A method and system for ultra-short-term wind speed prediction based on the fusion of CNN and GRU

Abstract

The present invention discloses a method and system for ultra-short-term wind speed prediction by integrating CNN and GRU. S1, obtain the operation data of the unit and meteorological data; S2, perform modeling data preparation, and the data preparation includes unit data preparation, meteorological data preparation and trend feature extraction; the unit data preparation includes target unit data reconstruction and neighboring unit selection; S3, construct a neighboring unit CNN module, a target unit CNN module, a meteorological GRU module and a trend feature GRU module according to the prepared data, fuse the features of the four modules, and perform model training and parameter optimization; S4, use the trained optimal model to predict the data to be predicted to obtain the wind speed prediction result. Combining the characteristics of the unit operation data with the characteristics of the meteorological data, the ultra-short-term wind speed prediction function of the wind turbine is realized.

Description

Technical Field

[0001] The present invention belongs to the field of wind turbines, and relates to a method and system for ultra-short-term wind speed prediction based on the fusion of CNN and GRU. Background Art

[0002] The operating environment of wind turbines is harsh and the wind conditions are complex. Especially in complex mountainous areas, the changes in wind speed and direction are diverse, and the fatigue impact on the turbines is becoming more and more obvious. Therefore, if accurate wind speed and direction predictions can be provided, feedforward control can be attempted to adjust the control strategy in advance for extreme wind conditions to cope with extreme wind condition changes. At present, the controller PLC of the turbine adjusts the control strategy based on data at the millisecond level. However, due to the volatility of wind speed, at least wind speed prediction values at the second level are required to meet the requirements of feedforward control. Summary of the Invention

[0003] An object of the present invention is to overcome the above-mentioned drawbacks of the prior art and provide a method and system for ultra-short-term wind speed prediction based on the fusion of CNN and GRU, which combines the characteristics of the operating data of the turbine and the characteristics of meteorological data to realize the ultra-short-term wind speed prediction function of the wind turbine.

[0004] To achieve the above object, the present invention adopts the following technical solutions:

[0005] An ultra-short-term wind speed prediction method based on the fusion of CNN and GRU includes the following processes:

[0006] S1. Obtain the operating data of the turbine and meteorological data;

[0007] S2. Prepare the modeling data. The data preparation includes the preparation of turbine data, the preparation of meteorological data, and the extraction of trend features. The preparation of turbine data includes the reconstruction of target turbine data and the selection of adjacent turbines;

[0008] S3. Construct an adjacent turbine CNN module, a target turbine CNN module, a meteorological GRU module, and a trend feature GRU module according to the prepared data, fuse the features of the four modules, and perform model training and parameter optimization;

[0009] S4. Use the trained optimal model to predict the data to be predicted and obtain the wind speed prediction result.

[0010] Preferably, in S1, the operating data of the turbine includes wind speed, wind direction, power, rotational speed, pitch angle, ambient temperature, and ambient humidity, and the meteorological data includes wind speed and wind direction.

[0011] Preferably, in S1, the time resolution of the meteorological data is 1 h, and the time resolution of the operating data of the turbine is 30 s.

[0012] Preferably, in S2, the process of reconstructing the target unit data is as follows: considering the data of different dimensions as different channels, reconstructing the data of each channel to form a matrix, completing the construction of the target unit data, and then selecting neighboring units.

[0013] Furthermore, there are two methods for selecting neighboring units, and either one can be chosen. One method is to use the method of unit clustering to group the units, taking each clustering result as a group, and the units within each group are neighboring units to each other. When predicting the target unit, refer to the wind speed characteristics of other units within the group where the target unit is located;

[0014] The other method is through correlation analysis. Select the units with a relatively high correlation with the wind speed of the target unit as neighboring units, calculate the correlation coefficient, sort the calculated correlation coefficients, and select the ones with larger correlation coefficients as the neighboring units of the target unit.

[0015] Preferably, in S2, the extraction of trend characteristics includes the trend characteristics of unit data and the trend characteristics of meteorological data. For the trend characteristics of unit data, take the past hour as the time window and extract the statistical characteristics within the time window; for meteorological data, extract the trend characteristics and calculate the difference between the last point and the first point within the past 10 hours.

