Performance prediction model construction method and performance prediction method of turbo expander

Abstract

The invention relates to the field of system equipment maintenance, and discloses a performance prediction model construction method and a performance prediction method of a turbo expander. The performance prediction model construction method comprises the steps of obtaining multiple groups of first historical operation data sets and historical performance data corresponding to the multiple groups of first historical operation data sets; the multiple groups of first historical operation data sets are obtained on the basis of a turbine expansion unit comprising multiple working modes; each group of first historical operation data set corresponds to one working condition type; the turbo expander unit comprises a plurality of turbo expanders and at least one heat exchanger; and training the initial performance prediction model based on the multiple groups of first historical operation data sets and the historical performance data and working condition types corresponding to the multiple groups of first historical operation data sets to obtain a performance prediction model. According to the invention, model training is carried out by using the operation data under different working conditions, so that the model can predict the performance of the turbo expander more accurately.

Description

Technical Field

[0001] The present invention relates to the field of system equipment maintenance, and in particular to a performance prediction model construction method and a performance prediction method for a turbine expander. Background Art

[0002] Turbine expanders play a crucial role in hydrogen liquefaction systems, and their performance directly impacts the efficiency and stability of the entire system. Turbine expander performance is affected by a variety of factors. In actual operation, turboexpanders may face complex operating conditions, leading to performance fluctuations and even system downtime. Therefore, analyzing and accurately predicting turboexpander performance are crucial for optimizing system operation, improving energy efficiency, and reducing downtime. Summary of the Invention

[0003] In view of this, the present invention provides a method for constructing a performance prediction model of a turbine expander and a performance prediction method, which construct a prediction model of the turbine expander, thereby realizing the prediction of the performance data of the turbine expander.

[0004] In a first aspect, the present invention provides a method for constructing a performance prediction model for a turboexpander, the method comprising:

[0005] Acquire multiple sets of first historical operating data sets and historical performance data corresponding to the multiple sets of first historical operating data sets; the multiple sets of first historical operating data sets are obtained based on a turboexpander unit having multiple operating modes; each set of first historical operating data sets corresponds to a type of operating condition; the turboexpander unit includes multiple turboexpanders and at least one heat exchanger;

[0006] Based on multiple sets of first historical operation data sets, and historical performance data and operating condition types corresponding to the multiple sets of first historical operation data sets, an initial performance prediction model is trained to obtain a performance prediction model.

[0007] Through the performance prediction model construction method for a turbine expander provided in this embodiment, in the process of building the model, multiple groups of first historical operation data sets based on different working modes and operating conditions (such as high-load conditions, low-load conditions, variable load conditions, etc.) and their corresponding historical performance data are obtained. This can more comprehensively cover various possible situations of the turbine expander in actual operation, learn the performance change laws under different operating conditions, and provide a solid foundation for training the performance prediction model, thereby significantly improving the performance prediction model's prediction accuracy for the turbine expander performance.

[0008] In an optional embodiment, in a turbine expander unit, multiple turbine expanders and heat exchangers are connected in series; the operating modes include a first operating mode and a second operating mode; when the input device of the turbine expander is not a heat exchanger, the turbine expander is in the first operating mode; when the input device of the turbine expander is a heat exchanger, the turbine expander is in the second operating mode.

[0009] Through the above-described embodiment, by utilizing the operating data of the turbine expander unit in the first operating mode and the second operating mode, it is possible to train the model based on the operating data obtained under different operating modes and different operating conditions to improve the accuracy of the model prediction. Specifically, when the turbine expander is in the first operating mode, that is, the input device of the turbine expander is not a heat exchanger, its operating data reflects the characteristics of multi-stage expansion and cooling. When the turbine expander is in the second operating mode, that is, the input device of the turbine expander is a heat exchanger, the operating data reflects the characteristics of rapid pre-cooling and coordinated cooling. During the model training process, the model can learn the operating rules under different operating modes and different operating conditions, thereby improving the comprehensiveness and accuracy of the prediction.

[0010] In an optional embodiment, obtaining multiple sets of first historical operation data sets includes:

[0011] Obtain multiple sets of first historical operation data;

[0012] Clustering is performed on multiple groups of first historical operation data to obtain multiple groups of first historical operation data sets and an operating condition type corresponding to each group of first historical operation data sets; the operating condition type includes at least one of a high load operating condition, a low load operating condition, and a variable load operating condition.

