Graphite waste acid gas treatment method and system based on artificial intelligence

Abstract

The invention belongs to the technical field of waste gas treatment, and discloses a graphite waste acid gas treatment method and system based on artificial intelligence, and the treatment method comprises the following steps: identifying main and auxiliary path gas sections, comparing the synchronization range of acid mist propulsion and spraying starting points, and measuring the deviation between the liquid drop track and the acid mist direction. Adjusting the advance response of the auxiliary path and the delay rhythm of the main path, tracking a boundary component trajectory, and connecting the moving direction and the response sequence of the acid mist in series to obtain an acid mist treatment response control sequence; according to the invention, by establishing a time sequence corresponding relation between an acid mist propelling section and a spraying response, synchronous control of trigger nodes in a channel, formation of an alignment mechanism by a moving path and a response rhythm, advanced response of an auxiliary path and delayed triggering of a main path are carried out, a peak shifting coordination structure is formed, and a component residual region is brought into a response interval in boundary expansion; the response chain is reconstructed according to the triggering sequence, the gas-liquid action process has the dynamic matching capacity, and the path collaboration is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of waste gas treatment. Specifically, it relates to a method and system for treating graphite waste acid gas based on artificial intelligence. Background Art

[0002] The technical field of waste gas treatment includes the collection, separation, purification, and reuse of the gases emitted during industrial production and energy conversion processes. The core content includes the control and treatment of gaseous pollutants such as sulfur oxides, nitrogen oxides, acidic gases, volatile organic compounds, and particulate matter. According to the characteristics, emission concentration, and emission volume of the pollutants, this technical field uses methods such as absorption, adsorption, catalytic oxidation, combustion, and biological treatment to treat gas pollutants, and is widely applied in industries such as chemical engineering, metallurgy, materials, and electric power. The waste gas treatment technology systematically focuses on pollutant reduction, resource recovery, and emission compliance. The technical path involves pollution source monitoring, purification device design, reaction process control, tail gas emission monitoring, and automatic regulation, emphasizing process continuity and system collaborative control.

[0003] The Chinese invention patent with the patent application number: CN202311746935.X discloses an accurate control system for waste gas treatment in a chemical industrial park, including: an on-line waste gas monitoring module, a PLC intelligent control module, and a waste gas purification and treatment module. The on-line waste gas monitoring module includes an on-line monitoring unit, a signal processing unit, and a signal transmission unit. The PLC intelligent control module includes a data storage unit and a processing and output unit. The waste gas purification and treatment module includes a first purification unit and a second purification unit arranged in sequence along the waste gas purification flow path. The on-line waste gas monitoring module is used to monitor the concentration values of the target components that pollute the atmosphere in the waste gas, and the concentration values are transmitted to the PLC intelligent control module, and the PLC intelligent control module controls the waste gas purification and treatment module to perform hierarchical purification operations.

[0004] The above-mentioned existing control systems treat waste gas through technological means such as identifying, classifying, analyzing the components, and optimizing the parameters of the waste gas, but they are not suitable for treating the acid mist gas generated during the treatment of graphite materials with strong acids in the graphite production process. Because this type of waste gas usually contains multiple pollution components such as sulfuric acid vapor, hydrogen fluoride gas, and graphite microparticles, the component content in the waste gas is complex and it is not easy to perform automated detection and control.

[0005] And the existing intelligent processing method for graphite waste acid gas is as follows: obtaining a gas sample through a gas collection system, collecting multiple data on the gas composition using spectroscopic analysis and sensor technology, inputting the obtained multi-dimensional data into a composition recognition model and a regulation model trained based on deep learning, and setting the optimal spray liquid ratio and the processing flow sequence based on the model judgment results; the technical means adopted include spectroscopic detection, multi-channel sensor data fusion, a convolutional neural network composition classification model, a parameter regulation system based on gradient descent, and a spray reaction system scheduling algorithm, constituting a complete intelligent gas processing solution.

