Burr fire detection method, device and equipment for industrial silicon smelting and storage medium

Abstract

The invention discloses a burnout detection method, device and equipment for industrial silicon smelting and a storage medium. The method comprises the following steps: acquiring and storing image data and sensor data in the smelting furnace in real time, wherein the sensor data comprises any one or more of temperature, airflow and furnace charge state; analyzing the image data through a preset machine learning model, and identifying the stabbing fire condition; recording image information and sensor information corresponding to the burnout condition; and according to the recorded image information and sensor information, analyzing real-time image information and sensor information, and predicting a probability value of occurrence of burnout. According to the technical scheme, through real-time data acquisition and machine learning analysis, burnout is effectively predicted and identified, the production safety and efficiency are improved, the heat loss and power consumption of industrial silicon smelting are reduced, meanwhile, equipment damage and production shutdown time caused by burnout are reduced, and the stability and economical efficiency of the smelting process are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent smelting, and in particular to a method, device, equipment and storage medium for detecting sparks in industrial silicon smelting. Background Art

[0002] In the process of industrial silicon smelting, stinging fire manifests itself as excessive viscosity of the charge, poor air permeability or short electrodes, which lead to the ejection of high-temperature hot air flow and the formation of a burning white flame. Stinging fire not only causes a large amount of heat loss and increases power consumption, but also causes uneven distribution of air flow in the furnace, thereby reducing the rate of chemical reactions. At the same time, stinging fire will also cause the furnace mouth temperature to rise, and operating conditions will become more difficult. If the stinging fire phenomenon is serious, it may also burn copper tiles or cooling water pipes, causing the furnace to shut down. Therefore, how to detect stinging fire in a timely manner has become a problem that the industry needs to solve urgently. Summary of the invention

[0003] The present invention provides a method, device, equipment and storage medium for detecting sparks in industrial silicon smelting, which are used to effectively predict and identify sparks through real-time data collection and machine learning analysis, thereby improving production safety and efficiency.

[0004] According to a first aspect of the present invention, a method for detecting sparks in industrial silicon smelting is provided, the method comprising:

[0005] Collect and store image data and sensor data in the smelting furnace in real time, wherein the sensor data includes any one or more of temperature, airflow and charge state;

[0006] Analyzing the image data through a preset machine learning model to identify the fire condition;

[0007] Recording image information and sensor information corresponding to the ignition condition;

[0008] According to the recorded image information and sensor information, the real-time image information and sensor information are analyzed to predict the probability value of the occurrence of a fire.

[0009] In one embodiment, it further includes:

[0010] Before identifying the fire image, the preset machine learning model is trained using the first historical fire image data set;

[0011] After the training, the preset machine learning model is tested using a second historical fire image data set;

[0012] After the test is passed, the preset machine learning model is evaluated through the third historical fire image data set;

[0013] When the evaluation result value is higher than a preset passing threshold, the real-time image data is processed using the preset machine learning model.

[0014] In one embodiment, analyzing the image data using a preset machine learning model includes:

[0015] Performing edge detection on the image data using a preset machine learning model;

[0016] Extracting image features from the edge-detected image, wherein the image features include any one or more of a color histogram, texture features, and edge information;

[0017] Perform fire-related pattern recognition on the image features.

[0018] In one embodiment, extracting image features from the image after edge detection includes:

[0019] Analyze the temporal change data of the image features of the fire, wherein the fire change data includes any one or more of the growth rate and the fluctuation condition;

[0020] The principal component analysis method is used to filter high-weight image features in fire detection to reduce data dimension.

[0021] In one embodiment, the using the first historical fire image data set to train the preset machine learning model includes:

[0022] By fusing a convolutional neural network and a recurrent neural network, the image data and the sensor data are processed simultaneously, wherein the processed image data and the sensor data are data sequences corresponding to moments in a time series;

[0023] The preset machine learning model is optimized by any one or more of hyperparameter adjustment, regularization and model integration.

[0024] In one embodiment, it further includes:

[0025] Analyze the probability of fire in different time periods in the future;

[0026] When the probability value of a fire occurring within a preset safety processing time period is greater than a preset alarm threshold, a fire warning is issued.

