Hydrogen chloride synthesis process control method, device and system, upper computer and medium

Abstract

The invention relates to a hydrogen chloride synthesis process control method, device and system, an upper computer and a medium. The method comprises the following steps: acquiring to-be-predicted data, wherein the to-be-predicted data comprises a three-channel pixel average value of a hydrogen chloride flame picture at a plurality of to-be-predicted moments in a time period; inputting to-be-predicted data into the hydrogen chloride concentration prediction model to obtain predicted hydrogen chloride concentration; inputting to-be-predicted data into the chlorine-hydrogen state prediction model to obtain a predicted chlorine-hydrogen state; determining that the predicted chlorine-hydrogen state is abnormal continuously for multiple times, and determining the ratio of chlorine to hydrogen according to the predicted hydrogen chloride concentration; and adjusting the flow of chlorine or hydrogen based on the ratio of chlorine to hydrogen so as to be within a preset range. By adopting the method, the hydrogen chloride concentration and the chlorine-hydrogen state can be predicted based on the flame color, the chlorine-hydrogen ratio can be automatically calculated according to the predicted hydrogen chloride concentration, and the flow of chlorine or hydrogen can be adjusted through feedback.

Description

Technical Field

[0001] The present application relates to the technical field of synthesis reaction control, and particularly to a control method for hydrogen chloride synthesis process, a control device for hydrogen chloride synthesis process, a control system for hydrogen chloride synthesis process, a computer device, a computer-readable storage medium, and a computer program product. Background Art

[0002] In the process of hydrogen chloride synthesis, the ratio of chlorine gas to hydrogen gas is mainly monitored by three methods: flowmeter detection, hydrogen chloride purity analysis, and flame color observation. Due to the particularity of the system, the flowmeter display of chlorine-hydrogen flow is not accurate and it is difficult to be used as the ultimate judgment basis for determining the chlorine-hydrogen ratio. And the hydrogen chloride purity analysis is usually an intermittent detection and analysis, which cannot achieve real-time determination. The flame color is a key indicator for evaluating the combustion status and is also a direct basis for reflecting the chlorine-hydrogen ratio. The following is the correspondence between the flame color and the combustion state:

[0003] Bluish-white flame: It indicates that the ratio of hydrogen gas to chlorine gas is appropriate and the combustion is sufficient, which is an ideal combustion state.

[0004] Yellow-green flame: Usually indicates that chlorine gas is in excess, and it is necessary to increase the hydrogen gas flow or reduce the chlorine gas supply.

[0005] White smoke flame: It may mean that hydrogen gas is in excess or the purity of chlorine gas is insufficient, and it is necessary to increase the chlorine gas flow or reduce the hydrogen gas supply.

[0006] To solve this problem, the commonly used method at present is the manual observation method. This method has limitations in timeliness and accuracy whether it is on-site observation through a sight glass or monitoring in the central control room, which is not conducive to precisely controlling the hydrogen chloride synthesis process.

[0007] CN116824170A proposes a method and system for analyzing the flame of a hydrogen chloride synthesis furnace, which uses a flame sensor to judge whether there is a flame in the hydrogen chloride synthesis furnace; collects the flame video and obtains the flame image by frame division; performs color detection on the obtained image under the RGB and HSI mixed color models to obtain the flame pixel point distribution matrix and get the preliminary image containing the suspected flame area; grayscales the preliminary image containing the suspected flame area, uses PSO search for global optimization, and substitutes the global optimal solution into the K-means clustering algorithm to cluster the image; multiplies the preliminary image matrix containing the suspected flame area by the clustered flame image matrix to obtain the final flame area; adjusts the gas concentration according to the RGB mean value and color temperature of the final flame area. This patent provides a method for identifying the flame color area, but does not give a specific method for automatically judging the flame color, judging over-hydrogen or over-chlorine according to the flame color, and subsequent regulation schemes.

[0008] Therefore, there is an urgent need for a precise automated intelligent recognition method to improve the efficiency and safety of hydrogen chloride synthesis, so as to achieve real-time monitoring of the flame color, automatically adjust the flow rates of hydrogen and chlorine, maintain the optimal combustion state, and thus improve production efficiency and reduce safety risks. Summary of the Invention

[0009] Based on this, it is necessary to provide a hydrogen chloride synthesis process control method, a hydrogen chloride synthesis process control device, a hydrogen chloride synthesis process control system, a host computer, a computer-readable storage medium, and a computer program product for the above technical problems.

[0010] In the first aspect, the present application provides a hydrogen chloride synthesis process control method, and the method includes:

[0011] Obtain data to be predicted, where the data to be predicted includes the average values of the three-channel pixels of hydrogen chloride flame pictures at multiple prediction times within a time period;

[0012] Input the data to be predicted into the hydrogen chloride concentration prediction model to obtain the predicted hydrogen chloride concentration;

[0013] Input the data to be predicted into the chlorine-hydrogen state prediction model to obtain the predicted chlorine-hydrogen state;

[0014] Determine that the predicted chlorine-hydrogen state is abnormal for multiple consecutive times, and determine the ratio of chlorine to hydrogen according to the predicted hydrogen chloride concentration;

[0015] Adjust the flow rate of chlorine or hydrogen based on the ratio of chlorine to hydrogen to make it within a preset range.

[0016] In one or more embodiments, the construction steps of the hydrogen chloride concentration prediction model include:

[0017] Construct a first data set, which includes a series of historical hydrogen chloride concentrations and the average values of the three-channel pixels of hydrogen chloride flame pictures corresponding to the historical hydrogen chloride concentration moments;

[0018] Construct a first training set based on the first data set;

[0019] Randomly draw multiple subsamples with replacement from the first training set, each subsample having the same size as the first training set, for constructing different decision trees, and the node splitting criterion for each decision tree is to minimize the mean square error;

[0020] Input the first training set into the random forest regression model and train it to obtain the hydrogen chloride concentration prediction model.

[0021] In one or more embodiments, the chlorine-hydrogen status prediction model includes a first chlorine-hydrogen status prediction model and a second chlorine-hydrogen status prediction model, and the prediction results of both include normal and abnormal; the step of inputting the data to be predicted into the chlorine-hydrogen status prediction model to obtain the predicted chlorine-hydrogen status includes:

[0022] Input the data to be predicted into the first chlorine-hydrogen status prediction model and the second chlorine-hydrogen status prediction model respectively to obtain corresponding prediction results;

[0023] Compare the prediction result of the first chlorine-hydrogen status prediction model with the prediction result of the second chlorine-hydrogen status prediction model;

[0024] If the prediction results of both are consistent, output the prediction result as the predicted chlorine-hydrogen status;

[0025] If the prediction results of both are inconsistent and one prediction result is normal while the other is abnormal, output the corresponding abnormal prediction result as the predicted chlorine-hydrogen status.

[0026] In one or more embodiments, the step of constructing the first data set includes:

[0027] Obtain the series of historical hydrogen chloride concentrations and the corresponding hydrogen chloride flame videos;

[0028] Perform frame segmentation on the hydrogen chloride flame videos to obtain a hydrogen chloride flame picture data set;

[0029] According to the sampling time of the historical hydrogen chloride concentration, obtain the hydrogen chloride flame pictures at the corresponding moments;

[0030] According to the hydrogen chloride flame pictures at the corresponding moments, obtain their target flame regions, use a mask to extract the non-zero pixel coordinates within the target flame regions, and calculate the average values of the three-channel pixels R mean 、G mean and B mean .