[0006] In the existing technology processing structure, a continuous process interaction logic has not been established for the relationship between acid mist and spray response. The response control basis usually comes from the component concentration value or the information collected in a single time period. The relationship between the advancing direction of the acid mist in the path and the spray response displacement has not been used as a rhythm adjustment reference, resulting in a fuzzy arrangement of the passage beats in the path control process; in the process of acid component treatment, the changes in the uncontacted areas at the airflow boundary have not been fully extracted, ignoring the non-functional reaction section after the acidic components remain at the edge of the droplets, and the reaction efficiency is limited by the judgment of the uniformity of the coverage structure; in a multi-path scenario, the main and secondary path control actions are often driven by fixed parameters, and a complementary linkage mechanism has not been formed, resulting in one party's response being advanced while the other party idles, weak path resource allocation ability, unbalanced timing execution sequence, easy misalignment of the response synchronization window, interruption of the gas-liquid action chain, and in severe cases, incomplete acid mist reaction or droplet waste, and the treatment area cannot maintain a closed-loop reaction process under a stable rhythm. Summary of the Invention

[0007] In order to solve the technical problems existing in the prior art, the present invention provides a method and a system for treating graphite waste acid gas based on artificial intelligence, enabling the gas-liquid action process to have dynamic matching ability, improving path coordination, and realizing in-depth treatment of waste acid gas.

[0008] To solve the above technical problems, the present invention provides the following technical solutions: A method for treating graphite waste acid gas based on artificial intelligence, comprising the following steps: S1: Identifying the gas movement segments of the main and secondary paths in the graphite acid mist passage, comparing the synchronous movement ranges of the acid mist advancing area and the atomizing spray starting point, analyzing the staggered lengths that appear synchronously between the two segments, and processing the response matching relationship between the main and secondary paths within the same time window to obtain a channel linkage structure type label; S2: Based on the channel linkage structure type label, measuring the angular deviation between the running trajectory of the droplet front edge and the forward movement route of the acid mist, and matching the flow overlap segment formed by the spray behavior and the acid mist direction in the secondary path to obtain a path contact feature offset identifier; S3: Based on the path contact feature offset identifier, compare the secondary path spray trigger rhythm with the acid mist arrival period, check the overlapping relationship between the front and back sequences, advance the start of the secondary path spray response before the acid mist movement action, and delay the main path spray to the next time beat position to obtain the secondary path early response scheduling record; S4: Based on the secondary path early response scheduling record, extend the analysis of the path changes of hydrofluoric acid and sulfuric acid in the non-fully covered section of the spray area, track the edge of the reaction zone where no droplet coverage is formed in two consecutive cycles, and locate the diffusion area where the acid mist does not enter the reaction state to obtain the component trigger response boundary section; S5: Based on the component trigger response boundary section, sort out the movement direction of the acid mist on the path section in the diffusion area where it has not entered the reaction state and the sequence of actions of the spray response points, and connect the start rhythms between the gas-liquid reaction nodes to obtain the acid mist treatment response control sequence.

[0009] The following is a further optimization of the above technical solution by the present invention: The channel linkage structure type labels include the main path gas movement section, the secondary path gas movement section, the staggered length, and the response matching relationship; the path contact feature offset identifiers include the angle deviation, the cross-sectional area distribution, and the flow direction overlapping section; the secondary path early response scheduling record includes the spray trigger rhythm, the acid mist arrival period, the start advancement position, and the time beat adjustment; the component trigger response boundary section includes the path change area, the reaction zone edge, and the diffusion area; the acid mist treatment response control sequence includes the movement direction, the response point order, the start rhythm, and the path relationship.

[0010] Further optimization: The specific steps of S1 are as follows: S101: Obtain the acid mist advancement section information in the main path and the secondary path in the graphite acid mist path, collect the starting position, flow range, and direction annotation of the acid mist advancement in the differential paths, and compare the time coverage difference and position offset of the acid mist movement starting area in the two path sections to obtain the path advancement starting position difference value; S102: Based on the path advancement starting position difference value, locate the atomizing spray starting positions of the main and secondary paths, extract the spatial contact area between the spray response surface and the acid mist front, and screen the spray response point sequence that coincides with the acid mist movement direction to obtain the path response synchronous coverage segment value; S103: Based on the path response synchronous coverage segment value, align the acid mist movement section and the spray response point sequence of the main and secondary paths within the same time window, analyze whether there is a front-back offset or starting misalignment in the response rhythm, and extract the rhythm corresponding state to obtain the channel linkage structure type label.

[0011] Further optimization: The specific steps of S2 are as follows: S201: Based on the channel linkage structure type tag, obtain the starting point and propulsion direction of the droplet front running trajectory, collect the moving direction line segment of the acid mist front within the same path segment, pair the droplet propulsion direction with the acid mist movement route, analyze the angular deviation trend of the two direction lines, and obtain the direction angle deviation value; S202: Based on the direction angle deviation value, extract the cross-position section of the spray streamline during the acid mist propulsion process, sort out the distribution density of the cross area within the continuous path length, calculate the ratio of the number of spray points to the number of acid mist propulsion points within the cross section, and obtain the coverage interval of the spray overlap section; S203: Based on the coverage interval of the spray overlap section, identify the path fitting area of the spray behavior in the acid mist movement direction in the sub-path, extract the serial number distribution of the continuous fitting section and the overlapping length of the spray response range, and construct a path contact state index according to the fitting rate and deviation value to obtain the path contact feature deviation identifier.

