Method and device for evaluating application effect of ship energy efficiency measures

Abstract

The invention provides a ship energy efficiency measure application effect evaluation method and device. The evaluation method comprises the steps of obtaining ship multi-source heterogeneous data; performing fusion processing on the multi-source heterogeneous data to obtain target data; dividing the target data into first data and second data; inputting the first data into a second oil consumption prediction model for oil consumption prediction to obtain a first prediction result; inputting the second data into a first oil consumption prediction model for oil consumption prediction to obtain a second prediction result; the first oil consumption prediction model and the second oil consumption prediction model are obtained by training a preset machine learning model; obtaining a first evaluation result according to the first prediction result and the first actual fuel consumption value; obtaining a second evaluation result according to the second prediction result and the second actual fuel consumption value; and obtaining a target evaluation result according to the first evaluation result and the second evaluation result, and outputting the target evaluation result. The efficiency and convenience of ship energy efficiency measure application effect evaluation can be improved, and the evaluation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of computer information processing, and in particular to a method and a device for evaluating the application effect of ship energy efficiency measures. Background Art

[0002] Ship technical energy efficiency measures include hull line optimization, ship lightweighting, air film drag reduction, coating drag reduction, high-efficiency propellers, hydrodynamic energy-saving devices, waste heat recovery, etc. Operational energy efficiency measures include weather route delineation, speed optimization, optimal shaft power, trim optimization, route optimization, hull cleaning, digital solutions, etc. During the operation of ships, multiple technologies and operational energy efficiency measures are usually applied at the same time. How to accurately evaluate the comprehensive effect of one or more energy-saving and carbon-reduction measures applied in actual ship operations has become a key to whether the shipping industry can achieve net zero emissions in the future.

[0003] The existing methods for evaluating the application effect of ship energy efficiency measures mainly include ship energy efficiency dynamic simulation method, ship model test method, and actual ship trial method. The ship energy efficiency dynamic simulation method has high computing resource requirements and requires high-performance computers to handle complex simulation tasks; the model accuracy is limited, and simplified assumptions and turbulence modeling may affect the accuracy of the results; verification and confirmation are difficult, and ensuring the effectiveness and accuracy of the model requires rich experience and data support; the user's skills are high, and professional knowledge and experience are required to set up and interpret the results. The ship model test method will produce a scale effect. Although it follows the law of similarity, it is difficult to completely eliminate the impact of scale differences in actual operation; the cost is high, and high-quality model production, maintenance, and operation of large-scale experimental facilities require a lot of capital investment; the time is long, and it often takes a long period from model production to completing a series of experiments; the scope is limited and cannot fully represent the behavior of full-size ships, especially when nonlinear effects are involved. The actual ship trial method is costly, involving fuel, manpower, time and potential risk costs; it is restricted by weather and needs to choose a suitable weather window; the amount of data is limited, and the amount of data obtained is relatively small compared to the dynamic simulation of ship energy efficiency and ship model testing; the trial conditions and sea conditions are small, and it is difficult to cover the actual operation of the ship. The existing evaluation methods are all based on simplified simulations and tests of the actual operation of the ship. The calculated application effects of energy efficiency measures have certain reference value, but it is difficult to fully consider the complex actual operation of the ship, and ships of different ship types and tonnages need to be re-modeled and calculated, which is costly and time-consuming, and it is difficult for non-professional staff to evaluate. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method and device for evaluating the application effect of ship energy efficiency measures, which can improve the efficiency and convenience of evaluating the application effect of ship energy efficiency measures and reduce the evaluation cost.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] A method for evaluating the effect of applying ship energy efficiency measures, comprising:

[0007] Acquire multi-source heterogeneous data of ships;

[0008] Performing fusion processing on the multi-source heterogeneous data to obtain target data;

[0009] Dividing the target data into first data before the energy efficiency measure is applied and second data after the energy efficiency measure is applied according to the time when the energy efficiency measure is applied;

[0010] The first data is input into a second fuel consumption prediction model to perform fuel consumption prediction, and obtain a first prediction result; the second fuel consumption prediction model is obtained by training a preset machine learning model according to the second data set;

[0011] Inputting the second data into a first fuel consumption prediction model to perform fuel consumption prediction, and obtaining a second prediction result; the first fuel consumption prediction model is obtained by training a preset machine learning model according to the first data set;

[0012] Obtaining a first evaluation result according to the first prediction result and the first actual fuel consumption value;

[0013] Obtaining a second evaluation result according to the second prediction result and the second actual fuel consumption value;

[0014] According to the first evaluation result and the second evaluation result, a target evaluation result is obtained and output.

[0015] Optionally, fusing the multi-source heterogeneous data to obtain target data includes:

[0016] Performing data cleaning on the multi-source heterogeneous data to obtain first intermediate data;

[0017] Converting the format of the first intermediate data to obtain second intermediate data;

[0018] The second intermediate data is subjected to data fusion processing to obtain target data.

[0019] Optionally, the training process of the first fuel consumption prediction model includes:

[0020] Acquire a first data set, the first data set comprising: a first training set and a first validation set;

[0021] Inputting the data in the first training set into an input layer of a preset machine learning model for processing to obtain a first output;

[0022] Inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output;

[0023] Inputting the second output into an output layer of a preset machine learning model for processing to obtain a first prediction result;

[0024] The preset machine learning model is tuned according to the first prediction result, the first validation set and the mean absolute percentage error to obtain a first fuel consumption prediction model.

[0025] Optionally, the training process of the second fuel consumption prediction model includes:

[0026] Acquire a second data set, wherein the second data set includes: a second training set and a second validation set;

[0027] Inputting the data in the second training set into an input layer of a preset machine learning model for processing to obtain a third output;

[0028] Inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output;

[0029] Inputting the fourth output into an output layer of a preset machine learning model for processing to obtain a second prediction result;

[0030] The preset machine learning model is tuned according to the second prediction result, the second validation set and the mean absolute percentage error to obtain a second fuel consumption prediction model.

