Combined cycle unit operation and maintenance system and method based on large language model

Abstract

The invention relates to the technical field of gas-steam combined cycle power generation, and particularly discloses a combined cycle unit operation and maintenance system and method based on a large language model, and the system comprises a screening module which is used for obtaining the overall operation parameters of a combined cycle unit, screening out key operation parameters of the unit according to the overall operation parameters of the combined cycle unit; the early warning module is used for predicting the change trend of the key operation parameters, determining a key operation parameter prediction value according to the change trend of the key operation parameters, and determining an early warning level according to the key operation parameter prediction value; the operation and maintenance module is used for determining a corresponding fault operation and maintenance strategy according to the key operation parameters when the early warning level exceeds a preset safety level; and the updating module is used for performing fault operation and maintenance on the combined cycle unit according to the fault operation and maintenance strategy, obtaining the operation and maintenance effect of the unit after the fault operation and maintenance, and updating the fault operation and maintenance strategy according to the operation and maintenance effect. Operation and maintenance of the unit are carried out based on the large language model, and accuracy and efficiency are improved.

Description

Technical Field

[0001] The present application relates to the technical field of gas-steam combined cycle power generation, and more specifically, to a combined cycle unit operation and maintenance system and method based on a large language model. Background Art

[0002] As the wave of digitalization and intelligence sweeps the world, the power industry is facing unprecedented transformation challenges. Traditional equipment monitoring and maintenance modes, especially alarm systems that rely on fixed thresholds and fault diagnosis based on expert experience, can no longer meet the needs of high-efficiency and high-reliability operation of modern power plants. These traditional methods usually issue alarms only after obvious abnormalities in the equipment, leaving a limited time window for operation and maintenance personnel to respond. This not only increases the risk of unplanned downtime, but may also result in significant economic losses and environmental impacts. For example, in gas-fired power plants, due to the high cost of fuel and the strict requirements of the power grid for continuous power supply, any unplanned downtime may have immeasurable consequences.

[0003] At the same time, traditional power plant fault diagnosis technology is limited by the limitations of the expert system rule base. The establishment of the rule base is highly dependent on the knowledge of experts in a specific field, which is not only time-consuming and laborious, but also often causes problems such as incomplete rule coverage and excessive complexity due to the subjectivity and limitations of expert experience, which seriously affects the efficiency and accuracy of fault diagnosis. Summary of the invention

[0004] The present invention provides a combined cycle unit operation and maintenance system and method based on a large language model, which is used to solve the problem of low efficiency and accuracy of combined cycle unit operation and maintenance in the prior art, including:

[0005] A screening module is used to obtain the overall operating parameters of the combined cycle unit and screen out the key operating parameters of the unit according to the overall operating parameters of the combined cycle unit;

[0006] An early warning module is used to predict the changing trend of key operating parameters, determine the predicted value of the key operating parameters according to the changing trend of the key operating parameters, and determine the early warning level according to the predicted value of the key operating parameters;

[0007] The operation and maintenance module is used to determine the corresponding fault operation and maintenance strategy according to key operating parameters when the warning level exceeds the preset safety level;

[0008] The update module is used to perform fault operation and maintenance on the combined cycle unit according to the fault operation and maintenance strategy, obtain the operation and maintenance effect of the unit after the fault operation and maintenance, and update the fault operation and maintenance strategy according to the operation and maintenance effect.

[0009] Furthermore, the screening module screens out key operating parameters of the unit according to the overall operating parameters of the combined cycle unit, including:

[0010] Determine the change trend of the overall operation parameters according to the overall operation parameters of the combined cycle unit, and draw the overall operation parameter change curve according to the change trend of the overall operation parameters;

[0011] Obtaining operating parameters of various components of the combined cycle unit, and drawing operating parameter variation curves according to the operating parameters of various components of the combined cycle unit;

[0012] Calculate the curve correlation coefficients of the overall operating parameter change curve and the component operating parameter change curve, and determine the influence of each component operating parameter according to the curve correlation coefficient;

[0013] A preset impact threshold is obtained, and component operating parameters whose impact is greater than the preset impact threshold are screened out to obtain key operating parameters.

