Smart energy management method and system applied to shipbuilding

Abstract

The invention discloses a smart energy management method and system applied to shipbuilding, and relates to the technical field of energy management. The method comprises the following steps: acquiring power consumption of a target ship manufacturing area in a current time period; calculating a generated power sequence corresponding to the target ship manufacturing area in the current time period according to the photovoltaic power generation mode, and calculating the total photovoltaic power generation amount and the power generation stability according to the generated power sequence; comparing the power consumption with the total photovoltaic power generation amount, if the power consumption is greater than or equal to the total photovoltaic power generation amount, judging whether the power generation stability exceeds a stability threshold, and if the power generation stability exceeds the stability threshold, switching to a power grid power generation mode; and if the power consumption is less than the total photovoltaic power generation amount, switching to a power grid power generation mode. According to the method, the power generation stability prediction precision is improved, so that accurate switching between photovoltaic power generation and power grid power generation is realized, and the energy management efficiency is improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of energy management, in particular to a smart energy management method and system applied to ship manufacturing. BACKGROUND

[0002] With the progress of energy technology and the application of intelligent management systems, ship manufacturing plants begin to introduce advanced energy management systems to realize efficient use of energy and energy saving and emission reduction through real-time monitoring, data analysis and optimized control; the system promotes the green development of the shipbuilding industry, continuously improves the energy management level of enterprises, and promotes the development of shipbuilding towards low carbon and high efficiency to meet the needs of market competition and sustainable development.

[0003] The prior art (publication number: CN108614507B) discloses an intelligent energy management method and an intelligent terminal, the method comprising: acquiring energy consumption information of each monitoring point within a first preset time, wherein the energy consumption information is collected by an intelligent collection device; when the time information is a working time period and the energy consumption information is greater than a first preset value, acquiring the first unit energy consumption, when the first unit energy consumption is greater than the preset unit energy consumption, giving a feedback result of "abnormal"; when the time information is a non-working time period and the energy consumption information is greater than a second preset value, giving a feedback result of "abnormal"; when the feedback result is "abnormal", acquiring the position information of the monitoring point corresponding to the abnormal energy consumption information, generating alarm information according to the feedback result and the position information, and sending the alarm information to the terminal. The technical scheme disclosed by the present application monitors and warns the energy consumption of the energy consumption equipment in the working time period and the non-working time period, so as to eliminate the hidden danger of the energy consumption equipment and save energy.

[0004] However, in the actual energy management process, photovoltaic power generation is preferentially used for power supply in the ship manufacturing process, and when the power generation stability and power supply are insufficient, it is difficult to accurately switch the power grid for power supply, thereby reducing the efficiency of energy management. SUMMARY

[0005] The purpose of the present application is to solve the problem of the prior art that in the actual energy management process, photovoltaic power generation is preferentially used for power supply in the ship manufacturing process, and when the power generation stability and power supply are insufficient, it is difficult to accurately switch the power grid for power supply, thereby reducing the efficiency of energy management, and to provide a smart energy management method and system applied to ship manufacturing.

[0006] The purpose of the present application can be achieved by the following technical solutions:

[0007] Firstly, a smart energy management method applied to ship manufacturing is proposed, which comprises:

[0008] acquiring the power consumption of the target ship manufacturing area in the current time period;

[0009] According to the photovoltaic power generation mode, a power generation sequence corresponding to the target ship manufacturing area in the current time period is calculated, and the total amount of photovoltaic power generation and the power generation stability are calculated according to the power generation sequence;

[0010] The power consumption and the total amount of photovoltaic power generation are compared, if the power consumption is greater than or equal to the total amount of photovoltaic power generation, it is judged whether the power generation stability exceeds the stability threshold, if the power generation stability exceeds the stability threshold, the grid power generation mode is switched to;

[0011] If the power consumption is less than the total amount of photovoltaic power generation, the grid power generation mode is switched to.

[0012] Optionally, the power generation sequence of the target area according to the time period comprises:

[0013] Meteorological image data of the target area according to the time period is obtained, and the meteorological image data is divided into a meteorological image sequence according to the time period; the meteorological image sequence is substituted into an optical flow algorithm to obtain a corresponding optical flow image sequence; the images in the meteorological image sequence and the optical flow image sequence are in one-to-one correspondence;

[0014] The meteorological image sequence is input into a target feature model to obtain a spatial feature sequence, and the optical flow image sequence is input into the target feature model to obtain a motion feature sequence;

[0015] The spatial feature sequence and the motion feature sequence are executed on a modal fusion model to obtain a fusion feature sequence;

[0016] The fusion feature sequence is input into a target power generation model to obtain a corresponding power generation sequence.

