Waste incineration plant hearth flame analysis method and device based on artificial neural network, electronic equipment and computer readable storage medium

Abstract

The invention discloses a waste incineration plant hearth flame analysis method based on an artificial neural network. The waste incineration plant hearth flame analysis method based on the artificial neural network comprises the steps that hearth flame image information of hearth flame in an incinerator is acquired; preprocessing the hearth flame image information, and taking the preprocessed hearth flame image information as a training sample; training the BP neural network model by using the training sample to generate an optimal BP neural network model; the preprocessed hearth flame image information is calculated through an optimal BP neural network model, and a hearth flame control command is generated; the system has the advantages of improving the combustion stability of the hearth, improving the reaction rate of a hearth control system, reducing the smoke emission, saving energy, reducing emission, reducing the working intensity of field personnel, improving the intelligent management of a power plant and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of waste incineration, and in particular to a method, device, electronic equipment and computer-readable storage medium for analyzing furnace flames in a waste incineration plant based on an artificial neural network. Background Art

[0002] With the rapid development of my country's economy and the acceleration of urbanization, waste disposal has become the top priority of national environmental protection issues, and waste incineration power generation has gradually become the preferred method of waste disposal in my country. The high efficiency and stability of furnace combustion has become the most important link in the waste incineration process.

[0003] Due to the unstable calorific value of garbage and the limitations of environmental protection supervision, the domestic furnace flame monitoring still continues to control the system through manual observation, but this method has many disadvantages due to the difficulty in controlling the level of on-site personnel. On the other hand, on-site personnel cannot make judgments quickly and accurately, which delays the control of the furnace flame system and increases flue gas emissions.

[0004] Therefore, a method, device, electronic device and computer-readable storage medium for analyzing furnace flame in a waste incineration plant based on an artificial neural network are provided to solve the above problems. Summary of the invention

[0005] The technical problem to be solved by the present invention is that the domestic furnace flame monitoring still continues to control the system by manual observation, but since the level of on-site personnel is difficult to control, a method, device, electronic device and computer-readable storage medium for analyzing the furnace flame of a waste incineration plant based on an artificial neural network are provided. The method for analyzing the furnace flame of a waste incineration plant based on an artificial neural network includes:

[0006] Acquire furnace flame image information of the furnace flame in the incinerator;

[0007] Preprocessing the furnace flame image information, and using the preprocessed furnace flame image information as a training sample;

[0008] Use the training samples to train the BP neural network model and generate the optimal BP neural network model;

[0009] The preprocessed furnace flame image information is calculated using the optimal BP neural network model to generate furnace flame control commands.

[0010] Furthermore, before obtaining the furnace flame image information of the furnace flame in the incinerator, the process also includes:

[0011] A high temperature monitoring CCD camera is installed before and after the incinerator to collect the flame video in the furnace through the high temperature monitoring CCD camera;

[0012] The furnace flame video in the furnace collected by the high temperature monitoring CCD camera is collected and encoded through a video splitter and an image video acquisition card to generate digitally encoded furnace flame image information.

[0013] Furthermore, the preprocessing of the furnace flame image information includes:

[0014] Discretize the collected furnace flame image information;

[0015] Quantify the gray value of the furnace flame image information after image discretization;

[0016] The furnace flame image information which has been quantized by gray value is compressed.

[0017] Furthermore, the training of the BP neural network model to generate the optimal BP neural network model includes:

[0018] Get the flame feature vector;

[0019] The self-organizing competition network function is selected as the BP neural network model. The input model of the BP neural network model is P = [p1, p2 ..., pr], the output model is A = [a1, a2, ..., ar], and the weighted sum of the input layer is n i =∑WP, the weighted vector sum of the output layer is g i =∑Wa, if neuron i wins, then a=1, if neuron i wins, then the rest of the neurons are 0;

[0020] Put the training samples into the BP neural network model and train the BP neural network model.

[0021] Furthermore, the obtaining of the flame feature vector includes: obtaining the average grayscale and the flame ratio;

[0022] The average grayscale formula is:

[0023]

[0024] Where r<i<L, r represents the lowest grayscale threshold, L represents the highest grayscale threshold, I represents the average grayscale of the image, p(i) represents the probability of the i-th grayscale, and m represents the actual maximum grayscale of the image;

[0025] The flame ratio formula is:

[0026]

[0027] Where A represents the flame ratio, S1 represents the assumed flame image area, and S2 represents the area of ​​the entire image.