[0016] Preferably, in S3, the data selected by neighboring unit selection is subjected to feature extraction through multiple CNN blocks, the target unit data is subjected to feature extraction through multiple CNN blocks, the meteorological data is subjected to feature extraction through multiple GRU blocks, and the trend feature data is subjected to feature extraction through multiple GRU blocks. The four parts of the extracted features are connected through a concatenate layer, and different weights and confidences are given to different features through an Attenion layer. Finally, the prediction ability of the model is enhanced through an FC layer and multiple DNN networks. The last layer of the DNN network is the output layer.

[0017] A very short-term wind speed prediction system integrating CNN and GRU includes:

[0018] A data acquisition module for acquiring unit operation data and meteorological data;

[0019] A modeling data preparation module for preparing modeling data. The data preparation includes unit data preparation, meteorological data preparation, and trend feature extraction; unit data preparation includes target unit data reconstruction and neighboring unit selection;

[0020] A model construction and training module for constructing a neighboring unit CNN module, a target unit CNN module, a meteorological GRU module, and a trend feature GRU module based on the prepared data, fusing the features of the four modules, and performing model training and parameter optimization;

[0021] A prediction module, which is used to predict the data to be predicted by using the trained optimal model to obtain the wind speed prediction result.

[0022] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the ultra-short-term wind speed prediction method based on the CNN+GRU fusion are implemented.

[0023] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the steps of the ultra-short-term wind speed prediction method based on the CNN+GRU fusion are implemented.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] The present invention first obtains the unit operation data and meteorological data; secondly, data preparation is carried out, including unit data preparation, meteorological data preparation, and trend feature extraction. Unit data preparation includes target unit data reconstruction and neighboring unit selection; subsequently, the prepared data is used for model construction and model training. Here, a total of four parts of models are constructed: the neighboring unit CNN module, the target unit CNN module, the meteorological GRU module, and the trend feature GRU module. The features of the four parts are fused, and model training and parameter optimization are carried out; finally, the trained optimal model is used to predict the data to be predicted to obtain the wind speed prediction result. The present invention realizes the ultra-short-term wind speed prediction function by considering more data sources, such as unit data, meteorological data, neighboring unit data, trend features, etc., fusing different data features, and constructing and training and optimizing the model. Brief Description of the Drawings

[0026] Figure 1 It is a flowchart of the prediction method of the present invention;

[0027] Figure 2 It is a schematic diagram of neighboring unit clustering of the present invention;

[0028] Figure 3 It is a schematic diagram of model construction of the present invention. Detailed Embodiments

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0030] It should be noted that the terms "front", "rear", "left", "right", "upper" and "lower" used in the following description refer to the directions in the accompanying drawings, and the terms "inner" and "outer" refer to the directions towards or away from the geometric center of a specific component respectively.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this invention belongs. The terms used in the specification of this invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0032] The ultra-short-term wind speed prediction method based on the fusion of CNN and GRU according to the present invention combines the characteristics of unit operation data and the characteristics of meteorological data to realize the ultra-short-term wind speed prediction function of wind turbines. First, obtain the original data, including unit operation data and meteorological data. Secondly, perform data preparation, including unit data preparation, meteorological data preparation, and trend feature extraction. Among them, for unit data preparation: the clustering method is used to select adjacent units for the unit data. In wind speed prediction, not only the wind speed situation of the target unit itself is considered, but also the correlation of the wind speeds of adjacent units is considered for feature fusion; for meteorological data preparation: it is corresponding to the unit data through the time index, and the original meteorological data is retained; for trend feature data preparation: it includes unit feature trends, meteorological feature trends, etc. Subsequently, build and train the model for the prepared data. Here, a total of four parts of the model are constructed: the adjacent unit CNN module, the target unit CNN module, the meteorological GRU module, and the trend feature GRU module. And the features of the four parts are fused, and the model is trained and the parameters are optimized. Finally, use the trained optimal model to predict the data to be predicted, obtain the wind speed prediction result, and finally realize the ultra-short-term wind speed prediction function.