[0017] Optionally, inputting the meteorological image sequence into the target feature model to obtain the spatial feature sequence and inputting the optical flow image sequence into the target feature model to obtain the motion feature sequence comprises:

[0018] The meteorological image sequence is determined as a first input feature matrix, the first input feature matrix is input into a down-sampling layer to obtain a first down-sampling feature matrix, the first down-sampling feature matrix is input into a two-dimensional convolution layer to obtain a first convolution feature matrix, the first convolution feature matrix is input into a batch normalization layer to obtain a first batch feature matrix, and the first batch feature matrix is input into a ReLU layer to obtain a first output feature matrix; the first output feature matrix is recorded as the spatial feature sequence;

[0019] The optical flow image sequence is determined as a second input feature matrix, the second input feature matrix is input to a down-sampling layer to obtain a second down-sampled feature matrix, the second down-sampled feature matrix is input to a two-dimensional convolution layer to obtain a second convolution feature matrix, the second convolution feature matrix is input to a batch normalization layer to obtain a second batch feature matrix, and the second batch feature matrix is input to a ReLU layer to obtain a second output feature matrix; the second output feature matrix is denoted as a motion feature sequence.

[0020] Optionally, the modal fusion model is performed on the spatial feature sequence and the motion feature sequence to obtain a fusion feature sequence, and the modal fusion model includes:

[0021] The spatial feature of the spatial feature sequence and the motion feature of the corresponding motion feature sequence are extracted respectively, the spatial feature is input to a global average pooling layer to obtain a first feature, the motion feature is input to a global pooling layer to obtain a second feature, and the first feature and the second feature are input to a concatenation layer to obtain a concatenated feature; the concatenated feature is sequentially input to a fully connected layer, a ReLU layer, a fully connected layer and a Sigmoid layer to obtain a third feature, the third feature corresponds to a spatial weight of the spatial feature and a motion weight of the motion feature respectively; the product of the spatial feature and the spatial weight and the product of the motion feature and the motion weight are summed to obtain a fusion feature; the spatial feature of the spatial feature sequence and the motion feature of the corresponding motion feature sequence are sequentially fused to obtain a fusion feature sequence.

[0022] Optionally, the training process of the target power generation model includes:

[0023] The fusion feature sequence and the power generation sequence are obtained from the database; the fusion feature sequence and the corresponding power generation sequence are integrated into a plurality of training data and test data;

[0024] The plurality of training data are input to an artificial intelligence model for training, and the test data are used to test the trained artificial intelligence model, so as to finally obtain a target power generation model with the fusion feature sequence and the corresponding power generation sequence as input and the power generation sequence as output.

[0025] Secondly, a smart energy management system applied to shipbuilding is proposed, and the system includes:

[0026] A data acquisition module acquires the power consumption of the target shipbuilding area in the current time period;

[0027] A photovoltaic power generation module calculates the corresponding power generation sequence of the target shipbuilding area in the current time period according to the photovoltaic power generation method, and calculates the total amount of photovoltaic power generation and the power generation stability according to the power generation sequence;

[0028] The energy management module compares the power consumption and the total amount of photovoltaic power generation, and if the power consumption is greater than or equal to the total amount of photovoltaic power generation, determines whether the power generation stability exceeds a stability threshold, and if the power generation stability exceeds the stability threshold, switches to the grid power generation mode; if the power consumption is less than the total amount of photovoltaic power generation, switches to the grid power generation mode.

[0029] Optionally, the time period determination module comprises an image data module, a feature conversion module, a feature fusion module and a power generation analysis module.

[0030] The image data module is configured to acquire meteorological image data of a target region according to a time period, divide the meteorological image data into a meteorological image sequence according to the time period, and obtain a corresponding optical flow image sequence by substituting the meteorological image sequence into an optical flow algorithm; the images in the meteorological image sequence and the optical flow image sequence are in a one-to-one correspondence.

[0031] The feature conversion module is configured to import the meteorological image sequence into a target feature model to obtain a spatial feature sequence, and import the optical flow image sequence into the target feature model to obtain a motion feature sequence.