[0027] According to a second aspect of the present invention, there is provided a spark detection device for industrial silicon smelting, comprising:

[0028] An acquisition module, used for real-time acquisition and storage of image data and sensor data in the smelting furnace, wherein the sensor data includes any one or more of temperature, airflow and charge state;

[0029] A recognition module, used to analyze the image data through a preset machine learning model to identify the fire condition;

[0030] A recording module, used to record image information and sensor information corresponding to the ignition status;

[0031] The analysis module is used to analyze the real-time image information and sensor information according to the recorded image information and sensor information, and predict the probability value of the occurrence of a fire.

[0032] In one embodiment, the acquisition module, the identification module, the recording module and the analysis module are controlled to execute any one of the above-mentioned methods for detecting sparks in industrial silicon smelting.

[0033] According to a third aspect of the present invention, there is provided an electronic device, the electronic device comprising: a communication interface, a processor, and a memory;

[0034] Wherein, the memory is used to store program instructions, and when the program instructions are executed by the processor that is communicatively connected to the memory through the communication interface, any of the above-mentioned spark detection methods for industrial silicon smelting is implemented.

[0035] According to a fourth aspect of the present invention, there is provided a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a computer (e.g., a processor in the computer), implement any of the above-mentioned methods for detecting sparks in industrial silicon smelting.

[0036] In summary, the present invention provides a method and device for detecting sparks in industrial silicon smelting, the method comprising: real-time acquisition and storage of image data and sensor data in a smelting furnace, the sensor data comprising any one or more of temperature, airflow and charge state; analyzing the image data through a preset machine learning model to identify the spark condition; recording the image information and sensor information corresponding to the spark condition; analyzing the real-time image information and sensor information based on the recorded image information and sensor information to predict the probability value of the occurrence of a spark. The technical solution of the present application effectively predicts and identifies sparks through real-time data acquisition and machine learning analysis, improves production safety and efficiency, reduces heat loss and power consumption in industrial silicon smelting, and at the same time reduces equipment damage and production downtime caused by sparks, ensuring the stability and economy of the smelting process.

[0037] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0038] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 A flow chart of a method for detecting sparks in industrial silicon smelting provided by an embodiment of the present invention;

[0041] Figure 2 A flow chart of another method for detecting sparks in industrial silicon smelting provided by an embodiment of the present invention;

[0042] Figure 3 A flowchart of step S12 of a method for detecting sparks in industrial silicon smelting provided by an embodiment of the present invention;

[0043] Figure 4 A flowchart of step S32 of a method for detecting sparks in industrial silicon smelting provided by an embodiment of the present invention;

[0044] Figure 5 A flow chart of step S21 of a method for detecting sparks in industrial silicon smelting provided by an embodiment of the present invention;

[0045] Figure 6 A flow chart of another method for detecting sparks in industrial silicon smelting provided by an embodiment of the present invention;

[0046] Figure 7 A structural diagram of a spark detection device for industrial silicon smelting provided by an embodiment of the present invention;

[0047] Figure 8 A structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0049] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0050] like Figure 1 As shown, the present invention provides a method for detecting sparks in industrial silicon smelting, and the method for detecting sparks in industrial silicon smelting comprises:

[0051] In step S11, image data and sensor data in the smelting furnace are collected and stored in real time, wherein the sensor data includes any one or more of temperature, airflow and charge state;

[0052] In step S12, the image data is analyzed by a preset machine learning model to identify the fire condition;

[0053] In step S13, the image information and sensor information corresponding to the ignition status are recorded;

[0054] In step S14, the real-time image information and sensor information are analyzed based on the recorded image information and sensor information to predict the probability value of a fire.

[0055] In one embodiment, image data and sensor data in the smelting furnace are collected in real time by sensors and cameras. The sensor data includes any one or more of temperature, airflow and charge state. The collected image data is analyzed by a preset machine learning model to identify the fire condition. The image data is processed by an intelligent algorithm to find potential fire problems in real-time monitoring. After the fire condition is identified, the corresponding image information and sensor information are recorded. The data is persistently processed to ensure that the identified fire events have corresponding records for subsequent analysis. According to the recorded image information and sensor information, the real-time collected image and sensor information are analyzed to predict the probability of future fire. By comparing and analyzing historical data with current data, probability prediction is performed to achieve the purpose of early warning, and active prevention function is possessed. Real-time data is processed and analyzed by a machine learning model to predict the fire condition, and real-time monitoring and early warning are achieved in a complex smelting environment, thereby improving the safety and efficiency of production.

