Unit working condition-parameter modeling method and system

Abstract

The invention belongs to the technical field of parameter optimization, and discloses a unit working condition-parameter modeling method and system. The method comprises the steps of constructing an optimal working condition table; collecting unit working condition data during operation of the unit, and marking the unit working condition data as real-time working condition data; according to the real-time working condition data, matching optimal operating parameters from the optimal working condition table; judging whether a trust instruction is generated or not according to the optimal operation parameters; if the trust instruction is generated, recommending a matched optimal operation parameter; if the trust instruction is not generated, weighting calculation is carried out on all the optimal operation parameters in the optimal working condition table, and the optimal operation parameters are obtained again and recommended; according to the method, the most similar unit working condition data can be matched from the optimal working condition table, so that the matching speed and the matching precision are improved; and intelligent recommendation of the optimal operation parameters is realized, so that the overall operation efficiency and the quality index of the unit are improved, and the adaptability, the robustness and the safety of the system are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of parameter optimization, and more specifically, to a unit operating condition-parameter modeling method and system. Background Art

[0002] Traditional thermal power unit operation optimization usually relies on historical operation data to generate optimal operating condition tables for recommending optimal operating parameters. This data-driven modeling method can better summarize the overall operating characteristics of the unit. However, in practical applications, the existing methods still have some problems:

[0003] 1. Due to the complexity and variability of the actual operating conditions of the unit, the similarity between the current operating conditions and the records in the optimal operating conditions table may be low, resulting in reduced rationality and accuracy of the recommended parameters.

[0004] 2. The optimal operating condition table cannot be updated adaptively, making it difficult to keep up with changes in the actual production environment, reducing the flexibility of the application.

[0005] A Chinese patent with publication number CN117873844A discloses a dynamic optimization operation monitoring method and system based on power plant energy efficiency analysis; including reconstructing the equipment status of the thermal system based on the historical operation data of the power plant or the historical database of equipment parameters; dividing the operating conditions, finding the optimal operating conditions under each operating condition of the unit, and establishing an initial optimal operating condition database with the minimum coal consumption rate of the unit as the target; comparing the real-time operating conditions and operating conditions with the operating conditions with the same or similar boundary conditions in the optimal operating condition database, and storing the comparison results in the optimal operating condition database; automatically issuing corresponding alarm prompt information for the operating data or equipment parameters in non-optimal operating conditions under the same or similar operating conditions, or automatically adjusting the unit operating status according to the various operating data or equipment parameters in the optimal operating conditions corresponding to the same or similar operating conditions in the optimal operating condition database to achieve the optimal operating condition.

[0006] Although the above technology can dynamically update the optimal operating condition database, it can only automatically adjust the unit operating status based on the various operating data or equipment parameters in the optimal operating condition corresponding to the same or similar operating conditions in the optimal operating condition database to achieve the optimal operating condition; for the operating conditions not stored in the optimal operating condition database, only an alarm prompt can be issued, and the optimal operating data cannot be automatically matched, which results in the unit being unable to reach the optimal operating state and even causing potential safety hazards or unit failure risks.

[0007] In view of this, the present invention proposes a unit operating condition-parameter modeling method and system to solve the above problems. Summary of the invention

[0008] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purpose, the present invention provides the following technical solutions: a unit operating condition-parameter modeling method, comprising:

[0009] Constructing an optimal operating condition table, which includes unit operating condition data and optimal operating parameters corresponding to the unit operating condition data;

[0010] Collect the unit operating data during operation and mark it as real-time operating data;

[0011] According to the real-time operating data, the best operating parameters are matched from the best operating table;

[0012] Determine whether to generate a trust instruction based on the optimal operating parameters;

[0013] If a trust instruction is generated, the matching optimal operating parameters are recommended; if no trust instruction is generated, all the optimal operating parameters in the optimal operating condition table are weightedly calculated, the optimal operating parameters are re-obtained and recommended.

[0014] Furthermore, the unit operating data includes power load and ambient temperature; the optimal operating parameters are operating parameters corresponding to the optimal operating efficiency and quality indicators of the unit under different unit operating data; the operating parameters include main steam temperature, main steam pressure, main steam flow, reheat steam temperature, feed water temperature and circulating water temperature;

[0015] The method for constructing the optimal operating condition table is as follows: collect historical unit operating condition data and the optimal operating parameters corresponding to the historical unit operating condition data, use multi-parameter steady-state screening method and fuzzy C-means clustering method to process the historical unit operating condition data, and construct the optimal operating condition table based on the processed historical unit operating condition data and the corresponding optimal operating parameters.

[0016] Furthermore, the improved salp swarm algorithm is used to match the best operating parameters from the best operating condition table, and the steps include:

[0017] Step 1: Preset the population size Z and the number threshold G;

[0018] Step 2: Generate an initialization population. The positions of individuals in the initialization population are defined in a one-dimensional search space. The positions of individuals correspond to the working condition labels one by one. The number of iterations T of the initialization population is 0.

[0019] Step 3: Determine the fitness function;

[0020] Step 4: Calculate the fitness of each individual in the initialized population, and take the position of the individual with the maximum fitness as the food source position;

[0021] Step 5: Update the leader position and follower position;

[0022] Step 6: Calculate the dynamic boundary, generate the reverse solution corresponding to each individual and add it to the population;

[0023] Step 7: Update the population, obtain a new population, and set T = T + 1;

[0024] Step 8: Loop steps 5 to 7 until the number of iterations corresponding to the new population is equal to the number threshold G. The loop ends, and the operating condition label corresponding to the food source position in the new population is obtained. The corresponding unit operating condition data is obtained according to the obtained operating condition label. The corresponding optimal operating parameters are matched from the optimal operating condition table according to the obtained unit operating condition data.