[0031] Optionally, inputting the first data into a second fuel consumption prediction model to perform fuel consumption prediction to obtain a first prediction result includes:

[0032] Inputting the first data into an input layer of a second fuel consumption prediction model for processing to obtain a third output;

[0033] Inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output;

[0034] The fourth output is input into the output layer of the preset machine learning model for processing to obtain a first prediction result.

[0035] Optionally, inputting the second data into a first fuel consumption prediction model to perform fuel consumption prediction to obtain a second prediction result includes:

[0036] Inputting the second data into the input layer of the first fuel consumption prediction model for processing to obtain a first output;

[0037] Inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output;

[0038] The second output is input into the output layer of the preset machine learning model for processing to obtain a second prediction result.

[0039] Optionally, obtaining a first evaluation result according to the first prediction result and the first actual fuel consumption value includes:

[0040] Obtaining a first evaluation result according to the first predicted fuel consumption, the first actual fuel consumption and the first mean absolute percentage error;

[0041] According to the second prediction result and the second fuel consumption value, a second evaluation result is obtained, including:

[0042] A second evaluation result is obtained according to the second predicted fuel consumption, the second actual fuel consumption and the second mean absolute percentage error.

[0043] Optionally, obtaining a target evaluation result according to the first evaluation result and the second evaluation result includes:

[0044] Obtaining an intermediate evaluation result according to the first evaluation result and the second evaluation result;

[0045] According to the intermediate evaluation results and indicator weights, the target evaluation results are obtained.

[0046] An embodiment of the present invention further provides a device for evaluating the application effect of ship energy efficiency measures, comprising:

[0047] Acquisition module, used to acquire multi-source heterogeneous data of ships;

[0048] A processing module is used to perform fusion processing on the multi-source heterogeneous data to obtain target data; divide the target data into first data before applying the energy efficiency measures and second data after applying the energy efficiency measures according to the time of applying the energy efficiency measures; input the first data into a second fuel consumption prediction model to perform fuel consumption prediction and obtain a first prediction result; the second fuel consumption prediction model is obtained by training a preset machine learning model according to a second data set; input the second data into the first fuel consumption prediction model to perform fuel consumption prediction and obtain a second prediction result; the first fuel consumption prediction model is obtained by training a preset machine learning model according to the first data set; obtain a first evaluation result based on the first prediction result and a first actual fuel consumption value; obtain a second evaluation result based on the second prediction result and the second actual fuel consumption value; obtain a target evaluation result based on the first evaluation result and the second evaluation result, and output it.

[0049] An embodiment of the present invention further provides a computing device readable storage medium, wherein the computing device readable storage medium stores a program, and when the program is executed by a processor, the method for evaluating the application effect of ship energy efficiency measures described in the present invention is implemented.

[0050] The above technical solution of the present invention has at least the following technical effects:

[0051] The above-mentioned evaluation method for the application effect of ship energy efficiency measures of the present invention obtains multi-source heterogeneous data of the ship; fuses the multi-source heterogeneous data to obtain target data; divides the target data into first data before the application of the energy efficiency measures and second data after the application of the energy efficiency measures according to the time of applying the energy efficiency measures; inputs the first data into the second fuel consumption prediction model to predict fuel consumption and obtain a first prediction result; the second fuel consumption prediction model is trained on the preset machine learning model according to the second data set; the second data is input into the first fuel consumption prediction model to predict fuel consumption and obtain a second prediction result; the first fuel consumption prediction model is trained on the preset machine learning model according to the first data set; the first evaluation result is obtained according to the first prediction result and the first actual fuel consumption value; the second evaluation result is obtained according to the second prediction result and the second actual fuel consumption value; the target evaluation result is obtained according to the first evaluation result and the second evaluation result, and output. It can improve the efficiency and convenience of the evaluation of the application effect of ship energy efficiency measures and reduce the evaluation cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flow chart of the method for evaluating the application effect of ship energy efficiency measures of the present invention;

[0053] Figure 2 is a schematic diagram of a machine learning model of a method for evaluating the application effect of ship energy efficiency measures of the present invention;

[0054] Figure 3 It is a schematic diagram of the implementation of the evaluation method of the application effect of ship energy efficiency measures of the present invention;

[0055] Figure 4 It is a schematic diagram of the device for evaluating the application effect of the ship energy efficiency measures of the present invention. DETAILED DESCRIPTION

[0056] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0057] like Figure 1 As shown, an embodiment of the present invention provides a method for evaluating the application effect of ship energy efficiency measures, including:

[0058] Step S1, acquiring multi-source heterogeneous data of ships;

[0059] Step S2, fusing the multi-source heterogeneous data to obtain target data;

[0060] Step S3, dividing the target data into first data before the energy efficiency measure is applied and second data after the energy efficiency measure is applied according to the time when the energy efficiency measure is applied;

[0061] Step S4, inputting the first data into a second fuel consumption prediction model to perform fuel consumption prediction and obtain a first prediction result; the second fuel consumption prediction model is obtained by training a preset machine learning model according to the second data set;

[0062] Step S5, inputting the second data into a first fuel consumption prediction model to perform fuel consumption prediction, and obtaining a second prediction result; the first fuel consumption prediction model is obtained by training a preset machine learning model according to the first data set;

[0063] Step S6, obtaining a first evaluation result according to the first prediction result and the first actual fuel consumption value;

[0064] Step S7, obtaining a second evaluation result according to the second prediction result and the second actual fuel consumption value;

[0065] Step S8, obtaining a target evaluation result based on the first evaluation result and the second evaluation result, and outputting the result.