[0014] Furthermore, the determining the influence degree of the operating parameters of each component according to the curve correlation coefficient includes:

[0015] The influence degree of the component operating parameters is determined according to the influence degree calculation formula, and the influence degree calculation formula is specifically as follows:

[0016]

[0017] Among them, P is the influence degree of component operation parameters, w is the preset importance weight, E max is the maximum value of the component operating parameter in the component operating parameter variation curve, E min is the minimum value of the component operating parameter in the component operating parameter variation curve, r is the curve correlation coefficient, r α is the preset allowable correlation coefficient, R is the preset range coefficient, and exp is the natural exponential function.

[0018] Furthermore, the early warning module determines the predicted value of the key operating parameter according to the change trend of the key operating parameter, including:

[0019] Obtaining historical operating parameters of each component of the combined cycle unit, and determining trend characteristic data of the historical operating parameters according to the historical operating parameters of each component;

[0020] An initial operating parameter prediction model is established based on a long short-term memory algorithm, and the initial operating parameter prediction model is trained according to trend characteristic data to obtain a trained operating parameter prediction model;

[0021] The key operating parameter trend characteristics are determined according to the changing trends of the key operating parameters, and the key operating parameter trend characteristics are input into the operating parameter prediction model to obtain the key operating parameter prediction values.

[0022] Furthermore, the warning module determines the warning level according to the predicted value of the key operating parameter, including:

[0023] Real-time monitoring of key operating parameters of the combined cycle unit, and calculation of the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter;

[0024] Determine whether the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than a first preset threshold value, and if the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than the first preset threshold value, set the preset third level as the warning level;

[0025] If the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is less than or equal to the first preset threshold, then determine whether the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than a second preset threshold;

[0026] If the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than the second preset threshold, the preset second level is set as the warning level;

[0027] If the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is less than or equal to the second preset threshold, the preset first level is set as the warning level.

[0028] Furthermore, the operation and maintenance module determines a corresponding fault operation and maintenance strategy according to key operation parameters, including:

[0029] Obtain the historical operating parameters of the combined cycle unit and the corresponding fault operation and maintenance strategy, obtain operation and maintenance cases, and establish an operation and maintenance knowledge base based on the fault operation and maintenance cases;

[0030] Calculating the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operation parameters, and if the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operation parameters is greater than a third preset threshold, generating a corresponding fault operation and maintenance strategy according to the operation and maintenance knowledge base;

[0031] If the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operation parameters is less than or equal to the third preset threshold, the prompt word text is set according to the key operation parameters, the prompt word text is input into the operation and maintenance strategy big language model, and the corresponding fault operation and maintenance strategy is output according to the operation and maintenance strategy big language model.

[0032] Furthermore, the calculating of the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operation parameters includes:

[0033] Obtain the operation parameters of each operation and maintenance case in the operation and maintenance knowledge base, calculate the absolute value of the difference between the key operation parameter and the corresponding operation parameter in each operation and maintenance case, and obtain the sum of the absolute values ​​of the difference in each operation and maintenance case;

[0034] Obtain the number of operating parameters of each operation and maintenance case in the operation and maintenance knowledge base, calculate the ratio of the number of key operating parameters to the number of operation and maintenance case operating parameters, and obtain the similarity weight;

[0035] The sum of the absolute values ​​of the differences in each operation and maintenance case is multiplied by the similarity weight value to obtain the similarity between the operation and maintenance case and the key operating parameters.

[0036] Furthermore, the step of setting the prompt word text according to the key operating parameters includes:

[0037] Determine the corresponding key operating components according to the key operating parameters of the combined cycle unit, and determine the historical failure frequency of each component according to the key operating components;

[0038] The fault types of key operating components whose historical fault frequencies are greater than a fourth preset threshold are screened out, and prompt word texts are set according to the fault types, warning levels and key operating parameters of each key operating component.

