Rapid preselection method and device for sustainable aviation fuel, terminal and medium

Abstract

The invention belongs to the field of aviation fuel, and discloses a rapid preselection method and device for continuous aviation fuel, a terminal and a medium. The method comprises the steps of obtaining preprocessed fuel data, expanding original fuel data into new fuel data through a data enhancement technology, and inputting the new fuel data into a neural network model for prediction to obtain target fuel data. According to the method, the oil product database and the chemical group property database are combined to obtain a brand new fuel database, and in addition, the sustainable aviation fuel is rapidly and efficiently screened through the method.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of aviation fuel, and particularly relates to a rapid pre-selection method, device, terminal and medium of sustainable aviation fuel. BACKGROUND

[0002] With the rapid development of global aviation industry, the carbon emissions generated by air transportation increase year by year, which has a significant impact on climate change. Therefore, developing and using sustainable aviation fuel has become an important way for the aviation industry to reduce carbon emissions and achieve green development. Sustainable aviation fuel has similar performance and safety to traditional aviation fuel, but the carbon emissions are significantly reduced, which helps to achieve the sustainable development of the aviation industry.

[0003] In actual application, a large amount of manpower and material resources are usually needed to ensure the safety and reliability of sustainable aviation fuel. Therefore, how to efficiently and quickly screen safe and reliable sustainable aviation fuel is a problem to be solved. SUMMARY

[0004] The application aims to provide a rapid pre-selection method, device, terminal and medium of sustainable aviation fuel, which can.

[0005] In a first aspect, the application provides a rapid pre-selection method of sustainable aviation fuel, which comprises the following steps:

[0006] obtaining preprocessed fuel data;

[0007] expanding the original fuel data into new fuel data through a data enhancement technique;

[0008] inputting the new fuel data into an integrated network model for prediction to obtain target fuel data.

[0009] Optionally, the step of obtaining preprocessed fuel data comprises:

[0010] obtaining fuel chemical group mass content;

[0011] obtaining physicochemical information of the chemical group;

[0012] combining the obtained fuel chemical group mass content and the physicochemical information of the chemical group to obtain a fuel database, and extracting preprocessed fuel data from the fuel database.

[0013] Optionally, the step of obtaining fuel chemical group mass content comprises:

[0014] obtaining mass content of C5-C25 hydrocarbons from an oil product database.

[0015] Optionally, the step of augmenting the original fuel data into new fuel data using data augmentation techniques includes:

[0016] The original fuel data is augmented using multivariate autoencoder (VAE) models to generate new fuel data.

[0017] Optionally, inputting the new fuel data into an ensemble learning model for prediction to obtain target fuel data includes:

[0018] Based on the new fuel data and the constructed decision tree, generate the first predicted fuel data;

[0019] Based on the new fuel data and the gradient boosting tree algorithm, a second predicted fuel data is generated;

[0020] Based on the new fuel data and support vector machine, a third predicted fuel data is generated;

[0021] Based on the new fuel data and neural network model, a fourth set of predicted fuel data is generated.

[0022] Target fuel data is generated based on the first to fourth predicted fuel data.

[0023] Optionally, generating target fuel data based on the first to fourth predicted fuel data includes:

[0024] The first to fourth predicted fuel data are integrated using a meta-model to generate target fuel data.

[0025] Optionally, the method further includes:

[0026] The error between the target fuel data and the actual data is calculated using the mean square error.

[0027] Secondly, this application also provides a rapid pre-selection device for sustainable aviation fuel, which is functional in implementing the method in the first aspect or any possible implementation thereof. Specifically, the device includes modules for implementing the method in the first aspect or any possible implementation thereof.

[0028] In one embodiment, the device includes: an acquisition module, an expansion module, and a prediction module;

[0029] The acquisition module is used to acquire preprocessed fuel data;

[0030] The expansion module is used to expand the original fuel data into new fuel data using data augmentation technology;

[0031] The prediction module is used to input the new fuel data into the ensemble learning model for prediction to obtain the target data.