[0025] Furthermore, in step 2, the population S0 is initialized to {X1, X2, ..., X Z}, X Z is the Zth individual, including Z individuals, of which the first individual is the leader and the remaining Z-1 individuals are followers; different digital labels are set for different unit operating condition data in the optimal operating condition table and marked as operating condition labels, and the operating condition labels correspond to the unit operating condition data one by one; the range of the operating condition label and the range of the one-dimensional search space are both [1, A];

[0026] In step 3, the fitness function is expressed as: f=XS; where f is the fitness and XS is the similarity; the method for obtaining the similarity includes:

[0027] Obtain the power load in the optimal operating condition table, calculate the mean and standard deviation of the power load in the optimal operating condition table, the mean is marked as the load mean, and the standard deviation is marked as the load standard deviation; obtain the ambient temperature in the optimal operating condition table, calculate the mean and standard deviation of the ambient temperature in the optimal operating condition table, the mean is marked as the temperature mean, and the standard deviation is marked as the temperature standard deviation;

[0028] The unit operating condition data corresponding to the operating condition label corresponding to the individual position is obtained, and marked as comparison operating condition data; the comparison operating condition data is standardized to obtain comparison standard data; the method for obtaining comparison standard data is: subtract the load mean from the power load in the comparison operating condition data, and then divide it by the load standard deviation to obtain the standardized power load; subtract the temperature mean from the ambient temperature in the comparison operating condition data, and then divide it by the temperature standard deviation to obtain the standardized ambient temperature; the standardized power load and the standardized ambient temperature are the comparison standard data; the real-time operating condition data is standardized and marked as real-time standard data; the method for obtaining real-time standard data is consistent with the method for obtaining comparison standard data; the Euler distance between the comparison standard data and the real-time standard data is calculated, and the reciprocal of the Euler distance is used as the similarity.

[0029] Furthermore, in step 5, the method for updating the leader position includes:

[0030] Sort the fitness of all individuals in the population from large to small, and place the individual position corresponding to the fitness next to the fitness corresponding to the food source position as the second position, and the individual position corresponding to the last fitness as the last position;

[0031]

[0032] Where X1 is the leader position, F is the food source position, d is the number of iterations, the number of iterations d is the degree of freedom parameter of the t distribution, t(d) is a random value in the t distribution with the number of iterations d, X′ is the second position, X″ is the last position, C1 is a random number in [-0.5,0.5], and C2 is a random number in [0,1];

[0033] Methods for updating follower locations include:

[0034]

[0035] Where, X v is the position of the vth individual, X v-1 is the position of the v-1th individual, v∈[2,Z].

[0036] Furthermore, in step 6, the dynamic boundary is [R1, R2], R1 is the operating condition label with the smallest value corresponding to the individual position in the population, R2 is the operating condition label with the largest value corresponding to the individual position in the population, 1≤R1<R2≤A;

[0037] The methods for generating the inverse solution include:

[0038]

[0039] In the formula, is the reverse solution of the ith individual, C3 is a random number in [0,1], i∈[1,Z], X i is the position of the i-th individual;

[0040] If the generated reverse solution is greater than or less than the dynamic boundary, it is marked as a transcendental solution, a random number function is used to randomly generate a value within the dynamic boundary, and the randomly generated value is assigned to the transcendental solution;

[0041] In step 7, the reverse solution of the individual is also taken as an individual, the fitness of all individuals in the population is calculated, the fitness of all individuals in the population is sorted from large to small, the first Z individuals are retained and taken as the new population, and the position of the individual with the largest fitness in the new population is taken as the new food source position.

[0042] Further, the method for determining whether to generate a trust instruction includes:

[0043] Input the real-time working condition data into the trained threshold prediction model, obtain the corresponding similarity threshold, and mark it as the real-time threshold; the training process of the threshold prediction model includes:

[0044] Pre-collect a group of real-time working condition data, set corresponding similarity thresholds for the a group of real-time working condition data, a is an integer greater than 1, and convert the real-time working condition data and the corresponding similarity thresholds into a corresponding set of feature vectors; obtain the similarity threshold corresponding to each group of real-time working condition data; and sequentially set corresponding similarity thresholds for the a group of real-time working condition data;

[0045] Each group of feature vectors is used as the input of the threshold prediction model. The threshold prediction model uses a group of predicted similarity thresholds corresponding to each group of real-time operating condition data as output, and uses the actual similarity threshold corresponding to each group of real-time operating condition data as the prediction target, where the actual similarity threshold is the pre-collected similarity threshold corresponding to the real-time operating condition data; minimizing the sum of prediction errors of all real-time operating condition data is used as the training target; the threshold prediction model is trained until the sum of prediction errors reaches convergence and the training is stopped; the threshold prediction model is a deep neural network model;

[0046] Compare the similarity corresponding to the obtained working condition label with the real-time threshold;

[0047] If the similarity is less than the real-time threshold, a trust instruction is generated;

[0048] If the similarity is greater than or equal to the real-time threshold, no trust instruction is generated.

[0049] Furthermore, the method for re-obtaining the optimal operating parameters includes:

[0050] Add the similarities corresponding to each group of unit operating condition data in the optimal operating condition table in turn to obtain the total similarity; multiply each power load in the optimal operating condition table by the corresponding similarity, and then add them in turn to obtain the total load value; multiply each ambient temperature in the optimal operating condition table by the corresponding similarity, and then add them in turn to obtain the total temperature value; divide the total load value by the sum of similarities to obtain the optimal power load; divide the total temperature value by the sum of similarities to obtain the optimal ambient temperature; use the optimal power load and the optimal ambient temperature as the re-acquired optimal operating parameters; and add the re-acquired optimal operating parameters and the corresponding real-time operating condition data into the optimal operating condition table.

[0051] Further, the credibility of the re-acquired optimal operating parameters is calculated, and it is determined whether to perform parameter optimization processing;

[0052] The optimal operating parameters retrieved are marked as analysis parameters; the analysis parameters and the corresponding real-time operating condition data are used as analysis data, and the analysis data is input into the trained trust analysis model to predict the corresponding credibility; the training process of the trust analysis model is consistent with the training process of the threshold prediction model, and both are deep neural network models;

[0053] Preset a credibility threshold and compare the predicted credibility with the credibility threshold;

[0054] If the predicted confidence is greater than or equal to the confidence threshold, there is no need to perform parameter optimization on the analysis parameters;

[0055] If the predicted credibility is less than the credibility threshold, the analysis parameters need to be optimized;

[0056] The parameter optimization processing methods include:

[0057] Obtain a parameter range, which includes the range corresponding to each parameter in the operating parameters; randomly combine the values ​​in each range within the parameter range to construct m parameter sets, wherein each parameter set includes a value in each range within the parameter range, and the m parameter sets are all different; set different digital labels for the m parameter sets and mark them as parameter labels, and the range of the parameter labels is [1, m]; according to the real-time operating condition data, the improved salp swarm algorithm is used again to screen out an optimal parameter set label from the m parameter set labels, and the parameter set corresponding to the screened optimal parameter set label is used as the parameter optimization processing result of the optimal operating parameters;

[0058] The step of selecting an optimal parameter set label from m parameter set labels is consistent with the step of matching the optimal operating parameters from the optimal operating condition table; the difference is that in the step of selecting an optimal parameter set label from m parameter set labels, the expression of the fitness function is: f′=KX; where f′ is fitness and KX is credibility.