[0066] In this embodiment, Figure 1As shown, in the evaluation method of the application effect of ship energy efficiency measures, first, multi-source heterogeneous data of the ship is obtained, and multi-source data that affect the ship energy efficiency results are selected, and these multi-source data are directly related to the ship energy efficiency results; then, the obtained multi-source heterogeneous data are screened, eliminated, unified in format, data fusion and other preliminary fusion processes are performed, and a basic data is formed according to the corresponding correlation relationship to obtain the target data; thirdly, the target data is divided into first data and second data according to the time of applying the energy efficiency measures; thirdly, the first data is processed by using the second fuel consumption prediction model to perform fuel consumption prediction and obtain a first prediction result; the second fuel consumption prediction model is obtained by training a preset machine learning model according to the second data set; the second data is processed by using the first fuel consumption prediction model to perform fuel consumption prediction and obtain a second prediction result; the first fuel consumption prediction model is used to train a preset machine learning model according to the second data set; the second data is processed by using the first fuel consumption prediction model to perform fuel consumption prediction and obtain a second prediction result; the first fuel consumption prediction model is used to train a preset machine learning model according to the first data set The preset machine learning model is trained; next, according to the first prediction result and the first actual fuel consumption value, the prediction result is compared with the actual result, and the error factor is combined to obtain a first evaluation result; according to the second prediction result and the second actual fuel consumption value, the prediction result is compared with the actual result, and the error factor is combined to obtain a second evaluation result; finally, according to the first evaluation result and the second evaluation result, appropriate evaluation data is selected to obtain the target evaluation result, and output it; the scheme of the present invention can realize the evaluation of the application effect of energy efficiency measures of ships under different speeds, different cargo drafts, and different sea conditions, and realize the comprehensive effect evaluation of the application of one or more technical energy efficiency and operational energy efficiency measures of ships; the method of the present invention has low cost, high efficiency, and short time consumption, and can provide non-professionals with a convenient ship energy efficiency measure effect evaluation tool, saving a lot of human resource costs.

[0067] In this embodiment, the acquired multi-source heterogeneous data includes: ship automatic identification system data, ship navigation data, ocean meteorological sea condition data and ship static data;

[0068] The static data of the ship mainly include basic technical parameters of the ship, such as main dimension parameters, main and auxiliary engine parameters, etc. The main indicators are as follows:

[0069]

[0070] The ship automatic identification system data is the information data on the position, speed, heading, etc. of the ocean ship generated by the ship automatic identification system. The main indicators are as follows:

[0071]

[0072] Marine meteorological sea condition data is marine meteorological reanalysis raster data with high temporal and spatial resolution, including sea condition data such as wind, waves, and currents. The main indicators are as follows:

[0073]

[0074]

[0075] Ship navigation data refers to various data generated during ship navigation, including ship navigation attitude data, ship main equipment operation data, fuel consumption data, hull data, bottom fouling data, etc. The main indicators are as follows:

[0076]

[0077]

[0078] In an optional embodiment of the present invention, in step S2, the multi-source heterogeneous data is fused to obtain target data, including:

[0079] Step S21, performing data cleaning on the multi-source heterogeneous data to obtain first intermediate data;

[0080] Step S22: convert the format of the first intermediate data to obtain second intermediate data.

[0081] Step S23: performing data fusion processing on the second intermediate data to obtain target data.

[0082] In this embodiment, the amount of ship-related data is large and the noise data is large, so the multi-source heterogeneous data needs to be preprocessed. First, the multi-source heterogeneous data is cleaned to remove abnormal data, erroneous data and redundant data that are obviously beyond the normal range, so as to obtain the first intermediate data. For example, the abnormal data that exceeds the speed threshold and exceeds the longitude and latitude range is preprocessed; then, the first intermediate data is converted into a data format, and the data collected by different sensors are uniformly converted into a standard format, for example, the weight data is converted into kg, and the time data is converted into the YYYY-MM-DD HH:MM:SS format, so as to obtain the second intermediate data; finally, the second intermediate data is subjected to data fusion processing, firstly, the ship automatic identification system data and the marine meteorological sea condition data are matched according to the time and longitude and latitude information, and the data are concatenated according to the time and longitude and latitude to obtain the intermediate fusion data; then, according to the ship's Maritime Mobile Identity (MMSI) and the International Maritime Organization identification number (IMO), the ship static data is matched for the intermediate fusion data, and the data are concatenated according to the MMSI and IMO information to obtain the target data.

[0083] In an optional embodiment of the present invention, in step S5, the training process of the first fuel consumption prediction model includes:

[0084] Step S511, obtaining a first data set, wherein the first data set includes: a first training set and a first validation set;

[0085] Step S512, inputting the data in the first training set into the input layer of a preset machine learning model for processing to obtain a first output;

[0086] Step S513, inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output;

[0087] Step S514, inputting the second output into the output layer of a preset machine learning model for processing to obtain a first prediction result;

[0088] Step S515: Tune the preset machine learning model according to the first prediction result, the first validation set and the mean absolute percentage error to obtain a first fuel consumption prediction model.

[0089] In this embodiment, a first data set is first obtained, where the first data set is objective data accumulated historically before the application of energy efficiency measures; the first data set is divided into a first training set and a first validation set according to a preset ratio, and the preset machine learning model is trained with the mean absolute percentage error as the model evaluation indicator to obtain a first fuel consumption prediction model.