[0032] Optionally, the acquisition module includes:

[0033] The first acquisition unit obtains the mass content of fuel chemical groups;

[0034] The second acquisition unit obtains the physicochemical information of chemical groups;

[0035] The preprocessing unit combines the obtained mass content of fuel chemical groups with the physicochemical information of the chemical groups to obtain a fuel database, and extracts preprocessed fuel data from the fuel database.

[0036] Optionally, the first acquisition unit includes:

[0037] The extraction unit obtains the mass content of C5-C25 hydrocarbons from the oil database.

[0038] Optionally, the expansion module includes:

[0039] The expansion unit uses multivariate autoencoder (VAE) models to expand the original fuel data and generate new fuel data.

[0040] Optionally, the prediction module includes:

[0041] The first prediction unit generates the first predicted fuel data based on the new fuel data and the constructed decision tree;

[0042] The second prediction unit generates second predicted fuel data based on the new fuel data and the gradient boosting tree algorithm;

[0043] The third prediction unit generates third predicted fuel data based on the new fuel data and the support vector machine;

[0044] The fourth prediction unit generates fourth predicted fuel data based on the new fuel data and neural network model;

[0045] The target prediction unit generates target fuel data based on the first to fourth predicted fuel data.

[0046] Optionally, the target prediction unit includes:

[0047] The integration unit integrates the first to fourth predicted fuel data using a meta-model to generate target fuel data.

[0048] Optionally, the device further includes:

[0049] The error module is used to calculate the error value between the target fuel data and the actual data using the mean square error.

[0050] The correction module is used to correct the target fuel data using the error value.

[0051] Thirdly, this application also provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in any of the above-mentioned embodiments.

[0052] Fourthly, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in one of the preceding methods.

[0053] Fifthly, this application also provides a computer program product that, when run on an electronic device, causes the electronic device to execute any of the implementation methods of the first aspect described above.

[0054] The beneficial effects of this application embodiment compared with the prior art are as follows: When acquiring preprocessed fuel data, this application constructs a new sustainable aviation fuel database, which combines an oil database with a chemical group property database to obtain a new fuel database that is more suitable for screening sustainable aviation fuels. Subsequently, the fuel data is input into a neural network model for training and prediction, and finally the desired target fuel data is obtained. This method can screen out safe and reliable sustainable aviation fuels more efficiently and quickly. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 A flowchart illustrating the rapid pre-selection method provided in this application;

[0057] Figure 2 This is a schematic diagram of the process for obtaining preprocessed fuel data provided in this application;

[0058] Figure 3 A flowchart illustrating the process of obtaining physicochemical information of chemical groups provided in this application;

[0059] Figure 4 A schematic diagram of the process for obtaining target fuel data is provided for this application;

[0060] Figure 5 This application provides a schematic diagram of the process for generating the fourth predicted fuel data.

[0061] Figure 6This is a flowchart illustrating the process of correcting target fuel data based on error values, as provided in this application.

[0062] Figure 7 A schematic diagram of the rapid pre-selection device provided in this application;

[0063] Figure 8 A schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0064] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, circuits, and methods are omitted so as not to obscure the description of this application with unnecessary detail.

[0065] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0066] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0067] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0068] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0069] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0070] It should be understood that the sequence number of each step in this embodiment does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of this application embodiment.

[0071] In practical applications, to ensure the safety and reliability of Sustainable Aviation Fuels (SAFs), each candidate fuel requires a rigorous evaluation process, including laboratory testing, engine testing, and flight testing. These tests require large fuel samples (at least 1000 liters), time (at least 3 years), and cost (at least US$10 million), posing a significant obstacle to the development and dissemination of SAFs. Therefore, the pre-screening tool of this invention aims to address this problem and provide a more efficient and economical evaluation method. Recent progress has been made in pre-screening methods for evaluating SAFs using small sample sizes, which helps reduce the costs and risks associated with SAF certification. Previous research has explored various techniques, such as gas chromatography, Fourier transform infrared (FTIR), and nuclear magnetic resonance (NMR), to predict the performance of conventional aviation fuels. However, these methods all have limitations and exhibit high prediction errors.