[0031] In one or more embodiments, the sampling time of the historical hydrogen chloride concentration is t HCl , and the acquisition time of the hydrogen chloride flame picture corresponding to this moment is t flame ;

[0032] The step of obtaining the hydrogen chloride flame pictures at the corresponding moments according to the sampling time of the historical hydrogen chloride concentration includes:

[0033] According to the t flame Retrieve the hydrogen chloride flame pictures at the corresponding moments from the hydrogen chloride flame picture data set,

[0034] Among them, H is the height from the flame in the synthesis furnace to the hydrogen chloride collection outlet, and v(h) is the instantaneous flow rate of hydrogen chloride gas at height h.

[0035] In one or more embodiments, the steps of constructing the first chlorine-hydrogen state prediction model include:

[0036] Obtain a second data set, which includes a series of historical chlorine-hydrogen states and the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the moments of the historical chlorine-hydrogen states;

[0037] Construct a second training set based on the second data set;

[0038] Randomly draw multiple subsamples with replacement from the second training set, each subsample having the same size as the second training set, for constructing different decision trees, and the node splitting criterion for each decision tree is the Gini index or information gain;

[0039] Input the second training set into a random forest classification model and train it to obtain the first chlorine-hydrogen state prediction model.

[0040] In one or more embodiments, the steps of constructing the second chlorine-hydrogen state prediction model include:

[0041] Obtain a second data set, which includes a series of historical chlorine-hydrogen states and the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the moments of the historical chlorine-hydrogen states;

[0042] Create a new feature column based on the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the moments of the historical chlorine-hydrogen states, and perform standardization processing on it to obtain a standardized feature column;

[0043] Construct a third training set based on the standardized feature column and the historical chlorine-hydrogen states;

[0044] Input the third training set into a logistic regression model and train it to obtain the second chlorine-hydrogen state prediction model.

[0045] In one or more embodiments, the anomalies include over-chlorine and over-hydrogen;

[0046] The step of determining the ratio of chlorine gas and hydrogen gas according to the predicted hydrogen chloride concentration includes: calculating the ratio of the volume of chlorine gas to the volume of hydrogen gas through the following formula:

[0047]

[0048] Among them, represents the predicted hydrogen chloride concentration, and when the predicted chlorine-hydrogen state is over-chlorine, v 过量represents the volume of chlorine gas. When predicting that the chlorine-hydrogen state is over-hydrogen, v 过量 represents the volume of hydrogen gas.

[0049] In a second aspect, the present application also provides a control device for the hydrogen chloride synthesis process. The device includes:

[0050] A data acquisition module for acquiring data to be predicted, where the data to be predicted includes the average value of the three-channel pixels of the hydrogen chloride flame pictures at multiple prediction moments within a time period;

[0051] A first prediction module for inputting the data to be predicted into a hydrogen chloride concentration prediction model to obtain the predicted hydrogen chloride concentration;

[0052] A second prediction module for inputting the data to be predicted into a chlorine-hydrogen state prediction model to obtain the predicted chlorine-hydrogen state;

[0053] A calculation module for determining that the predicted chlorine-hydrogen state is abnormal for multiple consecutive times, and determining the ratio of chlorine gas to hydrogen gas according to the predicted hydrogen chloride concentration;

[0054] An adjustment module for adjusting the flow rate of chlorine gas or hydrogen gas based on the ratio of chlorine gas to hydrogen gas so that it is within a preset range.

[0055] In a third aspect, the present application provides a control system for the hydrogen chloride synthesis process. The system includes: a sight glass, a camera, a chlorine gas flow control valve, a hydrogen gas flow control valve, and a host computer;

[0056] The sight glass is arranged on the synthesis furnace and is used to observe the hydrogen chloride flame in the synthesis furnace;

[0057] The camera is used to obtain a hydrogen chloride flame video through the sight glass;

[0058] The chlorine gas flow control valve is arranged on the chlorine gas supply pipe of the synthesis furnace;

[0059] The hydrogen gas flow control valve is arranged on the hydrogen gas supply pipe of the synthesis furnace;

[0060] The host computer is used to process the obtained hydrogen chloride flame video to obtain data to be predicted, where the data to be predicted includes the average value of the three-channel pixels of the hydrogen chloride flame pictures at multiple prediction moments within a time period. The host computer determines the predicted hydrogen chloride concentration and the predicted chlorine-hydrogen state according to the data to be predicted, determines that the predicted chlorine-hydrogen state is abnormal for multiple consecutive times, determines the ratio of chlorine gas to hydrogen gas according to the predicted hydrogen chloride concentration, and sends an instruction to the chlorine gas flow control valve or the hydrogen gas flow control valve based on the ratio of chlorine gas to hydrogen gas to adjust the flow rate of chlorine gas or hydrogen gas so that it is within a preset range.

[0061] In one or more embodiments, the system further includes a light-shielding box body, which is sleeved outside the sight glass and the camera for reducing the interference of external light.

[0062] In a fourth aspect, the present application provides a host computer, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of a hydrogen chloride synthesis process control method are implemented.

[0063] In a fifth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of a hydrogen chloride synthesis process control method are implemented.

[0064] In a sixth aspect, the present application provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps of the hydrogen chloride synthesis process control method are implemented.

[0065] One of the above technical solutions has the following advantages or beneficial effects: By predicting the hydrogen chloride concentration and the chlorine-hydrogen state based on the flame color, automatically calculating the chlorine-hydrogen ratio according to the predicted hydrogen chloride concentration, and feeding back and then adjusting the flow rates of chlorine or hydrogen. Compared with the traditional methods of manually observing the flame color to determine the chlorine-hydrogen state and detecting the hydrogen chloride concentration in the laboratory, this solution has advantages such as real-time and high efficiency, greatly reducing the labor cost and analysis cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is an application environment diagram of the hydrogen chloride synthesis process control method in one embodiment;

[0067] Figure 2 It is a flowchart of the hydrogen chloride synthesis process control method in one embodiment;

[0068] Figure 3 It is a flowchart of the hydrogen chloride synthesis process control method in a specific embodiment;

[0069] Figure 4 It is a schematic diagram of the shape of the target area obtained in a specific embodiment;

[0070] Figure 5 It is a schematic diagram of Data Table 1 in a specific embodiment;

[0071] Figure 6 It is a schematic diagram of the output result in a specific embodiment;

[0072] Figure 7 It is a structural block diagram of the hydrogen chloride synthesis process control device in one embodiment;

[0073] Figure 8Schematic diagram of the light-shielding box body;

[0074] Figure 9 Internal structure diagram of the host computer in one embodiment. Specific implementation manners

[0075] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0076] Referring to "embodiments" herein means that the specific features, structures or characteristics described in connection with the embodiments can be included in at least one embodiment of the present application. The phrase appears in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0077] The hydrogen chloride synthesis process control method provided by the present application can be applied to an application environment as Figure 1 shown. Among them, chlorine gas and hydrogen gas are introduced into the synthesis furnace 102 and burned to synthesize hydrogen chloride. The camera 803 acquires the hydrogen chloride flame video in the synthesis furnace 102, and the hydrogen chloride flame video is stored in the database 106 and communicates with the host computer 108 through the network.

[0078] In an application environment as Figure 2 shown, the host computer 108 can be a terminal, including but not limited to a personal computer, a laptop computer, a smart phone, a tablet computer, etc., and can also be implemented by an independent server or a server cluster composed of multiple servers.