[0059] A unit operating condition-parameter modeling system, implementing the unit operating condition-parameter modeling method, comprising:

[0060] The operating condition modeling module is used to construct an optimal operating condition table, which includes unit operating condition data and optimal operating parameters corresponding to the unit operating condition data;

[0061] The data acquisition module is used to collect the unit operating data during the operation of the unit and mark it as real-time operating data;

[0062] The working condition matching module is used to match the best operating parameters from the best working condition table according to the real-time working condition data;

[0063] A matching judgment module, used to judge whether to generate a trust instruction based on the optimal operating parameters;

[0064] The parameter recommendation module is used to recommend the matching optimal operating parameters if a trust instruction is generated; if no trust instruction is generated, weighted calculation is performed on all the optimal operating parameters in the optimal operating condition table, and the optimal operating parameters are re-obtained and recommended.

[0065] The technical effects and advantages of a unit operating condition-parameter modeling method and system of the present invention are as follows:

[0066] 1. By constructing an optimal operating condition table and using the improved salp swarm algorithm, the unit operating condition data with the greatest similarity to the real-time operating condition data can be accurately matched from the optimal operating condition table; even when the operating conditions are complex and changeable, the most similar unit operating condition data can still be matched from the optimal operating condition table, thereby improving the matching speed and matching accuracy; and setting up a trust instruction generation mechanism. When the similarity between the real-time operating condition data and the matched unit operating condition data is low, the similarity of each group of unit operating condition data in the optimal operating condition table is used as a weight for comprehensive weighted calculation, so that the optimal operating parameters can be automatically re-recommended; avoid the situation where the optimal operating parameters cannot be recommended when the real-time operating condition data is not stored in the optimal operating condition table; realize the intelligent recommendation of the optimal operating parameters, thereby improving the overall operating efficiency and quality indicators of the unit, and thereby enhancing the adaptability, robustness and safety of the system.

[0067] 2. By introducing the credibility analysis mechanism, the credibility corresponding to the re-acquired optimal operating parameters is calculated; and through threshold comparison, it is determined whether it is necessary to optimize the parameters of the re-acquired optimal operating parameters. If necessary, the parameter combination label matching method is adopted, and the improved salp swarm algorithm is combined again to accurately screen the best parameter set from multiple candidate parameter sets for matching to obtain the optimized optimal operating parameters; it can correct the calculation error of the optimal operating parameters, and realize the quantitative evaluation of the optimization process through the machine learning model, so as to ensure the recommendation quality of the optimal operating parameters; at the same time, it also deepens the optimization degree of the system algorithm to further realize the efficient and safe operation of the unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A schematic diagram of a unit operating condition-parameter modeling system according to Embodiment 1 of the present invention;

[0069] Figure 2 A schematic diagram of a unit operating condition-parameter modeling system according to Embodiment 2 of the present invention;

[0070] Figure 3 This is a schematic diagram of the optimal operating parameter recommendation interface of Example 2 of the present invention;

[0071] Figure 4 This is a flow chart of a unit operating condition-parameter modeling method according to Embodiment 3 of the present invention;

[0072] Figure 5 This is a schematic diagram of an electronic device according to Embodiment 4 of the present invention;

[0073] Figure 6 This is a schematic diagram of a storage medium according to Embodiment 5 of the present invention. DETAILED DESCRIPTION

[0074] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0075] Example 1

[0076] See also Figure 1 As shown, a unit operating condition-parameter modeling system described in this embodiment includes an operating condition modeling module, a data acquisition module, an operating condition matching module, a matching judgment module and a parameter recommendation module; each module is connected by wire and / or wireless means to realize data transmission between modules;

[0077] The operating condition modeling module is used to construct an optimal operating condition table, which includes unit operating condition data and optimal operating parameters corresponding to the unit operating condition data.

[0078] The optimal operating condition table is a table that records the optimal operating parameters corresponding to different units under different unit operating data.

[0079] The unit operating data includes power load and ambient temperature; power load is the generating power of the unit, and the unit is a thermal power unit; ambient temperature is the temperature of the environment in which the unit is located; the unit operating data is a key factor reflecting the operating status of the unit, and therefore directly affects the various operating parameters of the unit.

[0080] The optimal operating parameters are the operating parameters corresponding to the best operating efficiency and quality indicators under different unit operating conditions; the operating parameters include main steam temperature, main steam pressure, main steam flow, reheat steam temperature, feed water temperature and circulating water temperature; the main steam temperature is the temperature of the high-temperature and high-pressure main steam output by the boiler in the unit; the main steam pressure is the pressure of the high-temperature and high-pressure main steam output by the boiler in the unit, the main steam flow is the flow of the high-temperature and high-pressure main steam output by the boiler in the unit, the reheat steam temperature is the temperature of the medium-pressure reheat steam heated again after the high pressure of the turbine in the unit, the feed water temperature is the temperature of the feed water of the boiler in the unit, and the circulating water temperature is the temperature of the circulating water for cooling the turbine in the unit; the operating parameters reflect the thermodynamic performance of the unit and are the key factors affecting the power generation efficiency and safe operation of the unit. Therefore, reasonable control of the operating parameters is crucial to improving the economy and reliability of the unit.

[0081] The method for constructing the optimal operating condition table is as follows: collect historical unit operating condition data and the optimal operating parameters corresponding to the historical unit operating condition data, use multi-parameter steady-state screening method and fuzzy C-means clustering method to process the historical unit operating condition data, and construct the optimal operating condition table based on the processed historical unit operating condition data and the corresponding optimal operating parameters.

[0082] The historical unit operating data and the optimal operating parameters corresponding to the historical unit operating data are collected and obtained by technical personnel in this field during the historical unit operation process; the multi-parameter steady-state screening method and the fuzzy C-means clustering method are both existing technologies and will not be described in detail here; the multi-parameter steady-state screening method can eliminate non-steady-state data in the historical unit operating data to ensure the steady-state nature of the data; the fuzzy C-means clustering method can cluster the historical unit operating data, laying the foundation for constructing an optimal operating table that can cover a wider range of operating conditions, realizing the unification of clustering and connection, and simplifying the mapping relationship of parameters.

