110; otherwise, return to execute S770.

[0082] S7110, Output the optimal kernel function width (g) and penalty factor (c) in the optimal hyperparameter combination.

[0083] In this embodiment, to facilitate understanding of the comparison between the optimal result of the stress prediction model obtained in this invention, packaged into a data packet for stress monitoring, and the accuracy of stress prediction models in the prior art, Table 1 is a comparison table of the accuracy of the optimal result of the stress prediction model obtained in this invention, packaged into a data packet for stress monitoring, and the accuracy of stress prediction models in the prior art. As shown in Table 1, the accuracy of the optimal result of the stress prediction model obtained in this invention, packaged into a data packet for stress monitoring, is significantly better than the accuracy of prediction models in the prior art.

[0084] Table 1: Comparison of the accuracy of the optimal results of the stress prediction model obtained in this invention (encapsulated into a data package for stress monitoring) with existing stress prediction models. In one embodiment, in one embodiment, Figure 8 This is a structural block diagram of a stress monitoring device according to an embodiment of the present invention. This device is suitable for stress monitoring of critical components of a thermal power unit's steam turbine. The device can be implemented in hardware or software. It can be configured in an electronic device to implement a stress monitoring method according to an embodiment of the present invention. Figure 8 As shown, the device includes: a model building module 810, an initialization module 820, an optimization module 830, and a monitoring module 840.

[0085] Among them, the model building module 810 is used to build a stress prediction model based on historical operating data; The initialization module 820 is used to initialize at least two hyperparameters of the stress prediction model according to the chaotic mapping rule to obtain at least one set of hyperparameter combinations. The optimization module 830 is used to optimize the at least one set of hyperparameter combinations according to the pre-constructed fitness function and the preset gray wolf algorithm, and determine the optimal hyperparameter combination among the at least one set of hyperparameter combinations. The monitoring module 840 is used to take the optimal combination of hyperparameters as the optimal result of the stress prediction model, so that the optimal result of the stress prediction model is encapsulated into a data packet for stress monitoring.

[0086] In this embodiment of the invention, the model building module constructs a stress prediction model using historical operating data; the initialization module initializes at least two hyperparameters of the stress prediction model according to chaotic mapping rules, which can improve the diversity of the hyperparameter combination optimization process; the optimization module optimizes the at least one set of hyperparameter combinations using a pre-constructed fitness function and a preset gray wolf algorithm to obtain the optimal hyperparameter combination, thereby monitoring stress and improving the real-time performance and accuracy of stress monitoring. This enables online real-time monitoring of the stress of high-speed rotating turbine rotors, providing guidance for setting parameters such as the unit's temperature change rate and power change rate, ensuring the safe and stable operation of the unit and extending its lifespan.

[0087] In one embodiment, the initialization module 820 includes: The first generation unit is used to randomly generate a set of at least two hyperparameters corresponding to the stress prediction model; wherein the hyperparameters include at least the kernel function width (g) and the penalty factor (c); The second generation unit is used to generate N sets of chaotic hyperparameter combinations by mapping the randomly generated set of hyperparameters using a Logistic mapping. The determining module is used to transfer the component carriers corresponding to the N sets of chaotic hyperparameter combinations to the preset value ranges corresponding to the at least two hyperparameters respectively, to obtain at least one set of hyperparameter combinations; wherein, each set of hyperparameter combinations includes at least the kernel function width (g) and the penalty factor (c).

[0088] In one embodiment, the optimization module 830 includes: The first determining unit is used to determine, based on the pre-constructed fitness function, three sets of hyperparameter combinations in which the error between the true value of the first stress and the predicted value of the first stress reaches a first preset error threshold. The mutation unit is used to perform cross-mutation on at least two hyperparameters in the at least one set of hyperparameter combinations using a cross-operation method to obtain at least one set of first hyperparameter combinations after cross-mutation. The second determining unit is used to determine the second fitness value corresponding to the at least one set of first hyperparameter combinations based on the pre-constructed fitness function, and to determine the three sets of first hyperparameter combinations whose error between the second stress true value and the second stress predicted value reaches a second preset error threshold based on the second fitness value. The comparison unit is used to compare the three sets of hyperparameter combinations whose errors reach a first preset error threshold with the three sets of first hyperparameter combinations whose errors reach a second preset error threshold, and select the three sets of target hyperparameter combinations whose errors reach a third preset error threshold as the current optimal three sets of hyperparameter combinations. The iterative update unit is used to perform dynamic weighted averaging based on the levels corresponding to the three currently optimal hyperparameter combinations and the pre-configured fitness ratio coefficients to update the hyperparameter combinations. It returns the steps of the three hyperparameter combinations where the error between the first true stress value and the first predicted stress value reaches the first preset error threshold, as determined by the pre-constructed fitness function, and iteratively updates until the optimal hyperparameter combination is obtained.