[0012] Further optimization: The specific calculation formula for the ratio of the number of spray points to the number of acid mist propulsion points within the cross section is as follows: ; Wherein, represents the ratio of the number of spray points to the number of acid mist propulsion points within the cross section, represents the number of spray points within the cross section, represents the number of acid mist propulsion points within the cross section, represents the number of the th spray point within the cross section, represents the number of the acid mist propulsion point corresponding to the th spray point at the corresponding position on the path, represents the maximum value of the acid mist propulsion point numbers within the cross section, represents the minimum value of the acid mist propulsion point numbers within the cross section, represents the average value of the span of each acid mist propulsion point number within the cross section.

[0013] Further optimization: The specific steps of S3 are as follows: S301: Based on the path contact feature deviation identifier, collect the start time point and duration period in the spray response rhythm of the sub-path, obtain the time starting point and staying section where the acid mist arrives within the same path segment, compare their starting positions on the same time line, and obtain the difference between the response trigger and the air flow arrival order; S302: Based on the difference between the response trigger and the air flow arrival order, calculate the advance period value that the spray response trigger time in the sub-path is earlier than the acid mist propulsion starting point, align the advance period value with the path set beat sequence, analyze the position of the advance amount in the time structure, and obtain the sub-path response propulsion section number; S303: Advance the section number based on the sub-path response, delay the main-path spray response time to the next-level trigger section of the current beat sequence, adjust the position and time period of the main-path response start point, and identify the time overlap area between the front and rear responses to obtain the sub-path early response scheduling record.

[0014] Further optimization: The specific calculation formula for the advance time period when the spray response trigger time in the sub-path is earlier than the acid mist advancement start point is as follows: ; Wherein, represents the advance time period value when the spray response trigger time in the sub-path is earlier than the acid mist advancement start point, represents the spray response trigger time of the path segment , represents the acid mist advancement start point time corresponding to the path segment , represents the number of adjacent path segments of the path segment , represents the path segment number index from 1 to , represents the spatial length of the path segment , represents the beat sequence number corresponding to the path segment , represents the spray response trigger time in the path segment , represents the acid mist advancement start point time corresponding to the path segment .

[0015] Further optimization: The specific steps of S4 are as follows: S401: Based on the sub-path early response scheduling record, collect the path extension data of hydrofluoric acid and sulfuric acid in the non-covered section of the spray area, monitor the moving direction and continuous section of the two components at the boundary section, and classify the position change trend of the advancement line at the path edge to obtain the acid component boundary extension trajectory value; S402: Based on the acid component boundary extension trajectory value, compare the response areas where no droplet coverage is formed in the same path section in the current and the previous cycle, screen out the position segment sequence that is not covered in two consecutive cycles, and extract the continuously non-covered section to obtain the periodic spray fault section index; S403: Based on the periodic spray fault section index, extract the distribution of the residual components of hydrofluoric acid and sulfuric acid in the edge area of the spray response surface, process the spatial offset relationship between the component movement direction at the boundary and the droplet contact blank section, and identify the path segment area where the acid mist does not form a reaction to obtain the component trigger response boundary section.

[0016] Further optimization: The specific steps of S5 are as follows: S501: Based on the component-triggered response boundary section, collect the moving direction, advancing order, and flow velocity stable section of the acid mist within the corresponding path segment, sort out the starting position of the acid mist in the path and the advancing time sequence of the front section on the time axis, monitor the coincidence distribution of the acid mist movement behavior and the response section, and obtain the acid mist moving direction sorting set; S502: Based on the acid mist moving direction sorting set, obtain the start time of the spray response point and the corresponding node number, compare the advancing rhythm of the acid mist within the path segment with the triggering period of the spray response point, mark the sequence of the reaction nodes, and obtain the response node time sequence structure table; S503: Based on the response node time sequence structure table, connect the sorted gas-liquid reaction nodes in the path segment in the order of the start rhythm, extract the path segments and time series of the complete response chain, and combine and associate the path structures to obtain the acid mist treatment response control sequence.

