CN113862691B

CN113862691B - Control method and device for photovoltaic hydrogen production, storage medium and electronic equipment

Info

Publication number: CN113862691B
Application number: CN202111111060.7A
Authority: CN
Inventors: 刘斌; 孙鹤旭; 梅春晓; 董砚; 雷兆明; 廖文喆; 梁涛; 林涛; 井延伟; 白日新
Original assignee: Hebei Jiantou New Energy Co ltd; Hebei University of Technology; Hebei University of Science and Technology
Current assignee: Hebei Jiantou New Energy Co ltd; Hebei University of Technology; Hebei University of Science and Technology
Priority date: 2021-09-18
Filing date: 2021-09-18
Publication date: 2023-04-07
Anticipated expiration: 2041-09-18
Also published as: CN113862691A

Abstract

The invention discloses a photovoltaic hydrogen production control method and device, a storage medium and electronic equipment. Wherein, the method comprises the following steps: acquiring meteorological information of the geographical position of a photovoltaic component in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic component; determining a generating power predicted value of the photovoltaic module by adopting a photovoltaic generating power prediction model according to the meteorological information and the generating information, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm; controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: an energy management and control mode and a scheduling mode. The invention solves the technical problems of unstable power supply power and low hydrogen production efficiency of the water electrolysis hydrogen production system caused by large fluctuation of photovoltaic power generation power within short time.

Description

Control method and device for photovoltaic hydrogen production, storage medium and electronic equipment

Technical Field

The invention relates to the field of photovoltaic hydrogen production, in particular to a control method and device for photovoltaic hydrogen production, a storage medium and electronic equipment.

Background

The hydrogen production by photovoltaic power generation is the current hot research direction, and the main optimization aim is to obtain the maximum hydrogen production under the production condition of ensuring safety and stability. The dependence of a photovoltaic power generation system on meteorological environment is very strong, when weather conditions (such as irradiance) change violently, photovoltaic power generation power fluctuates greatly within a short time, however, when the water electrolysis hydrogen production equipment is used as a load and an adaptation process of the photovoltaic power fluctuation is gradual, the photovoltaic power generation power fluctuates greatly within a short time, for example, when the photovoltaic power generation system changes over 30% of instantaneous power generation power within 5 minutes, the water electrolysis hydrogen production system cannot bear the violent fluctuation, the hydrogen purity of the system changes, and a system controller sends out a warning and stops the system according to the response of a sensor. If the photovoltaic hydrogen production system is frequently stopped, the photovoltaic hydrogen production efficiency is seriously influenced.

In view of the above problems, no effective solution has been proposed.

Disclosure of Invention

The embodiment of the invention provides a control method and device for photovoltaic hydrogen production, storage equipment and electronic equipment, and aims to at least solve the technical problems of unstable power supply power and low hydrogen production efficiency of a water electrolysis hydrogen production system caused by large fluctuation of photovoltaic power generation power within a short time.

According to an aspect of an embodiment of the present invention, there is provided a control method for photovoltaic hydrogen production, including: acquiring meteorological information of the geographical position of a photovoltaic assembly in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic assembly; determining a generating power predicted value of the photovoltaic module by adopting a photovoltaic generating power prediction model according to the meteorological information and the generating information, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm; controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: an energy management and control mode and a scheduling mode.

Optionally, the acquiring meteorological information of the geographical position of the photovoltaic module in the photovoltaic hydrogen production system and the power generation information output by the dc converter of the photovoltaic module includes: adopt meteorological information collection subassembly to obtain above-mentioned meteorological information, wherein, above-mentioned meteorological information includes solar irradiance and following at least one: ambient temperature, humidity, wind direction, wind speed, photovoltaic panel temperature; the photovoltaic module is adopted to convert light energy into first direct current, the first direct current is converted into second direct current through the direct current converter, and the second direct current is output, so that power generation information output by the direct current converter is obtained.

Optionally, the meteorological information collecting assembly includes a target sensor group, where the target sensor group includes a solar irradiance sensor and at least one of: the device comprises an ambient temperature sensor, a photovoltaic panel temperature sensor, a humidity sensor, a wind direction sensor and a wind speed sensor.

Optionally, determining the predicted value of the generated power of the photovoltaic module by using a photovoltaic generated power prediction model according to the meteorological information and the power generation information includes: inputting the meteorological information and the power generation information into the photovoltaic power generation power prediction model, and performing power prediction on all direct current converters in the photovoltaic module by adopting the photovoltaic power generation power prediction model to obtain a power prediction result of each direct current converter; and superposing and integrating the plurality of power prediction results of all the direct current converters to obtain the generated power prediction value of the photovoltaic module.

Optionally, controlling an energy management and control mode and a scheduling mode of the photovoltaic hydrogen production system according to the predicted value of the generated power includes: when the predicted value of the generated power is larger than or equal to the rated power value of the photovoltaic hydrogen production system, an energy storage assembly in the photovoltaic hydrogen production system is adopted to absorb the excess power value in the predicted value of the generated power, wherein the energy storage assembly is used for carrying out energy scheduling on the photovoltaic hydrogen production system; and when the predicted value of the generated power is smaller than the rated power value of the photovoltaic hydrogen production system, the energy storage assembly is adopted to make up the power.