[0079] The synthesis furnace 102 synthesizes hydrogen chloride gas through the direct combustion of chlorine gas and hydrogen gas. In the prior art, the hydrogen chloride concentration can be obtained by sampling and detecting the hydrogen chloride gas, and the chlorine-hydrogen state can be determined by the flame color when chlorine gas and hydrogen gas are burned, including normal (the flame color is bluish-white), excessive chlorine (the flame color is yellowish-green), excessive hydrogen (the flame color is white), and the flow rates of chlorine gas or hydrogen gas are manually adjusted. This method has low timeliness and requires a large amount of labor costs. The host computer 108 determines the predicted hydrogen chloride concentration and the predicted chlorine-hydrogen state through real-time hydrogen chloride flame pictures, and determines the ratio of chlorine gas and hydrogen gas according to the two, and automatically adjusts the flow rates of chlorine gas or hydrogen gas to make them within the normal range.

[0080] In one embodiment, as Figure 2 shown, a hydrogen chloride synthesis process control method is provided, and this method is applied to Figure 1Taking the host computer 108 in as an example, the method includes the following steps:

[0081] S202, Obtain the data to be predicted, where the data to be predicted includes the three-channel pixel average values of hydrochloric acid flame pictures at multiple prediction moments within a time period.

[0082] Among them, the data to be predicted can be obtained by processing the data collected by the acquisition device. For example, through the device 803 as shown in Figure 1 , it can be a camera or the like to obtain data, and the data is processed by the host computer 108.

[0083] S204, Input the data to be predicted into the hydrochloric acid concentration prediction model to obtain the predicted hydrochloric acid concentration.

[0084] S206, Input the data to be predicted into the chlorine-hydrogen state prediction model to obtain the predicted chlorine-hydrogen state.

[0085] In this embodiment, the chlorine-hydrogen state prediction model may include one or more identical or different models that can predict the chlorine-hydrogen state. When there are multiple models that can predict the chlorine-hydrogen state, the prediction accuracy can be further improved. The predicted chlorine-hydrogen state includes normal and abnormal, where abnormal includes over-chlorine and over-hydrogen. Over-chlorine means excessive chlorine, and over-hydrogen means excessive hydrogen.

[0086] S208, Determine that the predicted chlorine-hydrogen state is abnormal for multiple consecutive times, and determine the ratio of chlorine to hydrogen according to the predicted hydrochloric acid concentration;

[0087] S210, Adjust the flow rate of chlorine or hydrogen based on the ratio of chlorine to hydrogen to make it within the preset range.

[0088] In the above-mentioned hydrochloric acid synthesis process control method of the embodiment, by obtaining the data to be predicted, using the hydrochloric acid concentration prediction model to obtain the predicted hydrochloric acid concentration, using the chlorine-hydrogen state prediction model to obtain the predicted chlorine-hydrogen state, the predicted chlorine-hydrogen state includes normal chlorine-hydrogen state and abnormal chlorine-hydrogen state. When the predicted chlorine-hydrogen state is abnormal for multiple consecutive times, determine the ratio of chlorine to hydrogen according to the predicted hydrochloric acid concentration, and automatically adjust the ratio of chlorine or hydrogen by the host computer, so that the ratio of the two is within the preset range to maintain the best combustion state.

[0089] Compared with the prior art, this method has higher timeliness, reduces the untimely and error in manual judgment, improves the hydrochloric acid synthesis efficiency, and can save a large amount of labor and material costs.

[0090] In this embodiment, to ensure operational safety, the volume ratio of hydrogen to chlorine is set between 1.08 - 1.05:1. When the ratio of the two is within this range, the hydrogen chloride concentration can reach 92 - 95% VOL.

[0091] In one embodiment, the steps for constructing the hydrogen chloride concentration prediction model include:

[0092] Construct a first data set, which includes a series of historical hydrogen chloride concentrations and the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the moments of the historical hydrogen chloride concentrations; construct a first training set based on the first data set; randomly draw multiple subsamples with replacement from the first training set, each subsample having the same size as the first training set, for constructing different decision trees, and the node splitting criterion for each decision tree is to minimize the mean squared error; input the first training set into a random forest regression model and perform training to obtain the hydrogen chloride concentration prediction model.

[0093] In this embodiment, the three-channel pixel average values of the hydrogen chloride flame pictures corresponding to the series of historical hydrogen chloride concentrations can be obtained by the host computer 108 after processing the hydrogen chloride flame pictures obtained at the moments corresponding to the historical hydrogen chloride concentrations.

[0094] The accuracy of the hydrogen chloride concentration prediction model constructed in this embodiment can reach over 95%.

[0095] In one embodiment, the steps for constructing the first data set include:

[0096] Obtain the series of historical hydrogen chloride concentrations and the corresponding hydrogen chloride flame videos;

[0097] Perform frame splitting on the hydrogen chloride flame videos to obtain a hydrogen chloride flame picture data set;

[0098] According to the sampling time of the historical hydrogen chloride concentrations, obtain the hydrogen chloride flame pictures at the corresponding moments;

[0099] According to the hydrogen chloride flame pictures at the corresponding moments, obtain their target flame regions, use a mask to extract the non-zero pixel coordinates within the target flame regions, and calculate the three-channel pixel average values R mean 、G mean and B mean .

[0100] In this embodiment, the historical hydrogen chloride concentration can be obtained through laboratory testing. The testing method can be manual testing or instrument testing. The hydrogen chloride flame video corresponding to the series of historical hydrogen chloride concentrations refers to the hydrogen chloride flame video within the time period corresponding to the series of historical hydrogen chloride concentrations. In this embodiment, the frame splitting process can be to extract frames from the hydrogen chloride flame video using a programming language. The time interval for obtaining pictures is adjusted according to the production situation, such as obtaining one picture per second, and the pictures are automatically saved, thereby obtaining a hydrogen chloride flame picture dataset.

[0101] The method for obtaining the target flame area can be to select the coordinate points of the flame area in the picture to form a closed interval, and this closed interval forms the required target flame area.

[0102] During the hydrogen chloride synthesis process, the observation of the flame usually needs to be carried out through a sight glass. However, as the reaction continues, the sight glass will gradually accumulate dirt of different degrees and in different areas. Due to production requirements, it is impossible to clean the dirt in real time, which brings difficulties to the overall extraction of the flame area. For this reason, the method provided in this example can appropriately adjust the extraction coordinates according to the actual dirt situation of the sight glass, so as to extract the area less affected by the sight glass. Compared with the clustering algorithm that directly identifies the flame color, this method has higher flexibility and accuracy.

[0103] In one embodiment, the sampling time of the historical hydrogen chloride concentration is t HCl , and the acquisition time of the hydrogen chloride flame picture corresponding to this moment is t flame ;

[0104] The step of obtaining the hydrogen chloride flame picture corresponding to the corresponding moment according to the sampling time of the historical hydrogen chloride concentration includes:

[0105] According to the t flame retrieve the hydrogen chloride flame picture corresponding to the corresponding moment from the hydrogen chloride flame picture dataset,

[0106] where H is the height from the flame in the synthesis furnace to the hydrogen chloride collection outlet, and v(h) is the instantaneous flow rate of the hydrogen chloride gas at height h.

[0107] Specifically, the specific definition of the flame in the synthesis furnace is the center of the flame, and the height h is based on the center of the flame as the reference point.

[0108] In one embodiment, the chlorine-hydrogen state prediction model includes a first chlorine-hydrogen state prediction model and a second chlorine-hydrogen state prediction model, and the prediction results of both include normal and abnormal; the step of inputting the data to be predicted into the chlorine-hydrogen state prediction model to obtain the predicted chlorine-hydrogen state includes: respectively inputting the data to be predicted into the first chlorine-hydrogen state prediction model and the second chlorine-hydrogen state prediction model to obtain corresponding prediction results; comparing the prediction results of the first chlorine-hydrogen state prediction model and the second chlorine-hydrogen state prediction model; if the prediction results of both are consistent, then output the prediction result as the predicted chlorine-hydrogen state; if the prediction results of both are inconsistent and one prediction result is normal and the other is abnormal, then output the corresponding abnormal prediction result as the predicted chlorine-hydrogen state.

