Turbo expander rotating speed adjusting method and system based on natural gas pressure fluctuation

Abstract

The invention relates to the technical field of expansion machine rotation speed regulation, in particular to a turbo expansion machine rotation speed regulation method and system based on natural gas pressure fluctuation, an expansion machine rotation speed prediction model is constructed by collecting a large amount of historical data, and a reliable prediction model is provided for achieving turbo expansion machine rotation speed prediction analysis. The real-time natural gas pressure fluctuation characteristic data and the real-time rotating speed influence factor data are combined to accurately predict the rotating speed of the target turbo expander, rotating speed fluctuation prediction data are generated, and the rotating speed fluctuation prediction data are compared with a preset first rotating speed fluctuation threshold value; and whether the rotating speed of the target turbo expander needs to be adjusted or not is judged, when adjustment needs to be conducted, a corresponding rotating speed adjusting scheme is accurately generated according to the rotating speed fluctuation prediction data in combination with an intelligent optimization algorithm, rotating speed adjusting operation is executed, and accurate detection and intelligent adjustment of the rotating speed of the turbo expander are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of expander speed regulation, and in particular to a turbine expander speed regulation method and system based on natural gas pressure fluctuation. Background Art

[0002] As turboexpanders are core equipment in natural gas liquefaction, energy recovery, and cryogenic engineering, their speed stability directly impacts system efficiency, equipment lifespan, and safety. During natural gas processing, factors such as pipeline pressure fluctuations and load variations can easily lead to abnormal turboexpander speeds, which can in turn cause mechanical vibration, reduced efficiency, and even equipment failure. Therefore, precise speed regulation technology is crucial to ensuring efficient and safe turboexpander operation.

[0003] Existing speed regulation technologies often acquire data and analyze it in combination with fixed thresholds. This makes it impossible to predict the speed of the turbine expander, and thus difficult to respond to fluctuations in natural gas in a timely manner, resulting in poor speed regulation results. At the same time, when selecting a speed regulation scheme, it is impossible to dynamically adjust according to real-time conditions, making it difficult to achieve global optimal regulation. Summary of the Invention

[0004] In response to the problems in the related art, the present invention provides a method and system for adjusting the speed of a turbine expander based on natural gas pressure fluctuations to overcome the above-mentioned technical problems existing in the existing related art.

[0005] To solve the above technical problems, the present invention provides the following technical solution: a method for adjusting the speed of a turbine expander based on natural gas pressure fluctuations, comprising the following steps:

[0006] S1. Collect historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data;

[0007] S2. Building an expander speed prediction model based on the historical natural gas pressure fluctuation characteristic data, the historical expander speed influencing factor data, and the historical expander speed fluctuation data;

[0008] S3. Collecting real-time natural gas pressure fluctuation characteristic data and real-time speed influencing factor data of the target turboexpander;

[0009] S4. Predicting and analyzing the speed of the target turbine expander based on the real-time natural gas pressure fluctuation characteristic data, the real-time speed influencing factor data, and the expander speed prediction model to generate speed fluctuation prediction data of the target turbine expander;

[0010] S5. Setting a speed fluctuation threshold, comparing the speed fluctuation prediction data with the speed fluctuation threshold, and generating speed adjustment instruction data;

[0011] If no adjustment is required, the speed adjustment operation is terminated;

[0012] If adjustment is required, a corresponding speed adjustment scheme is selected according to the speed fluctuation prediction data to perform the speed adjustment operation.

[0013] Preferably, the specific steps of collecting historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data are as follows:

[0014] S11, collect several sets of historical natural gas pressure fluctuation feature data within a single natural gas pressure fluctuation detection cycle in the working state of the turbine expander through the expander speed adjustment platform, and obtain the historical natural gas pressure fluctuation feature data set A = {a1, a2, ..., a i ,…,a k}, where a i represents the i-th group of natural gas pressure fluctuation characteristic data within a single natural gas pressure fluctuation detection cycle during the operation of the turboexpander in history, and k represents the total number of groups of natural gas pressure fluctuation characteristic data collected;

[0015] S12, collecting the turbine expander speed influencing factor data at the beginning of the natural gas pressure fluctuation detection period corresponding to each historical natural gas pressure fluctuation feature data in the historical natural gas pressure fluctuation feature data set online through the expander speed adjustment platform, and obtaining the historical expander speed influencing factor data set represents the data of factors influencing the speed of the turbine expander at the beginning of the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data;

[0016] The speed influencing factor data include but are not limited to nozzle opening, guide vane angle, impeller diameter, intake valve opening, rated speed and load power;

[0017] S13, online collecting the turbine expander speed fluctuation data within the natural gas pressure fluctuation detection period corresponding to each historical natural gas pressure fluctuation characteristic data in the historical natural gas pressure fluctuation characteristic data set through the expander speed adjustment platform, and obtaining the historical expander speed fluctuation data set in, represents the speed fluctuation data of the turbine expander during the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, and represents the j-th speed data of the turbine expander collected during the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, represents the rated speed of the turbine expander within the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, and n represents the total number of turbine expander speed data collected within the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data.