[0039] Furthermore, the updating module updates the fault operation and maintenance strategy according to the operation and maintenance effect, including:

[0040] Obtaining the overall operating parameters of the combined cycle unit after the fault operation and maintenance, calculating the difference between the overall operating parameters of the combined cycle unit after the fault operation and maintenance and the overall operating parameters of the combined cycle unit before the fault operation and maintenance, and obtaining the operation and maintenance effect evaluation parameters;

[0041] If the operation and maintenance effect evaluation parameter is greater than the fifth preset threshold, the fault operation and maintenance strategy and the corresponding key operation parameters are stored in the operation and maintenance knowledge base as a new operation and maintenance case;

[0042] If the operation and maintenance effect evaluation parameter is less than or equal to the fifth preset threshold, the operation and maintenance knowledge base is not updated.

[0043] In order to achieve the above object, the present invention also provides a combined cycle unit operation and maintenance method based on a large language model, comprising:

[0044] Obtain the overall operating parameters of the combined cycle unit, and screen out the key operating parameters of the unit according to the overall operating parameters of the combined cycle unit;

[0045] Predict the changing trend of key operating parameters, determine the predicted value of key operating parameters according to the changing trend of key operating parameters, and determine the warning level according to the predicted value of key operating parameters;

[0046] When the warning level exceeds the preset safety level, the corresponding fault operation and maintenance strategy is determined based on key operating parameters;

[0047] Perform fault operation and maintenance on the combined cycle unit according to the fault operation and maintenance strategy, obtain the operation and maintenance effect of the unit after the fault operation and maintenance, and update the fault operation and maintenance strategy according to the operation and maintenance effect.

[0048] The beneficial effects of the present invention are:

[0049] By applying the above technical solutions, the present invention combines deep learning and other technologies to perform trend warning diagnosis on the unit's data, discover possible signs of unit failure in advance, and achieve early warning of abnormal situations, effectively extending the preparation time window for operation and maintenance, avoiding losses caused by equipment damage or unplanned downtime, and accurately diagnosing the unit's fault data by establishing an operation and maintenance knowledge base and combining it with a large language model, thereby deriving corresponding operation and maintenance strategies, effectively improving the accuracy and efficiency of unit operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 A structural schematic diagram of a combined cycle unit operation and maintenance system based on a large language model proposed in an embodiment of the present invention is shown;

[0052] Figure 2 The overall flow chart of a combined cycle unit operation and maintenance method based on a large language model proposed in an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0054] The present application embodiment provides a combined cycle unit operation and maintenance system based on a large language model, such as Figure 1 As shown, including:

[0055] A screening module is used to obtain the overall operating parameters of the combined cycle unit and screen out the key operating parameters of the unit according to the overall operating parameters of the combined cycle unit; an early warning module is used to predict the changing trend of the key operating parameters, determine the predicted values ​​of the key operating parameters according to the changing trend of the key operating parameters, and determine the early warning level according to the predicted values ​​of the key operating parameters; an operation and maintenance module is used to determine the corresponding fault operation and maintenance strategy according to the key operating parameters when the early warning level exceeds the preset safety level; an update module is used to perform fault operation and maintenance on the combined cycle unit according to the fault operation and maintenance strategy, obtain the operation and maintenance effect of the unit after the fault operation and maintenance, and update the fault operation and maintenance strategy according to the operation and maintenance effect.

[0056] In this embodiment, the overall operating parameters of the combined cycle unit can be the power generation parameters of the unit. The key operating parameters of the unit are screened out through the screening module, and the early warning module issues early warnings for the changing trends of the key operating parameters. The operation and maintenance module generates a fault operation and maintenance strategy for the key operating parameters when the early warning level is too high. The fault operation and maintenance strategy is updated through the update module to ensure the safe operation of the unit.

[0057] In some embodiments of the present application, the screening module screens out key operating parameters of the unit based on the overall operating parameters of the combined cycle unit, including: determining the changing trend of the overall operating parameters based on the overall operating parameters of the combined cycle unit, and drawing an overall operating parameter change curve based on the changing trend of the overall operating parameters; obtaining operating parameters of each component of the combined cycle unit, and drawing an operating parameter change curve based on the operating parameters of each component of the combined cycle unit; calculating the curve correlation coefficient of the overall operating parameter change curve and the component operating parameter change curve, and determining the influence degree of each component operating parameter based on the curve correlation coefficient; obtaining a preset influence degree threshold, screening out component operating parameters whose influence degree is greater than the preset influence degree threshold, and obtaining key operating parameters.