[0083] The data acquisition module is used to collect the unit operating data during unit operation and mark it as real-time operating data.

[0084] The power load is obtained by a power meter installed at the generator terminals in the unit; the ambient temperature is obtained by a temperature sensor installed on the unit.

[0085] The operating condition matching module is used to match the best operating parameters from the best operating condition table according to the real-time operating condition data.

[0086] The improved salp swarm algorithm is used to match the best operating parameters from the best operating condition table. The steps include:

[0087] Step 1: Preset the population size Z and the number threshold G.

[0088] Step 2: Generate an initialization population. The positions of individuals in the initialization population are defined in a one-dimensional search space. The positions of individuals correspond to the working condition labels one by one. The number of iterations T of the initialization population is 0.

[0089] Step 3: Determine the fitness function.

[0090] Step 4: Calculate the fitness of each individual in the initialized population, and take the position of the individual with the maximum fitness as the food source position.

[0091] Step 5: Update the leader position and follower position.

[0092] Step 6: Calculate the dynamic boundary, generate the reverse solution corresponding to each individual and add it to the population.

[0093] Step 7: Update the population, obtain a new population, and set T = T + 1, that is, add one to the number of iterations corresponding to the population before the update as the number of iterations of the new population.

[0094] Step 8: Loop steps 5 to 7 until the number of iterations corresponding to the new population is equal to the number threshold G. The loop ends, and the operating condition label corresponding to the food source position in the new population is obtained. The corresponding unit operating condition data is obtained according to the obtained operating condition label. The corresponding optimal operating parameters are matched from the optimal operating condition table according to the obtained unit operating condition data.

[0095] It should be noted that the population size Z is determined by technical personnel in this field. In the process of optimizing historical operating parameters, multiple groups of different unit operating condition data are collected, and the multiple groups of different unit operating condition data are recorded in the optimal operating condition table; under the conditions of multiple groups of different unit operating condition data, multiple different population sizes are set for the same unit operating condition data, and the salp swarm algorithm is performed multiple times. After the same number of iterations, the corresponding operating condition labels are obtained, and the population size corresponding to the operating condition label that is the same as the operating condition label corresponding to the group of unit operating condition data is used as the population size corresponding to the group of unit operating condition data; and the population size corresponding to each group of unit operating condition data is obtained in this way, and the average of multiple population sizes is used as the preset population size Z.

[0096] The number threshold G is determined by a technician in this field. During the optimization process of historical operating parameters, multiple groups of different unit operating condition data are collected, and the multiple groups of different unit operating condition data are all recorded in the optimal operating condition table; under the conditions of multiple groups of different unit operating condition data, the salp swarm algorithm is performed multiple times on the same unit operating condition data to obtain multiple operating condition labels, wherein the number of iterations of the salp swarm algorithm is different each time, and the population size is the same and is Z; the number of iterations corresponding to the operating condition label that is the same as the operating condition label corresponding to the group of unit operating condition data is used as the number of iterations corresponding to the group of unit operating condition data; and the number of iterations corresponding to each group of unit operating condition data is obtained by analogy, and the average of multiple iteration numbers is used as the number threshold G.

[0097] In the above step 2, the population S0 is initialized to {X1, X2, ..., X Z}, X Z is the Zth individual, including Z individuals, of which the first individual is the leader and the remaining Z-1 individuals are followers; different digital labels are set for different unit operating condition data in the optimal operating condition table and marked as operating condition labels, and the operating condition labels correspond to the unit operating condition data one by one; the range of the operating condition labels is [1, A], and the range of the one-dimensional search space is also [1, A].

[0098] In the above step 3, the expression of the fitness function is: f=XS; where f is the fitness and XS is the similarity.

[0099] The similarity acquisition methods include:

[0100] Get the number of power loads in the optimal operating condition table and mark it as the number of data; add up all the power loads in the optimal operating condition table in turn, and then divide it by the number of data to get the load mean; add up the squares of the difference between each power load in the optimal operating condition table and the load mean in turn, and then divide it by the number of data to get the load variance; take the square root of the load variance to get the load standard deviation; add up all the ambient temperatures in the optimal operating condition table in turn, and then divide it by the number of data to get the temperature mean; add up the squares of the difference between each ambient temperature in the optimal operating condition table and the temperature mean in turn, and then divide it by the number of data to get the temperature variance; take the square root of the temperature variance to get the temperature standard deviation;

[0101] The unit operating condition data corresponding to the operating condition label corresponding to the individual position is obtained, and marked as comparison operating condition data; the comparison operating condition data is standardized to obtain comparison standard data; the method for obtaining comparison standard data is: subtract the load mean from the power load in the comparison operating condition data, and then divide it by the load standard deviation to obtain the standardized power load; subtract the temperature mean from the ambient temperature in the comparison operating condition data, and then divide it by the temperature standard deviation to obtain the standardized ambient temperature; the standardized power load and the standardized ambient temperature are the comparison standard data; the real-time operating condition data is standardized and marked as real-time standard data; the method for obtaining real-time standard data is consistent with the method for obtaining comparison standard data; the Euler distance between the comparison standard data and the real-time standard data is calculated, and the reciprocal of the Euler distance is used as the similarity.

[0102] The expression of the Euler distance between the standard data and the real-time working condition data is: Where O is the Euler distance, DD is the power load in the comparison working condition data, SD is the power load in the real-time working condition data, DW is the ambient temperature in the comparison working condition data, and SW is the ambient temperature in the real-time working condition data.

[0103] It should be noted that the purpose of standardizing the comparison operating data and the real-time operating data is to reduce the complexity of calculation: the dimensions and magnitudes of different data may vary greatly, and direct calculation will increase the complexity; standardization can unify the data into a dimensionless range of 0-1, simplifying the calculation; and speeding up the reasoning speed: after unifying the magnitude of indicators, operations such as calculating similarity will be more efficient, thereby speeding up the entire operating parameter optimization process.