[0018] Preferably, the specific steps of constructing the expander speed prediction model based on the historical natural gas pressure fluctuation characteristic data, the historical expander speed influencing factor data, and the historical expander speed fluctuation data are as follows:

[0019] S21. Setting a training data ratio, and dividing the historical natural gas pressure fluctuation feature dataset, the historical expander speed influencing factor dataset, and the historical expander speed fluctuation dataset according to the training data ratio to obtain historical natural gas pressure fluctuation feature training data, historical expander speed influencing factor training data, historical expander speed fluctuation training data, historical natural gas pressure fluctuation feature test data, historical expander speed influencing factor test data, and historical expander speed fluctuation test data;

[0020] S22. Build the initial GRU model and set the training error threshold and test accuracy threshold;

[0021] S23, setting a maximum number of training times, inputting the historical natural gas pressure fluctuation characteristic training data and the historical expander speed influencing factor training data as training data, and the historical expander speed fluctuation training data as training label data into the initial GRU model for training, and continuously adjusting the parameters of the initial GRU model according to the training results until the training error is less than the training error threshold or the number of training times exceeds the maximum number of training times, thereby obtaining a trained GRU model;

[0022] S24. Input the historical natural gas pressure fluctuation characteristic test data and the historical expander speed influencing factor test data as test data, and input the historical expander speed fluctuation test data as test label data into the trained GRU model for testing, and calculate the accuracy of the test results. If the accuracy of the test results is greater than or equal to the test accuracy threshold, the expander speed prediction model is obtained; otherwise, return to S23 for training until the accuracy of the test results is greater than or equal to the test accuracy threshold.

[0023] An expander speed prediction model is constructed by using historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data. This provides a reliable prediction model for realizing turbine expander speed prediction analysis and ensures the accuracy and reliability of turbine expander speed fluctuation prediction results.

[0024] Preferably, the specific steps of collecting the real-time natural gas pressure fluctuation characteristic data and the real-time speed influencing factor data of the target turboexpander are as follows:

[0025] S31, collecting natural gas pressure fluctuation characteristic data B of the target turboexpander in real time through a pressure sensor installed in the target turboexpander;

[0026] S32, obtain the real-time speed influencing factor data of the target turbine expander online through the expander speed adjustment platform, and obtain the real-time speed influencing factor data set C = {c1, c2, ..., c i ,…,c l}, where c i represents the i-th real-time speed influencing factor data of the target turboexpander, and l represents the total number of real-time speed influencing factor data;

[0027] The real-time speed influencing factor data includes but is not limited to nozzle opening, guide vane angle, impeller diameter, intake valve opening, rated speed and load power.

[0028] Preferably, the specific steps of performing prediction analysis on the speed of the target turbine expander based on the real-time natural gas pressure fluctuation characteristic data, the real-time speed influencing factor data, and the expander speed prediction model to generate the speed fluctuation prediction data of the target turbine expander are as follows:

[0029] S41. Input the real-time natural gas pressure fluctuation characteristic data B and the real-time speed influencing factor data set into the expander speed prediction model to perform prediction analysis on the speed of the target turbine expander, thereby generating speed fluctuation prediction data D of the target turbine expander.

[0030] Preferably, a speed fluctuation threshold is set, the speed fluctuation prediction data and the speed fluctuation threshold are numerically compared to generate speed adjustment instruction data; if adjustment is not required, the current speed adjustment operation is terminated; if adjustment is required, a corresponding speed adjustment scheme is selected according to the speed fluctuation prediction data to perform the speed adjustment operation. The specific steps are as follows:

[0031] S51, setting a first speed fluctuation threshold and a second speed fluctuation threshold, and performing a numerical comparison between the speed fluctuation prediction data D and the first speed fluctuation threshold;

[0032] If the speed fluctuation prediction data is less than or equal to the first speed fluctuation threshold, it indicates that the speed fluctuation of the target turbine expander is within a reasonable range, and the speed adjustment instruction data is output as not requiring adjustment. The speed adjustment instruction data is pushed to the expander speed adjustment platform, and the speed adjustment operation is terminated.

[0033] If the speed fluctuation prediction data is greater than the first speed fluctuation threshold, it means that the speed fluctuation of the target turbine expander is in an abnormal range, the output speed adjustment instruction data needs to be adjusted, and the corresponding speed adjustment scheme is selected according to the speed fluctuation prediction data.

[0034] Preferably, selecting a corresponding speed adjustment scheme according to the speed fluctuation prediction data includes the following steps:

[0035] S511, numerically comparing the speed fluctuation prediction data D with the second speed fluctuation threshold;

[0036] If the speed fluctuation prediction data is less than or equal to the second speed fluctuation threshold, it indicates that the speed fluctuation of the target turboexpander is slightly abnormal, and the nozzle opening of the target turboexpander is optimized by an intelligent optimization algorithm to obtain optimal nozzle opening data, and the nozzle opening of the target turboexpander is adjusted to the optimal nozzle opening data;

[0037] If the speed fluctuation prediction data is greater than the second speed fluctuation threshold, it means that the speed fluctuation of the target turbine expander is in a severe abnormal fluctuation. The intelligent optimization algorithm in steps S5111 to S5117 is used to optimize the intake valve opening of the target turbine expander to obtain the optimal intake valve opening data, and the nozzle opening of the target turbine expander is adjusted to the optimal nozzle opening data.