Optionally, before determining the predicted value of the generated power of the photovoltaic module according to the meteorological information and the power generation information by using a photovoltaic power generation prediction model, the method includes:

importing the meteorological information and the power generation information into a target program file to generate a data set; the target program file is established in advance based on target programming software, the data set comprises a first data set and a second data set, and the first data set and the second data set are obtained by converting the meteorological information and the power generation information respectively; carrying out format conversion on the time information in the data set to obtain target time information; merging the first data set and the second data set according to the target time information to obtain a merged data set; when detecting that all the loss values of the characteristic parameters in the merged data set are 0, determining that no abnormal value exists in the merged data set; and performing visual analysis on the merged data set without abnormal values, and determining whether all the characteristic parameters meet the preset requirements of the photovoltaic power generation power prediction model, wherein the merged data set meeting the preset requirements is a target data set.

Optionally, when detecting that a null value exists in the data set, performing padding processing on the null value; and when detecting that the target time information in the data set has time step loss, performing data supplement on the time step.

Optionally, before determining the predicted value of the generated power of the photovoltaic module according to the meteorological information and the power generation information by using a photovoltaic power generation prediction model, the method includes: carrying out initialization assignment on an initial prediction model before constructing the photovoltaic power generation power prediction model; selecting a data list of a target direct current converter with irradiance larger than zero in the target data set as sample data, and performing interval scaling and normalization processing on the sample data to obtain target sample data; the target sample data comprises training set data and verification set data; training the initial prediction model based on the training set data to obtain an intermediate prediction model; and evaluating whether the capacity value of the intermediate prediction model reaches a target capacity range or not based on the verification set data, and finishing the training of the intermediate prediction model if the capacity value reaches the target capacity range to obtain the photovoltaic power generation power prediction model.

According to another aspect of the embodiments of the present invention, there is also provided a control apparatus for photovoltaic hydrogen production, including: the acquisition module is used for acquiring meteorological information of the geographical position of a photovoltaic component in the photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic component; the determining module is used for determining a generating power predicted value of the photovoltaic module according to the meteorological information and the generating information by adopting a photovoltaic generating power prediction model, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm; and the control module is used for controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: an energy management and control mode and a scheduling mode.

According to another aspect of the embodiments of the present invention, there is also provided a non-volatile storage medium, which stores a plurality of instructions, the instructions being suitable for being loaded by a processor and executing any one of the above control methods for photovoltaic hydrogen production.

According to another aspect of the embodiments of the present invention, there is also provided an electronic device, including a memory and a processor, where the memory stores a computer program, and the processor is configured to execute the computer program to perform any one of the above control methods for photovoltaic hydrogen production.

In the embodiment of the invention, a photovoltaic hydrogen production control mode is adopted, and the meteorological information of the geographical position of the photovoltaic component in the photovoltaic hydrogen production system and the power generation information output by the direct current converter of the photovoltaic component are obtained; determining a generating power predicted value of the photovoltaic module by adopting a photovoltaic generating power prediction model according to the meteorological information and the generating information, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm; controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: the energy management and control mode and the scheduling mode achieve the purpose of accurately and timely predicting the power of the photovoltaic power generation, thereby realizing the technical effects of improving the power stability and the hydrogen production efficiency of the water electrolysis hydrogen production system, and further solving the technical problems of unstable power and low hydrogen production efficiency of the water electrolysis hydrogen production system caused by the large fluctuation of the photovoltaic power generation power in a short time.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

FIG. 1 is a flow diagram of a photovoltaic hydrogen production control method according to an embodiment of the invention;

FIG. 2 is a schematic diagram of a hardware configuration of an alternative photovoltaic hydrogen production control method according to an embodiment of the invention;

FIG. 3 is a schematic illustration of an alternative prior art grade and placement meteorological sensor;

FIG. 4 is a schematic diagram of an alternative data preprocessing flow according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of an alternative data pre-processing flow according to an embodiment of the invention;

FIG. 6 is a schematic diagram of an alternative photovoltaic power generation power prediction model construction method according to an embodiment of the invention;

FIG. 7 is a schematic diagram of an alternative photovoltaic power generation power prediction model construction method according to an embodiment of the invention;

FIG. 8 is a schematic diagram of an alternative training set and validation set loss function curve in accordance with an embodiment of the present invention;

fig. 9 is a schematic structural diagram of a control device for photovoltaic hydrogen production according to an embodiment of the invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

First, in order to facilitate understanding of the embodiments of the present invention, some terms or nouns referred to in the present invention will be explained as follows:

and (3) rolling layers: the method is mainly used for performing convolution operation on input data by applying a convolution kernel, namely, the convolution kernel performs sliding with a certain step length on multidimensional input, and calculates the inner product of the weight and the corresponding sliding.

A pooling layer: the method is mainly used for simplifying and extracting the data obtained from the previous layer.

Flattening layer: the method is mainly used for flattening the input tensor under the condition of keeping the 0 th axis and carrying out one-dimensional input.

Full connection layer: the fully-connected layer has different functions for different activation functions, for example, softmax is used for classification, sigmoid and ReLU can obtain fixed values, the activation functions of the fully-connected layer are ReLU functions, the number of neuron nodes of the two fully-connected layers is set respectively, and finally output is carried out.

One-dimensional convolution: the sliding window and multiplication summation is carried out in a single direction, and is commonly used in the field of sequence type data, such as time sequence. Wherein, the parameters of the filters in the convolution layer refer to the number of convolution kernels (i.e. filters), which are generally integers and represent the dimensionality of the output space; the kernel _ size parameter is used to specify the length of the one-dimensional convolution window. The parameter batch _ size refers to the number of samples to be trained per batch, and the size of the parameter influences the optimization degree and speed of the model and determines the gradient descending direction. When the batch _ size is larger, the descending direction is more accurate, the oscillation is smaller, but the excessively large batch _ size may cause the local optimal condition of the model; the smaller the batch _ size, the more random the model introduces and convergence is difficult to achieve. Increasing the size of the batch _ size within a reasonable range improves the utilization rate of the memory, the number of iterations required for completing one epoch is reduced, the processing speed is further increased, and the selection of the size of the batch _ size plays a crucial role in the performance of the model.