[0109] In this embodiment, for the same data to be predicted, corresponding prediction results can be obtained through the first chlorine-hydrogen state prediction model and the second chlorine-hydrogen state prediction model respectively. By comparing the predicted chlorine-hydrogen states of the two prediction models, the prediction accuracy is further improved.

[0110] In one embodiment, the construction steps of the first chlorine-hydrogen state prediction model include:

[0111] Obtain a second data set, which includes a series of historical chlorine-hydrogen states and the three-channel pixel average values of the hydrogen chloride flame pictures corresponding to the historical chlorine-hydrogen state at the corresponding moment; construct a second training set based on the second data set; randomly draw multiple subsamples with replacement from the second training set, and the size of each subsample is the same as that of the second training set, which is used to construct different decision trees, and the node splitting criterion for each decision tree is the Gini index or information gain; input the second training set into the random forest classification model and train it to obtain the first chlorine-hydrogen state prediction model.

[0112] In this embodiment, the historical chlorine-hydrogen state can be obtained through manual determination. The three-channel pixel average value of the hydrogen chloride flame picture corresponding to the historical chlorine-hydrogen state at the corresponding moment can be obtained by processing the hydrogen chloride flame picture obtained at the corresponding moment of the historical chlorine-hydrogen state by the upper computer 108. For the specific acquisition process, refer to "Obtain the hydrogen chloride flame picture corresponding to the corresponding moment according to the sampling time of the historical hydrogen chloride concentration" in obtaining the first data. The two processing processes are the same, and the only difference is that "the sampling time of the historical hydrogen chloride concentration" is replaced by "the sampling time of the historical chlorine-hydrogen state".

[0113] In this embodiment, the accuracy of the constructed first chlorine-hydrogen state prediction model can reach more than 97%.

[0114] In one embodiment, the construction steps of the second chlorine-hydrogen state prediction model include:

[0115] Obtain a second data set, which includes a series of historical chlorine-hydrogen states and the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the moments of the historical chlorine-hydrogen states;

[0116] Create a new feature column based on the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the moments of the historical chlorine-hydrogen states, and perform standardization processing on it to obtain a standardized feature column;

[0117] Construct a third training set based on the standardized feature column and the historical chlorine-hydrogen states;

[0118] Input the third training set into a logistic regression model and perform training to obtain the second chlorine-hydrogen state prediction model.

[0119] In this embodiment, the acquisition of the second data set is the same as that in the previous embodiment, and will not be elaborated here. The accuracy of the constructed second chlorine-hydrogen state prediction model can reach more than 97%.

[0120] The purpose of creating a new feature column based on the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the moments of the historical chlorine-hydrogen states is to improve the performance of the model by adding or transforming features. According to the domain knowledge of specific problems, create some meaningful new features, such as performing square sum and multiplication operations on color components.

[0121] In one embodiment, the abnormality includes over-chlorine and over-hydrogen; the step of determining the ratio of chlorine gas and hydrogen gas according to the predicted hydrogen chloride concentration includes: calculating the ratio of the volume of chlorine gas to the volume of hydrogen gas through the following formula:

[0122]

[0123] where, represents the predicted hydrogen chloride concentration, and v when the predicted chlorine-hydrogen state is over-chlorine 过量 represents the volume of chlorine gas, and when the predicted chlorine-hydrogen state is over-hydrogen, v 过量 represents the volume of hydrogen gas.

[0124] The calculation logic of the ratio of the volume of chlorine gas to the volume of hydrogen gas is as follows:

[0125] Assume that the purities of chlorine gas and hydrogen gas entering the synthesis furnace are both 100%. Among them, the standard cubic volume flow rate of chlorine gas is represented by v 1 and the standard cubic volume flow rate of hydrogen gas is represented by v 2 The concentration of hydrogen chloride is denoted as in units of %VOL.

[0126] It is known that the chemical equation for the reaction of chlorine gas and hydrogen gas is Under the same conditions, the volume ratio of gases is equal to the molar ratio, and chlorine and hydrogen react in a molar ratio of 1:1. When v 2 >v 1 it means that hydrogen is in excess and chlorine has completely reacted. Therefore, the volume v HCl of the generated hydrogen chloride is equal to the volume v 1 of chlorine.

[0127] The calculation of the hydrogen chloride concentration is based on the proportion of the volume of the product hydrogen chloride in the total volume of the gases after the reaction. The total volume of the gases after the reaction consists of the volume of hydrogen chloride and the volume of unreacted hydrogen gas. Therefore, according to the formula:

[0128]

[0129] Substitute v HCl = v 1 into it:

[0130]

[0131] So That is When chlorine is in excess, v 过量 represents chlorine, and conversely, when hydrogen is in excess, v 过量 represents hydrogen.

[0132] In this embodiment, according to the predicted hydrogen chloride concentration and the predicted chlorine-hydrogen state, the volume ratio of the excess gas and the non-excess gas is determined, and the flow rate of the excess gas is automatically adjusted by the host computer 108.

[0133] In a specific implementation manner, as Figure 3 shown, a method for controlling the hydrogen chloride synthesis process includes:

[0134] S301, regularly sampling and testing to obtain a series of historical hydrogen chloride concentrations and obtaining the hydrogen chloride flame video during the sampling period, where the number of a series of historical hydrogen chloride concentrations is not less than 500.

[0135] S302, performing frame extraction on the hydrogen chloride flame video to obtain a hydrogen chloride flame picture dataset. Specifically, one picture is extracted per second.

[0136] S303, according to the sampling time of the historical hydrogen chloride concentration, obtaining the hydrogen chloride flame picture at the corresponding moment. Among them, the sampling time of the historical hydrogen chloride concentration is t HCl , and the acquisition time of the hydrogen chloride flame picture corresponding to this moment is t flame . According to the t flame retrieve the hydrogen chloride flame picture at the corresponding moment from the hydrogen chloride flame picture dataset,

[0137] Wherein, H is the height from the flame in the synthesis furnace to the hydrogen chloride collection outlet, and v(h) is the instantaneous flow rate of the hydrogen chloride gas at height h.

[0138] S304. According to the hydrogen chloride flame picture at the corresponding moment, obtain its target flame area, and manually determine the chlorine-hydrogen state through this target flame area, that is, obtain the historical chlorine-hydrogen state.

[0139] Wherein, the specific steps of obtaining the target flame area according to the hydrogen chloride flame picture at the corresponding moment include: through the opencv built-in setMouseCallback function, use the mouse click to obtain the coordinate points of the corresponding target area of the picture, select more than or equal to 3 coordinate points, which can form a closed interval. For example, three points form a triangular area, four points form a quadrilateral area, and multiple points form a polygonal area. The obtained closed interval is the target flame area.

[0140] During the hydrogen chloride synthesis process, the observation of the flame usually needs to be carried out through a sight glass. However, as the reaction continues, the sight glass will gradually accumulate dirt in different degrees and different areas. Due to production requirements, it is impossible to clean the dirt in real time, which brings difficulties to the overall extraction of the flame area. The above method for obtaining the target flame area provided by the present invention can appropriately adjust the extraction coordinates according to the dirt condition of the actual sight glass, so as to extract the area less affected by the sight glass. Compared with the clustering algorithm that directly recognizes the flame color, this method of extracting the flame area coordinates has higher flexibility, and avoids the possible negative impacts caused by the dirty sight glass, improving the accuracy.