[0072] Based on this, this application proposes a rapid pre-selection method for sustainable aviation fuel, which will be described below with reference to embodiments.

[0073] like Figure 1 As shown, the rapid pre-selection method described in this application includes:

[0074] S101, Obtain pre-processed fuel data;

[0075] S102, the original fuel data is augmented using data enhancement technology to create new fuel data;

[0076] S103, the fuel data is input into the ensemble learning model for prediction to obtain the target fuel data.

[0077] This application improves the efficiency of obtaining fuel data by preprocessing the fuel type and data. The preprocessed fuel data is then input into an ensemble learning model for prediction, ultimately yielding the predicted target fuel data.

[0078] For example, a pre-trained ensemble learning model is first loaded. This model is typically trained on a large amount of historical data to learn how to extract useful information from the input fuel data and make accurate predictions. The pre-processed fuel data is then fed as input to the neural network model, which processes the input data through a series of computational layers (such as fully connected layers, convolutional layers, etc.) and generates the output target fuel data.

[0079] This application establishes a unique database of sustainable aviation fuels based on their chemical properties. Aviation fuel data is retrieved from this database and then used as input to a trained neural network model to ultimately obtain the target fuel data. This method enables faster and more efficient acquisition of aviation fuel data.

[0080] In embodiments achievable under this application, such as Figure 2 As shown, the acquisition of preprocessed fuel data includes:

[0081] S201, to obtain the mass content of fuel chemical groups;

[0082] S202, to obtain the mass content of fuel chemical groups;

[0083] S203 combines the obtained mass content of fuel chemical groups with the physicochemical information of the chemical groups to obtain a fuel database, and extracts pre-processed fuel data from the fuel database.

[0084] It should be noted that, based on the composition information of aviation fuel, aviation fuel is divided into conventional aviation fuel, blended fuel, and sustainable aviation fuel. The proportion of sustainable aviation fuel in the fuel database is small and the amount of data is insufficient to support training. This application uses data augmentation technology to expand the data in the fuel database, increase the proportion of sustainable aviation fuel in the dataset, and thus obtain new fuel data, thereby enhancing the model's ability to learn and generalize on sustainable aviation fuel data.

[0085] Aviation fuels are typically composed of hundreds of hydrocarbons with eight main chemical groups, such as paraffins, isoalkanes, and aromatics, with the carbon number of these compounds primarily between C6 and C21. Due to the important physicochemical properties of these hydrocarbons, the information in this application should be taken into account for the ASTM-regulated properties of aviation fuels.

[0086] For example, the mass content of chemical groups of different batches of fuels can be obtained from a database and combined with the physicochemical information of chemical groups obtained through important sampling methods to obtain fuel data containing rich composition information.

[0087] First, collect the physicochemical information of C6-C21 hydrocarbons. Hydrocarbons are common chemical groups in aviation fuels, and their physicochemical properties have a significant impact on the overall performance of the fuel. Key physicochemical information such as density, viscosity, heat of combustion, and auto-ignition point of hydrocarbons can be obtained from chemical databases, literature, or experimental data.

[0088] These hydrocarbon compounds are then classified into eight categories based on their chemical composition. The classification can be based on the hydrocarbon's structural type (e.g., straight-chain alkanes, branched alkanes, cycloalkanes, aromatics, etc.) or carbon chain length. For example, straight-chain alkanes of C6-C10 can be classified as Category I, branched alkanes of C11-C15 as Category II, and so on, up to Category VIII.

[0089] Next, importance sampling is employed to obtain key physicochemical information for each chemical group. Importance sampling is a statistical method that assigns sampling probability based on the importance (or contribution) of each sample. In this example, the importance of each hydrocarbon class can be determined based on its impact on fuel performance, and sampling can be conducted accordingly. Importance sampling allows for more efficient acquisition of key physicochemical information crucial for fuel performance prediction.

[0090] After obtaining key physicochemical information for each chemical group, the mass content of chemical groups in multiple batches of different fuels was retrieved from the fuel database. This data reflects the distribution of various hydrocarbon compounds in different batches of fuel.