[0033] As Figure 1 shown, it specifically includes the following processes:

[0034] First, starting from step S101, the original data is obtained. The original data includes unit operation data and meteorological data. The unit operation data includes dimensions such as wind speed, wind direction, power, rotational speed, pitch angle, ambient temperature, and ambient humidity. The meteorological data includes wind speed and wind direction. Regarding the time resolution requirement of the data, the time resolution of the meteorological data is 1 hour. If more refined data is required, data simulation needs to be carried out. The time resolution of the unit operation data depends on the acquisition frequency of the Scada system, and the common time resolutions are 1s, 5s, and 7s. To eliminate the volatility of the wind speed, here the original 1s data is downsampled at 30s time intervals, and the average value within each 30s time segment is used as the result after sampling. Finally, the time resolution of the unit operation data is 30s.

[0035] In step S102, the data preparation layer, data preparation for modeling is carried out. The data preparation includes unit data preparation, meteorological data preparation, and trend feature extraction.

[0036] Among them, the unit data preparation includes: target unit data reconstruction and neighboring unit selection. The main purpose of target unit data reconstruction is to construct historical data for a period of time to predict the wind speed for a future period of time, and construct a matrix to adapt to the input form of the CNN model. An effective data construction method is to regard data of different dimensions such as wind speed, wind speed, wind direction, power, rotational speed, pitch angle, ambient temperature, and ambient humidity as different channels, and perform data reconstruction on the data of each channel to form a matrix. For example, using the data of the previous 1 hour to predict the wind speed of the next 10 minutes. Assuming the time resolution of the unit data is 30s, there are 120 samples of 1 hour of historical data, and 20 samples for predicting the wind speed of 10 minutes. The matrix construction method is as follows.

[0037]

[0038] After completing the construction of the target unit data, considering the correlation of the wind speed of neighboring units, the historical wind speed information of neighboring units can be taken into account to improve the accuracy of wind speed prediction of the target unit. Therefore, it is necessary to select neighboring units of the target unit. An effective method is to use the method of unit clustering to group the units. Each clustering result is regarded as a group, and the units within each group are neighboring units to each other. When predicting the target unit, refer to the wind speed characteristics of other units within the group where the target unit is located. The schematic diagram of unit clustering is as Figure 2 shown.

[0039] In addition, another method can also be to use correlation analysis to regard the units with a relatively high correlation with the wind speed of the target unit as neighboring units, and calculate the correlation coefficient. The calculation formula is as follows:

[0040]

[0041] Among them, Cov(x,y) is the covariance of two variables, and σ x is the standard deviation of the variable. Sort the calculated correlation coefficients and select the ones with larger correlation coefficients as the neighboring units of the target unit. For example, if the number of neighboring units is 5, then construct the wind speeds of each unit as a matrix as the input of the CNN model, and the dimension of this matrix is 5×120×1.

[0042] The preparation of meteorological data is relatively easy. Directly select the meteorological data corresponding to the unit time. Since the time resolution of the meteorological data is 1 hour, select the data of the past 10 hours to predict the wind speed of the unit in the next 10 minutes. The meteorological data only has two dimensions, wind speed and wind direction, so the dimension of the constructed matrix is 10×2.

[0043] For the part of trend feature extraction, it includes the trend features of unit data and meteorological data. The trend features of unit data mainly take the past hour as the time window and extract the statistical features within the time window, such as the difference between the maximum value and the minimum value, standard deviation, mean, and trend feature (the difference between the last point and the first point). The meteorological data only extracts the trend feature, that is, calculates the difference between the last point and the first point within the past 10 hours.

[0044] In step S103, the model construction and model training layer, a total of four parts of the model are constructed: the neighboring unit CNN module, the target unit CNN module, the meteorological GRU module, and the trend feature GRU module. The model construction is as Figure 3 shown, that is, the neighboring unit data is subjected to feature extraction through multiple CNN blocks, the target unit data is subjected to feature extraction through multiple CNN blocks, the meteorological data is subjected to feature extraction through multiple GRU blocks, and the trend feature data is subjected to feature extraction through multiple GRU blocks. The four parts of the extracted features are connected through the concatenate layer, and different weights and confidences are given to different features through the Attention layer. Finally, the prediction ability of the model is enhanced through the FC layer and multiple DNN networks. The last DNN network is the output layer, including 20 output neurons corresponding to 20 predicted wind speed values within 10 minutes. After the network structure is built, the model training can be carried out. The improved adaptive gradient algorithm (Adadelta) is used for model training and parameter optimization, and the gradient update formula is as follows. After multiple steps of training, the optimal model can be obtained. After the model training is completed, the optimal model is saved for use during prediction.