[0090] Specifically, firstly, based on the first data set dataA1, the data is randomly divided into a first training set and a first validation set in a ratio of 7:3, and a machine learning model is selected as a prediction model. The model is trained with the first training set and the first validation set; the mean absolute percentage error (MAPE) is used as a model evaluation indicator, and the optimal model with a sample point fuel consumption MAPE less than 3% and a sub-segment and sub-speed summary fuel consumption MAPE less than 1% is selected as the first fuel consumption prediction model (modelA) for fuel consumption prediction before applying energy efficiency measures, and MAPE is used as the prediction error E of modelA. modelA，i，j , the format is as follows:

[0091]

[0092] Furthermore, the preset neural network is Figure 2 As shown, the first training set data is divided into small batches, and each batch of data is input into the preset neural network to obtain the output result of the model. For example, if the training set has 1000 samples and the single batch size is set to 100, 100 samples are used as a group of input models each time. The input layer of the preset machine learning model receives the training set data in batches and processes it to obtain the first output; then the first output is input into the loop processing layer of the preset machine learning model for processing to obtain the second output; the loop processing layer includes the forget gate rt, the input gate zt, and the current neuron state h′t. The forget gate rt is obtained by the following formula:

[0093] rt=sigmoid(Qr·[h t-1 , xt])

[0094] Among them, sigmoid is the activation function, Qr is the weight matrix, h t-1 is the previous neuron state, and xt is the input value of the current neuron;

[0095] The input gate zt is obtained by the following formula:

[0096] zt=sigmoid(Qz·[h t- 1 , xt])

[0097] Among them, sigmoid is the activation function, Qz is the weight matrix, h t-1 is the previous neuron state, and xt is the input value of the current neuron;

[0098] The current neuron state h′t is obtained by the following formula:

[0099] h′t=tanh(Q·[rt*h t-1 , xt])

[0100] Among them, tanh is the activation function, Q is the weight matrix, rt is the output value of the forget gate, and h t-1 is the previous neuron state, and xt is the input value of the current neuron;

[0101] The output value ht of the current neuron is obtained by the following formula:

[0102] ht=(1-zt)*h t-1 +zt*h′t])

[0103] Among them, zt is the input gate, h t-1 is the previous neuron state, h′t is the current neuron state;

[0104] By using the selected loss function, the model output is compared with the corresponding true label (the true value in the validation set) and the loss value is calculated to measure the quality of the current model prediction.

[0105] This example uses the mean absolute percentage error as the loss function:

[0106]

[0107] Among them, E modelA，i，j is the prediction error of model A under the conditions of speed i and draft j, n is the number of samples, y i，j is the actual observed value under (i, j) conditions, is the predicted value corresponding to the actual observed value under (i, j) conditions.

[0108] According to the calculated loss value, back propagation is performed using the chain rule to calculate the gradients corresponding to each parameter, and then the sample gradient is calculated using the stochastic gradient optimization algorithm. The sample gradient calculation formula for the weight parameter w is as follows:

[0109] w=w-η·∧ w E(w)

[0110] Among them, w is the weight parameter, η is the model learning rate, which is used to control the step size of each parameter update, ∧ w E(w) is the gradient of the loss function E with respect to w.

[0111] The model parameters are updated based on these gradients so that the model is continuously adjusted in the direction of reducing losses. This process is repeated for preset rounds to avoid insufficient training or overfitting. After each round of training, the validation set is used to evaluate the performance of the model in the current round, such as calculating the accuracy, loss value and other indicators on the validation set, so that the hyperparameters can be adjusted in time according to these feedbacks or the training can be stopped early. If the performance of the validation set no longer improves or even decreases, overfitting may have occurred, and the training can be stopped. After the training is completed, the second output is input into the output layer of the preset machine learning model for processing to obtain the first fuel consumption prediction model;

[0112] Similarly, if you select other machine learning models, such as decision tree, random forest, K nearest neighbor, elastic network regression, extreme gradient boosting tree, gradient boosting decision tree and other machine learning models, the training process is similar to the above example and can achieve the same effect.

[0113] In an optional embodiment of the present invention, in step S4, the training process of the second fuel consumption prediction model includes:

[0114] Step S411, obtaining a second data set, where the second data set includes: a second training set and a second validation set;

[0115] Step S412, inputting the data in the second training set into the input layer of a preset machine learning model for processing to obtain a third output;

[0116] Step S413, inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output;

[0117] Step S414, inputting the fourth output into the output layer of the preset machine learning model for processing to obtain a second prediction result;

[0118] Step S415: tune the preset machine learning model according to the second prediction result, the second validation set and the mean absolute percentage error to obtain a second fuel consumption prediction model.

[0119] In this embodiment, similar to the training process of the first fuel consumption prediction model, based on the second data set dataB1, the data is randomly divided into a second training set and a second validation set in a ratio of 7:3, and a machine learning model is selected as the prediction model. The model is trained with the second training set and the second validation set; the mean absolute percentage error (MAPE) is used as the model evaluation index, and the optimal model with a sample point fuel consumption MAPE of less than 3% and a sub-segment and sub-speed summary fuel consumption MAPE of less than 1% is selected as the second fuel consumption prediction model (modelB) for fuel consumption prediction before applying energy efficiency measures, and MAPE is used as the prediction error E of model modelB. modelB，i，j , the format is as follows:

[0120]

[0121] Furthermore, the preset neural network is Figure 2 As shown, the second training set data is divided into small batches, and each batch of data is input into the preset neural network to obtain the output result of the model. For example, if the training set has 1000 samples and the single batch size is set to 100, 100 samples are used as a group of input models each time. The input layer of the preset machine learning model receives the training set data in batches and processes it to obtain the third output; then the third output is input into the loop processing layer of the preset machine learning model for processing to obtain the fourth output; the loop processing layer includes the forget gate rt, the input gate zt, the current neuron state h′t, and the forget gate rt is obtained by the following formula:

[0122] rt=sigmoid(Qr·[h t-1 , xt])

[0123] Among them, sigmoid is the activation function, Qr is the weight matrix, h t-1 is the previous neuron state, and xt is the input value of the current neuron;

[0124] The input gate zt is obtained by the following formula:

[0125] zt=sigmoid(Qz·[h t- 1 , xt])

[0126] Among them, sigmoid is the activation function, Qz is the weight matrix, h t-1 is the previous neuron state, and xt is the input value of the current neuron;

[0127] The current neuron state h′t is obtained by the following formula:

[0128] h′t=tanh(Q·[rt*h t-1 , xt])

[0129] Among them, tanh is the activation function, Q is the weight matrix, rt is the output value of the forget gate, and h t-1 is the previous neuron state, and xt is the input value of the current neuron;

[0130] The output value ht of the current neuron is obtained by the following formula:

[0131] ht=(1-zt)*h t-1 +zt*h′t])

[0132] Among them, zt is the input gate, h t-1 is the previous neuron state, h′t is the current neuron state;

[0133] By using the selected loss function, the model output is compared with the corresponding true label (the true value in the validation set) and the loss value is calculated to measure the quality of the current model prediction.

[0134] This example uses the mean absolute percentage error as the loss function:

[0135]

[0136] Among them, E modelA，i，j is the prediction error of model A under the conditions of speed i and draft j, n is the number of samples, y i，j is the actual observed value under (i, j) conditions, is the predicted value corresponding to the actual observed value under (i, j) conditions.

[0137] According to the calculated loss value, back propagation is performed using the chain rule to calculate the gradients corresponding to each parameter, and then the sample gradient is calculated using the stochastic gradient optimization algorithm. The sample gradient calculation formula for the weight parameter w is as follows:

[0138] w=w-η·∧ w E(w)

[0139] Among them, w is the weight parameter, η is the model learning rate, which is used to control the step size of each parameter update, ∧ w E(w) is the gradient of the loss function E with respect to w.

[0140] The parameters of the model are updated based on these gradients so that the model is continuously adjusted in the direction of reducing losses. This process is repeated for preset rounds to avoid insufficient training or overfitting. After each round of training, the validation set is used to evaluate the performance of the model in the current round, such as calculating the accuracy, loss value and other indicators on the validation set, so that the hyperparameters can be adjusted in time according to these feedbacks or the training can be stopped early. If the performance of the validation set no longer improves or even decreases, overfitting may have occurred, and the training can be stopped. After the training is completed, the fourth output is input into the output layer of the preset machine learning model for processing to obtain a second fuel consumption prediction model;

[0141] Similarly, if you select other machine learning models, such as decision tree, random forest, K nearest neighbor, elastic network regression, extreme gradient boosting tree, gradient boosting decision tree and other machine learning models, the training process is similar to the above example and can achieve the same effect.

[0142] In an optional embodiment of the present invention, in step S4, the first data is input into a second fuel consumption prediction model to perform fuel consumption prediction to obtain a first prediction result, including:

[0143] Step S421, inputting the first data into the input layer of the second fuel consumption prediction model for processing to obtain a third output;

[0144] Step S422, inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output;

[0145] Step S423: input the fourth output into the output layer of the preset machine learning model for processing to obtain a first prediction result.

[0146] In this embodiment, Figure 3As shown, the target data includes ship data before and after the energy efficiency measures are applied. According to whether the energy efficiency measures are applied to the ship, the data before the energy efficiency measures are applied in the target data is used as the first data (dataA), and the data after the energy efficiency measures are applied in the target data is used as the second data (dataB); the first data is divided into small batches (batch), and each batch of data is input into the input layer of the second fuel consumption prediction model for processing to obtain the third output. For example, if the first data has 1000 samples, the single batch size is set to 100, and 100 samples are used as a group of input models each time. The input layer of the second fuel consumption prediction model receives the training set data in batches and processes it to obtain the third output; then the third output is input into the loop processing layer of the second fuel consumption prediction model for processing, and the loop processing layer includes a forget gate rt, an input gate zt, and a current neuron state h′t to obtain a fourth output; and then the fourth output is input into the output layer of the second fuel consumption prediction model for processing to obtain the first prediction result.

[0147] In an optional embodiment of the present invention, in step S5, the second data is input into the first fuel consumption prediction model to perform fuel consumption prediction to obtain a second prediction result, including:

[0148] Step S521, inputting the second data into the input layer of the first fuel consumption prediction model for processing to obtain a first output;

[0149] Step S522, inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output;

[0150] Step S523: input the second output into the output layer of a preset machine learning model for processing to obtain a second prediction result.

[0151] In this embodiment, Figure 3As shown, the target data includes ship data before and after the energy efficiency measures are applied. According to whether the energy efficiency measures are applied to the ship, the data before the energy efficiency measures are applied in the target data is used as the first data (dataA), and the data after the energy efficiency measures are applied in the target data is used as the second data (dataB); the second data is divided into small batches (batch), and each batch of data is input into the input layer of the first fuel consumption prediction model for processing to obtain the first output. For example, if the second data has 1000 samples, the single batch size is set to 100, and 100 samples are used as a group of input models each time. The input layer of the first fuel consumption prediction model receives the training set data in batches and processes it to obtain the first output; then the first output is input into the recurrent processing layer of the first fuel consumption prediction model for processing, and the recurrent processing layer includes a forget gate rt, an input gate zt, and a current neuron state h′t to obtain the second output; then the second output is input into the output layer of the first fuel consumption prediction model for processing to obtain the second prediction result.

[0152] In an optional embodiment of the present invention, in step S6, obtaining a first evaluation result according to the first prediction result and the first actual fuel consumption value includes:

[0153] Step S61, obtaining a first evaluation result according to the first predicted fuel consumption, the first actual fuel consumption and the first mean absolute percentage error;

[0154] In step S7, a second evaluation result is obtained according to the second prediction result and the second fuel consumption value, including:

[0155] Step S71, obtaining a second evaluation result according to the second predicted fuel consumption, the second actual fuel consumption and the second mean absolute percentage error.