[0038] Preferably, the step of optimizing the nozzle opening of the target turboexpander in S511 to obtain optimal nozzle opening data includes the following steps:

[0039] S5111, set the current number of iterations to t and the maximum number of iterations to t max The nozzle opening search interval is [e1, e2], N nozzle opening data are randomly generated in the nozzle opening search interval, and the opening search set Q = {q1, q2, ..., q i ,…,q N}, where q i represents the i-th opening search data, and N represents the total number of opening search data;

[0040] Wherein, e1 and e2 represent the lower limit and upper limit of the target turboexpander nozzle opening, respectively;

[0041] S5112, calculate the fitness value of each opening search data in the opening search set, arrange each opening search data in the opening search set from large to small according to the fitness value, and select the one with the highest fitness value The opening search data are used as the opening search subject, and the remaining opening search data are used as the opening search objects. Each opening search subject is randomly occupied by a corresponding number of opening search objects according to the size ratio of the fitness value, and forms an opening search group with the occupied opening search objects; the fitness value calculation formula is as follows:

[0042]

[0043] Among them, f i represents the fitness value of the i-th opening search data in the opening search set, x i represents the speed fluctuation prediction data when the target expander nozzle opening is adjusted to the nozzle opening data corresponding to the i-th opening search data in the opening search set, and φ represents the correction value;

[0044] S5113. The opening search objects in each opening search group are updated in position around the corresponding opening search subject in the nozzle opening search interval. The position update formula is as follows:

[0045] F i new =F i +U(0,θ·β i )·α i ,

[0046] Among them, F i new They represent the position of the i-th opening search object in the opening search set after the position is updated, F i represents the current position of the i-th opening search object in the opening search set, β i represents the distance between the i-th opening search object and the corresponding opening search subject in the opening search set, α i Indicates the direction vector of the i-th opening search object in the opening search set pointing to the corresponding opening search subject, U(0,θ·β i ) means obeying [0,θ·β i ] uniformly distributed random numbers between ;

[0047] S5114. Update the mutation rate of the current iteration of the aperture search set. The mutation rate update formula is as follows:

[0048]

[0049] Among them, δ t represents the mutation rate of the opening search set during the current iteration, and δ0 represents the initial mutation rate;

[0050] A corresponding number of opening search objects are randomly selected from each opening search group according to the updated mutation rate, and the selected opening search objects adopt a random walk strategy to update their positions in the nozzle opening search interval;

[0051] S5115. Calculate the fitness value of each opening search data in the opening search set after the position is updated. If the fitness value of the opening search data after the position is updated is greater than the original fitness value, replace the original position with the new position of the opening search data; otherwise, retain the original position.

[0052] If the fitness value of the opening search object is greater than the fitness value of the corresponding opening search subject, the opening search object becomes the new opening search subject, and the original opening search subject becomes the opening search object;

[0053] S5116. Calculate the total fitness value of each opening search group. The total fitness value calculation formula is as follows:

[0054]

[0055] Among them, M i Represents the total fitness value of the i-th open search group represents the weight of the open search object, x i represents the fitness value of the opening search subject in the i-th opening search group, y ij represents the fitness value of the jth opening search object in the i-th opening search group, and n represents the total number of opening search objects in the i-th opening search group;

[0056] The lowest fitness value of the openness search object in the openness search group with the lowest total fitness value is selected, and the remaining openness search groups occupy the openness search object according to probability. If there is no openness search object in the openness search group with the lowest total fitness value, the openness search group is deleted, and the openness search subjects in the openness search group are assigned to the openness search group with the highest total fitness value. The probability calculation formula is as follows:

[0057]

[0058] Among them, P i It represents the probability that the i-th openness search group occupies the openness search object with the lowest fitness value within the openness search group with the lowest total fitness value, and m represents the total number of openness search groups;

[0059] S5117: Determine whether the current number of iterations t is greater than or equal to the maximum number of iterations t max , if the current number of iterations t is greater than or equal to the maximum number of iterations t max, the opening search data with the highest fitness value is output as the optimal nozzle opening data; otherwise, the current iteration number t is increased by 1, and the process returns to S5113.

[0060] The present invention preliminarily analyzes whether the speed fluctuation is within a reasonable range by numerically comparing the speed fluctuation prediction data with the first speed fluctuation threshold, and determines whether speed adjustment is required based on the analysis result. When speed adjustment is required, the speed fluctuation prediction data and the second speed fluctuation threshold are numerically compared, and a speed adjustment scheme that meets the current situation is further analyzed to achieve adaptive adjustment of the speed adjustment scheme to avoid over-adjustment and under-adjustment. The nozzle opening or intake valve opening of the target turbine expander is optimized through an intelligent optimization algorithm. After multiple iterative optimizations, the nozzle opening or intake valve opening that minimizes the speed fluctuation prediction data of the turbine expander is searched to achieve intelligent adjustment of the turbine expander speed. At the same time, during the algorithm iteration process, the mutation rate is adaptively adjusted according to the number of iterations, thereby improving the search performance of the algorithm and thereby improving the accuracy of the optimization results.

[0061] The present invention also includes a turbine expander speed regulation system based on natural gas pressure fluctuations, comprising a historical data acquisition module, an expander speed prediction model construction module, a real-time data acquisition module, a speed fluctuation prediction module, a speed regulation instruction determination module, and a speed regulation scheme execution module;

[0062] The historical data acquisition module collects online, through the expander speed regulation platform, several sets of historical natural gas pressure fluctuation characteristic data within a single natural gas pressure fluctuation detection period in the turbine expander operating state, as well as corresponding speed influencing factor data and speed fluctuation data, to obtain historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data;

[0063] The expander speed prediction model building module trains and tests the initial GRU model based on the historical natural gas pressure fluctuation characteristic data, the historical expander speed influencing factor data, and the historical expander speed fluctuation data to obtain the expander speed prediction model;

[0064] The real-time data acquisition module collects the natural gas pressure fluctuation characteristic data of the target turbo expander in real time through the pressure sensor installed in the target turbo expander; and obtains the real-time speed influencing factor data of the target turbo expander online through the expander speed adjustment platform;