[0089] In one embodiment, the first determining unit includes: The first determining sub-unit is used to determine the true value of the first stress corresponding to the historical operating data; The prediction subunit is used to input the historical operating data into the stress prediction model to obtain the first stress prediction value; The second determining subunit is used to input the at least one set of hyperparameter combinations into the pre-constructed fitness function to determine the first fitness value corresponding to each of the at least one set of hyperparameter combinations. The third determining subunit is used to determine three sets of hyperparameter combinations based on the first fitness value where the error between the first stress true value and the first stress predicted value reaches a first preset error threshold.

[0090] In one embodiment, the variation unit includes: The range is used to determine the sub-unit, which is used to determine the upper and lower limit ranges of the values ​​of at least two hyperparameters in the at least one set of hyperparameter combinations; The ratio determination subunit is used to determine the first ratio corresponding to the upper and lower limits of the values ​​based on the current values ​​of the at least two hyperparameters and the upper and lower limits of the values, and to cross-change the first ratio to obtain the second ratio. The mutation determination subunit is used to determine at least two hyperparameters after crossover mutation based on the second ratio, the upper and lower limit range of the value, and the lower limit of the value corresponding to at least two hyperparameters, and to use the at least two hyperparameters after crossover mutation as at least one set of first hyperparameter combinations.

[0091] In one embodiment, the iterative update unit includes: The difference determination subunit is used to determine the level difference corresponding to the current optimal combination of three hyperparameters; The update subunit is used to dynamically weight the hyperparameter combination based on the level difference and the pre-configured fitness ratio coefficient. The pre-configured fitness ratio coefficient is the ratio of the three currently optimal hyperparameter combinations to the sum of the three target hyperparameter combinations whose errors reach the third preset error threshold.

[0092] In one embodiment, the pre-constructed fitness function is expressed by the formula: ,in, Represented as the true stress value, This is represented as a predicted value.

[0093] In one embodiment, Figure 9 This is a schematic diagram of an electronic device provided for an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The terminal device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0094] like Figure 9As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the terminal device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0095] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows terminal device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0096] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as stress monitoring methods.

[0097] In some embodiments, the stress monitoring method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on terminal device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the stress monitoring method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the stress monitoring method by any other suitable means (e.g., by means of firmware).

[0098] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0099] Computer programs used to implement the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable stress monitoring device, such that when executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer programs can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0100] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0101] To provide interaction with a user, the systems and techniques described herein can be implemented on a terminal device having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the terminal device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0102] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0103] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0104] In one embodiment, the present invention further includes a computer program product, the computer program product comprising a computer program that, when executed by a processor, implements the stress monitoring method described in any embodiment of the present invention.

[0105] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0106] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0107] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A stress monitoring method, characterized in that, include: A stress prediction model is constructed based on historical operating data; The stress prediction model is initialized according to the chaotic mapping rule to obtain at least one set of hyperparameter combinations. The optimal hyperparameter combination among the at least one set of hyperparameter combinations is determined by optimizing the at least one set of hyperparameter combinations based on the pre-constructed fitness function and the preset gray wolf algorithm. The optimal combination of hyperparameters is used as the optimal result of the stress prediction model, so that the optimal result of the stress prediction model is packaged into a data packet for stress monitoring. The initialization of at least two hyperparameters of the stress prediction model according to the chaotic mapping rule to obtain at least one set of hyperparameter combinations includes: Randomly generate a set of at least two hyperparameters corresponding to the stress prediction model; wherein the hyperparameters include at least the kernel width (g) and the penalty factor (c); The randomly generated set of hyperparameters is mapped using a logistic mapping to generate N sets of chaotic hyperparameter combinations; The component carriers corresponding to the N sets of chaotic hyperparameter combinations are mapped to the preset value ranges corresponding to the at least two hyperparameters respectively to obtain at least one set of hyperparameter combinations; wherein, each set of hyperparameter combinations contains at least two hyperparameters: the kernel function width (g) and the penalty factor (c); The step of optimizing the at least one set of hyperparameter combinations based on a pre-constructed fitness function and a preset gray wolf algorithm to determine the optimal hyperparameter combination among the at least one set of hyperparameter combinations includes: Based on the pre-constructed fitness function, determine three sets of hyperparameter combinations in which the error between the true value of the first stress and the predicted value of the first stress reaches a first preset error threshold. The crossover operation method is used to crossover mutate at least two hyperparameters in the at least one set of hyperparameter combinations to obtain at least one set of first hyperparameter combinations after crossover mutation; The second fitness value corresponding to the at least one set of first hyperparameter combinations is determined based on the pre-constructed fitness function, and the three sets of first hyperparameter combinations whose error between the second stress true value and the second stress predicted value reaches the second preset error threshold are determined based on the second fitness value. The three sets of hyperparameter combinations whose errors reach the first preset error threshold are compared with the three sets of first hyperparameter combinations whose errors reach the second preset error threshold, and the three sets of target hyperparameter combinations whose errors reach the third preset error threshold are selected as the current optimal three sets of hyperparameter combinations. The process involves dynamically weighting the hyperparameter combinations based on the levels corresponding to the three currently optimal hyperparameter combinations and the pre-configured fitness ratio coefficients to update the hyperparameter combinations. This process returns to the steps where the pre-constructed fitness function determines the three hyperparameter combinations whose error between the true first stress value and the predicted first stress value reaches a first preset error threshold. This process is iteratively updated until the optimal hyperparameter combination is obtained.