[0156] In this embodiment, Figure 3 As shown, first, the first predicted fuel consumption after applying the energy efficiency measures is compared and analyzed with the first actual fuel consumption of dataA before applying the measures to obtain a first evaluation result. The calculation formula for the first evaluation result of the energy efficiency measures is as follows:

[0157] f modelB，dataA，i，j =|FCP modelB，dataA，i，j -FCR dataA，i，j | / FCR dataA，i，j *100%-E modelB，i，j

[0158] Among them, f modelB，dataA，i，j is the energy saving effect (%) calculated by the second model modelB under the conditions of speed i and draft j in dataA for the application of energy efficiency measures to ships, FCP modelB，dataA，i，jis the total predicted fuel consumption of the second model modelB under the conditions of speed i and draft j in the first data dataA, FCR dataA，i，j is the total actual fuel consumption under the conditions of speed i and draft j in the first data dataA, E modelB，i，j is the prediction error of the second model modelB under the conditions of speed i and draft j.

[0159] Then, based on the second data dataB, the second predicted fuel consumption before the energy efficiency measure is applied is obtained through the first model modelA, and compared with the second actual fuel consumption after the energy efficiency measure is applied to the second data dataB. The second evaluation result calculation formula of the energy efficiency measure application is as follows:

[0160] f modelA，dataB，i，j =|FCR dataB，i，j -FCP modelA，dataB，i，j | / FCP modelA，dataB，i，j *100%-E modelA，i，j

[0161] Among them, f modelA，dataB，i，j is the energy saving effect (%) calculated by the first model A under the conditions of speed i and draft j in dataB when applying energy efficiency measures to ships, FCR dataB，i，j is the total predicted fuel consumption of the first model modelA under the conditions of speed i and draft j in dataB, FCP modelA，dataB，i，j is the total actual fuel consumption under the conditions of speed i and draft j in dataB, E modelA，i，j is the prediction error of the prediction model modelA under the conditions of speed i and draft j.

[0162] In an optional embodiment of the present invention, in step S8, obtaining a target evaluation result according to the first evaluation result and the second evaluation result includes:

[0163] Step S81, obtaining an intermediate evaluation result according to the first evaluation result and the second evaluation result;

[0164] Step S82, obtaining a target evaluation result according to the intermediate evaluation result and the indicator weight.

[0165] In this embodiment, first, the first evaluation result and the second evaluation result are compared, and the smaller one of the first evaluation result and the second evaluation result is taken as the intermediate evaluation result, and the final effect f of applying the energy efficiency measure under the conditions of speed i and draft j is calculated. i，j It is calculated by the following formula:

[0166] f i，j =min(f modelA，dataB，i，j , f modelB，dataA，i，j )

[0167] Then, the turnover ratio of a specific type of ship in the shipping industry sailing at speed i and draft j is used as the weight, and the weight ω i，j It can be given by the following table:

[0168]

[0169] Calculate the weighted average of the applied energy efficiency measures as the energy efficiency measure application target evaluation result f. The calculation formula for the target evaluation result f is:

[0170] f=∑∑f i，j ×ω i，j

[0171] The specific implementation process of the above method of the present invention is described below:

[0172] Step 111, acquiring multi-source heterogeneous data of ships;

[0173] Step 112, performing data cleaning on the multi-source heterogeneous data to obtain first intermediate data;

[0174] Step 113, converting the format of the first intermediate data to obtain second intermediate data;

[0175] Step 114, performing data fusion processing on the second intermediate data to obtain target data;

[0176] Step 115, dividing the target data into first data before the energy efficiency measure is applied and second data after the energy efficiency measure is applied according to the time when the energy efficiency measure is applied;

[0177] Step 116, obtaining a first data set, wherein the first data set includes: a first training set and a first validation set;

[0178] Step 117, inputting the data in the first training set into an input layer of a preset machine learning model for processing to obtain a first output;

[0179] Step 118, inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output;

[0180] Step 119, inputting the second output into an output layer of a preset machine learning model for processing to obtain a first prediction result;

[0181] Step 120, tuning a preset machine learning model according to the first prediction result, the first validation set, and the mean absolute percentage error to obtain a first fuel consumption prediction model;

[0182] Step 121, obtaining a second data set, wherein the second data set includes: a second training set and a second validation set;

[0183] Step 122, inputting the data in the second training set into an input layer of a preset machine learning model for processing to obtain a third output;

[0184] Step 123, inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output;

[0185] Step 124, inputting the fourth output into an output layer of a preset machine learning model for processing to obtain a second prediction result;

[0186] Step 125, tuning the preset machine learning model according to the second prediction result, the second validation set and the mean absolute percentage error to obtain a second fuel consumption prediction model;

[0187] Step 126, inputting the first data into the input layer of the second fuel consumption prediction model for processing to obtain a third output;

[0188] Step 127, inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output;

[0189] Step 128, inputting the fourth output into an output layer of a preset machine learning model for processing to obtain a first prediction result;

[0190] Step 129, inputting the second data into the input layer of the first fuel consumption prediction model for processing to obtain a first output;

[0191] Step 130, inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output;

[0192] Step 131, inputting the second output into an output layer of a preset machine learning model for processing to obtain a second prediction result;

[0193] Step 132, obtaining a first evaluation result according to the first predicted fuel consumption, the first actual fuel consumption and the first mean absolute percentage error;

[0194] Step 133, obtaining a second evaluation result according to the second prediction result and the second fuel consumption value, includes:

[0195] Step 134, obtaining a second evaluation result according to the second predicted fuel consumption, the second actual fuel consumption and the second mean absolute percentage error;

[0196] Step 135, obtaining an intermediate evaluation result according to the first evaluation result and the second evaluation result;

[0197] Step 136, obtaining a target evaluation result according to the intermediate evaluation result and the indicator weight.