[0045]

[0046] In step S104, the prediction layer obtains the optimal model and the data to be predicted, constructs the data in the same way as in the training phase, inputs the constructed data into the optimal model, and then the wind speed prediction result can be obtained, finally realizing the function of ultra-short-term wind speed prediction.

[0047] The following is the device embodiment of the present invention, which can be used to execute the method embodiment of the present invention. For the details not disclosed in the device embodiment, please refer to the method embodiment of the present invention.

[0048] In another embodiment of the present invention, an ultra-short-term wind speed prediction system based on the fusion of CNN and GRU is provided. This ultra-short-term wind speed prediction system based on the fusion of CNN and GRU can be used to implement the above-mentioned ultra-short-term wind speed prediction method based on the fusion of CNN and GRU. Specifically, this ultra-short-term wind speed prediction system based on the fusion of CNN and GRU includes a data acquisition module, a modeling data preparation module, a model construction and training module, and a prediction module.

[0049] Among them, the data acquisition module is used to acquire the unit operation data and meteorological data.

[0050] The modeling data preparation module is used to prepare the modeling data. The data preparation includes unit data preparation, meteorological data preparation, and trend feature extraction; the unit data preparation includes target unit data reconstruction and neighboring unit selection.

[0051] The model construction and training module is used to construct a neighboring unit CNN module, a target unit CNN module, a meteorological GRU module, and a trend feature GRU module according to the prepared data, fuse the features of the four modules, and perform model training and parameter optimization.

[0052] The prediction module is used to use the trained optimal model to predict the data to be predicted and obtain the wind speed prediction result.

[0053] In another embodiment of the present invention, a terminal device is provided. The terminal device includes a processor and a memory. The memory is used to store a computer program, and the computer program includes program instructions. The processor is used to execute the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding function. The processor described in the embodiment of the present invention can be used for the operation of the ultra-short-term wind speed prediction method based on CNN+GRU fusion, including: S1, obtaining the unit operation data and meteorological data; S2, preparing the modeling data, and the data preparation includes unit data preparation, meteorological data preparation, and trend feature extraction; the unit data preparation includes target unit data reconstruction and neighboring unit selection; S3, constructing a neighboring unit CNN module, a target unit CNN module, a meteorological GRU module, and a trend feature GRU module according to the prepared data, fusing the features of the four modules, and performing model training and parameter optimization; S4, using the trained optimal model to predict the prediction data to obtain the wind speed prediction result.

[0054] In another embodiment, the present invention also provides a computer-readable storage medium (Memory). The computer-readable storage medium is the memory device in the terminal device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and, of course, the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, and this storage space stores the operating system of the terminal. And, one or more instructions suitable for being loaded and executed by the processor are also stored in this storage space. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory.

[0055] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the ultra-short-term wind speed prediction method related to the CNN+GRU fusion in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor to perform the following steps: S1, obtaining unit operation data and meteorological data; S2, performing modeling data preparation, and the data preparation includes unit data preparation, meteorological data preparation, and trend feature extraction; the unit data preparation includes target unit data reconstruction and neighboring unit selection; S3, constructing a neighboring unit CNN module, a target unit CNN module, a meteorological GRU module, and a trend feature GRU module according to the prepared data, performing feature fusion on the four modules, and performing model training and parameter optimization; S4, using the trained optimal model to predict the prediction data to obtain a wind speed prediction result.

[0056] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0057] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0058] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0059] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus, causing a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one process or a plurality of processes and / or boxes Figure 1 one process or a plurality of processes and / or boxes Figure 1 steps for implementing the functions specified in one box or a plurality of boxes.

[0060] It should be noted that, in this document, relational terms such as first and second are used solely to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual such relationship or order between these entities or operations. Also, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0061] It should be understood that the above description is for illustrative purposes and not for limitation. Many embodiments and many applications other than the examples provided will be apparent to those skilled in the art upon reading the above description. Therefore, the scope of this patent should not be determined with reference to the above description, but should be determined with reference to the full scope of the foregoing claims and the equivalents thereof. For the sake of completeness, all articles and references, including patent applications and publications, are incorporated herein by reference. The omission of any aspect of the subject matter disclosed herein in the foregoing claims is not a waiver of that subject matter, nor should it be considered that the applicant has not considered that subject matter to be part of the disclosed inventive subject matter.