[0017] The present invention also provides a graphite waste acid gas treatment system based on artificial intelligence. The system is used to implement the above-mentioned graphite waste acid gas treatment method based on artificial intelligence. The system includes: The path recognition module obtains the gas advancing section positions of the main path and the secondary path in the graphite acid mist pipeline, compares the spatial coincidence area between the acid mist advancing area and the spray starting point, analyzes the synchronous relationship between the staggered length and the coincidence time period, and obtains the channel linkage structure type label; The contact recognition module, based on the channel linkage structure type label, measures the angle between the leading edge running track of the liquid droplet and the moving direction of the acid mist, records the continuous distribution section of the atomized spray streamline and the acid mist front intersection area, and matches the flow direction consistent section formed in the secondary channel to obtain the path contact feature offset identifier; The response scheduling module, based on the path contact feature offset identifier, compares the start time sequence of the spray rhythm of the secondary path with the acid mist advancing time period, checks whether there is a rhythm coincidence section in the front and back order, advances the spray response of the secondary path to the response node before the acid mist advances, and delays the spray rhythm of the main path to the next-level beat node to obtain the secondary path early response scheduling record; The boundary recognition module, based on the secondary path early response scheduling record, analyzes the path advancement changes of hydrofluoric acid and sulfuric acid outside the spray coverage, tracks the reaction edge path where no droplet coverage is formed in two consecutive cycles, and identifies the spatial area where the acid mist does not enter the neutralization reaction to obtain the component-triggered response boundary section; The path control module, based on the component-triggered response boundary section, sorts out the time sequence between the advancing direction of the acid mist in the path segment and the atomized response point, connects the start beats of all reaction nodes in series, and uses it as the rhythm main line of the path behavior output control to obtain the acid mist treatment response control sequence.

[0018] The present invention adopts the above technical solutions and has at least the following beneficial effects: In the present invention, by establishing a timing correspondence relationship between the acid mist propulsion section and the spray response, synchronous control of the trigger nodes in the path is carried out. The movement path and the response rhythm form a counterpoint mechanism. The secondary path responds in advance, and the main path is triggered later, constituting a peak-shifting coordination structure. The component residue area is incorporated into the response interval during the boundary expansion. The response chain is reconstructed according to the trigger order, and the path control forms a closed-loop structure driven by rhythm. The gas-liquid interaction process has the ability of dynamic matching, improving the path coordination. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic flow chart of the steps of the processing method in the embodiment of the present invention; Figure 2 It is a module diagram of the system in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] In the embodiments of the present invention, words such as "example" and "for example" are used to represent examples, illustrations or explanations; any embodiment or design solution described as an "example" in the present invention should not be construed as more preferred or more advantageous than other embodiments or design solutions; rather, the use of the word "example" is intended to present concepts in a specific manner; in addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two can be selected.

[0021] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meanings they express are the same; "(of)", "corresponding", and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meanings they express are the same.

[0022] In the embodiments of the present invention, sometimes subscripts such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meanings they express are the same.

[0023] To make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the drawings and specific embodiments.

[0024] Please refer to Figure 1 , a processing method for graphite waste acid gas based on artificial intelligence, comprising the following steps: S1: Identify the gas movement sections of the main and secondary paths in the graphite acid mist path, compare the synchronous movement ranges of the acid mist propulsion area and the atomizing spray starting point, analyze the staggered length that appears synchronously between the two sections, process the response matching relationship between the main and secondary paths within the same time window, and obtain the channel linkage structure type label; S2: Based on the channel linkage structure type label, measure the angular deviation between the running track of the droplet front edge and the forward movement route of the acid mist, record the continuous distribution process of the intersection area between the spray streamline and the acid mist front, match the flow overlap section formed by the spray behavior and the acid mist direction in the secondary path, and obtain the path contact feature offset identifier; S3: Based on the path contact feature offset identifier, compare the spray trigger rhythm of the secondary path with the arrival time period of the acid mist, check the overlap relationship between the front and back sequences, advance the spray response start of the secondary path before the acid mist movement action, and delay the main path spray to the next time beat position to obtain the secondary path early response scheduling record; S4: Based on the secondary path early response scheduling record, extend the analysis of the path changes of hydrofluoric acid and sulfuric acid in the section not fully covered in the spray area, track the edge of the reaction zone where no droplet coverage is formed within two consecutive cycles, incorporate the component spreading behavior that spreads along the outer edge of the spray into the analysis process, and locate the diffusion area where the acid mist does not enter the reaction state to obtain the component trigger response boundary section; S5: Based on the component trigger response boundary section, sort out the movement direction of the acid mist on the path section in the diffusion area where it has not entered the reaction state and the sequence of actions of the spray response points, connect the start rhythms between the gas-liquid reaction nodes, and use the path relationship of the formed response sequence as the source of the execution sequence output to obtain the acid mist treatment response control sequence.