[0065] The speed fluctuation prediction module inputs the real-time natural gas pressure fluctuation characteristic data and the real-time speed influencing factor data set into the expander speed prediction model to perform prediction analysis on the speed of the target turbine expander, thereby generating speed fluctuation prediction data of the target turbine expander;

[0066] The speed adjustment instruction determination module compares the speed fluctuation prediction data with a preset first speed fluctuation threshold value, generates speed adjustment instruction data according to the comparison result, and pushes the speed adjustment instruction data to the expander speed adjustment platform if adjustment is not required, thereby ending the current speed adjustment operation;

[0067] When the speed adjustment instruction data indicates that adjustment is required, the speed adjustment scheme execution module compares the speed fluctuation prediction data with a preset second speed fluctuation threshold, and selects a corresponding speed adjustment scheme to execute the speed adjustment operation based on the comparison result.

[0068] By means of the above technical solution, the present invention provides a method and system for adjusting the speed of a turbine expander based on natural gas pressure fluctuations, which has at least the following beneficial effects:

[0069] 1. The present invention constructs an expander speed prediction model by collecting a large amount of historical data, and accurately predicts the speed of the target turbine expander in combination with real-time natural gas pressure fluctuation characteristic data and real-time speed influencing factor data, generates speed fluctuation prediction data, and compares it with a preset first speed fluctuation threshold to determine whether the speed of the target turbine expander needs to be adjusted. When adjustment is required, a corresponding speed adjustment plan is accurately generated based on the speed fluctuation prediction data in combination with an intelligent optimization algorithm and the speed adjustment operation is executed, thereby realizing accurate detection and intelligent adjustment of the turbine expander speed.

[0070] 2. The present invention constructs an expander speed prediction model through historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data and historical expander speed fluctuation data, providing a reliable prediction model for realizing turbine expander speed prediction analysis, ensuring the accuracy and reliability of turbine expander speed fluctuation prediction results.

[0071] 3. The present invention performs a preliminary analysis on whether the speed fluctuation is within a reasonable range by numerically comparing the speed fluctuation prediction data with the first speed fluctuation threshold, and determines whether speed adjustment is required based on the analysis result. When speed adjustment is required, the speed fluctuation prediction data and the second speed fluctuation threshold are numerically compared to further analyze a speed adjustment scheme that meets the current situation, thereby realizing adaptive adjustment of the speed adjustment scheme and avoiding over-adjustment and under-adjustment.

[0072] 4. The nozzle opening or intake valve opening of the target turbine expander is optimized through an intelligent optimization algorithm. After multiple iterative optimizations, the nozzle opening or intake valve opening that minimizes the turbine expander speed fluctuation prediction data is searched, thereby realizing intelligent adjustment of the turbine expander speed. At the same time, during the algorithm iteration process, the mutation rate is adaptively adjusted according to the number of iterations, thereby improving the algorithm's search performance and thus improving the accuracy of the optimization results. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying any creative work.

[0074] Figure 1 A flow chart of the turbo expander speed regulation method provided by the present invention;

[0075] Figure 2 This is a module schematic diagram of the turbine expander speed regulation system provided by the present invention. DETAILED DESCRIPTION

[0076] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0077] Example 1 is as follows:

[0078] The existing speed regulation technology is difficult to respond to the fluctuation of natural gas in time and cannot dynamically select the speed regulation scheme according to the real-time situation. This embodiment proposes a turbine expander speed regulation method based on natural gas pressure fluctuation. Through the expander speed prediction model, the speed of the target turbine expander is accurately predicted in combination with real-time data, and the speed fluctuation prediction data is generated. It is compared with the preset first speed fluctuation threshold to determine whether the speed of the target turbine expander needs to be adjusted. When adjustment is required, the corresponding speed regulation scheme is accurately generated according to the speed fluctuation prediction data combined with the intelligent optimization algorithm and the speed regulation operation is executed, thereby realizing the accurate detection and intelligent adjustment of the turbine expander speed. Figure 1 As shown, the method includes the following steps:

[0079] S1. Collect historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data. As a specific implementation plan of this method, the detailed plan of this step is as follows:

[0080] S11, collect several sets of historical natural gas pressure fluctuation feature data within a single natural gas pressure fluctuation detection cycle in the working state of the turbine expander through the expander speed adjustment platform, and obtain the historical natural gas pressure fluctuation feature data set A = {a1, a2, ..., a i ,…,a k}, where a i represents the i-th group of natural gas pressure fluctuation characteristic data within a single natural gas pressure fluctuation detection cycle during the operation of the turboexpander in history, and k represents the total number of groups of natural gas pressure fluctuation characteristic data collected;

[0081] S12. Online collection of turbine expander speed influencing factor data corresponding to each historical natural gas pressure fluctuation characteristic data in the historical natural gas pressure fluctuation characteristic data set through the expander speed adjustment platform at the beginning of the natural gas pressure fluctuation detection period to obtain the historical expander speed influencing factor data set. represents the data of factors influencing the speed of the turbine expander at the beginning of the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data;

[0082] Data on factors affecting speed include but are not limited to nozzle opening, guide vane angle, impeller diameter, intake valve opening, rated speed, and load power;

[0083] S13, online collecting the turbine expander speed fluctuation data within the natural gas pressure fluctuation detection period corresponding to each historical natural gas pressure fluctuation feature data in the historical natural gas pressure fluctuation feature data set through the expander speed adjustment platform, and obtaining the historical expander speed fluctuation data set in, represents the speed fluctuation data of the turbine expander during the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, and represents the j-th speed data of the turbine expander collected during the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, represents the rated speed of the turbine expander within the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, and n represents the total number of turbine expander speed data collected within the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data.