[0028] Furthermore, the image discretization process specifically includes:

[0029] The collected images are spatially divided into several grids along the horizontal and vertical directions according to the preset sampling theorem. Each grid is represented by a grayscale value, and the positional relationship between pixels is represented by coordinates (x, y).

[0030] On the other hand, the present application provides a device for analyzing furnace flames in a waste incineration plant based on an artificial neural network, comprising:

[0031] An image acquisition module, used to acquire furnace flame image information of the furnace flame in the incinerator;

[0032] An image processing module is used to preprocess the furnace flame image information and use the preprocessed furnace flame image information as a training sample;

[0033] The model training module is used to train the BP neural network model using training samples to generate the optimal BP neural network model;

[0034] The model using module is used to calculate the pre-processed furnace flame image information using the optimal BP neural network model to generate a furnace flame control command.

[0035] On the other hand, the present application provides an electronic device, the electronic device comprising a memory and at least one processor, the memory storing instructions;

[0036] The at least one processor calls the instructions in the memory to enable the electronic device to execute each step of the method for analyzing furnace flames in a waste incineration plant based on an artificial neural network as described in any one of the above items.

[0037] On the other hand, the present application provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, implement the various steps of any of the above methods for analyzing furnace flames in a waste incineration plant based on an artificial neural network.

[0038] The implementation of the present invention has the following beneficial effects:

[0039] 1. The present invention monitors the furnace flame by self-learning, thereby saving labor costs; and can more accurately judge the combustion conditions of each area by performing regional color analysis on the collected images; it makes up for the shortcomings of the furnace safety monitoring system and has a positive feedback effect on the subsequent automatic control of the furnace flame. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is an example of a self-organizing competition network structure diagram;

[0041] Figure 2 Example of pre-training samples for self-organizing competitive neural networks;

[0042] Figure 3 An example of a sample after training for a self-organizing competitive neural network;

[0043] Figure 4 Provide an application example of self-organizing competitive neural network;

[0044] Figure 5 A flowchart of a method for analyzing furnace flames in a waste incineration plant based on an artificial neural network provided by an embodiment of the present invention;

[0045] Figure 6 A schematic diagram of the structure of a waste incineration plant furnace flame analysis device based on an artificial neural network provided in an embodiment of the present invention;

[0046] Figure 7 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Example

[0049] Please refer to the attached drawings of the specification. This embodiment provides a method, device, electronic device and computer-readable storage medium for analyzing furnace flames in a waste incineration plant based on an artificial neural network. The method for analyzing furnace flames in a waste incineration plant based on an artificial neural network includes:

[0050] Acquire furnace flame image information of the furnace flame in the incinerator;

[0051] The step of obtaining the furnace flame image information of the furnace flame in the incinerator also includes:

[0052] A high temperature monitoring CCD camera is installed before and after the incinerator to collect the flame video in the furnace through the high temperature monitoring CCD camera;

[0053] The furnace flame video in the furnace collected by the high temperature monitoring CCD camera is collected and encoded through a video splitter and an image video acquisition card to generate digitally encoded furnace flame image information.

[0054] Preprocessing the furnace flame image information, and using the preprocessed furnace flame image information as a training sample; the preprocessing of the furnace flame image information specifically includes discretizing the collected furnace flame image information; the image discretization process specifically includes: dividing the collected image into a plurality of grids in the horizontal and vertical directions according to a preset sampling theorem in space, each grid is represented by a gray value, and the positional relationship between pixel points is represented by coordinates (x, y).

[0055] The grayscale value of the furnace flame image information after image discretization is quantified; specifically, the sampled image has been discretized into pixels in space after computer processing, but the pixel value is still a continuous value (non-integer), and it is necessary to quantize the grayscale value to convert it into a positive integer. The sampled value is expressed with the closest positive integer and converted into a finite discrete value.

[0056] The furnace flame image information after gray value quantization is compressed, specifically including compression processing through encoding.