Primary epoch: the method is characterized in that all data of a training set are used for carrying out one-time complete training on a model, but the complete data set is not enough to be transmitted in a neural network once, and multiple iterations are needed, namely the epoch number is increased, so that the weight updating times in the neural network are also increased, but overfitting is caused due to excessive epochs, and therefore the most appropriate epoch size is found according to the data set through multiple tests. In the gradient descent algorithm, each iteration of the gradient descent is affected by the learning rate. If the learning rate is small, the number of iterations required to reach convergence can be very high; if the learning rate is large, each iteration may exceed a local minimum, and repeated jumps around the minimum may not converge.

Example 1

In accordance with an embodiment of the present invention, there is provided an embodiment of a method for photovoltaic hydrogen production control, it being noted that the steps illustrated in the flow chart of the accompanying figures may be carried out in a computer system, such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flow chart, in some cases, the steps illustrated or described may be carried out in an order different than presented herein.

Fig. 1 is a flow chart of a photovoltaic hydrogen production control method according to an embodiment of the invention, as shown in fig. 1, the method comprising the steps of:

step S102, acquiring meteorological information of the geographical position of a photovoltaic assembly in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic assembly;

step S104, determining a generating power predicted value of the photovoltaic module according to the meteorological information and the generating information by adopting a photovoltaic generating power prediction model, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm;

step S106, controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: an energy management and control mode and a scheduling mode.

In the embodiment of the invention, a photovoltaic hydrogen production control mode is adopted, and the meteorological information of the geographical position of the photovoltaic component in the photovoltaic hydrogen production system and the power generation information output by the direct current converter of the photovoltaic component are obtained; determining a generating power predicted value of the photovoltaic module by adopting a photovoltaic generating power prediction model according to the meteorological information and the generating information, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm; controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: the energy management and control mode and the scheduling mode achieve the purpose of accurately and timely predicting the power of the photovoltaic power generation, thereby realizing the technical effect of improving the power stability of the water electrolysis hydrogen production system, and further solving the technical problems of unstable power and low hydrogen production efficiency of the water electrolysis hydrogen production system caused by large-amplitude fluctuation of the photovoltaic power generation within a short time.

The embodiment of the invention relates to a power prediction system for hydrogen production by water electrolysis based on photovoltaic power generation and a short-time power prediction method based on a neural network, wherein the method at least can realize the following technical effects: setting a plurality of sensor groups according to different terrain conditions to acquire the meteorological information; training data are automatically acquired, and the operation flow is simplified; the photovoltaic power generation power prediction efficiency is improved, and the power prediction system can output the power generation power prediction value of one photovoltaic module in 1 minute. The embodiment of the invention relates to the power generation power prediction of a photovoltaic module based on the fluctuation of meteorological information and power generation information. The method provides a specific solution for stabilizing the production of the hydrogen production system on the basis of ensuring the hydrogen energy yield; moreover, more effective energy management of the photovoltaic hydrogen production system can be achieved through prediction of photovoltaic power generation direct current power; in addition, the method aims to utilize the power generation energy of the photovoltaic module to the maximum extent to carry out safe and reliable hydrogen production.

As an optional embodiment, fig. 2 is a schematic diagram of a hardware structure of an optional photovoltaic hydrogen production control method according to an embodiment of the present invention, and as shown in fig. 2, a photovoltaic power generation system and an electrolytic hydrogen production system upload power generation data and hydrogen production data to a monitoring system, the photovoltaic power generation system converts light energy into a first direct current by using a photovoltaic module, and converts the first direct current into a second direct current by using the direct current converter to output, so as to obtain power generation information output by the direct current converter, where the first direct current is a direct current directly converted from light energy, and the second direct current is a direct current suitable for the electrolytic hydrogen production system; the electrolytic hydrogen production system is used for electrolyzing water in an electrolytic cell based on the second direct current generated by the photovoltaic power generation system to generate hydrogen; the monitoring system and the meteorological data acquisition system upload the power generation information and the meteorological information to a database server, wherein the meteorological data acquisition system is used for acquiring the meteorological information by adopting a meteorological information acquisition assembly, and the meteorological information comprises solar irradiance and at least one of the following: ambient temperature, humidity, wind direction, wind speed, photovoltaic panel temperature; and the database server transmits the collected meteorological information and the collected power generation information to a power prediction server, and the power prediction server is used for determining a power generation power prediction value of the photovoltaic module by establishing a photovoltaic power generation power prediction model according to a convolutional neural network algorithm based on a depth algorithm theory.

In an alternative embodiment, acquiring meteorological information of a geographical location of a photovoltaic module in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic module includes:

step S202, acquiring the meteorological information by adopting a meteorological information acquisition assembly, wherein the meteorological information comprises solar irradiance and at least one of the following information: ambient temperature, humidity, wind direction, wind speed, photovoltaic panel temperature;

and step S204, converting light energy into a first direct current by adopting the photovoltaic module, converting the first direct current into a second direct current by the direct current converter and outputting the second direct current to obtain power generation information output by the direct current converter.

Optionally, the embodiment of the invention takes an independent photovoltaic hydrogen production system as an object, and the independent photovoltaic hydrogen production system is generally distributed in remote areas (such as mountainous regions and the like); the generated energy of the photovoltaic module is related to factors such as solar irradiance, temperature, geographical position and the like.