[0058] In some embodiments of the present application, determining the influence degree of each component operating parameter according to the curve correlation coefficient includes: determining the influence degree of the component operating parameter according to an influence degree calculation formula, and the influence degree calculation formula is specifically:

[0059]

[0060] Among them, P is the influence degree of component operation parameters, w is the preset importance weight, E max is the maximum value of the component operating parameter in the component operating parameter variation curve, E min is the minimum value of the component operating parameter in the component operating parameter variation curve, r is the curve correlation coefficient, r α is the preset allowable correlation coefficient, R is the preset range coefficient, and exp is the natural exponential function.

[0061] In this embodiment, the power generation parameter change curve is drawn by real-time monitoring of the power generation of the unit. The operating parameters of each component of the unit may be parameters such as combustion stability, exhaust pressure, and cooling water temperature. The change curve of the operating parameters of each component is drawn separately, and the curve correlation coefficient between the power generation parameter change curve and the change curve of the operating parameters of each component is determined according to the Pearson correlation coefficient algorithm. Thus, the key operating parameters are screened out by calculating the influence of the operating parameters of each component. The key operating parameters are parameters whose influence on the power generation parameters is greater than a preset influence threshold.

[0062] In some embodiments of the present application, the early warning module determines the predicted value of the key operating parameter according to the changing trend of the key operating parameter, including: obtaining the historical operating parameters of each component of the combined cycle unit, and determining the trend characteristic data of the historical operating parameters according to the historical operating parameters of each component; establishing an initial operating parameter prediction model based on the long short-term memory algorithm, and training the initial operating parameter prediction model according to the trend characteristic data to obtain a trained operating parameter prediction model; determining the key operating parameter trend characteristics according to the changing trend of the key operating parameters, and inputting the key operating parameter trend characteristics into the operating parameter prediction model to obtain the predicted value of the key operating parameter.

[0063] In this embodiment, according to the Long Short-Term Memory (LSTM) algorithm, a single-layer LSTM structure is adopted to establish an initial operating parameter prediction model. The initial operating parameter prediction model is trained by extracting trend feature data of historical operating parameters at preset time steps to obtain a trained operating parameter prediction model, and key operating parameters are predicted based on the trained operating parameter prediction model to obtain key operating parameter prediction values.

[0064] In some embodiments of the present application, the early warning module determines the early warning level according to the predicted value of the key operating parameter, including: real-time monitoring of the key operating parameters of the combined cycle unit, calculating the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter; judging whether the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than a first preset threshold value, if the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than the first preset threshold value, setting the preset third level as the early warning level; if the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is less than or equal to the first preset threshold value, judging whether the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than the second preset threshold value; if the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than the second preset threshold value, setting the preset second level as the early warning level; if the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is less than or equal to the second preset threshold value, setting the preset first level as the early warning level.

[0065] In this embodiment, the warning level is determined by the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter. The higher the absolute value of the difference, the greater the corresponding warning level. This embodiment sends a warning signal when the warning level is greater than level one, thereby achieving early warning of abnormal situations.

[0066] In some embodiments of the present application, the operation and maintenance module determines the corresponding fault operation and maintenance strategy based on the key operating parameters, including: obtaining the historical operating parameters of the combined cycle unit and the corresponding fault operation and maintenance strategy, obtaining operation and maintenance cases, and establishing an operation and maintenance knowledge base based on the fault operation and maintenance cases; calculating the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operating parameters, if the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operating parameters is greater than a third preset threshold, generating the corresponding fault operation and maintenance strategy based on the operation and maintenance knowledge base; if the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operating parameters is less than or equal to the third preset threshold, setting a prompt word text based on the key operating parameters, inputting the prompt word text into the operation and maintenance strategy large language model, and outputting the corresponding fault operation and maintenance strategy based on the operation and maintenance strategy large language model.

[0067] In this embodiment, an operation and maintenance knowledge base is constructed to store historical operation and maintenance cases, and corresponding fault operation and maintenance strategies are generated by calculating the similarity between the operation and maintenance cases and the key operating parameters to improve the operation and maintenance efficiency of the unit. When the similarity is less than or equal to the third preset threshold, a prompt word text is constructed through the key operating parameters, and the prompt word text is analyzed based on the large language model to generate a recommended fault operation and maintenance strategy.