[0013] Through the above implementation method, by clustering multiple groups of first historical operating data, the data are divided into different operating condition types (such as high-load operating conditions, low-load operating conditions, variable load operating conditions, etc.), so that the model can be trained for different operating condition types, ensuring that the model learns the operating data characteristics under different operating conditions, and improving the prediction accuracy of the performance prediction model.

[0014] In an optional embodiment, each group of first historical operating data contains multiple operating parameters; and the correlation between the operating parameters and the historical performance data meets a preset condition.

[0015] Through the above implementation, by ensuring that the correlation between operating parameters and historical performance data meets preset conditions, data that is directly related to performance and of high quality can be screened out. This helps reduce noise and redundancy in the data and improves the accuracy and reliability of subsequent data analysis.

[0016] In an optional embodiment, the method further includes:

[0017] Based on the multiple sets of first historical operating data sets, correlation analysis is performed on the operating parameters of each turbine expander in the turbine expander group to obtain correlation results between the operating parameters of each turbine expander.

[0018] In an optional embodiment, the method further includes:

[0019] Based on the correlation results between the operating parameters of each turbo expander, the operating strategy of each turbo expander is determined.

[0020] Through the above-described embodiment, correlation analysis of the operating parameters of each turboexpander in the turboexpander assembly based on multiple sets of first historical operating data sets can reveal the interrelationships between the operating parameters of each turboexpander. These correlation results can be used to formulate more scientific and efficient operating strategies, thereby optimizing equipment performance, improving operating efficiency, reducing energy consumption, and extending equipment life.

[0021] In an optional embodiment, the method further includes:

[0022] For each operating condition type, a correlation analysis is performed on the operating parameters corresponding to the operating condition type and the historical performance data to obtain the correlation results between the operating parameters corresponding to the operating condition type and the historical performance data;

[0023] Based on the correlation results between the operating parameters corresponding to the operating condition type and the historical performance data, the operating strategy of each turbo expander corresponding to the operating condition type is determined.

[0024] Through the above implementation, by analyzing the correlation between operating parameters and performance data under different operating conditions, the operating strategy of the expander under each operating condition is adjusted, thereby improving the performance and efficiency of the turboexpander. For example, under high-load conditions, the speed can be appropriately increased to improve adiabatic efficiency; under low-load conditions, the inlet pressure can be reduced to reduce energy consumption.

[0025] In a second aspect, the present invention provides a method for predicting performance of a turboexpander, the method comprising:

[0026] Acquire first operating data and an operating condition type corresponding to the first operating data; the first operating data is data of the target turbine expander at the current moment;

[0027] Based on the first operating data, the operating condition type, and the performance prediction model, the predicted performance data of the target turbine expander is determined; the performance prediction model is obtained through the first aspect or any implementation method of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0029] Figure 1 1 is a flow chart of a method for constructing a performance prediction model for a turboexpander according to an embodiment of the present invention;

[0030] Figure 2 1 is a flow chart of a method for constructing a performance prediction model for a turboexpander according to an embodiment of the present invention;

[0031] Figure 3 4 is a flow chart of a method for predicting performance of a turbo expander according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are 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 those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0033] In order to analyze the performance of a turbine expander and improve the accuracy of the performance prediction of a turbine expander, the embodiments of the present application provide a method for constructing a performance prediction model and a performance prediction method for a turbine expander. It should be noted that the execution subject of the method for constructing a performance prediction model and the performance prediction method for a turbine expander provided in the embodiments of the present application can be a device for constructing a performance prediction model for a turbine expander and a device for predicting the performance of a turbine expander. The device for constructing a performance prediction model for a turbine expander and the device for predicting the performance of a turbine expander can be implemented as part or all of an electronic device through software, hardware, or a combination of software and hardware. The electronic device can be a server or a terminal. The server in the embodiments of the present application can be a single server or a server cluster composed of multiple servers. The terminal in the embodiments of the present application can be a smart phone, a personal computer, a tablet computer, a wearable device, an intelligent robot, or other intelligent hardware devices. In the following method embodiments, the execution subject is an electronic device as an example for explanation.