[0091] Finally, by combining the chemical group mass content data obtained from the database with the physicochemical information of chemical groups obtained through important sampling methods, a fuel data set containing rich compositional information is obtained. Based on this, the input fuel data includes both the mass content information of the fuel's chemical groups and the physicochemical information of those chemical groups, providing comprehensive data support for subsequent fuel performance prediction and optimization.

[0092] The fuel database established by the above steps incorporates the properties of aviation fuel and includes sustainable fuels that are more in line with the needs of the aviation industry. Preprocessed fuel data can be obtained more efficiently from the fuel database.

[0093] In embodiments achievable under this application, obtaining the mass content of fuel chemical groups includes:

[0094] The mass content of C5-C25 hydrocarbons was obtained from the oil database.

[0095] For example, taking aviation fuel Jet A-1 as an example, obtaining the mass content of chemical groups from the fuel database is crucial to ensuring the safety and performance of the fuel.

[0096] Jet A-1 fuel's chemical composition meets the stringent standards of the International Air Transport Association (IATA) and the International Civil Aviation Organization (ICAO). These standards include parameters such as sulfur content, benzene content, freezing point, and flash point. For example, the sulfur content must be below 0.3%, the benzene content below 1%, the freezing point below -47°C, and the flash point not lower than 38°C. These parameters ensure the fuel's cleanliness, combustion efficiency, and safety.

[0097] Obtain the mass content of various chemical groups in Jet A-1 fuel from the fuel database. These chemical groups may include alkanes, alkenes, aromatics, etc. The mass content of each group will affect the properties of the fuel. For example, a high content of alkane groups can improve the combustion stability and calorific value of the fuel; while the content of olefins and aromatics needs to be controlled within a certain range to avoid negative impacts on the engine and emissions.

[0098] This step allows us to obtain key information about the chemical composition of oil products, which helps to better utilize the combination of information on the mass content of chemical groups with the physicochemical information of chemical groups to obtain a fuel database that is more suitable for the aviation industry.

[0099] In embodiments achievable under this application, the step of augmenting the original fuel data into new fuel data using data enhancement techniques includes:

[0100] The original fuel data is augmented using multivariate autoencoder (VAE) models to generate new fuel data.

[0101] For example, assuming the latent space has a dimension of 10, the encoder compresses the raw fuel data into a representation of 10 main features, samples thousands or even tens of thousands of points from the latent space, and generates a corresponding number of new fuel data samples through the decoder. VAEs allow sampling from a continuous distribution in the latent space, and the generated new fuel data has a high degree of diversity.

[0102] By utilizing multivariate autoencoder (VAE) models, this application effectively expands the original fuel data and generates new fuel data samples. This method not only increases the quantity of original fuel data but also introduces new feature combinations to increase the diversity of the original fuel data, which helps to improve the generalization ability and performance of the model in subsequent data analysis and modeling tasks.

[0103] In embodiments achievable under this application, such as Figure 3 As shown, obtaining the physicochemical information of chemical groups includes:

[0104] S301, classifying chemical groups according to chemical categories;

[0105] S302, using importance sampling to sample the classified chemical groups to obtain the physicochemical information of the chemical groups.

[0106] For example, suppose the chemical groups in the aviation fuel studied in this application are classified into alkanes, alkenes, and aromatics. Then, using importance sampling, different sampling probabilities are set based on the relative content of these three types of chemical groups in the fuel and their influence on performance. Higher sampling probabilities are set for groups with greater influence to obtain their physicochemical information more frequently, while lower sampling probabilities are set for groups with less influence. This application efficiently obtains the physicochemical information of key chemical groups, such as density, viscosity, heat of combustion, and auto-ignition point, through importance sampling. This physicochemical information will provide an important basis for subsequent fuel performance analysis and fuel formulation optimization. This step provides efficient and accurate fuel data for subsequent predictions.

[0107] For example, this application combines a database of petroleum products, including aviation fuel, gasoline, and diesel, with a database covering the physicochemical properties of C6-C21 hydrocarbons. By combining these two different database structures, a fuel database containing more comprehensive physicochemical information is formed.