[0032] The feature fusion module is configured to execute a modal fusion model on the spatial feature sequence and the motion feature sequence to obtain a fused feature sequence.

[0033] The power generation analysis module is configured to import the fused feature sequence into a target power generation model to obtain a corresponding power generation sequence.

[0034] Optionally, the feature conversion module comprises a meteorological feature module and a motion feature module.

[0035] The meteorological feature module is configured to determine the meteorological image sequence as a first input feature matrix, input the first input feature matrix into a down-sampling layer to obtain a first down-sampling feature matrix, input the first down-sampling feature matrix into a two-dimensional convolution layer to obtain a first convolution feature matrix, input the first convolution feature matrix into a batch normalization layer to obtain a first batch feature matrix, input the first batch feature matrix into a ReLU layer to obtain a first output feature matrix; the first output feature matrix is recorded as the spatial feature sequence.

[0036] The motion feature module is configured to determine the optical flow image sequence as a second input feature matrix, input the second input feature matrix into a down-sampling layer to obtain a second down-sampling feature matrix, input the second down-sampling feature matrix into a two-dimensional convolution layer to obtain a second convolution feature matrix, input the second convolution feature matrix into a batch normalization layer to obtain a second batch feature matrix, input the second batch feature matrix into a ReLU layer to obtain a second output feature matrix; the second output feature matrix is recorded as the motion feature sequence.

[0037] Optionally, the feature fusion module comprises:

[0038] The spatial features of the spatial feature sequence and the motion features of the corresponding motion feature sequence are extracted respectively, the spatial features are input into a global average pooling layer to obtain first features, the motion features are input into a global pooling layer to obtain second features, the first features and the second features are introduced into a concatenation layer to obtain concatenation features, the concatenation features are sequentially input into a fully connected layer, a ReLU layer, a fully connected layer and a Sigmoid layer to obtain third features, the third features correspond to spatial weights of the spatial features and motion weights of the motion features respectively, the product of the spatial features and the spatial weights and the product of the motion features and the motion weights are summed to obtain fusion features, and the spatial features of the spatial feature sequence and the motion features of the corresponding motion feature sequence are sequentially fused to obtain a fusion feature sequence.

[0039] Optionally, the training process of the target power generation model comprises:

[0040] The fusion feature sequence and the power generation sequence are obtained from the database, and the fusion feature sequence and the corresponding power generation sequence are integrated into a plurality of training data and test data.

[0041] The plurality of training data are introduced into the artificial intelligence model for training, the test data are used to test the trained artificial intelligence model, and finally the target power generation model with the input of the fusion feature sequence and the corresponding power generation sequence and the output of the power generation sequence is obtained.

[0042] A smart energy management system applied to shipbuilding is proposed, comprising:

[0043] A data acquisition module acquires the power consumption of the target shipbuilding area in the current time period.

[0044] A photovoltaic power generation module calculates the corresponding power generation sequence of the target shipbuilding area in the current time period according to the photovoltaic power generation method, and calculates the total amount of photovoltaic power generation and the power generation stability according to the power generation sequence.

[0045] An energy management module compares the power consumption and the total amount of photovoltaic power generation, if the power consumption is greater than or equal to the total amount of photovoltaic power generation, it is judged whether the power generation stability exceeds the stability threshold, if the power generation stability exceeds the stability threshold, it is switched to the grid power generation mode, if the power consumption is less than the total amount of photovoltaic power generation, it is switched to the grid power generation mode.

[0046] Optionally, the feature conversion module comprises a meteorological feature module and a motion feature module.

[0047] The weather feature module is configured to determine a first input feature matrix from the weather image sequence, input the first input feature matrix into a down-sampling layer to obtain a first down-sampled feature matrix, input the first down-sampled feature matrix into a two-dimensional convolution layer to obtain a first convolution feature matrix, input the first convolution feature matrix into a batch normalization layer to obtain a first batch feature matrix, and input the first batch feature matrix into a ReLU layer to obtain a first output feature matrix; the first output feature matrix is recorded as a spatial feature sequence.

[0048] The motion feature module is configured to determine a second input feature matrix from the optical flow image sequence, input the second input feature matrix into a down-sampling layer to obtain a second down-sampled feature matrix, input the second down-sampled feature matrix into a two-dimensional convolution layer to obtain a second convolution feature matrix, input the second convolution feature matrix into a batch normalization layer to obtain a second batch feature matrix, and input the second batch feature matrix into a ReLU layer to obtain a second output feature matrix; the second output feature matrix is recorded as a motion feature sequence.