[0034] According to an embodiment of the present invention, a method for constructing a performance prediction model of a turbine expander and an embodiment of a performance prediction method for a turbine expander are provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0035] In this embodiment, a method for constructing a performance prediction model of a turbo expander is provided, which can be used in the above-mentioned electronic devices, such as servers, etc. Figure 1 FIG. 1 is a flow chart of a method for constructing a performance prediction model for a turbo expander according to an embodiment of the present invention. Figure 1 As shown, the process includes the following S101-S103:

[0036] S101: Acquire multiple sets of first historical operation data sets and historical performance data corresponding to the multiple sets of first historical operation data sets.

[0037] Among them, multiple groups of first historical operation data sets are obtained based on a turbine expander unit including multiple operating modes; each group of first historical operation data sets corresponds to one operating condition type; the turbine expander unit includes multiple turbine expanders and at least one heat exchanger.

[0038] Specifically, the first historical operating data set includes multiple sets of historical operating data, each set of historical operating data including multiple operating parameters. Exemplary operating parameters may include inlet temperature, outlet temperature, inlet pressure, outlet pressure, flow rate, rotational speed, and the like. Performance data includes, but is not limited to, adiabatic efficiency, actual cooling capacity, expansion ratio, and the like. The first historical operating data set may correspond to a specific operating condition, such as a high-load condition, a low-load condition, or a variable-load condition.

[0039] Specifically, a turboexpander utilizes the high-speed airflow generated by the expansion of high-pressure gas to drive the rotation of its impeller, thereby achieving energy conversion and refrigeration. As the high-pressure gas undergoes adiabatic expansion within the expander and performs work, its energy enthalpy decreases, causing the gas itself to cool strongly, thereby achieving a cooling effect. The number of turboexpanders and heat exchangers included in a turboexpander unit, as well as the connection method (e.g., series or parallel) between multiple turboexpanders and heat exchangers, are not specifically limited herein. The operating modes included in a turboexpander unit can be categorized as high-load, low-load, and variable-load. Within a turboexpander unit, the turboexpander can operate in a first or second operating mode. The first operating mode can be understood as gradually cooling the hydrogen through the adiabatic expansion process of multiple turboexpander stages, ultimately achieving liquefaction. The second operating mode can be understood as rapidly pre-cooling the heat exchanger using an external cold source (e.g., liquid helium). After cooling the hydrogen in the heat exchanger, the hydrogen enters the turboexpander for further cooling.

[0040] In one possible implementation, multiple sensors can be installed in the turboexpander unit to collect multiple sets of first historical operating data. These sensors include temperature sensors, pressure sensors, and the like. For example, a temperature sensor can be installed at the inlet and outlet of the turboexpander to obtain the inlet and outlet temperatures. Another example is a pressure sensor installed at the inlet and outlet of the turboexpander to obtain the inlet and outlet pressures.

[0041] After obtaining the first historical operation data set, the performance data corresponding to the first historical operation data set can be further calculated based on the first historical operation data set. It should be noted that the implementation method of calculating the corresponding performance data based on the first historical data will be described in subsequent embodiments and will not be repeated here.

[0042] S102: Based on multiple sets of first historical operation data sets, and historical performance data and operating condition types corresponding to the multiple sets of first historical operation data sets, an initial performance prediction model is trained to obtain a performance prediction model.

[0043] Specifically, during the training of the initial performance model, the operating parameters and operating condition types in the first historical operating data are used as input features, and the performance data is used as output labels. The initial performance prediction model is trained using multiple sets of first historical operating data sets, as well as the historical performance data and operating condition types corresponding to the multiple sets of first historical operating data sets, to optimize the parameters in the model and obtain a trained performance prediction model.

[0044] Exemplarily, the initial performance prediction model may be a machine learning model, such as support vector regression (SVR), random forest, neural network, and the like.

[0045] Exemplarily, the first historical operating data set is divided into a training set and a test set. For example, 80% of the data set is used to train the performance prediction model, and 20% of the data set is used to evaluate the performance prediction model. During the performance prediction model evaluation process, the performance prediction model can be evaluated by calculating indicators such as the accuracy, precision, and recall rate of the model. If these indicators do not meet thresholds, the parameters in the performance prediction model are further adjusted to ensure that the performance prediction model can accurately predict the performance data of the turboexpander.