[0068] In some embodiments of the present application, the calculation of the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operation parameters includes: obtaining the operation parameters of each operation and maintenance case in the operation and maintenance knowledge base, calculating the absolute value of the difference between the key operation parameters and the corresponding operation parameters in each operation and maintenance case, and obtaining the sum of the absolute values ​​of the differences in each operation and maintenance case; obtaining the number of operation parameters of each operation and maintenance case in the operation and maintenance knowledge base, calculating the ratio of the number of key operation parameters to the number of operation parameters of the operation and maintenance case, and obtaining the similarity weight; multiplying the sum of the absolute values ​​of the differences in each operation and maintenance case by the similarity weight value to obtain the similarity between the operation and maintenance case and the key operation parameters.

[0069] In this embodiment, the initial similarity is obtained by calculating the sum of the absolute values ​​of the differences between the key operating parameters and the corresponding operating parameters in each operation and maintenance case, and the similarity weight is calculated by counting the ratio of the number of key operating parameters to the number of abnormal operating parameters in the operation and maintenance case. The similarity between the operation and maintenance case and the key operating parameters is obtained based on the similarity weight and the initial similarity.

[0070] In some embodiments of the present application, the prompt word text is set according to the key operating parameters, including: determining the corresponding key operating components according to the key operating parameters of the combined cycle unit, and determining the historical failure frequency of each component according to the key operating components; screening out the failure type of the key operating components whose historical failure frequency is greater than a fourth preset threshold, and setting the prompt word text according to the failure type, warning level and key operating parameters of each key operating component.

[0071] In this embodiment, by calculating the historical failure types of key operating components whose historical failure frequencies are greater than the fourth preset threshold, the current warning level and the current key operating parameters, a prompt word text is constructed, and the prompt word text is input into the operation and maintenance strategy language model, and then the recommended fault operation and maintenance strategy is output.

[0072] In some embodiments of the present application, the update module updates the fault operation and maintenance strategy according to the operation and maintenance effect, including: obtaining the overall operating parameters of the combined cycle unit after the fault operation and maintenance, calculating the difference between the overall operating parameters of the combined cycle unit after the fault operation and maintenance and the overall operating parameters of the combined cycle unit before the fault operation and maintenance, and obtaining the operation and maintenance effect evaluation parameter; if the operation and maintenance effect evaluation parameter is greater than the fifth preset threshold, the fault operation and maintenance strategy and the corresponding key operating parameters are stored as a new operation and maintenance case in the operation and maintenance knowledge base; if the operation and maintenance effect evaluation parameter is less than or equal to the fifth preset threshold, the operation and maintenance knowledge base is not updated.

[0073] In this embodiment, the operation and maintenance effect is evaluated based on the difference between the power generation power of the combined cycle unit after the fault operation and maintenance and the power generation power of the combined cycle unit before the fault operation and maintenance, and the fault operation and maintenance strategy and key operating parameters corresponding to the operation and maintenance effect evaluation parameter greater than the fifth preset threshold are stored in the operation and maintenance knowledge base to realize the update of the operation and maintenance knowledge base.

[0074] Based on the same technical concept, such as Figure 2 As shown, the present invention also provides a combined cycle unit operation and maintenance method based on a large language model, comprising:

[0075] S101, obtaining overall operating parameters of the combined cycle unit, and screening out key operating parameters of the unit according to the overall operating parameters of the combined cycle unit;

[0076] S102, predicting the changing trend of key operating parameters, determining the predicted value of the key operating parameters according to the changing trend of the key operating parameters, and determining the warning level according to the predicted value of the key operating parameters;

[0077] S103, when the warning level exceeds the preset safety level, determine the corresponding fault operation and maintenance strategy according to the key operation parameters;

[0078] S104, performing fault operation and maintenance on the combined cycle unit according to the fault operation and maintenance strategy, obtaining the operation and maintenance effect of the unit after the fault operation and maintenance, and updating the fault operation and maintenance strategy according to the operation and maintenance effect.