[0049] Optionally, the feature fusion module further comprises:

[0050] The spatial feature of the spatial feature sequence and the motion feature of the corresponding motion feature sequence are extracted respectively, the spatial feature is input into a global average pooling layer to obtain a first feature, the motion feature is input into a global pooling layer to obtain a second feature, the first feature and the second feature are introduced into a splicing layer to obtain spliced features, the spliced features sequentially pass through a full connection layer, a ReLU layer, a full connection layer and a Sigmoid layer to obtain a third feature, the third feature corresponds to a spatial weight of the spatial feature and a motion weight of the motion feature respectively, the product of the spatial feature and the spatial weight and the product of the motion feature and the motion weight are summed to obtain a fusion feature, and the spatial feature of the spatial feature sequence and the motion feature of the corresponding motion feature sequence are sequentially fused to obtain a fusion feature sequence.

[0051] Optionally, the power generation analysis module further comprises:

[0052] The fusion feature sequence and the power generation sequence are obtained from the database, the fusion feature sequence and the corresponding power generation sequence are integrated into a plurality of training data and test data, the plurality of training data are introduced into an artificial intelligence model for training, and the test data are used to test the trained artificial intelligence model; and finally, a target power generation model with the fusion feature sequence and the corresponding power generation sequence as input and the power generation sequence as output is obtained.

[0053] Optionally, the training process of the target power generation model comprises:

[0054] The fusion feature sequence and the power generation sequence are obtained from the database; and the fusion feature sequence and the corresponding power generation sequence are integrated into a plurality of training data and test data.

[0055] The plurality of training data is imported into the artificial intelligence model for training, and the test data is tested on the trained artificial intelligence model, and finally a target power generation model with the input being the fusion feature sequence and the corresponding power generation sequence and the output being the power generation sequence is obtained.

[0056] The beneficial effects of the present application are:

[0057] The present application provides a smart energy management method applied to shipbuilding. The power consumption of a target shipbuilding area in a current time period is obtained. The power generation sequence corresponding to the target shipbuilding area in the current time period is calculated according to a photovoltaic power generation mode, and the total amount of photovoltaic power generation and the power generation stability are calculated according to the power generation sequence. The power consumption and the total amount of photovoltaic power generation are compared. If the power consumption is greater than or equal to the total amount of photovoltaic power generation, it is determined whether the power generation stability exceeds a stability threshold. If the power generation stability exceeds the stability threshold, the power grid power generation mode is switched. If the power consumption is less than the total amount of photovoltaic power generation, the power grid power generation mode is switched. The present application improves the prediction accuracy of power generation stability, thereby realizing accurate switching between photovoltaic power generation and power grid power generation, and improving the energy management efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 A flowchart of the smart energy management method applied to shipbuilding provided by the present application is provided.

[0059] Figure 2 A structure diagram of the modal fusion model provided by the present application is provided.

[0060] Figure 3 A framework diagram of the smart energy management system applied to shipbuilding provided by the present application is provided. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0062] The present application provides a smart energy management method applied to shipbuilding. Referring to Figure 1 , Figure 1 A flowchart of the smart energy management method applied to shipbuilding provided by the present application is provided. The method comprises the following steps:

[0063] Obtaining the power consumption of the target ship manufacturing area in the current time period; calculating the power generation sequence corresponding to the target ship manufacturing area in the current time period according to the photovoltaic power generation mode, and calculating the total photovoltaic power generation and the power generation stability according to the power generation sequence; comparing the power consumption with the total photovoltaic power generation, if the power consumption is greater than or equal to the total photovoltaic power generation, determining whether the power generation stability exceeds the stability threshold, if the power generation stability exceeds the stability threshold, switching to the grid power generation mode; if the power consumption is less than the total photovoltaic power generation, switching to the grid power generation mode.

[0064] The intelligent energy management method applied to ship manufacturing provided by the embodiment of the application improves the power generation stability prediction accuracy, and realizes accurate switching between photovoltaic power generation and grid power generation, thereby improving the energy management efficiency.

[0065] Specifically, the time period includes one hour, half a day, one day, etc.; the target ship manufacturing area is a region of a ship manufacturing plant; the power consumption is the total power required for ship manufacturing; and the stability threshold is obtained by workers based on historical experience.