2. The method according to claim 1, characterized in that, The three hyperparameter combinations in which the error between the true value of the first stress and the predicted value of the first stress reaches a first preset error threshold, determined according to the pre-constructed fitness function, include: The at least one set of hyperparameter combinations is input into the pre-constructed fitness function to determine the first fitness value corresponding to each of the at least one set of hyperparameter combinations; Based on the first fitness value, three sets of hyperparameter combinations are determined where the error between the first actual stress value and the first predicted stress value reaches a first preset error threshold.

3. The method according to claim 1, characterized in that, The method of employing a crossover operation to perform crossover mutation on at least two hyperparameters in the at least one set of hyperparameter combinations to obtain at least one set of first hyperparameter combinations after crossover mutation includes: Determine the upper and lower limit ranges of the values ​​of at least two hyperparameters in the at least one set of hyperparameter combinations; Based on the current values ​​of the at least two hyperparameters and the upper and lower limits of the values, a first ratio corresponding to the upper and lower limits of the values ​​is determined, and the first ratio is cross-modified to obtain a second ratio; Based on the second ratio, the upper and lower limits of the value range, and the lower limits of the values ​​corresponding to the at least two hyperparameters, at least two hyperparameters after crossover mutation are determined, and the at least two hyperparameters after crossover mutation are used as at least one set of first hyperparameter combinations.

4. The method according to claim 1, characterized in that, The step of dynamically weighting and averaging the three optimal hyperparameter combinations based on their respective levels and pre-configured fitness ratios to update the hyperparameter combinations includes: Determine the level differences corresponding to the three optimal combinations of hyperparameters; The hyperparameter combination is updated by dynamically weighting the difference in level and the pre-configured fitness ratio coefficient. The pre-configured fitness ratio coefficient is the ratio of the three currently optimal hyperparameter combinations to the sum of the three target hyperparameter combinations whose errors reach the third preset error threshold.

5. The method according to claim 1, characterized in that, The pre-constructed fitness function is expressed by the formula: ,in, Represented as the true stress value, This is expressed as a predicted stress value.

6. A stress monitoring device, characterized in that, include: The model building module is used to build stress prediction models based on historical operational data. An initialization module is used to initialize at least two hyperparameters of the stress prediction model according to the chaotic mapping rule to obtain at least one set of hyperparameter combinations. The optimization module is used to optimize the at least one set of hyperparameter combinations according to the pre-constructed fitness function and the preset gray wolf algorithm, and determine the optimal hyperparameter combination among the at least one set of hyperparameter combinations; The monitoring module is used to take the optimal combination of hyperparameters as the optimal result of the stress prediction model, so that the optimal result of the stress prediction model is encapsulated into a data packet for stress monitoring. The initialization module includes: The first generation unit is used to randomly generate a set of at least two hyperparameters corresponding to the stress prediction model; wherein the hyperparameters include at least the kernel function width (g) and the penalty factor (c); The second generation unit is used to generate N sets of chaotic hyperparameter combinations by mapping the randomly generated set of hyperparameters using a Logistic mapping. The determining module is used to transfer the component carriers corresponding to the N sets of chaotic hyperparameter combinations to the preset value ranges corresponding to the at least two hyperparameters respectively, to obtain at least one set of hyperparameter combinations; wherein, each set of hyperparameter combinations includes at least two hyperparameters: the kernel function width (g) and the penalty factor (c); The optimization module includes: The first determining unit is used to determine, based on the pre-constructed fitness function, three sets of hyperparameter combinations in which the error between the true value of the first stress and the predicted value of the first stress reaches a first preset error threshold. The mutation unit is used to perform cross-mutation on at least two hyperparameters in the at least one set of hyperparameter combinations using a cross-operation method to obtain at least one set of first hyperparameter combinations after cross-mutation. The second determining unit is used to determine the second fitness value corresponding to the at least one set of first hyperparameter combinations based on the pre-constructed fitness function, and to determine the three sets of first hyperparameter combinations whose error between the second stress true value and the second stress predicted value reaches a second preset error threshold based on the second fitness value. The comparison unit is used to compare the three sets of hyperparameter combinations whose errors reach a first preset error threshold with the three sets of first hyperparameter combinations whose errors reach a second preset error threshold, and select the three sets of target hyperparameter combinations whose errors reach a third preset error threshold as the current optimal three sets of hyperparameter combinations. The iterative update unit is used to perform dynamic weighted averaging based on the levels corresponding to the three currently optimal hyperparameter combinations and the pre-configured fitness ratio coefficients to update the hyperparameter combinations. It returns the steps of the three hyperparameter combinations where the error between the first true stress value and the first predicted stress value reaches the first preset error threshold, as determined by the pre-constructed fitness function, and iteratively updates until the optimal hyperparameter combination is obtained.

7. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor to enable the at least one processor to perform the stress monitoring method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the stress monitoring method according to any one of claims 1-5.

Images