[0108] Suppose there exists an oil product database containing compositional information for various oil products such as aviation fuel, gasoline, and diesel, including the content of various hydrocarbon compounds, and the types and proportions of additives. This oil product database primarily focuses on the composition and formulation of the oil products. Simultaneously, there is another database covering the physicochemical properties of C6-C21 hydrocarbons. This hydrocarbon physicochemical property database records detailed physical and chemical properties of hydrocarbon compounds, such as density, viscosity, heat of combustion, auto-ignition point, flash point, and explosion limits. To create a fuel database containing more comprehensive physicochemical information, these two databases will be combined. The specific steps are as follows:

[0109] First, the correspondence between various hydrocarbon compounds in the oil product database and the corresponding hydrocarbon compounds in the physicochemical property database is identified by comparing their unique identifiers such as chemical structure, molecular formula, or CAS number.

[0110] After establishing the correspondence, the relevant physicochemical information from the physicochemical properties database is added to the oil product database to form the fuel database. In this way, the fuel database will not only contain its composition information but also the physicochemical properties of its key components.

[0111] During the integration process, the data is validated and corrected to ensure its accuracy and consistency. This may include checking the units, ranges, outliers, etc., and making necessary adjustments or additions.

[0112] To improve query efficiency and data management convenience, the integrated database can be optimized, such as by creating indexes and partitioning storage.

[0113] By following the steps above, the oil product database and the physicochemical property database can be combined to form a fuel database containing more comprehensive physicochemical information. This fuel database will provide a more complete and in-depth understanding of the composition and properties of various oil products, facilitating more efficient selection and use of aviation fuel and achieving sustainable development.

[0114] In embodiments achievable under this application, such as Figure 4 As shown, the step of inputting the new fuel data into the ensemble learning model for prediction to obtain the target fuel data includes:

[0115] S401, based on the new fuel data and the constructed decision tree, generates the first predicted fuel data;

[0116] S402, based on the new fuel data and the gradient boosting tree algorithm, generates second predicted fuel data;

[0117] S403 generates third predicted fuel data based on the new fuel data and support vector machine;

[0118] S404 generates fourth predicted fuel data based on new fuel data and neural network models;

[0119] S405, Target fuel data is generated based on the first to fourth predicted fuel data.

[0120] It should be noted that the ensemble learning model in this application mainly includes the following four basic learner models:

[0121] Random Forest: A machine learning method based on decision trees that improves model performance by constructing multiple decision trees and averaging or voting on their predictions. GBoost: A gradient boosting tree algorithm that sequentially trains multiple decision trees and uses gradient descent to minimize the loss function, gradually improving model performance. Support Vector Regression (SVR): Utilizes the idea of ​​Support Vector Machines (SVM) to fit data by finding the optimal hyperplane in the feature space, used to solve regression problems. Multi-Layer Neural Network (MLPRegressor): A machine learning algorithm based on artificial neural networks that learns the nonlinear relationship of input data through multi-layer neural networks to achieve regression tasks.

[0122] For example, assuming the new fuel data contains 100 samples, before steps S401 to S404, it is necessary to train decision trees, gradient boosting trees, support vector machines and neural network models using historical fuel data. Each model will output a predicted fuel dataset containing the predicted values ​​of 100 samples. The four predicted datasets are integrated to obtain the final target fuel dataset.

[0123] The ensemble model proposed in this application combines the advantages of multiple models and can provide more accurate prediction data than a single model.

[0124] In embodiments achievable under this application, generating target fuel data based on the first to fourth predicted fuel data includes:

[0125] The first to fourth predicted fuel data are integrated using a meta-model to generate target fuel data.

[0126] For example, suppose there are 1000 new fuel data samples and four basic models: decision tree, gradient boosting tree, support vector machine, and neural network. Based on the above four basic models, first to fourth predicted fuel data are generated. If linear regression is selected as the meta-model, the generated first to fourth predicted fuel data are used as the training set and validation set. For example, 80% of the data in the first to fourth predicted fuel data is used as the training set and 20% of the data is used as the validation set. Finally, the target fuel data is predicted using the meta-model.