[0198] The scheme of the present invention comprehensively utilizes multi-source heterogeneous data before and after the application of energy efficiency measures on operating ships, integrates ship AIS data, ship navigation data, marine meteorological sea conditions data and ship static data, and constructs a data model-based ship energy efficiency measures application effect evaluation method. By constructing two high-precision fuel consumption prediction models and using the method of exchanging data, the evaluation effect deviation caused by the difference in navigation conditions between the baseline group and the control group ships is eliminated, and the impact of more complex actual navigation conditions of ships on the evaluation of the application effect of energy efficiency technology is fully considered, and a calculation method for the overall effect of energy efficiency measures is proposed.

[0199] The solution of the present invention is based on the actual navigation data of the ship, fully considering the complex working conditions of the actual navigation of the ship, and using the machine learning method to establish a fuel consumption prediction model before and after the application of energy efficiency measures for ships; it can realize the evaluation of the application effect of energy efficiency measures of ships under different speeds, different drafts (cargo loads), and different sea conditions; it can realize the comprehensive effect evaluation of the application of one or more technical energy efficiency and operational energy efficiency measures of ships; the method of the present invention has low cost, high efficiency, and short time consumption, and can provide a convenient ship energy efficiency measure effect evaluation tool, saving a lot of human resource costs.

[0200] like Figure 4 As shown, an embodiment of the present invention further provides an evaluation device 30 for the application effect of ship energy efficiency measures, comprising:

[0201] An acquisition module 41 is used to acquire multi-source heterogeneous data of ships;

[0202] The processing module 42 is used to perform fusion processing on the multi-source heterogeneous data to obtain target data; divide the target data into first data before applying the energy efficiency measures and second data after applying the energy efficiency measures according to the time of applying the energy efficiency measures; input the first data into the second fuel consumption prediction model to perform fuel consumption prediction and obtain a first prediction result; the second fuel consumption prediction model is obtained by training a preset machine learning model according to the second data set; input the second data into the first fuel consumption prediction model to perform fuel consumption prediction and obtain a second prediction result; the first fuel consumption prediction model is obtained by training a preset machine learning model according to the first data set; obtain a first evaluation result based on the first prediction result and the first actual fuel consumption value; obtain a second evaluation result based on the second prediction result and the second actual fuel consumption value; obtain a target evaluation result based on the first evaluation result and the second evaluation result, and output it.

[0203] Optionally, fusing the multi-source heterogeneous data to obtain target data includes:

[0204] Performing data cleaning on the multi-source heterogeneous data to obtain first intermediate data;

[0205] Converting the format of the first intermediate data to obtain second intermediate data;

[0206] The second intermediate data is subjected to data fusion processing to obtain target data.

[0207] Optionally, the training process of the first fuel consumption prediction model includes:

[0208] Acquire a first data set, the first data set comprising: a first training set and a first validation set;

[0209] Inputting the data in the first training set into an input layer of a preset machine learning model for processing to obtain a first output;

[0210] Inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output;

[0211] Inputting the second output into an output layer of a preset machine learning model for processing to obtain a first prediction result;

[0212] The preset machine learning model is tuned according to the first prediction result, the first validation set and the mean absolute percentage error to obtain a first fuel consumption prediction model.

[0213] Optionally, the training process of the second fuel consumption prediction model includes:

[0214] Acquire a second data set, wherein the second data set includes: a second training set and a second validation set;

[0215] Inputting the data in the second training set into an input layer of a preset machine learning model for processing to obtain a third output;

[0216] Inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output;

[0217] Inputting the fourth output into an output layer of a preset machine learning model for processing to obtain a second prediction result;

[0218] The preset machine learning model is tuned according to the second prediction result, the second validation set and the mean absolute percentage error to obtain a second fuel consumption prediction model.

[0219] Optionally, inputting the first data into a second fuel consumption prediction model to perform fuel consumption prediction to obtain a first prediction result includes:

[0220] Inputting the first data into an input layer of a second fuel consumption prediction model for processing to obtain a third output;

[0221] Inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output;

[0222] The fourth output is input into the output layer of the preset machine learning model for processing to obtain a first prediction result.

[0223] Optionally, inputting the second data into a first fuel consumption prediction model to perform fuel consumption prediction to obtain a second prediction result includes:

[0224] Inputting the second data into the input layer of the first fuel consumption prediction model for processing to obtain a first output;

[0225] Inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output;

[0226] The second output is input into the output layer of the preset machine learning model for processing to obtain a second prediction result.

[0227] Optionally, obtaining a first evaluation result according to the first prediction result and the first actual fuel consumption value includes:

[0228] Obtaining a first evaluation result according to the first predicted fuel consumption, the first actual fuel consumption and the first mean absolute percentage error;

[0229] According to the second prediction result and the second fuel consumption value, a second evaluation result is obtained, including:

[0230] A second evaluation result is obtained according to the second predicted fuel consumption, the second actual fuel consumption and the second mean absolute percentage error.

[0231] Optionally, obtaining a target evaluation result according to the first evaluation result and the second evaluation result includes:

[0232] Obtaining an intermediate evaluation result according to the first evaluation result and the second evaluation result;

[0233] According to the intermediate evaluation results and indicator weights, the target evaluation results are obtained.

[0234] It should be noted that all implementation methods in the above method embodiments are applicable to the embodiments of the device and can achieve the same technical effects.

[0235] An embodiment of the present invention further provides a computing device readable storage medium, wherein a program is stored in the computing device readable storage medium, and when the program is executed by a processor, the method for evaluating the application effect of ship energy efficiency measures of the present invention is implemented. All implementations in the above method embodiments are applicable to the embodiments of the computing device readable storage medium, and can also achieve the same technical effects.