[0046] Through the performance prediction model construction method for a turbine expander provided in the embodiment of the present application, in the process of building the model, multiple groups of first historical operating data sets based on different operating conditions (such as high-load conditions, low-load conditions, variable load conditions, etc.) and their corresponding historical performance data are obtained, which can more comprehensively cover various possible situations of the turbine expander in actual operation, learn the performance change laws under different operating conditions, and provide a solid foundation for training the performance prediction model, thereby significantly improving the model's prediction accuracy of the turbine expander performance.

[0047] In some embodiments, a turboexpander assembly includes multiple turboexpanders and heat exchangers connected in series. The turboexpanders can operate in a first operating mode and a second operating mode. When the input device to the turboexpander is not a heat exchanger, the turboexpander is in the first operating mode; when the input device to the turboexpander is a heat exchanger, the turboexpander is in the second operating mode.

[0048] Specifically, when the turboexpander is in the first operating mode, the input device of the turboexpander is not a heat exchanger, i.e., the hydrogen directly enters the turboexpander for expansion and temperature reduction. When the turboexpander is in the second operating mode, the input device of the turboexpander is a heat exchanger, i.e., the hydrogen is first pre-cooled by the heat exchanger before entering the turboexpander for expansion and temperature reduction.

[0049] This application does not impose specific restrictions on the number of turboexpanders or heat exchangers in a turboexpander unit. For example, a turboexpander unit may include four turboexpanders and one heat exchanger. The four turboexpanders and one heat exchanger are connected in series, with the series order being turboexpander 1 - turboexpander 2 - heat exchanger - turboexpander 3 - turboexpander 4, and so on. In this turboexpander unit, the temperature of the hydrogen is gradually reduced through multi-stage expansion and precooling, ultimately achieving efficient liquefaction. Taking the series order of turboexpander 1 - turboexpander 2 - heat exchanger - turboexpander 3 - turboexpander 4 as an example, the hydrogen first enters turboexpander 1 for the first expansion and cooling. The cooled hydrogen then enters turboexpander 2 for the second expansion and cooling. The hydrogen is precooled by the heat exchanger, further reducing its temperature. The precooled hydrogen then enters turboexpander 3 for the third expansion and cooling. Finally, the hydrogen enters turboexpander 4 for the fourth expansion and cooling. Turbine expanders 1 and 2 are in the first operating mode, and turbo expanders 3 and 4 are in the second operating mode. This can also be understood as turbo expanders 1 and 2 operating in a two-stage expansion and pre-cooling mode, while turbo expanders 3 and 4 are in a heat exchanger-assisted liquefaction mode with a heat exchanger plus two sets of turbo expanders.

[0050] In an embodiment of the present application, by utilizing the operating data of the turbine expander unit in the first working mode and the second working mode, it is possible to train the model based on the operating data obtained under different working modes and different types of working conditions, so as to improve the accuracy of the model prediction. Specifically, when the turbine expander is in the first working mode, that is, the input device of the turbine expander is not a heat exchanger, its operating data reflects the characteristics of multi-stage expansion and cooling. When the turbine expander is in the second working mode, that is, the input device of the turbine expander is a heat exchanger, the operating data reflects the characteristics of rapid pre-cooling and coordinated cooling. During the model training process, the model can learn the operating rules under different working modes and different types of working conditions, thereby improving the comprehensiveness and accuracy of the prediction.

[0051] Of course, in other embodiments, the connection between the turbine expander and the heat exchanger in the turbine expander unit may also be parallel connection, or mixed connection, etc.

[0052] In some embodiments, in the above S101, multiple sets of first historical operation data sets are obtained by:

[0053] a1: Get multiple sets of first historical running data.

[0054] In one possible implementation, step a1 acquires multiple sets of first historical operating data by: first, operating the turboexpander unit according to preset operating parameters for a preset duration (e.g., one day). Then, controlling a valve or pump of the turboexpander unit to change the operating parameters of the turboexpander unit (e.g., operating the turboexpander unit in a low state for a predetermined duration). Finally, collecting the multiple sets of first historical operating data using a preset sensor.

[0055] In a possible implementation, the acquired first historical operation data is a data set that has undergone data preprocessing.

[0056] Due to sensor errors, electromagnetic interference, and other factors, the collected operating data may contain outliers and noise. Therefore, in the embodiments of the present application, the 3σ principle is used to remove outliers that significantly deviate from the normal range. For noisy data, a filtering algorithm, such as Kalman filtering, can be used for smoothing to improve data quality.