[0141] S305. According to the height and width of the target flame area, create a mask image with the same image size as it, the initial value is 0 representing black, fill the polygon on the mask image, the filling value is 255 representing white, use the mask to extract the non-zero pixel coordinates in the target flame area, and calculate the average value of the three-channel pixels of the hydrogen chloride flame picture corresponding to the historical hydrogen chloride concentration moment respectively, that is, R mean 、G mean and B mean , and their calculation formulas are as follows:

[0142] R mean =(R 1 +R 2 +R 3 +...+R n ) / n, R 1 、…R n represent the channel values of all non-zero red pixels;

[0143] G mean =(G1 +G 2 +G 3 +...+G n ) / n, G 1 …G n represents the channel values of all non - zero green pixels;

[0144] B mean＝ (B 1 +B 2 +B 3 +...+B n ) / n, B 1 …B n represents the channel values of all non - zero blue pixels.

[0145] When there are obvious color partitions in the obtained target flame area, obtain all pixel values within the target flame area, use an unsupervised learning algorithm to learn and classify all pixel values, obtain a self - classification result, extract the cluster with the most data points after classification as the target cluster, and output and save its pixel average value.

[0146] S306, construct a first data set based on a series of the historical hydrogen chloride concentrations and the three - channel pixel averages of the hydrogen chloride flame pictures at the corresponding times; construct a second data set based on a series of historical hydrogen - chlorine states and the three - channel pixel averages of the hydrogen chloride flame pictures at the corresponding times;

[0147] S307, perform offline modeling: including random forest regression models, random forest classification models, logistic regression models, etc.

[0148] S308, divide the first data set into a first training set and a first test set. The first training set and the first test set independently include a series of the historical hydrogen chloride concentrations and the three - channel pixel averages of the hydrogen chloride flame pictures at the corresponding times. When dividing the first training set and the first test set, it is preferably divided using the train_test_split function, with random_state being 42 and test_size being 0.2;

[0149] Randomly draw multiple subsamples with replacement from the first training set, each subsample having the same size as the first training set, for constructing different decision trees. The node splitting criterion for each decision tree is to minimize the mean squared error;

[0150] Take the historical hydrogen chloride concentration as the first target variable, the three - channel pixel averages of the hydrogen chloride flame pictures at the corresponding times as the first feature variables, input the first training set into the random forest regression model and train it, and use the first test set for testing to obtain a hydrogen chloride concentration prediction model as follows:

[0151]

[0152] Among them, y 1 is the predicted hydrogen chloride concentration, and A 1 represents the number of "trees" established, and f a (x) is the predicted hydrogen chloride concentration corresponding to the input data to be predicted by the a-th tree.

[0153] S309. Divide the second data set into a second training set and a second test set. The second training set and the second test set independently include a series of the historical chlorine-hydrogen states and the three-channel pixel averages of the hydrogen chloride flame pictures corresponding to their respective times;

[0154] Randomly draw multiple subsamples with replacement from the second training set. The size of each subsample is the same as that of the second training set and is used to construct different decision trees. The node splitting criterion for each decision tree is the Gini index or information gain;

[0155] Use the historical chlorine-hydrogen state as the second target variable, and the three-channel pixel average of the hydrogen chloride flame picture corresponding to its respective time as the second feature variable. Input the second training set into the random forest classification model for training and use the second test set for testing to obtain a chlorine-hydrogen state prediction model as follows:

[0156]

[0157] Among them, y 2 represents the predicted chlorine-hydrogen state, including the chlorine-hydrogen state being normal and the chlorine-hydrogen state being abnormal, where the abnormality includes over-chlorine and over-hydrogen; y 2 ∈{1, 2, 3}, 1 represents normal, 2 represents over-hydrogen, and 3 represents over-chlorine; x represents the input data to be predicted; {y (a) (x)} is the predicted chlorine-hydrogen state corresponding to the input data to be predicted by the a-th decision tree; A 2 represents the number of "trees" established, and mode represents taking the mode, that is, selecting the category with the most occurrences.

[0158] S310. Create a new feature column based on the three-channel pixel averages (original feature columns) of the hydrogen chloride flame pictures corresponding to the historical chlorine-hydrogen states, and perform standardization processing on it to obtain a standardized feature column. The standardization processing formula is as follows:

[0159]

[0160] Among them, T’ is the standardized feature column, T is the original feature column, is the mean of the original feature column, and O is the standard deviation of the original feature column;

[0161] Construct a third training set and a third test set based on the standardized feature columns and the historical chlorine-hydrogen status. The third training set and the third test set each independently include the standardized feature columns and the historical chlorine-hydrogen status. When constructing the third training set and the third test set, it is preferred to use the train_test_split function for partitioning, with random_state being 42 and test_size being 0.2;

[0162] Using the standardized feature columns as the third feature variables and the historical chlorine-hydrogen status as the third target variables, input the third training set into a logistic regression model and train it, and use the third test set for testing. The maximum number of iterations is 1000, and the regularization parameter C is 0.1, to obtain the second chlorine-hydrogen status prediction model as follows:

[0163] y 3 = arg max P(y=(1 or 2 or 3)|x; θ (1or2or3) )

[0164] where y 3 represents the predicted chlorine-hydrogen status, including the normal chlorine-hydrogen status and the abnormal chlorine-hydrogen status, where the abnormal includes over-chlorine and over-hydrogen; y 3 ∈{1, 2, 3}, 1 represents normal, 2 represents over-hydrogen, and 3 represents over-chlorine; θ is the parameter vector of the model, obtained through maximum likelihood estimation (MLE). The goal of MLE is to find a set of parameters that maximize the likelihood of the probability predicted by the model and the actual observed values; x represents the input data to be predicted; P(y=(1 or 2 or 3)|x; θ (1or2or3) ) represents the probability that the sample belongs to class 1 or 2 or 3 given the data x to be predicted and the model parameters θ.

[0165] Furthermore, the accuracy score function can also be used to calculate the accuracy of the model.

[0166] S311. Obtain the data to be predicted, which includes the average values of the three-channel pixels of the hydrochloric acid flame pictures at multiple prediction moments within a time period. The steps for obtaining the average values of the three-channel pixels of the hydrochloric acid flame pictures at the prediction moments include: video frame extraction to obtain images; obtaining the image coordinates of the target area; performing coordinate extraction on the images to obtain pixel data.

[0167] More specifically, the process of obtaining the three-channel pixel average values of the hydrogen chloride flame pictures at the multiple moments to be predicted includes: obtaining a hydrogen chloride flame video containing the moments to be predicted, performing frame extraction on the hydrogen chloride flame video containing the moments to be predicted to obtain a dataset of hydrogen chloride flame pictures containing the moments to be predicted, retrieving multiple hydrogen chloride flame pictures at the moments to be predicted from the dataset of hydrogen chloride flame pictures containing the moments to be predicted, and repeating S304 - 305 to process the multiple hydrogen chloride flame pictures at the moments to be predicted to obtain the three-channel pixel average values of the multiple hydrogen chloride flame pictures at the moments to be predicted.

[0168] S312, input the data to be predicted (i.e., Figure 3 the pixel data therein) into the hydrogen chloride concentration prediction model to obtain the predicted hydrogen chloride concentration, and output it to Data Table 1. The prediction accuracy of this model is as high as 99%.

[0169] S313, input the data to be predicted into the first chlorine-hydrogen state prediction model and the second chlorine-hydrogen state prediction model, and the two respectively obtain the predicted chlorine-hydrogen states and output them to Data Table 1. The prediction accuracies of the two models are respectively independently as high as 99%.