[0056] The technical solution in this embodiment solves the problem of difficulty in real-time detection of sparks in the industrial silicon smelting process. Through real-time data collection and machine learning analysis, sparks can be effectively predicted and identified, production safety and efficiency can be improved, heat loss and power consumption in industrial silicon smelting can be reduced, and equipment damage and production downtime caused by sparks can be reduced, ensuring the stability and economy of the smelting process.

[0057] In one embodiment, Figure 2 As shown, the following steps S21-S22 are also included:

[0058] In step S21, before identifying the fire image, the preset machine learning model is trained using the first historical fire image data set;

[0059] In step S22, after training, the preset machine learning model is tested using a second historical fire image data set;

[0060] In step S23, after the test is passed, the preset machine learning model is evaluated through the third historical fire image data set;

[0061] In step S24, when the evaluation result value is higher than a preset passing threshold, the real-time image data is processed using the preset machine learning model.

[0062] In one embodiment, the preset machine learning model must first be trained with historical data before it is actually applied. The purpose of training is to allow the preset machine learning model to learn how to identify the features and patterns of the thorny fire image, provide a basis for the preset machine learning model, and enable it to understand and process data. After the training is completed, the preset machine learning model needs to be tested to test its performance and accuracy. Verify whether the model can still perform well on data that has not been seen. Ensure that the model is not only effective for training data, but also generalizes to new data. After the test is passed, the preset machine learning model is further evaluated by another set of historical data. The purpose of the evaluation stage is to deeply detect the performance of the preset machine learning model in different scenarios or data sets, so as to further confirm the reliability and robustness of the model and ensure that the model can work stably on a wider range of data. If the evaluation result reaches the pre-set standard (i.e., passes the threshold), the preset machine learning model will be applied to the processing of real-time image data. After the model passes multiple tests and evaluations, it will be put into practical use to ensure that the final applied model has sufficient accuracy and stability.

[0063] In one embodiment, Figure 3 As shown, step S12 includes the following steps S31-S33:

[0064] In step S31, edge detection is performed on the image data using a preset machine learning model;

[0065] In step S32, image features are extracted from the edge-detected image, wherein the image features include any one or more of a color histogram, texture features, and edge information;

[0066] In step S33, fire-related pattern recognition is performed on the image features.

[0067] In one embodiment, a preset machine learning model is used to perform edge detection on image data to identify and extract edge information in the image. Edge detection is used to highlight areas with significant changes in the image, such as the outline and shape of an object, simplify the information in the image, remove unnecessary details, and highlight key information. After completing edge detection, image feature extraction is performed on the processed image. The goal of image feature extraction is to extract important information related to thorns from the image, such as color histogram, texture features, and edge information. Analyzing key image features related to thorns in the image is based on edge detection. By extracting image features, the image data can be analyzed more deeply to identify significant features related to thorns. Pattern recognition related to thorns is performed on the extracted image features. The purpose of thorn pattern recognition is to determine whether there is thorns in the image based on the extracted image features.

[0068] In one embodiment, Figure 4 As shown, step S32 includes the following steps S41-S42:

[0069] In step S41, analyzing the temporal change data of the image features of the fire, wherein the fire change data includes any one or more of the growth rate and the fluctuation condition;

[0070] In step S42, the principal component analysis method is used to filter out high-weight image features in fire detection to reduce the data dimension.

[0071] In one embodiment, the changes of the features of the fire image in the time dimension are analyzed. By monitoring the changes of the features of the fire image over time, the dynamic behavior of the fire phenomenon can be captured. The fire change data includes the growth rate and the fluctuation. These two features can reflect the development trend and stability of the fire phenomenon. The change data of the fire can reflect its state in different time periods. After completing the time series analysis, the image features in the fire detection are screened using the principal component analysis method. Principal component analysis is a statistical method that reduces the dimension of the data by identifying the most representative information in the data while retaining as much important information as possible. By screening out high-weight image features, the amount of data to be processed can be reduced, thereby improving the computational efficiency and the response speed of the model. The purpose of data dimensionality reduction is to reduce redundant information while maintaining key features, thereby simplifying subsequent processing steps. By analyzing the time change data of the features of the fire image, dynamic features related to the fire, such as the growth rate and the fluctuation, are extracted. The principal component analysis method is used to screen out high-weight image features to reduce the data dimension. By effectively reducing the complexity of the data while retaining key information, a more efficient fire detection process is achieved.