[0057] In addition, since the data collected by different types of sensors have different dimensions and numerical ranges, in order to facilitate subsequent data analysis, the Z-score normalization method is used to convert the operating data into a standard normal distribution with a mean of 0 and a standard deviation of 1.

[0058] For example, for a certain type of operating data, such as temperature, the normalization formula is:

[0059]

[0060] Among them, x new is the operating data after data preprocessing, that is, the first historical operating data, x is the operating data before data preprocessing, μ is the mean of this type of operating data, and σ is the standard deviation of this type of operating data.

[0061] In a possible implementation, each group of first historical operating data contains multiple operating parameters; and the correlation between the operating parameters and the historical performance data meets a preset condition.

[0062] The operating parameters include, but are not limited to, temperature, pressure, flow rate, rotation speed, etc. The performance data include, but are not limited to, adiabatic efficiency, expansion ratio, etc.

[0063] Optionally, the correlation between the operating parameter and the historical performance data can be represented by a correlation coefficient, such as the Pearson correlation coefficient. In this case, the correlation between the operating parameter and the historical performance data satisfies the preset condition when the correlation coefficient between the operating parameter and the historical performance data is greater than or equal to a preset threshold. The preset threshold can be set based on actual conditions and is not limited here.

[0064] The correlation coefficient reflects the degree of association between operating parameters and performance data, reflecting the degree of influence of operating parameters on performance data. For example, the inlet temperature of a turboexpander is negatively correlated with the adiabatic efficiency, that is, as the inlet temperature increases, the adiabatic efficiency decreases.

[0065] In addition, the correlation between the corresponding operating parameters of the turbine expanders in the turbine expander group can be calculated to determine the mutual influence of the operating parameters between the turbine expanders. Continuing with the turbine expander group (Turbine Expander 1 - Turbine Expander 2 - Heat Exchanger - Turbine Expander 3 - Turbine Expander 4) as an example, based on the correlation coefficients between the operating parameters of the four series-connected turbine expanders, the correlation between the operating parameters of the turbine expanders can be determined. For example, by calculating the correlation coefficients between the operating parameters of two adjacent turbine expanders, it is found that the outlet pressure and temperature of the previous turbine expander will significantly affect the inlet conditions of the next turbine expander, such as the inlet temperature and inlet pressure, and thus affect the performance data of the next turbine expander. Based on this, the operating control strategy of the series-connected turbine expanders can be optimized to improve the refrigeration performance of the entire turbine expander group.

[0066] In the embodiments of the present application, by ensuring that the correlation between operating parameters and historical performance data meets preset conditions, data that is directly related to performance and of high quality can be screened out. This helps reduce noise and redundancy in the data and improves the accuracy and reliability of subsequent data analysis.

[0067] a2: Clustering multiple groups of first historical operation data to obtain multiple groups of first historical operation data sets and the operating condition type corresponding to each group of first historical operation data sets.

[0068] The operating condition type includes at least one of a high load operating condition, a low load operating condition, and a variable load operating condition.

[0069] Specifically, clustering is an unsupervised learning method used to divide multiple groups of first historical operation data into several groups (i.e., multiple groups of first historical operation data sets) so that the data similarity within the same group is high, while the data similarity between different groups is low. By clustering these first historical operation data, similar operating states can be classified into the same group, thereby identifying different operating condition types. For example, a clustering algorithm can be used to cluster the multiple groups of first historical operation data, such as the K-means algorithm, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, and the like.

[0070] Specifically, a high-load operating condition refers to the operation of a turbine expander under a first preset operating condition. The first preset operating condition is a condition that approaches or reaches the maximum design load of the equipment. Here, the first preset operating condition can be understood as a preset operating parameter (such as the speed, etc.) reaching or exceeding a first threshold. At this time, the operating parameters of the turbine expander are generally at a relatively high level. A low-load operating condition refers to the operation of a turbine expander under a second preset operating condition. Here, the second preset operating condition can be understood as a preset operating parameter being less than a second threshold. It can also be understood that under a low-load operating condition, the turbine expander operates under conditions far below its designed load. At this time, the operating parameters of the turbine expander are generally at a relatively low level. A variable load operating condition refers to the operation of a turbine expander under conditions where the load conditions frequently fluctuate. Here, meeting the condition of frequent fluctuations means that the rate of change of a preset operating parameter is greater than a third threshold. At this time, the operating parameters of the equipment (such as speed, pressure, temperature, etc.) will fluctuate with changes in load.