[0170] For example, for one of the obtained flame pictures, according to the operation of S304, the target flame area of the quadrilateral area obtained is as Figure 4 shown, and its specific coordinates are selected as (451, 3), (628, 1), (549, 30), (393, 65). Figure 4 The area in Figure 5 is processed by S305 and then by S312 - S313, and the output result is output to Data Table 1, as mean shown, where "Filename" represents the name of the flame picture, the "Mean BGR Blue" column represents the output B mean value, the "Mean BGR Green" column represents the output G mean value, the "Mean BGR Red" column represents the output R

[0171] S314. Compare the prediction results of the first hydrogen chloride state prediction model and the second hydrogen chloride state prediction model through a comparison algorithm function. If the prediction results of the two are consistent, output the prediction result to a preset data table as the predicted hydrogen chloride state. If the prediction results of the two are inconsistent and one of the prediction results is normal while the other is abnormal, output the corresponding abnormal prediction result to a preset data table (not shown in the figure) as the predicted hydrogen chloride state. If the prediction results of the two are inconsistent and one of the prediction results is over-chlorine while the other is over-hydrogen, output "None" to the preset data table as the predicted hydrogen chloride state.

[0172] S315. Determine that the output prediction results (i.e., the results of the preset data table) are abnormal for multiple consecutive times (such as 3 times), and output the results to the display screen. As Figure 6 shown, the abnormal results of the color recognition of the flame and the corresponding predicted hydrogen chloride concentration are displayed in real time to remind the operator to make adjustments.

[0173] S316. Determine that the output prediction results (i.e., the results of the preset data table) are abnormal for multiple consecutive times (such as 3 times), and determine the ratio of chlorine gas to hydrogen gas according to the predicted hydrogen chloride concentration. The specific calculation formula is as follows:

[0174]

[0175] where, represents the predicted hydrogen chloride concentration. When the predicted hydrogen chloride state is over-chlorine, v 过量 represents the volume of chlorine gas. When the predicted hydrogen chloride state is over-hydrogen, v 过量 represents the volume of hydrogen gas.

[0176] For example, during the operation of a certain synthesis furnace, the flame color shows obvious yellow-green. After model prediction, the predicted result of its HCl purity is 82% VOL, and it is predicted to be in an over-chlorine state. The HCl purity data is automatically input into the above formula to obtain v 过量 ÷v 未过量 = 1 / 0.82 = 1.22. Since the predicted state is over-chlorine, v 过量 is automatically matched to v 氯气 , then v chlorine gas) / v hydrogen gas = 1.22 is automatically output.

[0177] S317. Adjust the flow rate of chlorine gas or hydrogen gas based on the ratio of chlorine gas to hydrogen gas to keep it within a preset range.

[0178] In this example, historical data can also be saved to a database and a historical data query function is provided. The operator can trace back the past changes in the flame color for process analysis and fault diagnosis.

[0179] It should be understood that for the foregoing method embodiments, although the steps in the flowcharts are shown sequentially according to the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the method embodiment may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0180] Based on the same inventive concept, an embodiment of the present application also provides a hydrogen chloride synthesis process control device for implementing the hydrogen chloride synthesis process control method involved above. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the hydrogen chloride synthesis process control device provided below can refer to the limitations on the hydrogen chloride synthesis process control method in the above text, and will not be repeated here.

[0181] In one embodiment, as Figure 7 shown, a hydrogen chloride synthesis process control device is provided, and the device includes:

[0182] A data acquisition module 701, configured to acquire data to be predicted, where the data to be predicted includes the average value of three-channel pixels of hydrogen chloride flame pictures at multiple prediction moments within a time period;

[0183] A first prediction module 702, configured to input the data to be predicted into a hydrogen chloride concentration prediction model to obtain a predicted hydrogen chloride concentration;

[0184] A second prediction module 703, configured to input the data to be predicted into a chlorine-hydrogen state prediction model to obtain a predicted chlorine-hydrogen state;

[0185] A calculation module 704, configured to determine that the predicted chlorine-hydrogen state is abnormal for multiple consecutive times, and determine the ratio of chlorine to hydrogen according to the predicted hydrogen chloride concentration;

[0186] An adjustment module 705, configured to adjust the flow rate of chlorine or hydrogen based on the ratio of chlorine to hydrogen so that it is within a preset range.

[0187] In one embodiment, the hydrogen chloride synthesis process control device further includes: a hydrogen chloride concentration prediction model construction module for constructing a hydrogen chloride concentration prediction model, specifically for: obtaining a first data set, which includes a series of historical hydrogen chloride concentrations and the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the moments of the historical hydrogen chloride concentrations; constructing a first training set based on the first data set; randomly extracting multiple sub-samples with replacement from the first training set, each sub-sample having the same size as the first training set, for constructing different decision trees, and the node splitting criterion for each decision tree is to minimize the mean square error; inputting the first training set into a random forest regression model and training it to obtain the hydrogen chloride concentration prediction model.

[0188] In one embodiment, the above hydrogen chloride concentration prediction model construction module is further used for: obtaining the series of historical hydrogen chloride concentrations and the corresponding hydrogen chloride flame videos;

[0189] Performing frame splitting on the hydrogen chloride flame video to obtain a hydrogen chloride flame picture data set;

[0190] Obtaining the hydrogen chloride flame picture corresponding to the corresponding moment according to the sampling time of the historical hydrogen chloride concentration;

[0191] According to the hydrogen chloride flame picture at the corresponding moment, obtaining its target flame area, using a mask to extract the non-zero pixel coordinates within the target flame area, and respectively calculating to obtain the three-channel pixel average values R mean 、G mean and B mean .

[0192] In one embodiment, the hydrogen chloride concentration prediction model construction module is further used for: obtaining the hydrogen chloride flame picture corresponding to the corresponding moment according to the sampling time of the historical hydrogen chloride concentration, including:

[0193] The sampling time of the historical hydrogen chloride concentration is t HCl , and the acquisition time of the hydrogen chloride flame picture corresponding to this moment is t flame , according to the t flame Retrieving the hydrogen chloride flame picture corresponding to the corresponding moment from the hydrogen chloride flame picture data set,

[0194]

[0195] where H is the height from the flame in the synthesis furnace to the hydrogen chloride collection outlet, and v(h) is the instantaneous flow rate of the hydrogen chloride gas at height h.

[0196] In one embodiment, the hydrogen chloride synthesis process control device further includes: a chlorine-hydrogen state prediction model construction module for constructing a chlorine-hydrogen state prediction model, specifically: constructing a first chlorine-hydrogen state prediction model and a second chlorine-hydrogen state prediction model, and the prediction results of both include normal and abnormal; respectively inputting the data to be predicted into the first chlorine-hydrogen state prediction model and the second chlorine-hydrogen state prediction model to obtain corresponding prediction results; comparing the prediction results of the first chlorine-hydrogen state prediction model and the second chlorine-hydrogen state prediction model; if the prediction results of both are consistent, output the prediction result as the predicted chlorine-hydrogen state; if the prediction results of both are inconsistent and one prediction result is normal and the other is abnormal, then output the corresponding abnormal prediction result as the predicted chlorine-hydrogen state.

[0197] In one embodiment, the chlorine-hydrogen state prediction model construction module is further configured to: obtain a second data set, which includes a series of historical chlorine-hydrogen states and the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the historical chlorine-hydrogen states at corresponding times; construct a second training set based on the second data set; randomly draw multiple subsamples with replacement from the second training set, and the size of each subsample is the same as that of the second training set, for constructing different decision trees, and the node splitting criterion for each decision tree is the Gini index or information gain; input the second training set into a random forest classification model and train it to obtain the first chlorine-hydrogen state prediction model.