[0127] Meta-models can optimally integrate the prediction results of ensemble learning models, capture the complementarity between different base models, and generate more accurate target fuel data. This application improves the predictive performance of the model through this step.

[0128] In embodiments achievable under this application, such as Figure 5 As shown, generating the fourth predicted fuel data based on the new fuel data and the neural network model includes:

[0129] S501, determine the fuel data to be retained based on the input vector of the current time step and the hidden state data of the previous time step;

[0130] S502 updates the retained fuel data in real time at each time step to obtain the latest fuel data;

[0131] S503, determine the output fuel data based on the input vector of the current time step and the hidden state data of the previous time step;

[0132] S504, calculate the hidden state data for the current time step based on the output fuel data and activation function;

[0133] S505: Based on the output fuel data and the hidden state data at the current time step, obtain the fourth fuel data.

[0134] For example, the neural model uses an input gate to control how much fuel data is passed to the storage unit. The input gate is calculated as a weighted sum of the input vector X at the current time step and the hidden state H at the previous time step, activated by a sigmoid function, resulting in a value between 0 and 1. This value determines how much fuel data will be passed to the storage unit.

[0135] The output of the input gate is multiplied by the stored state C from the previous time step to obtain the updated stored state, which represents the latest fuel data. This updated stored state contains the input information for the current time step and the memory from the previous time step.

[0136] The neural network model uses a forget gate to determine how much of the previous memory will be forgotten. The forget gate is calculated by a weighted sum of the input vector at the current time step and the hidden state at the previous time step, activated by a sigmoid function, resulting in a value between 0 and 1. This value determines how much of the previous memory will be forgotten.

[0137] The output of the forget gate is multiplied by the memory state of the previous time step to obtain the updated memory state. This updated memory state contains the current time step and the previously memorized input information, while forgetting some of the previous memory.

[0138] Neural network models use an output gate to determine how much memory will be output to the next time step. The output gate is calculated by weighting the input vector of the current time step and the hidden states of the previous time step, and then activating it with a sigmoid function to obtain a value between 0 and 1. This value determines how much memory will be output.

[0139] The output of the output gate is multiplied by the updated memory state, activated by the activation function (tanh), to obtain the hidden state at the current time step. The hidden state contains the output at the current time step.

[0140] Finally, by connecting a fully connected layer, the fourth fuel data is predicted based on the hidden state and the output gate output at the current time step.

[0141] This application uses a neural network model to efficiently predict the safest and most reliable target fuel for the current situation.

[0142] In embodiments achievable under this application, the hidden state data for the current time step includes: data for the current time step and information for all previous time steps.

[0143] For example, suppose we are using a recurrent neural network (RNN) to predict a fuel consumption sequence. In this sequence, each time step represents a specific time period (such as one hour, one day, etc.), and the input vector for each time step may contain relevant data such as fuel consumption, temperature, and pressure during that time period.

[0144] As the RNN processes data, a hidden state is generated at each time step. This hidden state reflects not only the input information at the current time step but also the accumulated information from all previous time steps. Specifically, when the RNN reaches the t-th time step, its hidden state h_t will be a function of the t-th input x_t and the (t-1)-th hidden state h_t-1. This function is typically implemented using nonlinear transformations (such as activation functions) and weight matrices.

[0145] The hidden state is essentially a "memory" unit that stores all relevant information from the beginning of the sequence to the current time step. This allows RNNs to leverage historical information to predict future fuel consumption or other relevant metrics. In practical applications, capturing this time dependency is crucial for fuel management. For example, when predicting fuel consumption, considering historical information such as previous consumption patterns and environmental factors can improve prediction accuracy. Similarly, understanding past experience and trends is essential when optimizing fuel formulations or adjusting fuel usage strategies. Therefore, integrating this information through the hidden states in RNNs provides strong support for fuel management.

[0146] In embodiments achievable under this application, such as Figure 6 As shown, the method further includes:

[0147] S601, calculates the error between the target fuel data and the actual data using mean square error;

[0148] S602 uses error values ​​to correct the target fuel data.