[0079] By applying the above technical solutions, the present invention uses a screening module to obtain the overall operating parameters of the combined cycle unit, and screens out the key operating parameters of the unit according to the overall operating parameters of the combined cycle unit; an early warning module is used to predict the change trend of the key operating parameters, determine the predicted value of the key operating parameters according to the change trend of the key operating parameters, and determine the early warning level according to the predicted value of the key operating parameters; an operation and maintenance module is used to determine the corresponding fault operation and maintenance strategy according to the key operating parameters when the early warning level exceeds the preset safety level; an update module is used to perform fault operation and maintenance on the combined cycle unit according to the fault operation and maintenance strategy, obtain the operation and maintenance effect of the unit after the fault operation and maintenance, and update the fault operation and maintenance strategy according to the operation and maintenance effect. The operation and maintenance of the unit based on the large language model improves accuracy and efficiency.

[0080] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present invention can be implemented by hardware, or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present invention.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A combined cycle unit operation and maintenance system based on a large language model, characterized in that: include: A screening module is used to obtain the overall operating parameters of the combined cycle unit and screen out the key operating parameters of the unit according to the overall operating parameters of the combined cycle unit; An early warning module is used to predict the changing trend of key operating parameters, determine the predicted value of the key operating parameters according to the changing trend of the key operating parameters, and determine the early warning level according to the predicted value of the key operating parameters; The operation and maintenance module is used to determine the corresponding fault operation and maintenance strategy according to key operating parameters when the warning level exceeds the preset safety level; The update module is used to perform fault operation and maintenance on the combined cycle unit according to the fault operation and maintenance strategy, obtain the operation and maintenance effect of the unit after the fault operation and maintenance, and update the fault operation and maintenance strategy according to the operation and maintenance effect.

2. The combined cycle unit operation and maintenance system based on a large language model according to claim 1, characterized in that: The screening module screens out the key operating parameters of the unit according to the overall operating parameters of the combined cycle unit, including: Determine the change trend of the overall operation parameters according to the overall operation parameters of the combined cycle unit, and draw the overall operation parameter change curve according to the change trend of the overall operation parameters; Obtaining operating parameters of various components of the combined cycle unit, and drawing operating parameter variation curves according to the operating parameters of various components of the combined cycle unit; Calculate the curve correlation coefficients of the overall operating parameter change curve and the component operating parameter change curve, and determine the influence of each component operating parameter according to the curve correlation coefficient; A preset impact threshold is obtained, and component operating parameters whose impact is greater than the preset impact threshold are screened out to obtain key operating parameters.

3. The combined cycle unit operation and maintenance system based on a large language model according to claim 2, characterized in that: Determining the influence degree of the operating parameters of each component according to the curve correlation coefficient includes: The influence degree of the component operating parameters is determined according to the influence degree calculation formula, and the influence degree calculation formula is specifically as follows: Among them, P is the influence degree of component operation parameters, w is the preset importance weight, E max is the maximum value of the component operating parameter in the component operating parameter variation curve, E min is the minimum value of the component operating parameter in the component operating parameter variation curve, r is the curve correlation coefficient, r α is the preset allowable correlation coefficient, R is the preset range coefficient, and exp is the natural exponential function.

4. The combined cycle unit operation and maintenance system based on a large language model according to claim 1, characterized in that: The early warning module determines the predicted value of the key operating parameter according to the change trend of the key operating parameter, including: Obtaining historical operating parameters of each component of the combined cycle unit, and determining trend characteristic data of the historical operating parameters according to the historical operating parameters of each component; An initial operating parameter prediction model is established based on a long short-term memory algorithm, and the initial operating parameter prediction model is trained according to trend characteristic data to obtain a trained operating parameter prediction model; The key operating parameter trend characteristics are determined according to the changing trends of the key operating parameters, and the key operating parameter trend characteristics are input into the operating parameter prediction model to obtain the key operating parameter prediction values.