[0072] In one embodiment, Figure 5 As shown, step S21 includes the following steps S51-S52:

[0073] In step S51, the image data and the sensor data are processed simultaneously by fusing the convolutional neural network and the recurrent neural network, wherein the processed image data and the sensor data are data sequences corresponding to moments in a time series;

[0074] In step S52, the preset machine learning model is optimized by any one or more of hyperparameter adjustment, regularization and model integration.

[0075] In one embodiment, the image data and sensor data are processed simultaneously by fusing a convolutional neural network (CNN) and a recurrent neural network (RNN). The convolutional neural network is mainly used to process image data. It can automatically extract image features and identify fire-stab patterns by extracting spatial features in the image, and is suitable for the analysis of flame images; while the recurrent neural network is good at processing time series data, can capture the time dependency in the data, and is used to process time series data and analyze the change trend of flames. The long short-term memory network (LSTM) is a variant of RNN, which is particularly suitable for capturing long-term dependencies in time series. The image data and sensor data are processed as corresponding data sequences in a time series. The image data and sensor data at each moment are synchronized, reflecting the different dimensional information at the same time point, and can better capture the changes in the fire-stab phenomenon in time and space. The spatial features of the image and the temporal features of the sensor data can be used simultaneously to improve the accuracy and comprehensiveness of the detection. After completing the data processing, the system needs to optimize the preset machine learning model. The three optimization methods mentioned here are hyperparameter adjustment, regularization, and model integration. The optimal combination of model performance parameters is found by adjusting the model's hyperparameters (such as learning rate, batch size, etc.). The model is prevented from overfitting by adding regularization terms, and its generalization ability is improved so that it can perform well on new data. Combining the prediction results of multiple models, the stability and accuracy of the overall model are usually improved by methods such as voting or averaging. Through the technical solution of this embodiment, the performance of the model can be significantly improved to ensure its accuracy and robustness in practical applications.

[0076] By integrating convolutional neural networks and recurrent neural networks, image data and sensor data are processed simultaneously, making full use of the characteristics of these two types of data to improve the comprehensiveness of detection. The model is optimized through methods such as hyperparameter adjustment, regularization, and model integration to ensure good performance and adaptability. This ensures that complex industrial data can be processed efficiently and accurately, and accurate detection and prediction of fire phenomena can be achieved.

[0077] In one embodiment, Figure 6 As shown, the following steps S61-S62 are also included:

[0078] In step S61, the probability values ​​of fire ignition occurring in different time periods in the future are analyzed;

[0079] In step S62, when the probability value of a fire occurring within a preset safety processing time period is greater than a preset alarm threshold, a fire warning is issued.

[0080] In one embodiment, the probability of fire in the future time period is analyzed, and the upcoming time period is predicted using existing data and models, and the possibility of fire in these time periods is calculated. The probability value reflects the risk level of fire incidents in the future. By performing probability analysis on the future, the risks in different time periods can be evaluated in advance, thereby providing a basis for subsequent decision-making. After completing the probability analysis, the analysis results are judged. If the probability value of fire in the preset safety processing time period (i.e., the time range with enough time to take countermeasures) exceeds a pre-set alarm threshold, the system will trigger the early warning mechanism. If the probability value is greater than the alarm threshold, an early warning is immediately issued to remind the operator of the possible risks and take necessary measures. It ensures that the system can respond in a timely manner when the risk is really imminent, avoiding unnecessary alarms.

[0081] The probability of fire in different time periods in the future is analyzed and the risk is predicted. Conditional judgment is made based on the preset safe processing time period and alarm threshold. When the risk of fire exceeds the threshold, a fire warning is issued. The timeliness and accuracy of the warning are ensured, and timely response is made when the risk is high, thereby protecting the safety and efficiency of the production process.