[0149] For example, suppose there is an RNN-based fuel prediction model that predicts future target fuel data based on preprocessed fuel data and other relevant factors.

[0150] The RNN model generates target fuel data, i.e., the model's predicted value, based on the input data. Next, the mean squared error (MSE) between the target fuel data and the actual data is calculated, yielding an error value. This error value reflects the accuracy of the model's prediction.

[0151] Based on the calculated MSE, the model is revised. For example, the weights of the RNN may be adjusted, or an optimization algorithm may be used to minimize the MSE. This process may involve multiple iterations until the MSE reaches a preset threshold (e.g., 10%), indicating that the error is now within a certain range. After revision, the MSE is recalculated and compared to the value before revision. If the MSE has decreased, the revision is effective, and the model's prediction accuracy has improved.

[0152] This process not only allows for the evaluation of model performance but also enables targeted corrections based on error values, thereby improving the accuracy of target fuel data predictions. This is of great significance for fuel management, cost control, and sustainable development.

[0153] The methods of the embodiments of this application have been described above with reference to the accompanying drawings. It should be noted that all numerical values ​​appearing above are merely illustrative and do not constitute specific limitations on this application. It should also be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially, these steps are not necessarily executed in the order shown in the figures. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Furthermore, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with other steps or at least a portion of the steps or stages of other steps. The system of the embodiments of this application will be described below with reference to the accompanying drawings. For brevity, appropriate omissions will be made when describing the system below; relevant content can be referred to in the description of the methods above and will not be repeated.

[0154] Figure 7 This is a schematic diagram of the structure of a rapid pre-selection device for sustainable aviation fuel provided in an embodiment of this application.

[0155] like Figure 7 As shown, the device 700 includes a data acquisition module 701, an expansion module 702, and a prediction module 703. The device 700 is capable of performing any of the rapid pre-selection methods for sustainable aviation fuel described above. For example, the data acquisition module 701 can be used to perform step S101, the expansion module 702 can be used to perform step S102, and the prediction module 703 can be used to perform step S103.

[0156] In embodiments achievable under this application, the acquisition module includes:

[0157] The first acquisition unit obtains the mass content of fuel chemical groups;

[0158] The second acquisition unit obtains the physicochemical information of chemical groups;

[0159] The preprocessing unit combines the obtained mass content of fuel chemical groups with the physicochemical information of the chemical groups to obtain a fuel database, and extracts preprocessed fuel data from the fuel database.

[0160] In embodiments achievable under this application, the first acquisition unit includes:

[0161] The extraction unit obtains the mass content of C5-C25 hydrocarbons from the oil database.

[0162] In embodiments achievable under this application, the expansion module includes:

[0163] The expansion unit uses multivariate autoencoder (VAE) models to expand the original fuel data and generate new fuel data.

[0164] In embodiments achievable under this application, the prediction module includes:

[0165] The first prediction unit generates the first predicted fuel data based on the new fuel data and the constructed decision tree;

[0166] The second prediction unit generates second predicted fuel data based on the new fuel data and the gradient boosting tree algorithm;

[0167] The third prediction unit generates third predicted fuel data based on the new fuel data and the support vector machine;

[0168] The fourth prediction unit generates fourth predicted fuel data based on the new fuel data and neural network model;

[0169] The target prediction unit generates target fuel data based on the first to fourth predicted fuel data.

[0170] In embodiments achievable under this application, the target prediction unit includes:

[0171] The integration unit integrates the first to fourth predicted fuel data using a meta-model to generate target fuel data.

[0172] In embodiments achievable under this application, the apparatus further includes:

[0173] The error module is used to calculate the error value between the target fuel data and the actual data using the mean square error.

[0174] The correction module is used to correct the target fuel data using the error value.

[0175] Figure 8This is a schematic diagram of the structure of the electronic device provided in an embodiment of this application. For example... Figure 8 As shown, the electronic device 800 of this embodiment includes: at least one processor 801 ( Figure 8 Only one is shown), memory 802, and computer program 803 stored in memory 802 and executable on at least one processor 801, which executes the steps in the above embodiments when the processor 801 executes the computer program 803.