Optionally, the first direct current is direct current obtained by directly converting light energy, and the second direct current is direct current suitable for the electrolytic hydrogen production system, that is, the second direct current can be directly used as a power supply source of the electrolytic hydrogen production system.

In an alternative embodiment, the meteorological information acquisition assembly comprises a target sensor set, wherein the target sensor set comprises a solar irradiance sensor, and at least one of: the device comprises an ambient temperature sensor, a photovoltaic panel temperature sensor, a humidity sensor, a wind direction sensor and a wind speed sensor.

Optionally, fig. 3 is a schematic diagram of an optional slope and meteorological sensor arrangement in the prior art, and the inclination angles of the solar panels are different according to different terrain conditions, so that the solar irradiance is different, so that the data error generated by a single sensor is large, and the final prediction deviation is large. Therefore, the embodiment of the invention sets a multi-sensor group according to different terrain conditions, the number of the sensors can be configured according to the number of the current transformers at most, wherein the target sensor group comprises solar irradiance sensors and at least one of the following sensors: the system comprises an environment temperature sensor, a photovoltaic panel temperature sensor, a humidity sensor, a wind direction sensor and a wind speed sensor, wherein the selection of the sensors in the meteorological information acquisition assembly is related to the geographical position of the photovoltaic hydrogen production system.

In an alternative embodiment, determining the predicted value of the generated power of the photovoltaic module according to the meteorological information and the power generation information by using a photovoltaic power generation prediction model comprises:

step S302, inputting the meteorological information and the power generation information into the photovoltaic power generation power prediction model, and performing power prediction on all the direct current converters in the photovoltaic module by adopting the photovoltaic power generation power prediction model to obtain a power prediction result of each direct current converter;

and step S304, overlapping and integrating the plurality of power prediction results of all the direct current converters to obtain a generated power prediction value of the photovoltaic module.

In this embodiment, the photovoltaic power generation prediction model is used to perform power prediction on all the dc converters in the photovoltaic module, so as to better integrate the influence of meteorological features on all the dc converters, output the power generation prediction values of all the converters, and store the prediction results. And superposing and integrating all the prediction results to obtain the predicted value of the generated power of the photovoltaic module.

In an optional embodiment, controlling an energy management and control manner and a scheduling manner of the photovoltaic hydrogen production system according to the predicted value of the generated power includes:

step S402, when the predicted value of the generated power is larger than or equal to the rated power value of the photovoltaic hydrogen production system, an energy storage assembly in the photovoltaic hydrogen production system is adopted to absorb the excess power value in the predicted value of the generated power, wherein the energy storage assembly is used for carrying out energy scheduling on the photovoltaic hydrogen production system;

and S404, when the predicted value of the generated power is smaller than the rated power value of the photovoltaic hydrogen production system, performing power compensation by using the energy storage assembly.

Optionally, the photovoltaic hydrogen production system further includes an energy storage assembly for performing energy scheduling on the photovoltaic hydrogen production system, wherein the energy storage assembly may be but is not limited to a super capacitor, a lithium battery, and a lead-acid battery. When the predicted value of the generated power is larger than or equal to the rated power value of the photovoltaic hydrogen production system, the energy storage assembly is adopted to absorb the excess power value in the predicted value of the generated power; and when the predicted value of the generated power is smaller than the rated power value of the photovoltaic hydrogen production system, the energy storage assembly is adopted to make up the power. According to the change condition of the generated power prediction value, the energy storage assembly can be flexibly adjusted, the stability of the operation of the photovoltaic hydrogen production system is guaranteed to a great extent, and the stable output of the photovoltaic hydrogen production system is further guaranteed.

As an alternative embodiment, fig. 4 is a schematic diagram of an alternative data preprocessing flow according to an embodiment of the present invention, and as shown in fig. 4, before determining the predicted value of the generated power of the photovoltaic module according to the meteorological information and the power generation information by using a photovoltaic power generation prediction model, the method further includes:

step S502, importing the meteorological information and the power generation information into a target program file to generate a data set; the target program file is established in advance based on target programming software, the data set comprises a first data set and a second data set, and the first data set and the second data set are obtained by converting the meteorological information and the power generation information respectively;

step S504, converting the format of the time information in the data set to obtain target time information;

step S506, merging the first data set and the second data set according to the target time information to obtain a merged data set;

step S508, when it is detected that all the loss values of the characteristic parameters in the merged data set are 0, determining that no abnormal value exists in the merged data set;

step S510, performing a visual analysis on the merged data set without an abnormal value, and determining whether all the characteristic parameters meet a preset requirement of the photovoltaic power generation power prediction model, where the merged data set meeting the preset requirement is a target data set.

Optionally, the photovoltaic power generation power model needs meteorological information and power generation information as data input, a Pathon tool is used to integrate the meteorological information and the power generation information, and the form is automatically filled, and the specific procedure is as follows:

# Merge data

Plant1＝pd.merge(g1,w1,how＝‘left’,on＝[‘date_time’])