[0236] Those skilled in the art will appreciate that the units and algorithm steps of each example 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 performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0237] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0238] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0239] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0240] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0241] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0242] In addition, it should be noted that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it is understood that all or any steps or components of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or a network of computing devices in hardware, firmware, software or a combination thereof, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.

[0243] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general-purpose device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code for implementing a method or device. That is to say, such a program product also constitutes the present invention, and a storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

[0244] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for evaluating the application effect of ship energy efficiency measures, characterized in that: include: Acquire multi-source heterogeneous data of ships; Performing fusion processing on the multi-source heterogeneous data to obtain target data; Dividing the target data into first data before the energy efficiency measure is applied and second data after the energy efficiency measure is applied according to the time when the energy efficiency measure is applied; The first data is input into a second fuel consumption prediction model to perform fuel consumption prediction, and obtain a first prediction result; the second fuel consumption prediction model is obtained by training a preset machine learning model according to the second data set; Inputting the second data into a first fuel consumption prediction model to perform fuel consumption prediction, and obtaining a second prediction result; the first fuel consumption prediction model is obtained by training a preset machine learning model according to the first data set; Obtaining a first evaluation result according to the first prediction result and the first actual fuel consumption value; Obtaining a second evaluation result according to the second prediction result and the second actual fuel consumption value; According to the first evaluation result and the second evaluation result, a target evaluation result is obtained and output.

2. The method for evaluating the application effect of ship energy efficiency measures according to claim 1, characterized in that: The multi-source heterogeneous data are fused to obtain target data, including: Performing data cleaning on the multi-source heterogeneous data to obtain first intermediate data; Converting the format of the first intermediate data to obtain second intermediate data; Perform data fusion processing on the second intermediate data to obtain target data.

3. The method for evaluating the application effect of ship energy efficiency measures according to claim 1, characterized in that: The training process of the first fuel consumption prediction model includes: Acquire a first data set, the first data set comprising: a first training set and a first validation set; Inputting the data in the first training set into an input layer of a preset machine learning model for processing to obtain a first output; Inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output; Inputting the second output into an output layer of a preset machine learning model for processing to obtain a first prediction result; The preset machine learning model is tuned according to the first prediction result, the first validation set and the mean absolute percentage error to obtain a first fuel consumption prediction model.

4. The method for evaluating the application effect of ship energy efficiency measures according to claim 1, characterized in that: The training process of the second fuel consumption prediction model includes: Acquire a second data set, wherein the second data set includes: a second training set and a second validation set; Inputting the data in the second training set into an input layer of a preset machine learning model for processing to obtain a third output; Inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output; Inputting the fourth output into an output layer of a preset machine learning model for processing to obtain a second prediction result; The preset machine learning model is tuned according to the second prediction result, the second validation set and the mean absolute percentage error to obtain a second fuel consumption prediction model.

5. The method for evaluating the application effect of ship energy efficiency measures according to claim 4, characterized in that: Inputting the first data into a second fuel consumption prediction model to perform fuel consumption prediction to obtain a first prediction result includes: Inputting the first data into an input layer of a second fuel consumption prediction model for processing to obtain a third output; Inputting the third output into a loop processing layer of a preset machine learning model for processing to obtain a fourth output; The fourth output is input into the output layer of the preset machine learning model for processing to obtain a first prediction result.

6. The method for evaluating the application effect of ship energy efficiency measures according to claim 3 is characterized in that: Inputting the second data into the first fuel consumption prediction model to perform fuel consumption prediction, and obtaining a second prediction result, includes: Inputting the second data into the input layer of the first fuel consumption prediction model for processing to obtain a first output; Inputting the first output into a loop processing layer of a preset machine learning model for processing to obtain a second output; The second output is input into the output layer of the preset machine learning model for processing to obtain a second prediction result.

7. The method for evaluating the application effect of ship energy efficiency measures according to claim 1, characterized in that: According to the first prediction result and the first actual fuel consumption value, a first evaluation result is obtained, including: Obtaining a first evaluation result according to the first predicted fuel consumption, the first actual fuel consumption and the first mean absolute percentage error; According to the second prediction result and the second fuel consumption value, a second evaluation result is obtained, including: A second evaluation result is obtained according to the second predicted fuel consumption, the second actual fuel consumption and the second mean absolute percentage error.

8. The method for evaluating the application effect of ship energy efficiency measures according to claim 1, characterized in that: According to the first evaluation result and the second evaluation result, a target evaluation result is obtained, including: Obtaining an intermediate evaluation result according to the first evaluation result and the second evaluation result; According to the intermediate evaluation results and indicator weights, the target evaluation results are obtained.

9. A device for evaluating the effect of application of ship energy efficiency measures, characterized in that: include: Acquisition module, used to acquire multi-source heterogeneous data of ships; A processing module, used for fusing the multi-source heterogeneous data to obtain target data; The target data is divided into first data before the energy efficiency measures are applied and second data after the energy efficiency measures are applied according to the time when the energy efficiency measures are applied; the first data is input into a second fuel consumption prediction model to perform fuel consumption prediction, and a first prediction result is obtained; the second fuel consumption prediction model is obtained by training a preset machine learning model according to the second data set; Inputting the second data into a first fuel consumption prediction model to perform fuel consumption prediction, and obtaining a second prediction result; the first fuel consumption prediction model is obtained by training a preset machine learning model according to the first data set; Obtaining a first evaluation result according to the first prediction result and the first actual fuel consumption value; Obtaining a second evaluation result according to the second prediction result and the second actual fuel consumption value; According to the first evaluation result and the second evaluation result, a target evaluation result is obtained and output.

10. A computing device readable storage medium, characterized in that: The computing device readable storage medium stores a program, which implements the method according to any one of claims 1 to 7 when executed by a processor.

Images