[0071] For example, the performance characteristics of a turboexpander vary for different operating conditions. The performance characteristics of a turboexpander can be analyzed separately for each operating condition to provide a basis for targeted optimization. For example, under high-load conditions, the turboexpander has a large flow rate and noticeable temperature fluctuations. Under low-load conditions, the flow rate is small and the temperature is relatively stable. Under variable load conditions, various operating parameters fluctuate significantly. For different operating conditions, corresponding operation optimization strategies can be formulated. For example, under high-load conditions, the speed can be appropriately increased to improve adiabatic efficiency; under low-load conditions, the speed can be reduced to reduce energy consumption.

[0072] In an embodiment of the present application, by clustering multiple groups of first historical operating data, the data is divided into different operating condition types (such as high-load operating conditions, low-load operating conditions, variable load operating conditions, etc.), so that the model can be trained for different operating condition types, ensuring that the model learns the operating data characteristics under different operating conditions and improving the prediction accuracy of the model.

[0073] In some embodiments, the first historical operation data set includes multiple sets of first historical operation data, each set of first historical operation data corresponds to performance data, and the historical performance data corresponding to the first historical operation data set is the performance data corresponding to each set of first historical operation data.

[0074] In one possible implementation, when the first historical operating data includes the temperature and pressure of the inlet and outlet, and the performance data includes the adiabatic efficiency, the adiabatic efficiency is calculated by using the temperature and pressure of the inlet and outlet of the turbine expander, the ideal gas state equation and the adiabatic expansion process equation.

[0075] For example, the formula for calculating adiabatic efficiency is as follows:

[0076]

[0077] Among them, η s is the adiabatic efficiency, h1 and h2 are the actual inlet and outlet enthalpies, and h 2s is the enthalpy after isentropic expansion. The enthalpy can be calculated using the inlet and outlet temperatures and pressures combined with the thermodynamic properties table of hydrogen.

[0078] In one possible implementation, when the first historical operating data includes the temperature and pressure of the inlet and outlet, and the performance data includes the adiabatic efficiency, the enthalpy difference of the hydrogen at the inlet and outlet of the turbine expander is determined based on the temperature and pressure of the inlet and outlet combined with the thermodynamic property table of hydrogen according to the law of conservation of energy.

[0079] In a possible implementation, when the first historical operation data includes the inlet and outlet pressures, and the performance data includes the expansion ratio, the expansion ratio is calculated based on the inlet and outlet pressures.

[0080] Exemplarily, the expansion ratio is calculated as follows:

[0081]

[0082] Wherein, π is the expansion ratio, P1 is the inlet pressure, and P2 is the outlet pressure. For example, the pressures at the inlet and outlet of the turboexpander can be measured by a pressure sensor.

[0083] In some embodiments, the method for constructing a performance prediction model for a turboexpander provided in this application further includes the following:

[0084] Based on the multiple sets of first historical operating data sets, correlation analysis is performed on the operating parameters of each turbine expander in the turbine expander group to obtain correlation results between the operating parameters of each turbine expander.

[0085] Optionally, based on multiple sets of first historical operation data sets, correlation coefficients between the turbine expanders are calculated to obtain correlation results between the operating parameters of the turbine expanders.

[0086] Furthermore, based on the correlation results between the operating parameters of each turbo expander, an operating strategy for each turbo expander is determined, wherein the operating strategy includes but is not limited to adjusting the operating parameters.

[0087] Continuing with the turboexpander unit (Turboexpander 1 - Turboexpander 2 - Heat Exchanger - Turboexpander 3 - Turboexpander 4) described above, analysis of the operating data of the four series-connected turboexpanders revealed that for every 1K increase in the outlet temperature of the preceding turboexpander, the inlet temperature of the following turboexpander increased by an average of 0.8K, and the adiabatic efficiency of the following turboexpander decreased by an average of approximately 1.5%. This indicates that the outlet temperature of the preceding turboexpander must be strictly controlled to ensure the performance of the subsequent turboexpanders.