As an optional embodiment, fig. 5 is a schematic diagram of another optional data preprocessing process according to an embodiment of the present invention, in the embodiment of the present invention, an implementation program of photovoltaic power generation power prediction is mainly written on a programming software PyCharm, and a specific process is as shown in fig. 5, a python program file is established, the weather information and the power generation information are read into the program file in the PyCharm, a read statement is selected according to a storage form of data in a database, the weather information and the power generation information are imported into a pandas library, a data set is generated, wherein the data set is stored in a csv format, and data reading is performed using pd. Converting the original date-time format in the data set into the date-time format in the pandas through a pd.to _ datetime function; checking whether a null value exists in a data set, checking the null value by using an isna function, and filling the null value when the null value in the data set is detected; checking whether a lost time step exists in a data set, and supplementing data to the time step when detecting that the target time information in the data set has the lost time step; the method comprises the steps that meteorological data and power generation data information in a data set belong to two data sets, namely a first data set and a second data set, merging processing is conducted on the first data set and the second data set, namely repeated parameter columns in the two data sets are renamed or deleted, and the two data sets are spliced according to a date-time sequence on the left side by using a merge function; performing lost value check on the merged data set by using an isna function and a sum function, wherein if the lost values of the parameters of each column are 0, the merged data set has no abnormal value; performing visual analysis according to the preprocessed data set, importing the data into a seaborn data visualization module, setting image backgrounds, colors and the like by using a set function, and sequentially outputting the hourly distribution condition of each characteristic parameter in the data set in one day; and carrying out visual analysis on the merged data set without abnormal values, and determining whether all the characteristic parameters meet the preset requirements of the photovoltaic power generation power prediction model, wherein the merged data set meeting the preset requirements is a target data set. In an optional embodiment, when detecting that a null value exists in the data set, performing padding processing on the null value; and when detecting that the target time information in the data set has time step loss, supplementing data to the time step.

As an alternative embodiment, fig. 6 is a schematic diagram of an alternative photovoltaic power generation prediction model construction method according to an embodiment of the present invention, and as shown in fig. 6, before determining a generated power predicted value of the photovoltaic module according to the meteorological information and power generation information by using a photovoltaic power generation prediction model, the method includes:

step S602, before the photovoltaic power generation power prediction model is constructed, an initial value is assigned to the initial prediction model;

step S604, selecting a data list of a target direct current converter with irradiance larger than zero in the target data set as sample data, and performing interval scaling processing and normalization processing on the sample data to obtain target sample data; the target sample data comprises training set data and verification set data;

step S606, training the initial prediction model based on the training set data to obtain an intermediate prediction model;

step S608, evaluating whether the capability value of the intermediate prediction model reaches a target capability range based on the verification set data, and if the capability value reaches the target capability range, ending the training of the intermediate prediction model to obtain the photovoltaic power generation power prediction model.

Optionally, the target dc converter is any converter in the photovoltaic module.

The photovoltaic power generation system in the embodiment of the invention can realize the technical effect of outputting one photovoltaic power generation power predicted value in 1 minute. According to the time scale of photovoltaic power generation power prediction, a prediction point is output for 15 minutes in the current short-term power prediction, and a traditional neural network prediction method is adopted. Research literature materials show that the power fluctuation of the alkaline water electrolysis hydrogen production electrolytic cell is 10% of rated power within 5 minutes time scale, the fault shutdown condition can not occur, and when the time scale is increased according to the proportion, the alkaline water electrolysis hydrogen production electrolytic cell can normally operate within 15 minutes time scale, for example, the power fluctuation is 30%. However, under the cloudy weather condition of photovoltaic power generation, fluctuation is severe, and in an extreme case, the fluctuation is 70% within a few minutes, an electrolytic cell of the hydrogen production system by alkaline electrolysis cannot bear the severe fluctuation, the hydrogen production system by photovoltaic is frequently shut down, and the yield of hydrogen production by photovoltaic is seriously influenced. In the embodiment of the invention, one photovoltaic power generation power predicted value is output in 1 minute, and the photovoltaic hydrogen production system can be efficiently managed and scheduled according to the output photovoltaic power generation power predicted value, so that the stability of photovoltaic hydrogen production is greatly ensured, and the output efficiency of products is ensured. Even if the photovoltaic power generation system works in cloudy weather and the output power of the photovoltaic power station fluctuates severely, the output power can be predicted, and the photovoltaic hydrogen production system can rapidly adjust the system energy according to the predicted value because the time scale is 1 minute, so that the fluctuation degree of the photovoltaic hydrogen production system is greatly reduced, and the condition of frequent shutdown is avoided.

As an optional embodiment, a neural network algorithm in deep learning is used as a basic model to construct a photovoltaic power generation power prediction system, a convolutional neural network is mainly used as a power prediction model base, the convolutional neural network is a deep neural network model, a supervised learning mechanism is adopted, most problems can be adapted, the convolutional neural network has strong superiority in the aspects of mining local features and extracting global training features, and compared with other neural network models, the convolutional neural network has better fault tolerance and higher performance. The embodiment of the invention is programmed in a Python environment, and the convolutional neural network framework is mainly constructed by using a Keras module in a TensorFlow2.0 as a background in a Python3.7 environment. Fig. 7 is a schematic diagram of another alternative photovoltaic power generation power prediction model construction method according to an embodiment of the present invention, and as shown in fig. 7, the photovoltaic power generation power prediction model construction method includes the following method steps:

step S701, defining a converter power prediction model class for a target converter, wherein the target converter is any one of a plurality of converters in the data set and serves as an example converter;

step S702, carrying out initialization assignment on the initial prediction model;

step S703, preparing input data, selecting a data list of a target direct current converter with irradiance larger than zero in the target data set as sample data, and performing interval scaling and normalization processing on the sample data to obtain target sample data as input data; the target sample data comprises training set data and verification set data.

Optionally, the training set data is used to train the initial prediction model to obtain an intermediate prediction model; and the verification set data is used for evaluating whether the capacity value of the intermediate prediction model reaches a target capacity range, and if the capacity value reaches the target capacity range, finishing the training of the intermediate prediction model to obtain the photovoltaic power generation power prediction model.