[0104] In the above step 5, the method for updating the leader position includes:

[0105] Sort the fitness of all individuals in the population from large to small, and place the individual position corresponding to the fitness next to the fitness corresponding to the food source position as the second position, and the individual position corresponding to the last fitness as the last position;

[0106]

[0107] Where X1 is the leader position, F is the food source position, d is the number of iterations, the number of iterations d is the degree of freedom parameter of the t distribution, t(d) is a random value in the t distribution with the number of iterations d, X′ is the second position, X″ is the last position, C1 is a random number in [-0.5,0.5], and C2 is a random number in [0,1].

[0108] It should be noted that the t-distribution is a continuous probability distribution. The shape of the t-distribution is related to the degree of freedom parameter. Different degrees of freedom parameters will result in different shapes of the t-distribution. The value of t(d) is randomly obtained according to the probability of the t-distribution. The t-distribution is introduced in updating the leader position, and the perturbation ability of the t-distribution is used to improve the ability of global optimization.

[0109] Methods for updating follower locations include:

[0110]

[0111] Where, X v is the position of the vth individual, X v-1 is the position of the v-1th individual, v∈[2,Z]; that is, the current position of the v-1th individual is added to the current position of the v-1th individual, and then multiplied by half to give the updated position of the v-1th individual.

[0112] In the above step 6, the dynamic boundary is [R1, R2], R1 is the working condition label with the smallest value corresponding to the individual position in the population, R2 is the working condition label with the largest value corresponding to the individual position in the population, 1≤R1<R2≤A; the dynamic boundary can make up for the defect that the fixed boundary cannot save the search experience, which is conducive to shortening the search time.

[0113] The methods for generating the inverse solution include:

[0114]

[0115] In the formula, is the reverse solution of the ith individual, C3 is a random number in [0,1], i∈[1,Z], X i is the position of the i-th individual.

[0116] If the generated reverse solution is greater than or less than the dynamic boundary, it is marked as a transcendental solution, a random number function is used to randomly generate a value within the dynamic boundary, and the randomly generated value is assigned to the transcendental solution.

[0117] In the above step 7, the reverse solution of the individual is also taken as an individual, the fitness of all individuals in the population is calculated, the fitness of all individuals in the population is sorted from large to small, the first Z individuals are retained and used as the new population, and the position of the individual with the largest fitness in the new population is taken as the new food source position.

[0118] It should be understood that when matching the best operating parameters from the optimal operating condition table, the reason for using the improved salp swarm algorithm is that the algorithm is highly flexible: by presetting the population size and number threshold parameters, the convergence speed and accuracy requirements of the salp swarm algorithm can be flexibly adjusted; strong diversity: the concept of reverse solution is introduced into the salp swarm algorithm, which can increase the diversity of the population and avoid falling into local optimality; fast convergence speed: by updating the positions of leaders and followers, combined with the adjustment of dynamic boundaries, the salp swarm algorithm can quickly approach the global optimal solution; wide applicability: the salp swarm algorithm can be applied to different types of operating condition tables and has strong versatility.

[0119] The matching judgment module is used to judge whether to generate a trust instruction based on the optimal operating parameters.

[0120] The method for determining whether to generate a trust instruction includes:

[0121] Input the real-time working condition data into the trained threshold prediction model, obtain the corresponding similarity threshold, and mark it as the real-time threshold; the training process of the threshold prediction model includes:

[0122] A group of real-time operating condition data is collected in advance, and corresponding similarity thresholds are set for the group of real-time operating condition data, where a is an integer greater than 1, and the real-time operating condition data and the corresponding similarity thresholds are converted into a corresponding set of feature vectors; the similarity thresholds corresponding to the real-time operating condition data are collected by technical personnel in the process of optimizing historical operating parameters, and the power load and ambient temperature therein are adjusted multiple times under the condition of a group of unit operating condition data, and the unit is controlled to operate according to the optimal operating parameters corresponding to the group of unit operating condition data. When the operating efficiency and quality indicators of the unit change for the first time, the corresponding power load and ambient temperature are recorded, and the similarity between the recorded power load and ambient temperature and the power load and ambient temperature in the unit operating condition data before adjustment is calculated, and the similarity is used as the similarity corresponding to the group of unit operating condition data; and so on, multiple similarities corresponding to each group of unit operating condition data are obtained, and the average of the multiple similarities corresponding to each group of unit operating condition data is used as the similarity threshold corresponding to each group of unit operating condition data;

[0123] Obtain the similarity threshold corresponding to each group of real-time working condition data; set the corresponding similarity threshold for group a of real-time working condition data in turn;

[0124] Each set of feature vectors is used as the input of the threshold prediction model. The threshold prediction model uses a set of predicted similarity thresholds corresponding to each set of real-time operating condition data as output, and uses the actual similarity threshold corresponding to each set of real-time operating condition data as the prediction target. The actual similarity threshold is the similarity threshold corresponding to the real-time operating condition data collected in advance; the training target is to minimize the sum of the prediction errors of all real-time operating condition data; wherein the calculation formula of the prediction error is ηK =(β K -ε K ) 2 , where η K is the prediction error, K is the group number of the feature vector corresponding to the real-time working condition data, β K is the prediction similarity threshold corresponding to the Kth group of real-time operating condition data, ε K is the actual similarity threshold corresponding to the Kth group of real-time operating condition data; the threshold prediction model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

[0125] The above threshold prediction model is specifically a deep neural network model; it includes an input layer, a hidden layer and an output layer; each hidden layer includes multiple neurons, each neuron is connected to the neurons in the next layer, and the connection contains weights, which determine the importance and influence of data transmission in the neural network; an activation function is applied to each neuron between the hidden layer and the output layer, and the activation function reflects nonlinearity, allowing the network to learn more complex patterns and features.

[0126] Compare the similarity corresponding to the obtained working condition label with the real-time threshold;

[0127] If the similarity is less than the real-time threshold, a trust instruction is generated, indicating that the real-time operating condition data is similar to the acquired unit operating condition data, and therefore the corresponding optimal operating parameters are consistent;

[0128] If the similarity is greater than or equal to the real-time threshold, no trust instruction is generated; this indicates that the real-time operating condition data is significantly different from the acquired unit operating condition data, and therefore the corresponding optimal operating parameters are inconsistent. When the optimal operating parameters corresponding to the acquired unit operating condition data are used to control the unit operation, the unit's operating efficiency and quality indicators will be reduced, and the optimal operating parameters need to be re-acquired.