[0066] In an implementation manner, obtaining the power generation sequence of the target area according to the time period includes:

[0067] Obtaining meteorological image data of the target area according to the time period, dividing the meteorological image data into a meteorological image sequence according to the time period, and substituting the meteorological image sequence into an optical flow algorithm to obtain a corresponding optical flow image sequence; the images in the meteorological image sequence and the optical flow image sequence are in one-to-one correspondence;

[0068] Importing the meteorological image sequence into a target feature model to obtain a spatial feature sequence, and importing the optical flow image sequence into the target feature model to obtain a motion feature sequence;

[0069] Performing a modal fusion model on the spatial feature sequence and the motion feature sequence to obtain a fusion feature sequence;

[0070] Importing the fusion feature sequence into a target power generation model to obtain a corresponding power generation sequence.

[0071] Specifically, the optical flow algorithm can be a Farneback algorithm, a Horn-Schunck optical flow algorithm, etc.; the spatial feature specifically includes the thickness of a cloud layer, the coverage area of the cloud layer, etc.; the motion feature specifically includes the change rate of the cloud layer and the motion trajectory of the cloud layer, etc.; the fusion of the spatial feature and the motion feature specifically can be the relationship between the moving direction of the cloud layer and the wind speed, etc.

[0072] In an implementation manner, importing the meteorological image sequence into a target feature model to obtain a spatial feature sequence, and importing the optical flow image sequence into the target feature model to obtain a motion feature sequence include:

[0073] The meteorological image sequence is determined as a first input feature matrix, the first input feature matrix is input to a down-sampling layer to obtain a first down-sampled feature matrix, the first down-sampled feature matrix is input to a two-dimensional convolution layer to obtain a first convolution feature matrix, the first convolution feature matrix is input to a batch normalization layer to obtain a first batch feature matrix, and the first batch feature matrix is input to a ReLU layer to obtain a first output feature matrix; the first output feature matrix is denoted as a spatial feature sequence;

[0074] The optical flow image sequence is determined as a second input feature matrix, the second input feature matrix is input to a down-sampling layer to obtain a second down-sampled feature matrix, the second down-sampled feature matrix is input to a two-dimensional convolution layer to obtain a second convolution feature matrix, the second convolution feature matrix is input to a batch normalization layer to obtain a second batch feature matrix, and the second batch feature matrix is input to a ReLU layer to obtain a second output feature matrix; the second output feature matrix is denoted as a motion feature sequence.

[0075] In an implementation manner, referring to Figure 2 , Figure 2 A structure diagram of a modal fusion model provided by the embodiment of the application is shown in FIG. 1. The modal fusion model is used to perform modal fusion on the spatial feature sequence and the motion feature sequence to obtain a fusion feature sequence.

[0076] The spatial feature of the spatial feature sequence and the motion feature corresponding to the motion feature sequence are extracted respectively, the spatial feature is input to a global average pooling layer to obtain a first feature, the motion feature is input to a global pooling layer to obtain a second feature, and the first feature and the second feature are input to a concatenation layer to obtain a concatenated feature; the concatenated feature is sequentially input to a full connection layer, a ReLU layer, a full connection layer and a Sigmoid layer to obtain a third feature, the third feature corresponds to a spatial weight of the spatial feature and a motion weight of the motion feature respectively; the product of the spatial feature and the spatial weight and the product of the motion feature and the motion weight are summed to obtain a fusion feature; the spatial feature of the spatial feature sequence and the motion feature corresponding to the motion feature sequence are sequentially fused to obtain a fusion feature sequence.

[0077] In an implementation manner, in the modal fusion model, for two input spatial features and motion features, the spatial feature is denoted as a k ∈R B×C×H×W , the motion feature is denoted as a y ∈R B×C×H×W , is input to a global average pooling to obtain global information and a reduced dimension feature denoted as and is obtained by and concatenation a c ∈R B×2C ; then, a c ∈R B×2CThe spatial weight is denoted as w1, and the motion feature is denoted as w2=1-w1, and the sum of w1 and w2 is equal to 1, which are obtained by sequentially inputting into the full connection layer, the ReLU layer, the full connection layer and the Sigmoid layer; finally, the fusion feature is obtained, and all the fusion features are combined to obtain a fusion feature sequence;

[0078] The calculation formula of the fusion feature is:

[0079]

[0080] Wherein, a represents the fusion feature, GAP() represents the global pooling function, concat() represents the concatenation function, and FC() represents the full connection function.