Optionally, the above-mentioned preparation input data may be, but is not limited to: defining a function named prepare _ input _ data, setting data to be input, selecting a data list of target current transformers and irradiance greater than 0 as a sample list, and reading the length of the sample list by using a len function. Importing a MinMaxScale function from a skleann.preprocessing module to perform interval scaling based on the maximum and minimum values so as to improve the convergence speed and precision of the model; and carrying out normalization processing on the sample data by using a MinMaxScaler function, distributing the sample data according to seven to three proportions, wherein the first 70 percent of the sample data is used as training set data, and the last 30 percent of the sample data is used as verification set data. And after normalization is finished, readjusting the data, respectively putting the data into a newly-built list X and a newly-built list Y, wherein X represents the input of a data set, Y represents a label of the data set, and the label is stored as an object attribute by using an array function, so that the shape of an input array is ensured.

Step S704, a convolutional neural network model is constructed and trained, and the initial prediction model is trained by the training set data to obtain an intermediate prediction model; and evaluating whether the capacity value of the intermediate prediction model reaches a target capacity range or not based on the verification set data, and ending the training of the intermediate prediction model if the capacity value reaches the target capacity range to obtain the photovoltaic power generation power prediction model.

Optionally, the convolutional neural network in the embodiment of the present invention mainly includes, in addition to the input layer and the output layer, a convolutional layer, a maximum pooling layer, a flat layer, and a full-link layer on both sides. The above-mentioned building of the convolutional neural network model further includes designing a convolutional layer, a pooling layer, a flat layer, etc. of the convolutional neural network model, setting an activation function, etc., for example, setting an output space dimension of the convolutional layer to be 10 (i.e., the number of filters in the convolutional neural network model is 10), and specifying a length of a 1D convolutional window; taking a maximum value pooling layer, taking the maximum value in the window as output, and setting the size of a pooling window; adding a flat layer behind the maximum pooling layer; and the activation function of the full connection layer is a ReLU function, the number of the neuron nodes of the full connection layer is set based on the ReLU function, and finally, an intermediate prediction model is output.

Optionally, the constructed intermediate prediction model is trained, and factors affecting the training result and precision include parameters of network initialization that cannot be learned in training, in addition to the data set. For example, in the one-dimensional convolution and convolution neural network, the main factors influencing the model performance are the convolution kernel size, the batch size, the training times, the learning rate and the like, and the optimal combination of each parameter can be found through repeated experiments to achieve the optimal performance of the model as far as possible. For the size of the convolution kernel, a general setting principle is that the smaller the convolution kernel is, the smaller the required parameters and calculation amount are, the larger the convolution kernel is, and the size of the convolution kernel is generally selected to be 3, and then the value of the convolution kernel is gradually and slowly increased according to the condition of the data set. For the batch size, the batch _ size =32 is generally set, the data length in the data set gradually increases, and for data having a length of about 2000, it is generally better to obtain 128 training results for the batch _ size. For the training times, the model with too small setting is not completely trained, and overfitting can be caused by too large setting, so that the value of epoch is generally in the range of 100-1000. For the learning rate, the learning rate range of the general neural network is between 0.01 and 0.1, and the learning rate is gradually adjusted in the range, so that the photovoltaic power generation power prediction model is obtained.

Step S705, displaying a loss image of the photovoltaic power generation power prediction model.

Optionally, the loss images of the training set and the verification set of the convolutional neural network model are drawn through a 2D meeting map library matplotlib of python and are graphically displayed. For example, fig. 8 is a schematic diagram of a loss function curve of an optional training set and validation set according to an embodiment of the present invention, as shown in fig. 8, an error of the training set is between 0.01 and 0.02, and an error of the validation set is between 0.015 and 0.03, and a smaller error indicates more accurate prediction, so that a prediction result of the photovoltaic power generation prediction model is obtained to be relatively accurate.

Step S706, judging whether the photovoltaic power generation power prediction model belongs to one-step prediction or multi-step prediction;

in step S707, if the determination result is the one-step prediction, the one-step prediction image is output.

Optionally, a prediction function named forecast is defined, the prediction type is determined to be single step prediction in the conditions of the defined prediction function, the prepared input data, the convolutional neural network model, the lost image display and the like are placed into the prediction function, a condition function is set, if the given prediction type is one step prediction, image output of the one step prediction is performed, and if the given type is not the one step prediction, the image of multi-step prediction is output.

Alternatively, when performing multi-step prediction on the photovoltaic power generation power prediction model, for example, using 16 data of 15 minutes in the past 4 hours to predict the output of the next 15 minutes, a function named multi step for forecast needs to be defined. Determining target data of the verification set (such as data of the last third of the sample data); assigning unassigned self _ predictions in initialization to a new list; inputting data of a first verification set, taking the data at-self.X// 3 in the sample data set as the first input, wherein the shape of the array is (T, D); using while function, when the data length of self-evaluation _ predictions is smaller than the length of target data of a verification set, continuously acquiring data, converting an array last _ x of 1 multiplied by 1 into a scalar, performing model prediction, and assigning a predicted value to pred; adding pred to a new self.identification _ predictions list using an apend function for updating the list until the data length of the self.identification _ predictions is greater than or equal to the target data length of the verification set, and stopping the loop; roll function is used to scroll last _ x-1 length along a given axis, default axis = None because there is no assignment to axis, data is flattened before scrolling and then restored, so that a new input is formed, and finally pred's data content is assigned to last _ x-1 to store the original data.