[0084] S2. Constructing an expander speed prediction model based on historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data. As a specific implementation plan of this method, the detailed plan of this step is as follows:

[0085] S21. Setting a training data ratio, dividing the historical natural gas pressure fluctuation feature dataset, the historical expander speed influencing factor dataset, and the historical expander speed fluctuation dataset according to the training data ratio to obtain historical natural gas pressure fluctuation feature training data, historical expander speed influencing factor training data, historical expander speed fluctuation training data, historical natural gas pressure fluctuation feature test data, historical expander speed influencing factor test data, and historical expander speed fluctuation test data;

[0086] S22. Build the initial GRU model and set the training error threshold and test accuracy threshold;

[0087] S23, setting a maximum number of training times, inputting historical natural gas pressure fluctuation feature training data and historical expander speed influencing factor training data as training data, and historical expander speed fluctuation training data as training label data into the initial GRU model for training, and continuously adjusting the parameters of the initial GRU model according to the training results until the training error is less than the training error threshold or the number of training times is greater than the maximum number of training times, thereby obtaining a trained GRU model;

[0088] S24. Input the historical natural gas pressure fluctuation characteristic test data and the historical expander speed influencing factor test data as test data, and input the historical expander speed fluctuation test data as test label data into the trained GRU model for testing, and calculate the accuracy of the test results. If the accuracy of the test results is greater than or equal to the test accuracy threshold, the expander speed prediction model is obtained; otherwise, return to S23 for training until the accuracy of the test results is greater than or equal to the test accuracy threshold.

[0089] S3, collecting real-time natural gas pressure fluctuation characteristic data and real-time speed influencing factor data of the target turboexpander, as a specific implementation plan of the method, the detailed plan of this step is as follows:

[0090] S31, collecting natural gas pressure fluctuation characteristic data B of the target turboexpander in real time through a pressure sensor installed in the target turboexpander;

[0091] S32, obtain the real-time speed influencing factor data of the target turbine expander online through the expander speed adjustment platform, and obtain the real-time speed influencing factor data set C = {c1, c2, ..., c i ,…,c l}, where c irepresents the i-th real-time speed influencing factor data of the target turboexpander, and l represents the total number of real-time speed influencing factor data;

[0092] The real-time speed influencing factor data includes but is not limited to nozzle opening, guide vane angle, impeller diameter, intake valve opening, rated speed and load power.

[0093] S4. Predicting and analyzing the target turboexpander speed based on the real-time natural gas pressure fluctuation characteristic data, the real-time speed influencing factor data, and the expander speed prediction model to generate target turboexpander speed fluctuation prediction data. S41. Inputting the real-time natural gas pressure fluctuation characteristic data B and the real-time speed influencing factor data set into the expander speed prediction model to predict and analyze the target turboexpander speed to generate target turboexpander speed fluctuation prediction data D.

[0094] S5. Setting a speed fluctuation threshold, comparing the speed fluctuation prediction data with the speed fluctuation threshold, and generating speed adjustment instruction data;

[0095] If no adjustment is required, the speed adjustment operation is terminated;

[0096] If adjustment is required, the corresponding speed adjustment scheme is selected according to the speed fluctuation prediction data to perform the speed adjustment operation. As a specific implementation plan of this method, the detailed scheme of this step is as follows:

[0097] S511, comparing the speed fluctuation prediction data D with the second speed fluctuation threshold;

[0098] If the speed fluctuation prediction data is less than or equal to the second speed fluctuation threshold, it indicates that the speed fluctuation of the target turbo expander is in a slightly abnormal fluctuation, and the nozzle opening of the target turbo expander is optimized by an intelligent optimization algorithm to obtain optimal nozzle opening data, and the nozzle opening of the target turbo expander is adjusted to the optimal nozzle opening data;

[0099] If the speed fluctuation prediction data is greater than the second speed fluctuation threshold, it means that the speed fluctuation of the target turbine expander is in a severe abnormal fluctuation. The intelligent optimization algorithm in steps S5111 to S5117 is used to optimize the intake valve opening of the target turbine expander to obtain the optimal intake valve opening data, and the nozzle opening of the target turbine expander is adjusted to the optimal nozzle opening data.

[0100] In S511, the nozzle opening of the target turbo expander is optimized to obtain the optimal nozzle opening data, including the following steps:

[0101] S5111, set the current number of iterations to t and the maximum number of iterations to t maxThe nozzle opening search interval is [e1, e2], N nozzle opening data are randomly generated in the nozzle opening search interval, and the opening search set Q = {q1, q2, ..., q i ,…,q N}, where q i represents the i-th opening search data, and N represents the total number of opening search data;

[0102] Wherein, e1 and e2 represent the lower limit and upper limit of the target turboexpander nozzle opening, respectively;

[0103] S5112. Calculate the fitness value of each opening search data in the opening search set, arrange each opening search data in the opening search set from large to small according to the fitness value, and select the one with the highest fitness value. The opening search data are used as the opening search subject, and the remaining opening search data are used as the opening search objects. Each opening search subject is randomly occupied by a corresponding number of opening search objects according to the size ratio of the fitness value, and forms an opening search group with the occupied opening search objects; the fitness value calculation formula is as follows:

[0104]

[0105] Among them, f i represents the fitness value of the i-th opening search data in the opening search set, x i represents the speed fluctuation prediction data when the target expander nozzle opening is adjusted to the nozzle opening data corresponding to the i-th opening search data in the opening search set, and φ represents the correction value;

[0106] S5113. The opening search objects in each opening search group are updated in position around the corresponding opening search subject in the nozzle opening search interval. The position update formula is as follows:

[0107] F i new =F i +U(0,θ·β i )·α i ,

[0108] Among them, F i new They represent the position of the i-th opening search object in the opening search set after the position is updated, F i represents the current position of the i-th opening search object in the opening search set, β i represents the distance between the i-th opening search object and the corresponding opening search subject in the opening search set, α iIndicates the direction vector of the i-th opening search object in the opening search set pointing to the corresponding opening search subject, U(0,θ·β i ) means obeying [0,θ·β i ] uniformly distributed random numbers between ;

[0109] S5114. Update the mutation rate of the current iteration of the open search set. The mutation rate update formula is as follows:

[0110]

[0111] Among them, δ t It represents the mutation rate of the current iteration of the opening search set, and δ0 represents the initial mutation rate;

[0112] According to the updated mutation rate, a corresponding number of opening search objects are randomly selected from each opening search group. The selected opening search objects adopt a random walk strategy to update their positions in the nozzle opening search interval.

[0113] S5115. Calculate the fitness value of each opening search data in the opening search set after the position is updated. If the fitness value of the opening search data after the position is updated is greater than the original fitness value, replace the original position with the new position of the opening search data; otherwise, retain the original position.

[0114] If the fitness value of the opening search object is greater than the fitness value of the corresponding opening search subject, the opening search object becomes the new opening search subject, and the original opening search subject becomes the opening search object;

[0115] S5116. Calculate the total fitness value of each opening search group. The total fitness value calculation formula is as follows:

[0116]

[0117] Among them, M i Represents the total fitness value of the i-th open search group, represents the weight of the open search object, x i represents the fitness value of the opening search subject in the i-th opening search group, y ij represents the fitness value of the jth opening search object in the i-th opening search group, and n represents the total number of opening search objects in the i-th opening search group;

[0118] The lowest fitness value of the openness search object in the openness search group with the lowest total fitness value is selected, and the remaining openness search groups occupy the openness search object according to probability. If there is no openness search object in the openness search group with the lowest total fitness value, the openness search group is deleted, and the openness search subjects in the openness search group are assigned to the openness search group with the highest total fitness value. The probability calculation formula is as follows:

[0119]

[0120] Among them, P i It represents the probability that the i-th openness search group occupies the openness search object with the lowest fitness value within the openness search group with the lowest total fitness value, and m represents the total number of openness search groups;

[0121] S5117: Determine whether the current number of iterations t is greater than or equal to the maximum number of iterations t max , if the current number of iterations t is greater than or equal to the maximum number of iterations t max , the opening search data with the highest fitness value is output as the optimal nozzle opening data; otherwise, the current iteration number t is increased by 1, and the process returns to S5113.

[0122] The second embodiment is as follows:

[0123] See also Figure 2 , a turbine expander speed regulation system based on natural gas pressure fluctuation, including a historical data acquisition module, an expander speed prediction model construction module, a real-time data acquisition module, a speed fluctuation prediction module, a speed regulation instruction determination module and a speed regulation scheme execution module;

[0124] The historical data acquisition module collects online several sets of natural gas pressure fluctuation characteristic data and corresponding speed influencing factor data and speed fluctuation data within a single natural gas pressure fluctuation detection cycle in the working state of the turbine expander through the expander speed regulation platform, and obtains historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data;

[0125] The expander speed prediction model building module trains and tests the initial GRU model based on historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data to obtain the expander speed prediction model;

[0126] The real-time data acquisition module collects the natural gas pressure fluctuation characteristic data of the target turbo expander in real time through the pressure sensor installed in the target turbo expander; and obtains the real-time speed influencing factor data of the target turbo expander online through the expander speed adjustment platform;

[0127] The speed fluctuation prediction module inputs the real-time natural gas pressure fluctuation characteristic data and the real-time speed influencing factor data set into the expander speed prediction model to perform prediction analysis on the speed of the target turbine expander and generate speed fluctuation prediction data of the target turbine expander;

[0128] The speed adjustment instruction determination module compares the speed fluctuation prediction data with a preset first speed fluctuation threshold value, and generates speed adjustment instruction data based on the comparison result. If adjustment is not required, the speed adjustment instruction data is pushed to the expander speed adjustment platform, ending the speed adjustment operation.

[0129] When the speed adjustment instruction data indicates that adjustment is required, the speed adjustment scheme execution module compares the speed fluctuation prediction data with a preset second speed fluctuation threshold, and selects a corresponding speed adjustment scheme to execute the speed adjustment operation according to the comparison result.