[0088] In this way, correlation analysis of the operating parameters of each turboexpander in the turboexpander unit, based on multiple sets of first historical operating data sets, can reveal the interrelationships between the operating parameters of each turboexpander. These correlation results can be used to formulate more scientific and efficient operating strategies, thereby optimizing equipment performance, improving operating efficiency, reducing energy consumption, and extending equipment life.

[0089] In some embodiments, the method for constructing a performance prediction model for a turboexpander provided in this application further includes the following:

[0090] First, for each operating condition type, a correlation analysis is performed on the operating parameters corresponding to the operating condition type and the historical performance data to obtain the correlation results between the operating parameters corresponding to the operating condition type and the historical performance data.

[0091] Specifically, for each operating condition type, the correlation between the operating parameter corresponding to the operating condition type and the historical performance data can be a positive correlation, a negative correlation, or the like. For example, the speed of a turboexpander is positively correlated with its adiabatic efficiency. For example, the relationship between the operating parameter and the historical performance data can be characterized by a correlation coefficient, such as the Pearson correlation coefficient.

[0092] Then, based on the correlation results between the operating parameters corresponding to the operating condition type and the historical performance data, the operating strategy of each turbo expander corresponding to the operating condition type is determined.

[0093] For example, under high load conditions, within a certain range, the speed can be increased to improve the insulation efficiency. Under low load conditions, the speed can be reduced to reduce energy consumption.

[0094] In the embodiment of the present application, by analyzing the correlation between operating parameters and performance data under different working conditions, the operating strategy of the expander under the working condition type is adjusted, thereby improving the performance of the turbine expander and improving the efficiency of the turbine expander.

[0095] Figure 2 The following is a flow chart of a method for constructing a prediction model. The specific implementation steps of the method for constructing a prediction model include the following steps S201-S207:

[0096] S201: Obtain original historical operation data.

[0097] S202: Preprocessing the original historical operation data to obtain preprocessed original historical operation data.

[0098] The implementation of data preprocessing has been described in the above embodiments and will not be repeated here.

[0099] S203: Calculate performance data corresponding to the pre-processed original historical operation data.

[0100] The calculation method of the performance data corresponding to the operating data has been described in the above embodiment and will not be repeated here.

[0101] S204: Determine multiple groups of first historical operation data and multiple groups of second historical operation data based on correlation analysis between the pre-processed original historical operation data and the corresponding performance data.

[0102] Here, the correlation analysis based on the pre-processed original historical operation data and the corresponding performance data refers to the correlation analysis between the operation parameters in the pre-processed original historical operation data and the corresponding performance data.

[0103] Optionally, the result of the correlation analysis is characterized based on a Pearson correlation coefficient, wherein the Pearson correlation coefficient between the operating parameters and the performance data in the first historical operating data and the second historical operating data satisfies a preset threshold.

[0104] For example, the Pearson correlation coefficient shows that the speed of the turboexpander is positively correlated with the adiabatic efficiency, indicating that increasing the speed can improve the adiabatic efficiency within a certain range.

[0105] S205: Clustering the multiple groups of first historical operation data and the multiple groups of second historical operation data respectively to obtain multiple groups of first historical operation data sets and multiple groups of second historical operation data sets.

[0106] For example, the K-Means algorithm is used to divide the operating conditions into three categories, namely high-load conditions, low-load conditions, and variable-load conditions. Under high-load conditions, the inlet and outlet pressures and flows of the turbine expander are large, and the temperature changes are also more obvious. Under low-load conditions, the pressure and flow are small, and the temperature is relatively stable. Under variable-load conditions, various data (such as temperature) fluctuate greatly. For different operating conditions, corresponding operation optimization strategies need to be formulated separately. For example, under high-load conditions, the speed is appropriately increased to improve the adiabatic efficiency; under low-load conditions, the inlet pressure is reduced to reduce energy consumption.

[0107] S206: Constructing a prediction model based on the multiple sets of first historical operation data sets and the multiple sets of second historical operation data sets.

[0108] The prediction model includes a performance prediction model, which is trained based on multiple sets of first historical operation data sets and is used to predict performance data of the turbine expander.

[0109] For example, when the prediction model is a performance prediction model, an SVR algorithm is employed, using turboexpander operating parameters (such as inlet temperature, pressure, and speed) as input variables and a performance indicator (adiabatic efficiency) as output variable to construct the performance prediction model. The performance prediction model is trained and optimized using a large amount of historical data, enabling the performance prediction model to accurately predict turboexpander performance data under different operating conditions.