[0129] The parameter recommendation module is used to recommend the matching optimal operating parameters if a trust instruction is generated; if no trust instruction is generated, weighted calculation is performed on all the optimal operating parameters in the optimal operating condition table, and the optimal operating parameters are re-obtained and recommended.

[0130] Methods for regaining optimal operating parameters include:

[0131] Add the similarities corresponding to each group of unit operating condition data in the optimal operating condition table in turn to obtain the total similarity; multiply each power load in the optimal operating condition table by the corresponding similarity, and then add them in turn to obtain the total load value; multiply each ambient temperature in the optimal operating condition table by the corresponding similarity, and then add them in turn to obtain the total temperature value; divide the total load value by the sum of similarities to obtain the optimal power load; divide the total temperature value by the sum of similarities to obtain the optimal ambient temperature; use the optimal power load and the optimal ambient temperature as the re-acquired optimal operating parameters; and add the re-acquired optimal operating parameters and the corresponding real-time operating condition data into the optimal operating condition table.

[0132] This embodiment constructs an optimal operating condition table and adopts an improved salp group algorithm to accurately match the unit operating condition data with the greatest similarity to the real-time operating condition data from the optimal operating condition table; it is achieved that even when the operating conditions are complex and changeable, the most similar unit operating condition data can still be matched from the optimal operating condition table, thereby improving the matching speed and matching accuracy; and a trust instruction generation mechanism is set. When the similarity between the real-time operating condition data and the matched unit operating condition data is low, the similarity of each group of unit operating condition data in the optimal operating condition table is used as a weight for comprehensive weighted calculation, so that the optimal operating parameters can be automatically re-recommended; avoid the situation where the optimal operating parameters cannot be recommended when the real-time operating condition data is not stored in the optimal operating condition table; realize intelligent recommendation of the optimal operating parameters, thereby improving the overall operating efficiency and quality indicators of the unit, and thereby enhancing the adaptability, robustness and safety of the system.

[0133] Example 2

[0134] See also Figure 2 As shown, this embodiment further improves the design on the basis of Embodiment 1. Due to certain limitations in the similarity calculation process, it is impossible to fully and accurately reflect the correlation between the real-time operating condition data and the unit operating condition data, that is, the weights corresponding to different unit operating condition data in the optimal operating condition table also have certain deviations, resulting in errors in the re-acquired optimal operating parameters. It is necessary to further analyze the re-acquired optimal operating parameters to ensure that the recommended optimal operating parameters are the optimal operating parameters corresponding to the real-time operating condition data, so as to achieve efficient and safe operation of the unit; therefore, this embodiment provides a unit operating condition-parameter modeling system, which also includes a parameter optimization module;

[0135] The parameter optimization module is used to calculate the credibility of the re-acquired optimal operating parameters and determine whether to perform parameter optimization processing.

[0136] The re-acquired optimal operating parameters are marked as analysis parameters; the analysis parameters and the corresponding real-time operating condition data are used as analysis data, and the analysis data are input into the trained trust analysis model to predict the corresponding credibility; the training process of the trust analysis model is consistent with the training process of the threshold prediction model, both of which are deep neural network models; the credibility corresponding to the analysis data is determined by a technician in the field of historical operating parameter optimization, and b groups of real-time operating condition data are collected, where b is an integer greater than 1; under the condition of a group of real-time operating condition data, the corresponding optimal operating parameters are finely adjusted in descending order during the operation of the unit, and the corresponding unit operating efficiency is recorded in turn, that is, the unit operating efficiency corresponding to the multiple groups of analysis data corresponding to the group of real-time operating condition data is recorded; the recorded operating efficiency of each unit is divided by the unit operating efficiency before the optimal operating parameters are adjusted to obtain the efficiency ratio; the efficiency ratio corresponding to each unit operating efficiency is used as the credibility corresponding to the multiple groups of analysis data corresponding to the group of real-time operating condition data; and the credibility corresponding to the multiple groups of analysis data corresponding to the b groups of real-time operating condition data is obtained by analogy.

[0137] Preset a credibility threshold and compare the predicted credibility with the credibility threshold; the credibility threshold is preset by technicians in this field according to the actual performance requirements of the unit;

[0138] If the predicted credibility is greater than or equal to the credibility threshold, there is no need to optimize the analysis parameters, which means that the analysis parameters are the optimal operating parameters of the real-time operating condition data.

[0139] If the predicted credibility is less than the credibility threshold, the analysis parameters need to be optimized; this indicates that the analysis parameters are not the best operating parameters for the real-time operating data and need to be optimized to obtain accurate best operating parameters.

[0140] The parameter optimization processing methods include:

[0141] Obtain a parameter range, the parameter range includes the range corresponding to each parameter in the operating parameters; the parameter range is determined according to the technical specifications of the unit; the values ​​in each range within the parameter range are randomly combined to construct m parameter sets, wherein each parameter set includes a value in each range within the parameter range, and the m parameter sets are all different; different digital labels are set for the m parameter sets, and are marked as parameter labels, and the range of the parameter labels is [1, m]; according to the real-time operating condition data, the improved salp swarm algorithm is used again to screen out an optimal parameter set label from the m parameter set labels, and the parameter set corresponding to the screened optimal parameter set label is used as the parameter optimization processing result of the optimal operating parameters;

[0142] The step of selecting an optimal parameter set label from m parameter set labels is consistent with the step of matching the optimal operating parameters from the optimal operating condition table; the difference is that in the step of selecting an optimal parameter set label from m parameter set labels, the expression of the fitness function is: f′=KX; where f′ is fitness and KX is credibility.

[0143] The recommended interface for optimal operating parameters is as follows: Figure 3 shown.

[0144] This embodiment introduces a credibility analysis mechanism to calculate the credibility corresponding to the re-acquired optimal operating parameters; and through threshold comparison, it is determined whether it is necessary to perform parameter optimization processing on the re-acquired optimal operating parameters. If necessary, a parameter combination label matching method is adopted, and the improved salp swarm algorithm is combined again to accurately screen the best parameter set from multiple candidate parameter sets for matching to obtain the optimized optimal operating parameters; the calculation error of the optimal operating parameters can be corrected, and the quantitative evaluation of the optimization process can be achieved through a machine learning model, thereby ensuring the recommendation quality of the optimal operating parameters; at the same time, the optimization degree of the system algorithm is deepened to further achieve efficient and safe operation of the unit.