5. The combined cycle unit operation and maintenance system based on a large language model according to claim 4, characterized in that: The warning module determines the warning level according to the predicted value of the key operating parameter, including: Real-time monitoring of key operating parameters of combined cycle units, and calculation of the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter; Determine whether the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than a first preset threshold value, and if the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than the first preset threshold value, set the preset third level as the warning level; If the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is less than or equal to the first preset threshold, then determine whether the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than a second preset threshold; If the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is greater than the second preset threshold, the preset second level is set as the warning level; If the absolute value of the difference between the real-time value of the key operating parameter and the corresponding predicted value of the key operating parameter is less than or equal to the second preset threshold, the preset first level is set as the warning level.

6. The combined cycle unit operation and maintenance system based on a large language model according to claim 1, characterized in that: The operation and maintenance module determines the corresponding fault operation and maintenance strategy according to the key operation parameters, including: Obtain the historical operating parameters of the combined cycle unit and the corresponding fault operation and maintenance strategy, obtain operation and maintenance cases, and establish an operation and maintenance knowledge base based on the fault operation and maintenance cases; Calculating the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operation parameters, and if the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operation parameters is greater than a third preset threshold, generating a corresponding fault operation and maintenance strategy according to the operation and maintenance knowledge base; If the similarity between the operation and maintenance cases in the operation and maintenance knowledge base and the key operation parameters is less than or equal to the third preset threshold, the prompt word text is set according to the key operation parameters, the prompt word text is input into the operation and maintenance strategy big language model, and the corresponding fault operation and maintenance strategy is output according to the operation and maintenance strategy big language model.

7. The combined cycle unit operation and maintenance system based on a large language model according to claim 6, characterized in that: The similarity between the operation and maintenance cases and the key operation parameters in the operation and maintenance knowledge base is calculated, including: Obtain the operation parameters of each operation and maintenance case in the operation and maintenance knowledge base, calculate the absolute value of the difference between the key operation parameter and the corresponding operation parameter in each operation and maintenance case, and obtain the sum of the absolute values ​​of the difference in each operation and maintenance case; Obtain the number of operating parameters of each operation and maintenance case in the operation and maintenance knowledge base, calculate the ratio of the number of key operating parameters to the number of operation and maintenance case operating parameters, and obtain the similarity weight; The sum of the absolute values ​​of the differences in each operation and maintenance case is multiplied by the similarity weight value to obtain the similarity between the operation and maintenance case and the key operating parameters.

8. The combined cycle unit operation and maintenance system based on a large language model according to claim 6, characterized in that: The step of setting the prompt word text according to the key operating parameters includes: Determine the corresponding key operating components according to the key operating parameters of the combined cycle unit, and determine the historical failure frequency of each component according to the key operating components; The fault types of key operating components whose historical fault frequencies are greater than a fourth preset threshold are screened out, and prompt word texts are set according to the fault types, warning levels and key operating parameters of each key operating component.

9. The combined cycle unit operation and maintenance system based on a large language model according to claim 6, characterized in that: The updating module updates the fault operation and maintenance strategy according to the operation and maintenance effect, including: Obtaining the overall operating parameters of the combined cycle unit after the fault operation and maintenance, calculating the difference between the overall operating parameters of the combined cycle unit after the fault operation and maintenance and the overall operating parameters of the combined cycle unit before the fault operation and maintenance, and obtaining the operation and maintenance effect evaluation parameters; If the operation and maintenance effect evaluation parameter is greater than the fifth preset threshold, the fault operation and maintenance strategy and the corresponding key operation parameters are stored in the operation and maintenance knowledge base as a new operation and maintenance case; If the operation and maintenance effect evaluation parameter is less than or equal to the fifth preset threshold, the operation and maintenance knowledge base is not updated.

10. A combined cycle unit operation and maintenance method based on a large language model, characterized in that: include: Obtain the overall operating parameters of the combined cycle unit, and screen out the key operating parameters of the unit according to the overall operating parameters of the combined cycle unit; Predict the changing trend of key operating parameters, determine the predicted value of key operating parameters according to the changing trend of key operating parameters, and determine the warning level according to the predicted value of key operating parameters; When the warning level exceeds the preset safety level, the corresponding fault operation and maintenance strategy is determined based on key operating parameters; Perform fault operation and maintenance on the combined cycle unit according to the fault operation and maintenance strategy, obtain the operation and maintenance effect of the unit after the fault operation and maintenance, and update the fault operation and maintenance strategy according to the operation and maintenance effect.

Images