[0198] In one embodiment, the chlorine-hydrogen state prediction model construction module is further configured to: obtain a second data set, which includes a series of historical chlorine-hydrogen states and the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the historical chlorine-hydrogen states at corresponding times;

[0199] Create a new feature column based on the three-channel pixel average values of hydrogen chloride flame pictures corresponding to the historical chlorine-hydrogen states at corresponding times, and perform standardization processing on it to obtain a standardized feature column;

[0200] Construct a third training set based on the standardized feature column and the historical chlorine-hydrogen states;

[0201] Input the third training set into a logistic regression model and train it to obtain the second chlorine-hydrogen state prediction model.

[0202] For the specific limitations of the hydrogen chloride synthesis process control device, reference can be made to the limitations of the hydrogen chloride synthesis process control method in the above text, which will not be elaborated here. Each module in the above hydrogen chloride synthesis process control device can be implemented in whole or in part by software, hardware, and their combinations. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to each of the above modules.

[0203] In addition, in the implementation manner of the hydrogen chloride synthesis process control device in the above example, the logical division of each program module is only for illustration. In actual applications, according to needs, for example, considering the configuration requirements of the corresponding hardware or the convenience of software implementation, the above functions can be assigned to different program modules to complete, that is, the internal structure of the hydrogen chloride synthesis process control device is divided into different program modules to complete all or part of the functions described above.

[0204] In one embodiment, as Figure 8 shown, a hydrogen chloride synthesis process control system is provided. The system includes: a sight glass 802, a camera 803, a chlorine flow regulating valve, a hydrogen flow regulating valve, and a host computer;

[0205] The sight glass 802 is arranged on the synthesis furnace and is used to observe the hydrogen chloride flame in the synthesis furnace;

[0206] The camera 803 is used to obtain the hydrogen chloride flame video through the sight glass 802;

[0207] The chlorine flow regulating valve is arranged on the chlorine supply pipe of the synthesis furnace;

[0208] The hydrogen flow regulating valve is arranged on the hydrogen supply pipe of the synthesis furnace;

[0209] The host computer is used to process the obtained hydrogen chloride flame video to obtain the data to be predicted. The data to be predicted includes the three-channel pixel average values of the hydrogen chloride flame pictures at multiple predicted moments within a time period. The host computer determines the predicted hydrogen chloride concentration and the predicted chlorine-hydrogen state according to the data to be predicted. If it is determined that the predicted chlorine-hydrogen state is abnormal for multiple consecutive times, the ratio of chlorine to hydrogen is determined according to the predicted hydrogen chloride concentration, and an instruction is sent to the chlorine flow regulating valve or the hydrogen flow regulating valve based on the ratio of chlorine to hydrogen to adjust the flow rate of chlorine or hydrogen so that it is within the preset range.

[0210] In one implementation manner, the system further includes a light-shielding box 801, which is covered outside the sight glass 802 and the camera 803 to reduce the interference of external light.

[0211] The light-shielding box body 801 is made of fireproof materials and coated with light-shielding materials on the inner side to prevent external light from entering the synthesis furnace and affecting the flame color. The light-shielding box body 801 can be polygonal, such as Figure 8 as shown, it can be pentagonal. The pentagonal light-shielding box body includes three rectangular faces 801a, 801b, and 801c and two triangular faces 801d. One of the rectangular faces 801a is installed vertically, and one end of each of the other two rectangular faces 801b and 801c is connected to the upper end and the lower end of the rectangular face 801a respectively, and the other ends of the two are perpendicular to each other and connected. Plane mirrors are installed on the inner side walls of the two rectangular faces 801b and 801c. The two triangular faces 801d seal the two side faces of the pentagon respectively. A sight glass opening is provided on the vertically installed rectangular face 801a for installing the sight glass 802 with a diameter of L meters. The distance from the center of the sight glass to the bottom edge of the rectangular face 801a is H 1 meters, and the lengths of the rectangular faces 801b and 801c are both H 2 meters.

[0212] According to the principle of light reflection, at a position meters above the sight glass 802, a monitoring camera installation area with the same diameter of L meters is demarcated, and the camera 803 is installed through an opening according to the size of the camera. The camera 803 and the sight glass 802 are located on the same side. The flame in the synthesis furnace is transmitted through the sight glass 802, and through the plane mirror arranged inside the light-shielding box body 801, using the principle of plane mirror reflection, it is reflected to the camera 803, so as to obtain a clear image, effectively avoiding external light from entering the synthesis furnace and reducing the interference with the observation of the flame color in the sight glass 802.

[0213] In addition, if the selected camera 803 has a small size, an artificial observation port 804 can be opened at a similar position at the same time. An observation mirror is installed in the artificial observation port 804, which is composed of two lenses glued together, with a convex lens on one side and a concave lens on the other side, similar to the structure of a door peephole. The area of this artificial observation port is small, further reducing the entry of external light. At the same time, with the help of the optical principle, the operator can clearly observe the scene inside the synthesis furnace through the artificial observation port 804 on the outside, realizing the combination of artificial observation and the camera, and improving the flexibility and accuracy of observation.

[0214] In addition, the light-shielding box body 801 can also be hexagonal. At this time, the camera 803 and the sight glass 802 can be installed on two opposite faces of the light-shielding box body 801 respectively, and they are arranged opposite to each other. The camera 803 can directly obtain the flame video through the sight glass 802.

[0215] Above, no matter which structure the light-shielding box body 801 adopts, it is necessary to ensure that all installation interfaces are sealed to avoid external light sources from affecting the flame color in the synthesis furnace.

[0216] The light-shielding box body provided in this embodiment can not only automatically monitor the flame through the camera, but also retain the condition of manual visual observation. In this way, it not only ensures the accuracy of the hydrogen chloride synthesis process method based on flame color recognition, but also retains the possibility of on-site manual observation.

[0217] In a specific embodiment, L can be 0.1m, and H 1 can be 0.1m, and H 2 can both be 0.43m.

[0218] In one embodiment, a host computer is provided. The host computer can be a mobile terminal or a server. The internal structure diagram of the host computer can be as Figure 9 shown. The host computer includes a processor, a memory, an input / output interface, a communication interface, and a display unit. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface and the display unit are connected to the system bus through the input / output interface. Among them, the processor of the host computer is used to provide computing and control capabilities. The memory of the host computer includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the host computer is used to exchange information between the processor and external devices. The communication interface of the host computer is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be implemented through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it realizes a hydrogen chloride synthesis process control method. The display unit of the host computer is used to form a visually visible picture, which can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen.

[0219] Those skilled in the art can understand that Figure 9 the structure shown in

[0220] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it realizes the steps in the above method embodiments.

[0221] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, it realizes the steps in the above method embodiments.

[0222] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memories can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0223] It should be noted that in the embodiments of the present application, some industry-existing solutions such as certain software, components, models, etc. may be mentioned. They should be regarded as exemplary. The purpose is only to illustrate the feasibility in the implementation of the technical solutions of the present application, but it does not mean that the applicant has already or necessarily used this solution. The methods described in the present application are all fully self-developed executable software algorithms. All software algorithms are implemented with the help of general high-level languages, such as C++ language, Python language, etc. The development environments such as Visual Studio Community Edition and Pycharm Community Edition are all free and publicly available software, and do not involve software licensing and other intellectual property issues.

[0224] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification. In the above embodiments, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0225] As used in the embodiments herein, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or (module) units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0226] As used herein, "a plurality of" means two or more. "And / or" describes the association relationship of associated objects and indicates that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after.