[0082] In one embodiment, Figure 7 FIG. 1 is a block diagram of a spark detection device for industrial silicon smelting according to an exemplary embodiment. Figure 7 As shown, the spark detection device for industrial silicon smelting includes a collection module 71 , an identification module 72 , a recording module 73 and an analysis module 74 .

[0083] The acquisition module 71 is used to acquire and store image data and sensor data in the smelting furnace in real time, wherein the sensor data includes any one or more of temperature, airflow and charge state;

[0084] The recognition module 72 is used to analyze the image data through a preset machine learning model to identify the fire condition;

[0085] The recording module 73 is used to record the image information and sensor information corresponding to the ignition status;

[0086] The analysis module 74 is used to analyze the real-time image information and sensor information according to the recorded image information and sensor information, and predict the probability value of the occurrence of a fire.

[0087] The acquisition module 71, the identification module 72, the recording module 73 and the analysis module 74 included in the block diagram of the industrial silicon smelting spark detection device are controlled to execute the industrial silicon smelting spark detection method described in any of the above embodiments.

[0088] like Figure 8As shown, the present invention provides an electronic device 800, which includes: a communication interface, a processor 801, and a memory 802;

[0089] Among them, the memory 802 is used to store program instructions. When the program instructions are executed by the processor 801 that is communicatively connected to the memory 802 through the communication interface, the image data and sensor data in the smelting furnace are collected and stored in real time, and the sensor data include any one or more of temperature, airflow and charge state; the image data is analyzed by a preset machine learning model to identify the fire condition; the image information and sensor information corresponding to the fire condition are recorded; based on the recorded image information and sensor information, the real-time image information and sensor information are analyzed to predict the probability value of the occurrence of fire.

[0090] The present invention provides a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, image data and sensor data in a smelting furnace are collected and stored in real time, wherein the sensor data include any one or more of temperature, airflow, and charge state; the image data are analyzed by a preset machine learning model to identify a sparking condition; image information and sensor information corresponding to the sparking condition are recorded; and real-time image information and sensor information are analyzed based on the recorded image information and sensor information to predict the probability value of a sparking condition.

[0091] It should be understood that the specific features, operations and details described hereinabove about the method of the present invention may also be similarly applied to the device and system of the present invention, or, vice versa. In addition, each step of the method of the present invention described above may be performed by the corresponding parts or units of the device or system of the present invention.

[0092] It should be understood that each module / unit of the device of the present invention can be implemented in whole or in part by software, hardware, firmware or a combination thereof. Each module / unit can be embedded in the processor of the computer device in the form of hardware or firmware or independent of the processor, or can be stored in the memory of the computer device in the form of software for the processor to call to perform the operation of each module / unit. Each module / unit can be implemented as an independent component or module, or two or more modules / units can be implemented as a single component or module.

[0093] In one embodiment, a computer device is provided, which includes a memory and a processor, and the memory stores computer instructions executable by the processor, and the computer instructions instruct the processor to execute each step of the method of the embodiment of the present invention when executed by the processor. The computer device can be a server, a terminal, or any other electronic device with necessary computing and / or processing capabilities in a broad sense. In one embodiment, the computer device may include a processor, a memory, a network interface, a communication interface, etc. connected through a system bus. The processor of the computer device can be used to provide necessary computing, processing and / or control capabilities. The memory of the computer device may include a non-volatile storage medium and an internal memory. An operating system, a computer program, etc. may be stored in or on the non-volatile storage medium. The internal memory can provide an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface and the communication interface of the computer device can be used to connect and communicate with external devices through a network. The steps of the method of the present invention are executed by the processor.

[0094] The present invention may be implemented as a computer-readable storage medium having a computer program stored thereon, which causes the steps of the method of an embodiment of the present invention to be executed when executed by a processor. In one embodiment, the computer program is distributed on a plurality of computer devices or processors coupled to a network so that the computer program is stored, accessed, and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, may be performed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations may be performed by one or more computer devices or processors, and one or more other method steps / operations may be performed by one or more other computer devices or processors. One or more computer devices or processors may perform a single method step / operation, or perform two or more method steps / operations.