[0130] Those skilled in the art will appreciate that all or part of the steps in the above-mentioned embodiment methods can be accomplished by instructing the relevant hardware through a program. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0132] The preferred embodiments of the invention disclosed above are intended only to help illustrate the invention. These preferred embodiments do not exhaust all details, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for adjusting the speed of a turbine expander based on natural gas pressure fluctuations, characterized in that: The steps include: S1. Collect historical natural gas pressure fluctuation characteristic data, historical expander speed influencing factor data, and historical expander speed fluctuation data; S2. Building an expander speed prediction model based on the historical natural gas pressure fluctuation characteristic data, the historical expander speed influencing factor data, and the historical expander speed fluctuation data; S3. Collecting real-time natural gas pressure fluctuation characteristic data and real-time speed influencing factor data of the target turboexpander; S4. Predicting and analyzing the speed of the target turbine expander based on the real-time natural gas pressure fluctuation characteristic data, the real-time speed influencing factor data, and the expander speed prediction model to generate speed fluctuation prediction data of the target turbine expander; S5. Setting a speed fluctuation threshold, comparing the speed fluctuation prediction data with the speed fluctuation threshold, and generating speed adjustment instruction data; If no adjustment is required, the speed adjustment operation is terminated; If adjustment is required, a corresponding speed adjustment scheme is selected according to the speed fluctuation prediction data to perform the speed adjustment operation.

2. The method for adjusting the speed of a turboexpander according to claim 1, wherein: Said S1 comprises the following steps: S11, collect several sets of historical natural gas pressure fluctuation feature data within a single natural gas pressure fluctuation detection cycle in the working state of the turbine expander through the expander speed adjustment platform, and obtain the historical natural gas pressure fluctuation feature data set A = {a1, a2, ..., a i ,…,a k }, where a i represents the i-th group of natural gas pressure fluctuation characteristic data within a single natural gas pressure fluctuation detection cycle during the operation of the turboexpander in history, and k represents the total number of groups of natural gas pressure fluctuation characteristic data collected; S12, collecting the turbine expander speed influencing factor data at the beginning of the natural gas pressure fluctuation detection period corresponding to each historical natural gas pressure fluctuation feature data in the historical natural gas pressure fluctuation feature data set online through the expander speed adjustment platform, and obtaining the historical expander speed influencing factor data set represents the data of factors influencing the speed of the turbine expander at the beginning of the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data; The speed influencing factor data include but are not limited to nozzle opening, guide vane angle, impeller diameter, intake valve opening, rated speed and load power; S13, online collecting the turbine expander speed fluctuation data within the natural gas pressure fluctuation detection period corresponding to each historical natural gas pressure fluctuation characteristic data in the historical natural gas pressure fluctuation characteristic data set through the expander speed adjustment platform, and obtaining the historical expander speed fluctuation data set in, represents the speed fluctuation data of the turbine expander during the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, and represents the j-th speed data of the turbine expander collected during the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, represents the rated speed of the turbine expander within the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data, and n represents the total number of turbine expander speed data collected within the natural gas pressure fluctuation detection period corresponding to the i-th historical natural gas pressure fluctuation characteristic data.

3. The method for adjusting the speed of a turboexpander according to claim 2, wherein: The S2 comprises the following steps: S21. Setting a training data ratio, and dividing the historical natural gas pressure fluctuation feature dataset, the historical expander speed influencing factor dataset, and the historical expander speed fluctuation dataset according to the training data ratio to obtain historical natural gas pressure fluctuation feature training data, historical expander speed influencing factor training data, historical expander speed fluctuation training data, historical natural gas pressure fluctuation feature test data, historical expander speed influencing factor test data, and historical expander speed fluctuation test data; S22. Build the initial GRU model and set the training error threshold and test accuracy threshold; S23, setting a maximum number of training times, inputting the historical natural gas pressure fluctuation characteristic training data and the historical expander speed influencing factor training data as training data, and the historical expander speed fluctuation training data as training label data into the initial GRU model for training, and continuously adjusting the parameters of the initial GRU model according to the training results until the training error is less than the training error threshold or the number of training times exceeds the maximum number of training times, thereby obtaining a trained GRU model; S24. Input the historical natural gas pressure fluctuation characteristic test data and the historical expander speed influencing factor test data as test data, and input the historical expander speed fluctuation test data as test label data into the trained GRU model for testing, and calculate the accuracy of the test results. If the accuracy of the test results is greater than or equal to the test accuracy threshold, the expander speed prediction model is obtained; otherwise, return to S23 for training until the accuracy of the test results is greater than or equal to the test accuracy threshold.

4. The method for adjusting the speed of a turboexpander according to claim 3, wherein: The S3 includes the following steps: S31, collecting natural gas pressure fluctuation characteristic data B of the target turboexpander in real time through a pressure sensor installed in the target turboexpander; S32, obtain the real-time speed influencing factor data of the target turbine expander online through the expander speed adjustment platform, and obtain the real-time speed influencing factor data set C = {c1, c2, ..., c i ,…,c l }, where c i represents the i-th real-time speed influencing factor data of the target turboexpander, and l represents the total number of real-time speed influencing factor data; The real-time speed influencing factor data includes but is not limited to nozzle opening, guide vane angle, impeller diameter, intake valve opening, rated speed and load power.

5. The method for adjusting the speed of a turboexpander according to claim 4, wherein: The S4 includes: The real-time natural gas pressure fluctuation characteristic data B and the real-time speed influencing factor data set are input into the expander speed prediction model to perform prediction analysis on the speed of the target turbine expander to generate speed fluctuation prediction data D of the target turbine expander.