It should be noted that for simplicity of description, the above-mentioned method embodiments are shown as a series of combinations of acts, but those skilled in the art will recognize that the present invention is not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the invention. Further, those skilled in the art will appreciate that the embodiments described in this specification are presently preferred and that no acts or modules are required by the invention.

Through the above description of the embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but the former is a better implementation mode in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, or a network device) to execute the method according to the embodiments of the present invention.

Example 2

According to an embodiment of the present invention, there is also provided an embodiment of an apparatus for implementing the above control method for photovoltaic hydrogen production, fig. 9 is a schematic structural diagram of a control apparatus for photovoltaic hydrogen production according to an embodiment of the present invention, and as shown in fig. 9, the control apparatus for photovoltaic hydrogen production includes: an obtaining module 80, a determining module 82, a control module 84, wherein:

the acquiring module 80 is configured to acquire meteorological information of a geographical location where a photovoltaic module in the photovoltaic hydrogen production system is located, and power generation information output by a direct current converter of the photovoltaic module; the determining module 82 is configured to determine a predicted value of the generated power of the photovoltaic module according to the meteorological information and the power generation information by using a photovoltaic power generation power prediction model, where the photovoltaic power generation power prediction model is constructed and trained based on a preset neural network algorithm; the control module 84 is configured to control an operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, where the operation mode at least includes: an energy management and control mode and a scheduling mode.

It should be noted that the above modules may be implemented by software or hardware, for example, for the latter, the following may be implemented: the modules can be located in the same processor; alternatively, the modules may be located in different processors in any combination.

It should be noted here that the acquiring module 80, the determining module 82, and the control module 84 correspond to steps S102 to S106 in embodiment 1, and the modules are the same as the corresponding steps in the implementation example and application scenarios, but are not limited to the disclosure in embodiment 1. It should be noted that the modules described above may be executed in a computer terminal as part of an apparatus.

It should be noted that, reference may be made to the relevant description in embodiment 1 for alternative or preferred embodiments of this embodiment, and details are not described here again.

The photovoltaic hydrogen production control device may further include a processor and a memory, where the obtaining module 80, the determining module 82, the control module 84, and the like are stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions.

The processor comprises a kernel, and the kernel calls the corresponding program unit from the memory, and the kernel can be set to be one or more. The memory may include volatile memory in a computer readable medium, random Access Memory (RAM) and/or nonvolatile memory such as Read Only Memory (ROM) or flash memory (flash RAM), and the memory includes at least one memory chip.

According to an embodiment of the present application, there is also provided an embodiment of a non-volatile storage medium. Optionally, in this embodiment, the non-volatile storage medium includes a stored program, and when the program runs, the apparatus in which the non-volatile storage medium is located is controlled to execute any one of the above control methods for photovoltaic hydrogen production.

Optionally, in this embodiment, the nonvolatile storage medium may be located in any one of computer terminals in a computer terminal group in a computer network, or in any one of mobile terminals in a mobile terminal group, and the nonvolatile storage medium includes a stored program.

Optionally, the device in which the non-volatile storage medium is controlled to execute the following functions when the program runs: acquiring meteorological information of the geographical position of a photovoltaic assembly in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic assembly; determining a generating power predicted value of the photovoltaic module by adopting a photovoltaic generating power prediction model according to the meteorological information and the generating information, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm; controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: an energy management and control mode and a scheduling mode.

There is also provided, in accordance with an embodiment of the present application, an embodiment of a computer program product, which, when executed on a data processing device, is adapted to execute a program that initializes the steps of the control method for photovoltaic hydrogen production as described above.

Optionally, the computer program product is adapted to perform a program for initializing the following method steps when executed on a data processing device: acquiring meteorological information of the geographical position of a photovoltaic assembly in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic assembly; determining a generating power predicted value of the photovoltaic module by adopting a photovoltaic generating power prediction model according to the meteorological information and the generating information, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm; controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: an energy management and control mode and a scheduling mode.

There is also provided, according to an embodiment of the present application, an embodiment of an electronic device, including a memory and a processor, where the memory stores a computer program, and the processor is configured to execute the computer program to perform any one of the above control methods for photovoltaic hydrogen production.

The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.

In the above embodiments of the present invention, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the embodiments provided in the present application, it should be understood that the disclosed technology can be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units may be a logical division, and in actual implementation, there may be another division, for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, units or modules, and may be in an electrical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on a plurality of units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable non-volatile storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a non-volatile storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned nonvolatile storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims

1. A control method for photovoltaic hydrogen production is characterized by comprising the following steps:

acquiring meteorological information of the geographical position of a photovoltaic assembly in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic assembly, wherein the meteorological information comprises solar irradiance;

determining a generating power predicted value of the photovoltaic module by adopting a photovoltaic generating power prediction model according to the meteorological information and the generating information, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm, and the time scale of the photovoltaic generating power prediction model for outputting the generating power predicted value is 1 minute;

controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: an energy management and control mode and a scheduling mode;

the method for acquiring the power generation information output by the direct current converter of the photovoltaic module comprises the following steps: the photovoltaic module is adopted to convert light energy into a first direct current, the first direct current is converted into a second direct current through the direct current converter, and the second direct current is output to obtain power generation information output by the direct current converter; the second direct current is directly used as a power supply source of the electrolytic hydrogen production system;

determining a generating power predicted value of the photovoltaic module according to the meteorological information and the generating information by adopting a photovoltaic generating power prediction model, wherein the determining comprises the following steps: inputting the meteorological information and the power generation information into the photovoltaic power generation power prediction model, and performing power prediction on all direct current converters in the photovoltaic module by adopting the photovoltaic power generation power prediction model to obtain a power prediction result of each direct current converter; superposing and integrating a plurality of power prediction results of all the direct current converters to obtain a power generation power prediction value of the photovoltaic module;

controlling an energy management and control mode and a scheduling mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the energy management and control mode and the scheduling mode comprise the following steps: when the predicted value of the generated power is larger than or equal to the rated power value of the photovoltaic hydrogen production system, absorbing the excess power value in the predicted value of the generated power by using an energy storage assembly in the photovoltaic hydrogen production system, wherein the energy storage assembly is used for carrying out energy scheduling on the photovoltaic hydrogen production system; and when the predicted value of the generated power is smaller than the rated power value of the photovoltaic hydrogen production system, the energy storage assembly is adopted to make up the power.