[0057] Use the training samples to train the BP neural network model and generate the optimal BP neural network model;

[0058] The training of the BP neural network model to generate the optimal BP neural network model includes:

[0059] Obtaining flame feature vectors; specifically including: obtaining average grayscale and flame ratio;

[0060] The average grayscale reflects the average brightness of the flame. The flame image is stored in the computer in the form of grayscale, which can directly reflect the changes in the radiation energy signal in the furnace and is an important indicator of the combustion state. The average grayscale formula is:

[0061]

[0062] Where r<i<L, r represents the lowest grayscale threshold, L represents the highest grayscale threshold, I represents the average grayscale of the image, p(i) represents the probability of the i-th grayscale, and m represents the actual maximum grayscale of the image;

[0063] The flame ratio reflects the area ratio of the effective flame area in the image. The flame ratio formula is:

[0064]

[0065] Where A represents the flame ratio, S1 represents the assumed flame image area, and S2 represents the area of ​​the entire image.

[0066] The self-organizing competition network function is selected as the BP neural network model. The input model of the BP neural network model is P = [p1, p2 ..., pr], the output model is A = [a1, a2, ..., ar], and the weighted sum of the input layer is n i =∑WP, the weighted vector sum of the output layer is g i =∑Wa, if neuron i wins, then a=1, if neuron i wins, then the rest of the neurons are 0;

[0067] Specifically, if a neuron wins, it will add a positive value to the weight of the input layer connected to it, increasing its chance of winning. Otherwise, it will add a negative value, reducing its chance of winning. By inputting samples into the trained network, the samples are divided into different categories. The trained network can always successfully classify the input samples into the category that is closest to the identified category.

[0068] Put the training samples into the BP neural network model to train the BP neural network model;

[0069] Specifically, a burning stable picture from a previous project was selected, and the drying area, burning area and ember area were cut out from it, and divided into three pictures as the training basis, and a total of 30 groups of RGB grayscale values ​​were obtained as training samples. Five frames of image features were continuously selected at each moment to provide 25 vector network input samples. Figure 2 Figure 3 The figure shows the changes of sample vectors and network weights of the self-organizing competitive neural network before and after training, where x represents sample data and o represents network weights. After training, it can be clearly seen that the network has a good classification and recognition function, defining samples as dry area, burning area and burnt area.

[0070] The preprocessed furnace flame image information is calculated using the optimal BP neural network model to generate furnace flame control commands.

[0071] In addition, this embodiment uses the trained model to test a power plant, extracts 5 groups of continuous data per minute and inputs them into the trained neural network model for analysis. After continuous training, the competitive response is gathered into three areas, namely the drying area, the combustion area and the burnout area. It shows that the output of the trained model can already reflect the combustion situation of the boiler and distinguish the states under different working conditions. When running at full load, 90% of the sample points should be concentrated in the combustion area, and the drying area and the burnout area should be as fire-free as possible, that is, the samples should be gathered in these two areas as little as possible. If the number of samples gathered in the drying area increases, it is necessary to notify the control system to increase the combustion air volume and reduce the drying air volume. If the number of samples gathered in the burnout area increases, it is necessary to notify the control system to increase the combustion air volume and reduce the burnout air volume. Vice versa.

[0072] The above describes the method for analyzing the furnace flame of a garbage incineration plant based on an artificial neural network in an embodiment of the present invention. The following describes the device for analyzing the furnace flame of a garbage incineration plant based on an artificial neural network in an embodiment of the present invention. Figure 6 In the embodiment of the present invention, the furnace flame analysis device of a garbage incineration plant based on an artificial neural network includes the following for the above embodiment:

[0073] An image acquisition module 601 is used to acquire furnace flame image information of a furnace flame in an incinerator;

[0074] An image processing module 602 is used to pre-process the furnace flame image information and use the pre-processed furnace flame image information as a training sample;

[0075] The model training module 603 is used to train the BP neural network model using the training samples to generate the optimal BP neural network model;

[0076] The model using module 604 is used to calculate the pre-processed furnace flame image information using the optimal BP neural network model to generate a furnace flame control command.

[0077] above Figure 6 The waste incineration plant furnace flame analysis device based on artificial neural network in the embodiment of the present invention is described in detail from the perspective of modular functional entities, and the electronic equipment in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0078] Figure 77 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device 700 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 710 (for example, one or more processors) and a memory 720, and one or more storage media 730 (for example, one or more mass storage devices) storing application programs 733 or data 732. Among them, the memory 720 and the storage medium 730 can be temporary storage or permanent storage. The program stored in the storage medium 730 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the electronic device 700. Furthermore, the processor 710 can be configured to communicate with the storage medium 730 to execute a series of instruction operations in the storage medium 730 on the electronic device 700.