[0110] S207: Evaluate the test results using the prediction model.

[0111] For example, the performance prediction model built using SVR was used to predict turboexpander performance data for a specific period of time. The predicted performance data was compared with actual performance data during actual operation. The results showed that the prediction error for adiabatic efficiency was within ±3%, and the prediction error for actual cooling capacity was within ±5%. This demonstrates the high accuracy of the performance prediction model and provides reliable decision-making support for operators.

[0112] Figure 3 Schematic diagram of a performance prediction method for a turbine expander provided in an embodiment of the present application. Figure 3 As shown, the performance prediction method of the turbo expander includes the following S301-S302:

[0113] S301: Acquire first operating data and the operating condition type corresponding to the first operating data.

[0114] The first operating data is the data of the target turbine expander at the current moment.

[0115] S302: Determine predicted performance data of the target turboexpander based on the first operating data, the operating condition type, and the performance prediction model.

[0116] The performance prediction model is obtained by the above-mentioned method for constructing a performance prediction model of a turbine expander.

[0117] Specifically, the first operating data and the operating condition type are input into a performance prediction model to determine predicted performance data for the target turboexpander. The predicted performance data may be performance data at a certain point in the future. This predicted performance data at the future point in time can help operators understand the operating status of the equipment and adjust the turboexpander's operating strategy in a timely manner.

[0118] The above mainly introduces the solution provided in the embodiment of the present application from the perspective of method.

[0119] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for constructing a performance prediction model for a turboexpander, characterized in that: The method comprises: Acquiring multiple sets of first historical operating data sets and historical performance data corresponding to the multiple sets of first historical operating data sets; the multiple sets of first historical operating data sets are obtained based on a turboexpander unit having multiple operating modes; each set of the first historical operating data sets corresponds to a type of operating condition; the turboexpander unit includes multiple turboexpanders and at least one heat exchanger; Based on the multiple sets of first historical operation data sets, and the historical performance data and operating condition types corresponding to the multiple sets of first historical operation data sets, an initial performance prediction model is trained to obtain a performance prediction model.

2. The method according to claim 1, characterized in that In the turbine expander unit, the multiple turbine expanders and the heat exchanger are connected in series; the operating mode includes a first operating mode and a second operating mode; when the input device of the turbine expander is not the heat exchanger, the turbine expander is in the first operating mode; when the input device of the turbine expander is the heat exchanger, the turbine expander is in the second operating mode.

3. The method according to claim 1 or 2, characterized in that The obtaining of multiple sets of first historical operation data sets includes: Obtain multiple sets of first historical operation data; Clustering the multiple groups of first historical operation data to obtain the multiple groups of first historical operation data sets and the operating condition type corresponding to each group of the first historical operation data sets; the operating condition type includes at least one of a high load operating condition, a low load operating condition, and a variable load operating condition.

4. The method according to claim 3, characterized in that Each group of the first historical operating data contains multiple operating parameters; the correlation between the operating parameters and the historical performance data meets a preset condition.

5. The method according to claim 4, characterized in that The method further comprises: Based on the multiple groups of first historical operating data sets, correlation analysis is performed on the operating parameters of each of the turbine expanders in the turbine expander group to obtain correlation results between the operating parameters of each of the turbine expanders.

6. The method according to claim 5, characterized in that The method further comprises: Based on the correlation results between the operating parameters of each of the turbo expanders, an operating strategy for each of the turbo expanders is determined.

7. The method according to claim 4, characterized in that The method further comprises: For each operating condition type, performing a correlation analysis on the operating parameters corresponding to the operating condition type and the historical performance data to obtain a correlation result between the operating parameters corresponding to the operating condition type and the historical performance data; Based on the correlation results between the operating parameters corresponding to the operating condition type and the historical performance data, the operating strategy of each of the turbine expanders corresponding to the operating condition type is determined.

8. A method for predicting the performance of a turboexpander, characterized in that: The method comprises: Acquire first operating data and an operating condition type corresponding to the first operating data; the first operating data is data of a target turbine expander at a current moment; Based on the first operating data, the operating condition type, and a performance prediction model, the predicted performance data of the target turbine expander is determined; the performance prediction model is obtained by the performance prediction model construction method of the turbine expander according to any one of claims 1-7.

Images