[0176] The processor 801 can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0177] In some embodiments, memory 802 may be an internal storage unit of electronic device 800, such as a hard disk or memory of electronic device 800. In other embodiments, memory 802 may be an external storage device of electronic device 800, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on electronic device 800. Furthermore, memory 802 may include both internal and external storage units of electronic device 800. Memory 802 is used to store operating system, application programs, boot loader data, and other programs, such as program code of computer programs. Memory 802 may also be used to temporarily store data that has been output or will be output.

[0178] It should be noted that the information interaction and execution process between the above-mentioned units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0179] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units is merely an example. In practical applications, the above functions can be assigned to different functional units or modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0180] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps described in the various method embodiments above.

[0181] This application provides a computer program product that, when run on a computer, can implement the methods described above.

[0182] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographic device / electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0183] It should be understood that the sequence numbers of the steps in the above embodiments do not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. In the description, specific details such as particular system structures and technologies are set forth for illustrative purposes rather than for limiting purposes, so as to provide a thorough understanding of the embodiments of this application. However, those skilled in the art should understand that this application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of this application with unnecessary details.

[0184] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0185] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0186] Furthermore, in the description of this application and the appended claims, the terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0187] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0188] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0189] In the embodiments provided in this application, it should be understood that the disclosed apparatus, computer equipment, and methods can be implemented in other ways. For example, the apparatus and computer equipment embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0190] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A rapid pre-selection method for sustainable aviation fuel, characterized in that, The method includes: Acquire preprocessed fuel data; The original fuel data is augmented using data augmentation technology to create new fuel data; The new fuel data is input into the ensemble learning model for prediction to obtain the target fuel data.

2. The rapid pre-selection method for sustainable aviation fuel according to claim 1, characterized in that, The acquisition of preprocessed fuel data includes: Obtain the mass content of fuel chemical groups; To obtain physicochemical information of chemical groups; The obtained mass content of fuel chemical groups is combined with the physicochemical information of the chemical groups to obtain a fuel database, from which preprocessed fuel data is extracted.

3. The rapid pre-selection method for sustainable aviation fuel according to claim 2, characterized in that, The method of obtaining the mass content of fuel chemical groups includes: The mass content of hydrocarbons from C5 to C25 was obtained from the oil database.

4. The rapid pre-selection method for sustainable aviation fuel according to claim 2, characterized in that, The process of augmenting the original fuel data into new fuel data using data augmentation techniques includes: The original fuel data is augmented using multivariate autoencoder (VAE) models to generate new fuel data.

5. The rapid pre-selection method for sustainable aviation fuel according to claim 1, characterized in that, The step of inputting the new fuel data into the ensemble learning model for prediction to obtain the target fuel data includes: Based on the new fuel data and the constructed decision tree, generate the first predicted fuel data; Based on the new fuel data and the gradient boosting tree algorithm, a second predicted fuel data is generated; Based on the new fuel data and support vector machine, a third predicted fuel data is generated; Based on the new fuel data and neural network model, a fourth set of predicted fuel data is generated. Target fuel data is generated based on the first to fourth predicted fuel data.

6. The rapid pre-selection method for sustainable aviation fuel according to claim 5, characterized in that, The step of generating target fuel data based on the first to fourth predicted fuel data includes: The first to fourth predicted fuel data are integrated using a meta-model to generate target fuel data.

7. The rapid pre-selection method for sustainable aviation fuel according to claim 1, characterized in that, The method further includes: The error between the target fuel data and the actual data is calculated using the mean square error. The target fuel data is corrected using the error value.

8. A rapid pre-selection device for sustainable aviation fuel, characterized in that, The device includes: a data acquisition module, an expansion module, and a prediction module; The acquisition module is used to acquire preprocessed fuel data; The expansion module is used to expand the original fuel data into new fuel data using data augmentation technology; The prediction module is used to input the new fuel data into the ensemble learning model for prediction to obtain the target data.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images