[0081] In an implementation manner, the modal fusion model realizes effective fusion of the spatial feature sequence and the motion feature sequence, thereby significantly improving the completeness and accuracy of feature expression. The global average pooling layer extracts the global information of the spatial feature and the motion feature, which can effectively capture the key information in the feature sequence and realize dimension reduction, reduce the calculation complexity, and at the same time, retain the core semantics of the feature. The design of the concatenation layer integrates the dimension-reduced spatial feature and motion feature, providing a basis for subsequent feature weighted fusion. After the processing of the full connection layer, the ReLU layer, the full connection layer and the Sigmoid layer, the weights of the spatial feature and the motion feature are generated. This dynamic weighting method can adaptively adjust the proportion of the spatial feature and the motion feature in the fusion process according to the characteristics of the input data, reducing the information loss or redundancy problem. The fusion feature sequence obtained by weighted summation can better reflect the spatial and motion information in the original data, improving the data representation ability of the model and providing guarantee for subsequent analysis.

[0082] In an implementation manner, the total amount of photovoltaic power generation and the power generation stability are calculated according to the power generation sequence, including:

[0083] The total amount of photovoltaic power generation is obtained by summing the power generation sequence, and the power generation stability is calculated by variance according to the power generation sequence and the current time period.

[0084] In an implementation manner, the fusion feature sequence is introduced into the target power generation model to obtain a corresponding power generation sequence, including:

[0085] The fusion feature sequence and the power generation sequence are obtained from the database; the fusion feature specifically includes cloud layer thickness, cloud layer movement trajectory, etc.; the power generation sequence is obtained by experts according to the fusion feature sequence, specifically, the greater the thickness of the cloud layer, the worse the light transmittance of the cloud layer, the stronger the attenuation ability of the sunlight, and the lower the power generation equipment receives the power generation; the movement direction and the movement intensity of the cloud layer are obtained through the cloud layer movement trajectory, the movement direction of the cloud layer indicates the cloud coverage trend, and the movement intensity of the cloud layer indicates the cloud sun-shading change rate; the greater the cloud sun-shading change rate, the more likely to cause the power generation to rise or fall sharply, and the fusion feature sequence and the corresponding power generation sequence are integrated into a plurality of training data and test data;

[0086] The plurality of training data is imported into the artificial intelligence model for training, and the test data is used to test the trained artificial intelligence model; specifically, the fusion feature sequence and the corresponding power generation sequence in the test data are input into the trained artificial intelligence model, and the power generation sequence is output. Whether the absolute value of the difference between the power generation sequence and the power generation sequence recorded in the test data is within an acceptable range; yes, the test data passes the test, and the next set of test data is tested; no, the related parameters of the artificial intelligence model need to be adjusted, and the test data is continuously used for testing; until a set proportion of test data passes the test; finally, a target power generation model with input fusion feature sequence and corresponding power generation sequence and output power generation sequence is obtained; wherein the artificial intelligence model is a PerfCNN-LSTM model.

[0087] Specifically, the database is used by the staff to store data, and the preset power consumption threshold, the preset sensitivity threshold and the preset power generation stability threshold are obtained from the database.

[0088] Based on the same inventive concept, the embodiments of the present application also provide a smart energy management system applied to shipbuilding. Referring to Figure 3 , Figure 3 A framework diagram of a smart energy management system applied to shipbuilding is provided for the embodiments of the present application, comprising:

[0089] The data acquisition module acquires the power consumption of the target shipbuilding area in the current time period;

[0090] The photovoltaic power generation module calculates the power generation sequence corresponding to the target shipbuilding area in the current time period according to the photovoltaic power generation method, and calculates the total amount of photovoltaic power generation and the power generation stability according to the power generation sequence;

[0091] The energy management module compares the power consumption and the total photovoltaic power generation amount, if the power consumption is greater than or equal to the total photovoltaic power generation amount, it is judged whether the power generation stability exceeds the stability threshold, if the power generation stability exceeds the stability threshold, it is switched to the grid power generation mode, if the power consumption is less than the total photovoltaic power generation amount, it is switched to the grid power generation mode.

[0092] The intelligent energy management system applied to ship manufacturing provided by the embodiment of the present application improves the power generation stability prediction accuracy, and then realizes the accurate switching of photovoltaic power generation and grid power generation, and improves the energy management efficiency.

[0093] It should be noted that in this paper, such as the term "including", "including" or any other variant is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or equipment.