[0079] The electronic device 700 may also include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input and output interfaces 750, and / or one or more operating systems 731, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 7 The structure of the electronic device shown does not constitute a limitation on the electronic device, and may include more or less components than shown in the figure, or combine some components, or arrange the components differently.

[0080] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of a method for analyzing furnace flames in a waste incineration plant based on an artificial neural network.

[0081] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0082] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0083] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for analyzing the furnace flame of a waste incineration plant based on an artificial neural network, characterized in that, the method for analyzing the furnace flame of a waste incineration plant based on an artificial neural network includes: Obtaining the furnace flame image information of the furnace flame in the incinerator; Preprocessing the furnace flame image information, and using the preprocessed furnace flame image information as a training sample; Training the BP neural network model with the training sample to generate an optimal BP neural network model; Calculating the preprocessed furnace flame image information using the optimal BP neural network model to generate a furnace flame control command.

2. The method for analyzing the furnace flame of a waste incineration plant based on an artificial neural network according to claim 1, characterized in that, before obtaining the furnace flame image information of the furnace flame in the incinerator, it further includes, Setting a high-temperature resistant monitoring CCD camera before and after the furnace of the incinerator, and collecting the furnace flame video in the furnace through the high-temperature resistant monitoring CCD camera; Collecting and encoding the furnace flame video in the furnace collected by the high-temperature resistant monitoring CCD camera through a video splitter and an image video capture card to generate digital encoded furnace flame image information.

3. The method for analyzing the furnace flame of a waste incineration plant based on an artificial neural network according to claim 2, characterized in that, the preprocessing of the furnace flame image information includes, Discretizing the collected furnace flame image information; Quantifying the gray value of the furnace flame image information after image discretization; Compressing the furnace flame image information after gray value quantification.

4. The method for analyzing the furnace flame of a waste incineration plant based on an artificial neural network according to claim 3, characterized in that, the training of the BP neural network model to generate an optimal BP neural network model includes, Obtaining a flame feature vector; The self-organizing competition network function is selected as the BP neural network model. The input model of the BP neural network model is P = [p1, p2 ..., pr], the output model is A = [a1, a2, ..., ar], and the weighted sum of the input layer is n i =∑WP, the weighted vector sum of the output layer is g i =∑Wa, if neuron i wins, then a=1, if neuron i wins, then the rest of the neurons are 0; Putting the training sample into the BP neural network model and training the BP neural network model.

5. The method for analyzing the furnace flame of a waste incineration plant based on an artificial neural network according to claim 4, characterized in that, the obtaining of the flame feature vector includes: obtaining the average gray value and the flame ratio; The formula for the average gray value is: where r < i < L, r represents the lowest threshold of the gray level, L represents the highest threshold of the gray level, I represents the average gray value of the image, p(i) represents the probability of the i-th gray level, and m represents the actual maximum gray level of the image; The formula for the flame ratio is: where A represents the flame ratio, S1 represents the area of the assumed flame image, and S2 represents the area of the entire image.

6. The method for analyzing the furnace flame of a waste incineration plant based on an artificial neural network according to claim 5, characterized in that, the specific process of image discretization includes: Dividing the collected picture into several grids in the spatial direction according to the preset sampling theorem along the horizontal and vertical directions, each grid is represented by a gray value, and the position relationship between pixel points is represented by coordinates (x, y).

7. An apparatus for analyzing the furnace flame of a waste incineration plant based on an artificial neural network, characterized in that, including: An image acquisition module, used to acquire furnace flame image information of the furnace flame in the incinerator; An image processing module is used to preprocess the furnace flame image information and use the preprocessed furnace flame image information as a training sample; The model training module is used to train the BP neural network model using training samples to generate the optimal BP neural network model; The model using module is used to calculate the pre-processed furnace flame image information using the optimal BP neural network model to generate a furnace flame control command.

8. An electronic device, comprising a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the electronic device executes each step of the method for analyzing furnace flame of a waste incineration plant based on an artificial neural network as described in any one of claims 1-6.

9. A computer-readable storage medium having instructions stored thereon, It is characterized in that When the instructions are executed by the processor, the various steps of the method for analyzing furnace flames in a waste incineration plant based on an artificial neural network as claimed in any one of claims 1 to 6 are implemented.

Images