[0145] Example 3

[0146] See also Figure 4 As shown, the parts not described in detail in this embodiment refer to the description of embodiment 1 and embodiment 2, and a unit operating condition-parameter modeling method is provided, the method comprising:

[0147] Constructing an optimal operating condition table, which includes unit operating condition data and optimal operating parameters corresponding to the unit operating condition data;

[0148] Collect the unit operating data during operation and mark it as real-time operating data;

[0149] According to the real-time operating data, the best operating parameters are matched from the best operating table;

[0150] Determine whether to generate a trust instruction based on the optimal operating parameters;

[0151] If a trust instruction is generated, the matching optimal operating parameters are recommended; if no trust instruction is generated, all the optimal operating parameters in the optimal operating condition table are weightedly calculated, the optimal operating parameters are re-obtained and recommended.

[0152] Example 4

[0153] See also Figure 5As shown, the present application also provides an electronic device 500. The electronic device 500 may include one or more processors and one or more memories. The memories store computer readable codes, and when the computer readable codes are run by one or more processors, a unit operating condition-parameter modeling method as described above may be executed.

[0154] The method or system according to the embodiment of the present application can also be used by Figure 5 The electronic device architecture shown in FIG. Figure 5 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, ROM 503, RAM 504, a communication port 505 connected to a network, an input / output 506, a hard disk 507, etc. The storage device in the electronic device 500, such as ROM 503 or hard disk 507, may store a unit operating condition-parameter modeling method provided in the present application. Furthermore, the electronic device 500 may also include a user interface 508. Of course, Figure 5 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 5 One or more components of an electronic device are shown.

[0155] Example 5

[0156] See also Figure 6 As shown, one embodiment of the present application discloses a computer-readable storage medium 600. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are executed by the processor, a unit operating condition-parameter modeling method according to an embodiment of the present application described with reference to the above figures can be executed. The storage medium 600 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0157] In addition, according to the embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: a unit operating condition-parameter modeling method. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0158] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0159] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A unit operating condition-parameter modeling method, characterized in that: include: Constructing an optimal operating condition table, which includes unit operating condition data and optimal operating parameters corresponding to the unit operating condition data; Collect the unit operating data during operation and mark it as real-time operating data; According to the real-time operating data, the best operating parameters are matched from the best operating table; Determine whether to generate a trust instruction based on the optimal operating parameters; If a trust instruction is generated, the matching optimal operating parameters are recommended; if no trust instruction is generated, all the optimal operating parameters in the optimal operating condition table are weightedly calculated, the optimal operating parameters are re-obtained and recommended.

2. A unit operating condition-parameter modeling method according to claim 1, characterized in that: The unit operating data includes power load and ambient temperature; the optimal operating parameters are the operating parameters corresponding to the optimal operating efficiency and quality indicators of the unit under different unit operating data; the operating parameters include main steam temperature, main steam pressure, main steam flow, reheat steam temperature, feed water temperature and circulating water temperature; The method for constructing the optimal operating condition table is as follows: collect historical unit operating condition data and the optimal operating parameters corresponding to the historical unit operating condition data, use multi-parameter steady-state screening method and fuzzy C-means clustering method to process the historical unit operating condition data, and construct the optimal operating condition table based on the processed historical unit operating condition data and the corresponding optimal operating parameters.

3. A unit operating condition-parameter modeling method according to claim 2, characterized in that: The improved salp swarm algorithm is used to match the best operating parameters from the best operating condition table. The steps include: Step 1: Preset the population size Z and the number threshold G; Step 2: Generate an initialization population. The positions of individuals in the initialization population are defined in a one-dimensional search space. The positions of individuals correspond to the working condition labels one by one. The number of iterations T of the initialization population is 0. Step 3: Determine the fitness function; Step 4: Calculate the fitness of each individual in the initialized population, and take the position of the individual with the maximum fitness as the food source position; Step 5: Update the leader position and follower position; Step 6: Calculate the dynamic boundary, generate the reverse solution corresponding to each individual and add it to the population; Step 7: Update the population, obtain a new population, and set T = T + 1; Step 8: Loop steps 5 to 7 until the number of iterations corresponding to the new population is equal to the number threshold G. The loop ends, and the operating condition label corresponding to the food source position in the new population is obtained. The corresponding unit operating condition data is obtained according to the obtained operating condition label. The corresponding optimal operating parameters are matched from the optimal operating condition table according to the obtained unit operating condition data.

4. A unit operating condition-parameter modeling method according to claim 3, characterized in that: In step 2, the population S0 is initialized to {X1, X2, ..., X Z }, X Z is the Zth individual, including Z individuals, of which the first individual is the leader and the remaining Z-1 individuals are followers; different digital labels are set for different unit operating condition data in the optimal operating condition table and marked as operating condition labels, and the operating condition labels correspond to the unit operating condition data one by one; the range of the operating condition label and the range of the one-dimensional search space are both [1, A]; In step 3, the fitness function is expressed as: f=XS; In the formula, f is fitness, XS is similarity; the similarity acquisition method includes: Obtain the power load in the optimal operating condition table, calculate the mean and standard deviation of the power load in the optimal operating condition table, the mean is marked as the load mean, and the standard deviation is marked as the load standard deviation; obtain the ambient temperature in the optimal operating condition table, calculate the mean and standard deviation of the ambient temperature in the optimal operating condition table, the mean is marked as the temperature mean, and the standard deviation is marked as the temperature standard deviation; The unit operating condition data corresponding to the operating condition label corresponding to the individual position is obtained, and marked as comparison operating condition data; the comparison operating condition data is standardized to obtain comparison standard data; the method for obtaining comparison standard data is: subtract the load mean from the power load in the comparison operating condition data, and then divide it by the load standard deviation to obtain the standardized power load; subtract the temperature mean from the ambient temperature in the comparison operating condition data, and then divide it by the temperature standard deviation to obtain the standardized ambient temperature; the standardized power load and the standardized ambient temperature are the comparison standard data; the real-time operating condition data is standardized and marked as real-time standard data; the method for obtaining real-time standard data is consistent with the method for obtaining comparison standard data; the Euler distance between the comparison standard data and the real-time standard data is calculated, and the reciprocal of the Euler distance is used as the similarity.