6. The method for adjusting the speed of a turboexpander according to claim 5, wherein: The S5 comprises the following steps: Setting a first speed fluctuation threshold and a second speed fluctuation threshold, and performing a numerical comparison between the speed fluctuation prediction data D and the first speed fluctuation threshold; If the speed fluctuation prediction data is less than or equal to the first speed fluctuation threshold, it indicates that the speed fluctuation of the target turbine expander is within a reasonable range, and the speed adjustment instruction data is output as not requiring adjustment. The speed adjustment instruction data is pushed to the expander speed adjustment platform, and the speed adjustment operation is terminated. If the speed fluctuation prediction data is greater than the first speed fluctuation threshold, it means that the speed fluctuation of the target turbine expander is in an abnormal range, the output speed adjustment instruction data needs to be adjusted, and the corresponding speed adjustment scheme is selected according to the speed fluctuation prediction data.

7. The method for adjusting the speed of a turboexpander according to claim 6, wherein: The selecting of a corresponding speed adjustment scheme according to the speed fluctuation prediction data comprises the following steps: Comparing the speed fluctuation prediction data D with the second speed fluctuation threshold; If the speed fluctuation prediction data is less than or equal to the second speed fluctuation threshold, it indicates that the speed fluctuation of the target turboexpander is slightly abnormal, and the nozzle opening of the target turboexpander is optimized by an intelligent optimization algorithm to obtain optimal nozzle opening data, and the nozzle opening of the target turboexpander is adjusted to the optimal nozzle opening data; If the speed fluctuation prediction data is greater than the second speed fluctuation threshold, it means that the speed fluctuation of the target turbine expander is in a severe abnormal fluctuation. The intelligent optimization algorithm in steps S5111 to S5117 is used to optimize the intake valve opening of the target turbine expander to obtain the optimal intake valve opening data, and the nozzle opening of the target turbine expander is adjusted to the optimal nozzle opening data.

8. The method for adjusting the speed of a turboexpander according to claim 7, wherein: Optimizing the nozzle opening of the target turboexpander by using an intelligent optimization algorithm to obtain optimal nozzle opening data includes the following steps: S5111, build the opening search set, set the current number of iterations to t, the maximum number of iterations to t max The nozzle opening search interval is [e1, e2], N nozzle opening data are randomly generated in the nozzle opening search interval, and the opening search set Q = {q1, q2, ..., q i ,…,q N }, where q i represents the i-th opening search data, and N represents the total number of opening search data; Wherein, e1 and e2 represent the lower limit and upper limit of the target turboexpander nozzle opening, respectively; S5112, calculate the fitness value of each opening search data in the opening search set, arrange each opening search data in the opening search set from large to small according to the fitness value, and select the one with the highest fitness value The opening search data are used as the opening search subject, and the remaining opening search data are used as the opening search objects. Each opening search subject is randomly occupied by a corresponding number of opening search objects according to the size ratio of the fitness value, and forms an opening search group with the occupied opening search objects; the fitness value calculation formula is as follows: Among them, f i represents the fitness value of the i-th opening search data in the opening search set, x i represents the speed fluctuation prediction data when the target expander nozzle opening is adjusted to the nozzle opening data corresponding to the i-th opening search data in the opening search set, and φ represents the correction value; S5113. The opening search objects in each opening search group are updated in position around the corresponding opening search subject in the nozzle opening search interval. The position update formula is as follows: F i new =F i +U(0,θ·β i )·a i , Among them, F i new They represent the position of the i-th opening search object in the opening search set after the position is updated, F i represents the current position of the i-th opening search object in the opening search set, β i represents the distance between the i-th opening search object and the corresponding opening search subject in the opening search set, α i Indicates the direction vector of the i-th opening search object in the opening search set pointing to the corresponding opening search subject, U(0,θ·β i ) means obeying [0,θ·β i ] uniformly distributed random numbers between ; S5114. Update the mutation rate of the current iteration of the aperture search set. The mutation rate update formula is as follows: Among them, δ t represents the mutation rate of the opening search set during the current iteration, and δ0 represents the initial mutation rate; A corresponding number of opening search objects are randomly selected from each opening search group according to the updated mutation rate, and the selected opening search objects adopt a random walk strategy to update their positions in the nozzle opening search interval; S5115. Calculate the fitness value of each opening search data in the opening search set after the position is updated. If the fitness value of the opening search data after the position is updated is greater than the original fitness value, replace the original position with the new position of the opening search data; otherwise, retain the original position. If the fitness value of the opening search object is greater than the fitness value of the corresponding opening search subject, the opening search object becomes the new opening search subject, and the original opening search subject becomes the opening search object; S5116. Calculate the total fitness value of each opening search group. The total fitness value calculation formula is as follows: Among them, M i Represents the total fitness value of the i-th open search group, Indicates the weight of the open search object, x i represents the fitness value of the opening search subject in the i-th opening search group, y ij represents the fitness value of the jth opening search object in the i-th opening search group, and n represents the total number of opening search objects in the i-th opening search group; The lowest fitness value of the openness search object in the openness search group with the lowest total fitness value is selected, and the remaining openness search groups occupy the openness search object according to probability. If there is no openness search object in the openness search group with the lowest total fitness value, the openness search group is deleted, and the openness search subjects in the openness search group are assigned to the openness search group with the highest total fitness value. The probability calculation formula is as follows: Among them, P i It represents the probability that the i-th openness search group occupies the openness search object with the lowest fitness value within the openness search group with the lowest total fitness value, and m represents the total number of openness search groups; S5117: Determine whether the current number of iterations t is greater than or equal to the maximum number of iterations t max , if the current number of iterations t is greater than or equal to the maximum number of iterations t max , the opening search data with the highest fitness value is output as the optimal nozzle opening data; otherwise, the current iteration number t is increased by 1, and the process returns to S5113.

9. A system for implementing the turboexpander speed regulation method according to any one of claims 1 to 8.

Images