[0227] As used herein, "first / second" is merely to distinguish similar objects and does not represent a specific order for the objects. It can be understood that "first / second" can be interchanged in a specific order or sequence when permitted. It should be understood that the objects distinguished by "first / second" can be interchanged appropriately so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0228] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for controlling a hydrogen chloride synthesis process, characterized in that: The method comprises: Acquire data to be predicted, the data to be predicted comprising three-channel pixel averages of hydrogen chloride flame images at multiple moments to be predicted within a time period; Inputting the data to be predicted into a hydrogen chloride concentration prediction model to obtain a predicted hydrogen chloride concentration; Inputting the data to be predicted into a chlorine-hydrogen state prediction model to obtain a predicted chlorine-hydrogen state; Determining that the predicted hydrogen chloride state is abnormal for multiple consecutive times, and determining the ratio of chlorine to hydrogen based on the predicted hydrogen chloride concentration; The flow rate of chlorine or hydrogen is adjusted based on the ratio of chlorine to hydrogen to be within a preset range.

2. The method according to claim 1, characterized in that: The steps of constructing the hydrogen chloride concentration prediction model include: Constructing a first data set, the first data set comprising a series of historical hydrogen chloride concentrations and three-channel pixel averages of hydrogen chloride flame images at times corresponding to the historical hydrogen chloride concentrations; Constructing a first training set based on the first data set; Randomly extracting multiple subsamples with replacement from the first training set, each subsample having the same size as the first training set, for constructing different decision trees, wherein the node splitting criterion of each decision tree is to minimize the mean square error; Inputting the first training set into a random forest regression model and performing training to obtain the hydrogen chloride concentration prediction model; and / or, The chlorine-hydrogen state prediction model includes a first chlorine-hydrogen state prediction model and a second chlorine-hydrogen state prediction model, and the prediction results of both models include normal and abnormal. The step of inputting the data to be predicted into the chlorine-hydrogen state prediction model to obtain the predicted chlorine-hydrogen state includes: respectively inputting the data to be predicted into the first chlorine hydrogen state prediction model and the second chlorine hydrogen state prediction model to obtain corresponding prediction results; comparing the prediction result of the first chlorine-hydrogen state prediction model with the prediction result of the second chlorine-hydrogen state prediction model; If the prediction results of the two are consistent, the prediction result is output as the predicted chlorine-hydrogen state; If the two prediction results are inconsistent and one prediction result is normal and the other prediction result is abnormal, the corresponding abnormal prediction result is output as the predicted chlorine-hydrogen state.

3. The method according to claim 2, characterized in that The step of constructing the first data set comprises: Obtaining the series of historical hydrogen chloride concentrations and the corresponding hydrogen chloride flame videos; Performing frame processing on the hydrogen chloride flame video to obtain a hydrogen chloride flame image dataset; According to the sampling time of the historical hydrogen chloride concentration, a hydrogen chloride flame image at a corresponding time is obtained; According to the hydrogen chloride flame image at the corresponding time, the target flame area is obtained, and the non-zero pixel coordinates in the target flame area are extracted using a mask, and the three-channel pixel average values ​​R are calculated respectively. mean , G mean and B mean .

4. The method according to claim 3, characterized in that: The sampling time of the historical hydrogen chloride concentration is t HCl The acquisition time of the hydrogen chloride flame image corresponding to this moment is t flame ; The step of obtaining the hydrogen chloride flame image at the corresponding time according to the sampling time of the historical hydrogen chloride concentration comprises: According to the flame Retrieve the hydrogen chloride flame image at the corresponding time from the hydrogen chloride flame image data set, Wherein, H is the height from the flame in the synthesis furnace to the hydrogen chloride collection outlet, and v(h) is the instantaneous flow velocity of hydrogen chloride gas at the height h.

5. The method according to any one of claims 2 to 4, characterized in that: The steps of constructing the first chlorine hydrogen state prediction model include: Acquire a second data set, the second data set comprising a series of historical chlorine-hydrogen states and three-channel pixel averages of a hydrogen chloride flame image at a time corresponding to the historical chlorine-hydrogen states; Constructing a second training set based on the second data set; Randomly extracting multiple subsamples with replacement from the second training set, each subsample having the same size as the second training set, for constructing different decision trees, wherein a node splitting criterion of each decision tree is a Gini index or information gain; Inputting the second training set into a random forest classification model and performing training to obtain the first chlorine hydrogen state prediction model; And / or, the step of constructing the second chlorine hydrogen state prediction model comprises: Acquire a second data set, the second data set comprising a series of historical chlorine-hydrogen states and three-channel pixel averages of a hydrogen chloride flame image at a time corresponding to the historical chlorine-hydrogen states; Creating a new feature column based on the average values ​​of three-channel pixels of the hydrogen chloride flame image at the time corresponding to the historical hydrogen chloride state, and standardizing the new feature column to obtain a standardized feature column; Constructing a third training set based on the standardized feature column and the historical chlorine-hydrogen state; The third training set is input into a logistic regression model and trained to obtain the second hydrogen chlorine state prediction model.

6. The method according to any one of claims 1 to 5, characterized in that: The anomalies include excessive chlorine and excessive hydrogen; The step of determining the ratio of chlorine to hydrogen according to the predicted hydrogen chloride concentration comprises: The ratio of the volume of chlorine to the volume of hydrogen is calculated by the following formula: in, It represents the predicted concentration of hydrogen chloride. When the predicted hydrogen chloride state is overchlorinated, vexcess represents the volume of chlorine. When the predicted hydrogen chloride state is overhydrogenated, vexcess represents the volume of hydrogen.

7. A hydrogen chloride synthesis process control device, characterized in that: The device comprises: A data acquisition module, used to acquire data to be predicted, the data to be predicted including three-channel pixel averages of hydrogen chloride flame images at multiple moments to be predicted within a time period; A first prediction module, used for inputting the data to be predicted into a hydrogen chloride concentration prediction model to obtain a predicted hydrogen chloride concentration; A second prediction module is used to input the data to be predicted into a chlorine-hydrogen state prediction model to obtain a predicted chlorine-hydrogen state; a calculation module, for determining that the predicted chlorine-hydrogen state is abnormal for multiple consecutive times, and determining the ratio of chlorine to hydrogen according to the predicted hydrogen chloride concentration; The regulating module is used to regulate the flow rate of chlorine or hydrogen based on the ratio of chlorine to hydrogen so as to make it within a preset range.

8. A hydrogen chloride synthesis process control system, characterized in that: The system comprises: a sight glass, a camera, a chlorine flow regulating valve, a hydrogen flow regulating valve and a host computer; A sight glass, arranged on the synthesis furnace, for observing the hydrogen chloride flame in the synthesis furnace; A camera, used to obtain a video of the hydrogen chloride flame through the viewing mirror; A chlorine gas flow regulating valve is arranged on the chlorine gas supply pipe of the synthesis furnace; A hydrogen flow regulating valve is arranged on the hydrogen supply pipe of the synthesis furnace; The host computer is used to process the acquired hydrogen chloride flame video to obtain data to be predicted, wherein the data to be predicted includes three-channel pixel averages of hydrogen chloride flame images at multiple moments to be predicted within a time period, and the host computer determines a predicted hydrogen chloride concentration and a predicted hydrogen chloride state according to the data to be predicted, determines that the predicted hydrogen chloride state is abnormal for multiple consecutive times, determines the ratio of chlorine to hydrogen according to the predicted hydrogen chloride concentration, and sends an instruction to the chlorine flow regulating valve or the hydrogen flow regulating valve based on the ratio of chlorine to hydrogen to adjust the flow of chlorine or hydrogen to make it within a preset range.

9. The system according to claim 8, characterized in that The system also includes a light shielding box, which is arranged outside the view mirror and the camera to reduce interference from external light.

10. A host computer, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images