[0094] Although the embodiments of the present application have been shown and described, it can be understood by those of ordinary skill in the art that various changes, modifications, replacements and variations can be made to these embodiments without departing from the principles and spirits of the present application.

Claims

1. A smart energy management method applied to shipbuilding, characterized in that, The method comprises: acquiring the power consumption of the target ship manufacturing area in the current time period; calculating the power generation sequence of the target ship manufacturing area in the current time period according to the photovoltaic power generation mode, and calculating the total amount of photovoltaic power generation and the power generation stability according to the power generation sequence; comparing the power consumption and the total amount of photovoltaic power generation, if the power consumption is greater than or equal to the total amount of photovoltaic power generation, determining whether the power generation stability exceeds the stability threshold, if the power generation stability exceeds the stability threshold, switching to the grid power generation mode, and if the power consumption is less than the total amount of photovoltaic power generation, switching to the grid power generation mode.

2. The intelligent energy management method applied to shipbuilding according to claim 1, characterized in that, The calculation of the power generation sequence of the target ship manufacturing area in the current time period according to the photovoltaic power generation mode comprises: acquiring meteorological image data of the target area according to the current time period, dividing the meteorological image data into a meteorological image sequence according to the current time period, and substituting the meteorological image sequence into an optical flow algorithm to obtain a corresponding optical flow image sequence; the images in the meteorological image sequence and the optical flow image sequence are in one-to-one correspondence; introducing the meteorological image sequence into a target feature model to obtain a spatial feature sequence, and introducing the optical flow image sequence into the target feature model to obtain a motion feature sequence; executing a modal fusion model on the spatial feature sequence and the motion feature sequence to obtain a fusion feature sequence; introducing the fusion feature sequence into a target power generation model to obtain a corresponding power generation sequence.

3. The intelligent energy management method applied to shipbuilding according to claim 2, characterized in that, The introduction of the meteorological image sequence into the target feature model to obtain the spatial feature sequence, and the introduction of the optical flow image sequence into the target feature model to obtain the motion feature sequence comprise: determining the meteorological image sequence as a first input feature matrix, inputting the first input feature matrix into a down-sampling layer to obtain a first down-sampling feature matrix, inputting the first down-sampling feature matrix into a two-dimensional convolution layer to obtain a first convolution feature matrix, inputting the first convolution feature matrix into a batch normalization layer to obtain a first batch feature matrix, inputting the first batch feature matrix into a ReLU layer to obtain a first output feature matrix; the first output feature matrix is recorded as the spatial feature sequence; determining the optical flow image sequence as a second input feature matrix, inputting the second input feature matrix into a down-sampling layer to obtain a second down-sampling feature matrix, inputting the second down-sampling feature matrix into a two-dimensional convolution layer to obtain a second convolution feature matrix, inputting the second convolution feature matrix into a batch normalization layer to obtain a second batch feature matrix, inputting the second batch feature matrix into a ReLU layer to obtain a second output feature matrix; the second output feature matrix is recorded as the motion feature sequence.

4. The intelligent energy management method applied to shipbuilding according to claim 2, characterized in that, The execution of the modal fusion model on the spatial feature sequence and the motion feature sequence to obtain the fusion feature sequence comprises: The spatial features of the spatial feature sequence and the motion features of the motion feature sequence corresponding thereto are extracted respectively, the spatial features are input into a global average pooling layer to obtain first features, the motion features are input into a global pooling layer to obtain second features, and the first features and the second features are introduced into a concatenation layer to obtain concatenated features; the concatenated features sequentially pass through a fully connected layer, a ReLU layer, a fully connected layer, and a Sigmoid layer to obtain third features, which correspond to spatial weights of the spatial features and motion weights of the motion features respectively; the product of the spatial features and the spatial weights and the product of the motion features and the motion weights are summed to obtain fused features; the spatial features of the spatial feature sequence and the motion features of the motion feature sequence corresponding thereto are sequentially fused to obtain a fused feature sequence.

5. The intelligent energy management method applied to shipbuilding according to claim 2, characterized in that, The training process of the target power generation model comprises: obtaining the fused feature sequence and the power generation power sequence from the database; integrating the fused feature sequence and the power generation power sequence corresponding thereto into a plurality of training data and test data; introducing the plurality of training data into an artificial intelligence model for training, testing the trained artificial intelligence model with the test data, and finally obtaining a target power generation model with the input being the fused feature sequence and the corresponding power generation power sequence and the output being the power generation power sequence.