[0095] It will be understood by those skilled in the art that the method steps of the present invention can be completed by instructing related hardware such as a computer device or a processor through a computer program, and the computer program can be stored in a non-temporary computer-readable storage medium, and the steps of the present invention are executed when the computer program is executed. Depending on the circumstances, any reference to memory, storage, database or other media herein may include non-volatile and / or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (through real-time data acquisition and machine learning analysis, effectively predict and identify spurs, improve production safety and efficiency, reduce heat loss and power consumption of industrial silicon smelting, and reduce equipment damage and production downtime caused by spurs, ensuring the stability and economy of the smelting process PROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid state disk, etc. Examples of volatile memory include random access memory (RAM), external cache memory, etc.

[0096] The various technical features described above can be combined arbitrarily. Although all possible combinations of these technical features are not described, any combination of these technical features should be considered to be covered by this specification as long as there is no contradiction in such combination.

[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting sparks in industrial silicon smelting, characterized in that: include: Collect and store image data and sensor data in the smelting furnace in real time, wherein the sensor data includes any one or more of temperature, airflow and charge state; Analyzing the image data through a preset machine learning model to identify the fire condition; Recording image information and sensor information corresponding to the ignition condition; According to the recorded image information and sensor information, the real-time image information and sensor information are analyzed to predict the probability value of the occurrence of a fire.

2. The method for detecting sparks in industrial silicon smelting according to claim 1, characterized in that: Also includes: Before identifying the fire image, the preset machine learning model is trained using the first historical fire image data set; After the training, the preset machine learning model is tested using a second historical fire image data set; After the test is passed, the preset machine learning model is evaluated through the third historical fire image data set; When the evaluation result value is higher than a preset passing threshold, the real-time image data is processed using the preset machine learning model.

3. The method for detecting sparks in industrial silicon smelting according to claim 1, characterized in that: The analyzing the image data by using a preset machine learning model includes: Performing edge detection on the image data using a preset machine learning model; Extracting image features from the edge-detected image, wherein the image features include any one or more of a color histogram, texture features, and edge information; Perform fire-related pattern recognition on the image features.

4. The method for detecting sparks in industrial silicon smelting according to claim 3, characterized in that: The step of extracting image features from the edge detected image comprises: Analyze the temporal change data of the image features of the fire, wherein the fire change data includes any one or more of the growth rate and the fluctuation condition; The principal component analysis method is used to filter high-weight image features in fire detection to reduce data dimension.

5. The method for detecting sparks in industrial silicon smelting according to claim 2, characterized in that: The using the first historical fire image data set to train the preset machine learning model includes: By fusing a convolutional neural network and a recurrent neural network, the image data and the sensor data are processed simultaneously, wherein the processed image data and the sensor data are data sequences corresponding to moments in a time series; The preset machine learning model is optimized by any one or more of hyperparameter adjustment, regularization and model integration.

6. The method for detecting sparks in industrial silicon smelting according to claim 1, characterized in that: Also includes: Analyze the probability of fire in different time periods in the future; When the probability value of a fire occurring within a preset safety processing time period is greater than a preset alarm threshold, a fire warning is issued.

7. A spark detection device for industrial silicon smelting, characterized in that: include: An acquisition module, used for real-time acquisition and storage of image data and sensor data in the smelting furnace, wherein the sensor data includes any one or more of temperature, airflow and charge state; A recognition module, used to analyze the image data through a preset machine learning model to identify the fire condition; A recording module, used to record image information and sensor information corresponding to the ignition status; The analysis module is used to analyze the real-time image information and sensor information according to the recorded image information and sensor information, and predict the probability value of the occurrence of a fire.

8. The spark detection device for industrial silicon smelting as claimed in claim 7, characterized in that: The acquisition module, the identification module, the recording module and the analysis module are controlled to execute the method for detecting sparks in industrial silicon smelting as described in any one of claims 1-6.

9. An electronic device, characterized in that: include: Communication interface, processor, memory; Wherein, the memory is used to store program instructions, and when the program instructions are executed by the processor that is communicatively connected to the memory through the communication interface, the electronic device implements the method for detecting sparks in industrial silicon smelting as described in any one of claims 1 to 6.

10. A computer-readable storage medium having program instructions stored thereon, characterized in that: When the program instructions are executed by a computer, the computer implements the method for detecting sparks in industrial silicon smelting as described in any one of claims 1 to 6.

Images