Claims

1. A method for predicting ultra-short-term wind speed by fusing CNN and GRU, characterized in that, It includes the following processes: S1. Obtain the unit operation data and meteorological data; The unit operation data includes wind speed, wind direction, power, rotational speed, pitch angle, ambient temperature, and ambient humidity, and the meteorological data includes wind speed and wind direction; S2. Prepare the modeling data. The data preparation includes unit data preparation, meteorological data preparation, and trend feature extraction; The unit data preparation includes target unit data reconstruction and neighboring unit selection; S3. Construct a neighboring unit CNN module, a target unit CNN module, a meteorological GRU module, and a trend feature GRU module based on the prepared data, fuse the features of the four modules, and perform model training and parameter optimization; The data selected by the neighboring unit undergoes feature extraction through multiple CNN blocks, the target unit data undergoes feature extraction through multiple CNN blocks, the meteorological data undergoes feature extraction through multiple GRU blocks, and the trend feature data undergoes feature extraction through multiple GRU blocks. The four parts of the extracted features are connected through a concatenate layer, and different weights and confidences are given to different features through an Attention layer. Finally, the prediction ability of the model is enhanced through an FC layer and multiple DNN networks. The last DNN network is the output layer; S4. Use the trained optimal model to predict the data to be predicted and obtain the wind speed prediction result.

2. The ultra-short-term wind speed prediction method based on the fusion of CNN and GRU according to claim 1, wherein In S1, the time resolution of the meteorological data is 1h, and the time resolution of the unit operation data is 30s.

3. The ultra-short-term wind speed prediction method based on the CNN+GRU fusion according to claim 1, wherein In S2, the process of target unit data reconstruction is as follows: regard the data of different dimensions as different channels, reconstruct the data of each channel to form a matrix, complete the construction of the target unit data, and then perform neighboring unit selection.

4. The ultra-short-term wind speed prediction method based on the CNN+GRU fusion according to claim 3, characterized in that, There are two methods for neighboring unit selection. Either one can be chosen. One method is to use the method of unit clustering to group the units, take each clustering result as a group, and the units within each group are neighboring units to each other. When predicting the target unit, refer to the wind speed characteristics of other units within the group where the target unit is located; The other method is through correlation analysis. Select the units with a relatively high wind speed correlation with the target unit as neighboring units, calculate the correlation coefficient, sort the calculated correlation coefficients, and select the ones with larger correlation coefficients as the neighboring units of the target unit.

5. The ultra-short-term wind speed prediction method based on the CNN+GRU fusion according to claim 1, wherein In S2, the trend feature extraction includes the trend features of unit data and meteorological data. The trend feature of unit data takes the past hour as the time window and extracts the statistical features within the time window; For the meteorological data, extract the trend feature by calculating the difference between the last point and the first point within the past 10 hours.

6. A very short-term wind speed prediction system integrating CNN and GRU, characterized in that, It includes: A data acquisition module for obtaining the unit operation data and meteorological data; The unit operation data includes wind speed, wind direction, power, rotational speed, pitch angle, ambient temperature, and ambient humidity, and the meteorological data includes wind speed and wind direction; A modeling data preparation module for preparing the modeling data. The data preparation includes unit data preparation, meteorological data preparation, and trend feature extraction; The unit data preparation includes target unit data reconstruction and neighboring unit selection; The model construction and training module is used to construct the adjacent unit CNN module, the target unit CNN module, the meteorological GRU module and the trend feature GRU module according to the prepared data, fuse the features of the four modules, and perform model training and parameter optimization; The data selected from the adjacent units is subjected to feature extraction through multiple CNN blocks, the data of the target unit is subjected to feature extraction through multiple CNN blocks, the meteorological data is subjected to feature extraction through multiple GRU blocks, and the trend feature data is subjected to feature extraction through multiple GRU blocks. The four parts of the extracted features are connected through the concatenate layer, and different weights and confidences are given to different features through the Attention layer. Finally, the prediction ability of the model is increased through the FC layer and multiple DNN networks. The last DNN network is the output layer; The prediction module is used to predict the data to be predicted by using the trained optimal model to obtain the wind speed prediction result.

7. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the ultra-short-term wind speed prediction method based on CNN+GRU fusion according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the ultra-short-term wind speed prediction method based on CNN+GRU fusion according to any one of claims 1 to 5.

Images