5. A unit operating condition-parameter modeling method according to claim 4, characterized in that: In step 5, the method for updating the leader position includes: Sort the fitness of all individuals in the population from large to small, and place the individual position corresponding to the fitness next to the fitness corresponding to the food source position as the second position, and the individual position corresponding to the last fitness as the last position; Where X1 is the leader position, F is the food source position, d is the number of iterations, the number of iterations d is the degree of freedom parameter of the t distribution, t(d) is a random value in the t distribution with the number of iterations d, X′ is the second position, X″ is the last position, C1 is a random number in [-0.5,0.5], and C2 is a random number in [0,1]; Methods for updating follower positions include: Where, X v is the position of the vth individual, X v-1 is the position of the v-1th individual, v∈[2,Z].

6. A unit operating condition-parameter modeling method according to claim 5, characterized in that: In step 6, the dynamic boundary is [R1, R2], R1 is the operating condition label with the smallest value corresponding to the individual position in the population, R2 is the operating condition label with the largest value corresponding to the individual position in the population, 1≤R1<R2≤A; The methods for generating the inverse solution include: In the formula, is the reverse solution of the ith individual, C3 is a random number in [0,1], i∈[1,Z], X i is the position of the i-th individual; If the generated reverse solution is greater than or less than the dynamic boundary, it is marked as a transcendental solution, a random number function is used to randomly generate a value within the dynamic boundary, and the randomly generated value is assigned to the transcendental solution; In step 7, the reverse solution of the individual is also taken as an individual, the fitness of all individuals in the population is calculated, the fitness of all individuals in the population is sorted from large to small, the first Z individuals are retained and taken as the new population, and the position of the individual with the largest fitness in the new population is taken as the new food source position.

7. A unit operating condition-parameter modeling method according to claim 6, characterized in that: The method for determining whether to generate a trust instruction includes: Input the real-time working condition data into the trained threshold prediction model, obtain the corresponding similarity threshold, and mark it as the real-time threshold; the training process of the threshold prediction model includes: Pre-collect a group of real-time working condition data, set corresponding similarity thresholds for the a group of real-time working condition data, a is an integer greater than 1, and convert the real-time working condition data and the corresponding similarity thresholds into a corresponding set of feature vectors; obtain the similarity threshold corresponding to each group of real-time working condition data; and sequentially set corresponding similarity thresholds for the a group of real-time working condition data; Each group of feature vectors is used as the input of the threshold prediction model. The threshold prediction model uses a group of predicted similarity thresholds corresponding to each group of real-time operating condition data as output, and uses the actual similarity threshold corresponding to each group of real-time operating condition data as the prediction target, where the actual similarity threshold is the pre-collected similarity threshold corresponding to the real-time operating condition data; minimizing the sum of prediction errors of all real-time operating condition data is used as the training target; the threshold prediction model is trained until the sum of prediction errors reaches convergence and the training is stopped; the threshold prediction model is a deep neural network model; Compare the similarity corresponding to the obtained working condition label with the real-time threshold; If the similarity is less than the real-time threshold, a trust instruction is generated; If the similarity is greater than or equal to the real-time threshold, no trust instruction is generated.

8. A unit operating condition-parameter modeling method according to claim 7, characterized in that: The method for re-obtaining the optimal operating parameters comprises: Add the similarities corresponding to each group of unit operating condition data in the optimal operating condition table in turn to obtain the total similarity; multiply each power load in the optimal operating condition table by the corresponding similarity, and then add them in turn to obtain the total load value; multiply each ambient temperature in the optimal operating condition table by the corresponding similarity, and then add them in turn to obtain the total temperature value; divide the total load value by the sum of similarities to obtain the optimal power load; divide the total temperature value by the sum of similarities to obtain the optimal ambient temperature; use the optimal power load and the optimal ambient temperature as the re-acquired optimal operating parameters; and add the re-acquired optimal operating parameters and the corresponding real-time operating condition data into the optimal operating condition table.

9. A unit operating condition-parameter modeling method according to claim 8, characterized in that: The method further comprises: calculating the credibility of the re-acquired optimal operating parameters, and determining whether to perform parameter optimization processing; The optimal operating parameters retrieved are marked as analysis parameters; the analysis parameters and the corresponding real-time operating condition data are used as analysis data, and the analysis data is input into the trained trust analysis model to predict the corresponding credibility; the training process of the trust analysis model is consistent with the training process of the threshold prediction model, and both are deep neural network models; A confidence threshold is preset, and the confidence of the prediction is compared with the confidence threshold; if the confidence of the prediction is greater than or equal to the confidence threshold, there is no need to perform parameter optimization processing on the analysis parameters; if the confidence of the prediction is less than the confidence threshold, it is necessary to perform parameter optimization processing on the analysis parameters; The parameter optimization processing methods include: Obtain a parameter range, which includes the range corresponding to each parameter in the operating parameters; randomly combine the values ​​in each range within the parameter range to construct m parameter sets, wherein each parameter set includes a value in each range within the parameter range, and the m parameter sets are all different; set different digital labels for the m parameter sets and mark them as parameter labels, and the range of the parameter labels is [1, m]; according to the real-time operating condition data, the improved salp swarm algorithm is used again to screen out an optimal parameter set label from the m parameter set labels, and the parameter set corresponding to the screened optimal parameter set label is used as the parameter optimization processing result of the optimal operating parameters; The step of selecting an optimal parameter set label from m parameter set labels is consistent with the step of matching the optimal operating parameters from the optimal operating condition table; the difference is that in the step of selecting an optimal parameter set label from m parameter set labels, the expression of the fitness function is: f′=KX; where f′ is fitness and KX is credibility.

10. A unit operating condition-parameter modeling system, implementing a unit operating condition-parameter modeling method according to any one of claims 1 to 9, characterized in that: include: The operating condition modeling module is used to construct an optimal operating condition table, which includes unit operating condition data and optimal operating parameters corresponding to the unit operating condition data; The data acquisition module is used to collect the unit operating data during the operation of the unit and mark it as real-time operating data; The working condition matching module is used to match the best operating parameters from the best working condition table according to the real-time working condition data; A matching judgment module, used to judge whether to generate a trust instruction based on the optimal operating parameters; The parameter recommendation module is used to recommend the matching optimal operating parameters if a trust instruction is generated; if no trust instruction is generated, weighted calculation is performed on all the optimal operating parameters in the optimal operating condition table, and the optimal operating parameters are re-obtained and recommended.

Images