6. A smart energy management system applied to shipbuilding, characterized in that, The system comprises: a data acquisition module configured to acquire an electricity consumption of a target ship manufacturing area in a current time period; a photovoltaic power generation module configured to calculate a corresponding power generation power sequence of the target ship manufacturing area in the current time period according to a photovoltaic power generation mode, and calculate a total amount of photovoltaic power generation and a power generation stability according to the power generation power sequence; an energy management module configured to compare the electricity consumption with the total amount of photovoltaic power generation, determine whether the power generation stability exceeds a stability threshold if the electricity consumption is greater than or equal to the total amount of photovoltaic power generation, and switch to a power grid power generation mode if the power generation stability exceeds the stability threshold, or switch to the power grid power generation mode if the electricity consumption is less than the total amount of photovoltaic power generation.

7. The intelligent energy management system applied to shipbuilding according to claim 6, characterized in that, The power supply module comprises an image data module, a feature conversion module, a feature fusion module, and a power generation analysis module: The image data module is configured to acquire meteorological image data of a target area according to a current time period, divide the meteorological image data into a meteorological image sequence according to the current time period, and obtain a corresponding optical flow image sequence by substituting the meteorological image sequence into an optical flow algorithm; the images in the meteorological image sequence and the optical flow image sequence are in a one-to-one correspondence; The feature conversion module is configured to introduce the meteorological image sequence into a target feature model to obtain a spatial feature sequence, and introduce the optical flow image sequence into the target feature model to obtain a motion feature sequence; The feature fusion module is configured to execute a modal fusion model on the spatial feature sequence and the motion feature sequence to obtain a fused feature sequence; The power generation analysis module is configured to introduce the fused feature sequence into a target power generation model to obtain a corresponding power generation power sequence.

8. The intelligent energy management system applied to shipbuilding according to claim 7, characterized in that, The feature conversion module comprises a meteorological feature module and a motion feature module: The weather feature module is configured to determine a first input feature matrix from the weather image sequence, input the first input feature matrix into a down-sampling layer to obtain a first down-sampled feature matrix, input the first down-sampled feature matrix into a two-dimensional convolution layer to obtain a first convolution feature matrix, input the first convolution feature matrix into a batch normalization layer to obtain a first batch feature matrix, input the first batch feature matrix into a ReLU layer to obtain a first output feature matrix, and record the first output feature matrix as a spatial feature sequence. The motion feature module is configured to determine a second input feature matrix from the optical flow image sequence, input the second input feature matrix into a down-sampling layer to obtain a second down-sampled feature matrix, input the second down-sampled feature matrix into a two-dimensional convolution layer to obtain a second convolution feature matrix, input the second convolution feature matrix into a batch normalization layer to obtain a second batch feature matrix, input the second batch feature matrix into a ReLU layer to obtain a second output feature matrix, and record the second output feature matrix as a motion feature sequence.

9. The intelligent energy management system for shipbuilding of claim 7, wherein, The feature fusion module comprises: The spatial feature of the spatial feature sequence and the motion feature of the corresponding motion feature sequence are extracted respectively, the spatial feature is input into a global average pooling layer to obtain a first feature, the motion feature is input into a global pooling layer to obtain a second feature, the first feature and the second feature are introduced into a concatenation layer to obtain a concatenated feature, the concatenated feature sequentially passes through a full connection layer, a ReLU layer, a full connection layer and a Sigmoid layer to obtain a third feature, the third feature corresponds to a spatial weight of the spatial feature and a motion weight of the motion feature respectively, the product of the spatial feature and the spatial weight and the product of the motion feature and the motion weight are summed to obtain a fusion feature, and the spatial feature of the spatial feature sequence and the motion feature of the corresponding motion feature sequence are sequentially fused to obtain a fusion feature sequence.

10. The intelligent energy management system for shipbuilding of claim 7, wherein, The training process of the target power generation model comprises: The fusion feature sequence and the power generation sequence are obtained from the database, and the fusion feature sequence and the corresponding power generation sequence are integrated into a plurality of training data and test data; The plurality of training data are introduced into an artificial intelligence model for training, the test data are used to test the trained artificial intelligence model, and finally a target power generation model is obtained, wherein the input is the fusion feature sequence and the corresponding power generation sequence, and the output is the power generation sequence.

Images