2. The method of claim 1, wherein obtaining meteorological information of a geographical location of a photovoltaic module in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic module comprises:

acquiring the meteorological information by adopting a meteorological information acquisition component, wherein the meteorological information further comprises one of the following: ambient temperature, humidity, wind direction, wind speed, photovoltaic panel temperature.

3. The method of claim 2, wherein the meteorological information acquisition assembly comprises a target sensor set, wherein the target sensor set comprises a solar irradiance sensor, and at least one of: the device comprises an ambient temperature sensor, a photovoltaic panel temperature sensor, a humidity sensor, a wind direction sensor and a wind speed sensor.

4. The method of claim 1, prior to determining a generated power prediction value for the photovoltaic module from the meteorological information and the power generation information using a photovoltaic power generation prediction model, comprising:

importing the meteorological information and the power generation information into a target program file to generate a data set; the target program file is established in advance based on target programming software, the data set comprises a first data set and a second data set, and the first data set and the second data set are obtained by converting the meteorological information and the power generation information respectively;

carrying out format conversion on the time information in the data set to obtain target time information;

merging the first data set and the second data set according to the target time information to obtain a merged data set;

when all the loss values of the characteristic parameters in the merged data set are detected to be 0, determining that no abnormal value exists in the merged data set;

and carrying out visual analysis on the merged data set without abnormal values, and determining whether all the characteristic parameters meet the preset requirements of the photovoltaic power generation power prediction model, wherein the merged data set meeting the preset requirements is a target data set.

5. The method of claim 4, comprising: when detecting that a null value exists in the data set, filling the null value; and when detecting that the target time information in the data set has time step loss, performing data supplement on the time step.

6. The method of claim 4, wherein prior to determining a generated power prediction value for the photovoltaic module from the meteorological information and the power generation information using a photovoltaic power generation prediction model, comprising:

carrying out initialization assignment on an initial prediction model before constructing the photovoltaic power generation power prediction model;

selecting a data list of a target direct current converter with irradiance larger than zero in the target data set as sample data, and carrying out interval scaling processing and normalization processing on the sample data to obtain target sample data; wherein the target sample data comprises training set data and validation set data;

training the initial prediction model based on the training set data to obtain an intermediate prediction model;

and evaluating whether the capacity value of the intermediate prediction model reaches a target capacity range or not based on the verification set data, and ending the training of the intermediate prediction model if the capacity value reaches the target capacity range to obtain the photovoltaic power generation power prediction model.

7. A control device for photovoltaic hydrogen production, which is used for executing the control method for photovoltaic hydrogen production of any one of claims 1 to 6, and is characterized by comprising:

the system comprises an acquisition module, a control module and a control module, wherein the acquisition module is used for acquiring meteorological information of the geographical position of a photovoltaic assembly in a photovoltaic hydrogen production system and power generation information output by a direct current converter of the photovoltaic assembly, and the meteorological information comprises solar irradiance;

the determining module is used for determining a generating power predicted value of the photovoltaic module according to the meteorological information and the generating information by adopting a photovoltaic generating power prediction model, wherein the photovoltaic generating power prediction model is constructed and trained on the basis of a preset neural network algorithm, and the time scale of the photovoltaic generating power prediction model for outputting the generating power predicted value is 1 minute;

and the control module is used for controlling the operation mode of the photovoltaic hydrogen production system according to the predicted value of the generated power, wherein the operation mode at least comprises the following steps: an energy management and control mode and a scheduling mode;

the acquisition module is further used for converting light energy into a first direct current by using the photovoltaic module, converting the first direct current into a second direct current by using the direct current converter and outputting the second direct current to obtain power generation information output by the direct current converter; the second direct current is directly used as a power supply source of the electrolytic hydrogen production system;

the determining module is further configured to input the meteorological information and the power generation information into the photovoltaic power generation power prediction model, and perform power prediction on all the direct current converters in the photovoltaic module by using the photovoltaic power generation power prediction model to obtain a power prediction result of each direct current converter; superposing and integrating a plurality of power prediction results of all the direct current converters to obtain a power generation power prediction value of the photovoltaic module;

the control module is further configured to, when the predicted value of the generated power is greater than or equal to a rated power value of the photovoltaic hydrogen production system, absorb an excess power value in the predicted value of the generated power by using an energy storage assembly in the photovoltaic hydrogen production system, where the energy storage assembly is used for energy scheduling of the photovoltaic hydrogen production system; and when the predicted value of the generated power is smaller than the rated power value of the photovoltaic hydrogen production system, the energy storage assembly is adopted to make up the power.

8. A non-volatile storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the method for photovoltaic hydrogen production control of any one of claims 1 to 6.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the computer program to execute the control method for photovoltaic hydrogen production according to any one of claims 1 to 6.