[0025] The channel linkage structure type label includes the main path gas movement section, the secondary path gas movement section, the staggered length, and the response matching relationship; the path contact feature offset identifier includes the angular deviation, the cross-sectional area distribution, and the flow overlap section; the secondary path early response scheduling record includes the spray trigger rhythm, the acid mist arrival time period, the start advancement position, and the time beat adjustment; the component trigger response boundary section includes the path change area, the edge of the reaction zone, and the diffusion area; the acid mist treatment response control sequence includes the movement direction, the response point order, the start rhythm, and the path relationship.

[0026] In this embodiment, the specific steps of S1 are as follows: S101: Obtain the acid mist advancement section information in the main path and the secondary path in the graphite acid mist path, collect the starting position, flow range, and direction annotation of the acid mist advancement in the differential paths, and compare the time coverage difference and position offset of the acid mist movement starting area in the two path sections to obtain the path advancement starting position difference value.

[0027] First, collect the channel sections on the graphite processing production line with typical acid mist emission characteristics. In the actual device, the main path through which the acid mist flows can be the exhaust pipe for heat treatment tail gas, and the secondary path may be the auxiliary channel for exhaust balance. Record the physical coordinates of the acid mist propulsion starting points on these two types of paths and the triggering times of the pressure sensors within the corresponding time periods. By collecting the flow direction labels of each channel section, present the propulsion direction in the form of a line segment vector. Taking the main path direction as the positive x-axis direction and the secondary path with an upward deviation of 20° from the intersection angle with the main path as an example, mark the flow direction differences under different working conditions. Subsequently, identify the propulsion lengths and the coordinate differences of the propulsion starting points per unit time for each path section, with the collection range controlled within the first 30% area of the total length of the reaction section. Further record the air pressure readings, temperature values, and acid concentration data corresponding to the propulsion starting points. By marking the initial pressure value Pa1 at the beginning of the propulsion section and the initial value Pb1 of the secondary path, if the difference between the two values is greater than 30 Pa, it is determined that the propulsion starting points of the two paths are out of sync and enter the position offset state. Extract five consecutive groups of propulsion starting point coordinate values according to the recorded time interval Δt = 2 s. By judging the coincidence rate of the propulsion starting point coordinate points of the main and secondary paths in the spatial projection, a judgment reference value of less than 20% can be regarded as having a position offset. If the coincidence rate of three or more measurement points is lower than this reference, confirm the existence of the offset state. Then, extract the direction vectors of each propulsion section, and combine the size of the included angle θ between the two channel directions to judge whether the propulsion direction difference is in a deviated state. If the included angle is greater than 25°, it is processed according to the offset path. By combining three types of data: the spatial coordinate difference of the propulsion starting points, the included angle of the propulsion direction vectors, and the air pressure difference, judge the differences between the starting points of the main and secondary paths, and extract the synchronous distribution of the position, direction, and propulsion state of the offset section. Finally, obtain the position difference value of the path propulsion starting point.

[0028] S102: Based on the position difference value of the path propulsion starting point, locate the atomizing spray starting positions of the main and secondary paths, extract the spatial contact area between the spray response surface and the acid mist front, and screen the spray response point sequence that coincides with the acid mist movement direction to obtain the path response synchronous coverage segment value.

[0029] In the acid drainage section of graphite processing, confirm the physical layout positions of the main path and auxiliary path spray devices, collect the coordinate values and device numbers of the atomization starting points of each spray trigger device in the channel, arrange the spray trigger positions in the flow direction order, record the trigger timestamps corresponding to each group of nozzles, and align them with the starting time period of the acid mist front advancement. Judge whether there is a situation where the atomization trigger time is earlier than the acid mist front advancement time. If the forward shift time is within 1 second, it is recognized as the initial synchronization point. Subsequently, extract the overlapping area between the spray response surface and the acid mist advancement area within the spatial range, obtain the two-dimensional plane contact section by corresponding coordinates, delimit the boundary of the spray coverage range, and perform area overlapping processing with the projection boundary of the acid mist flow area to form a contact area projection map. If the overlap rate exceeds the set benchmark of 0.6, it is determined as an effective contact area. The basis for setting this benchmark is that the industrial conventional acid mist diffusion radius is within 250 mm and the spray atomization radius is 300 mm, and the coincidence degree of the average action boundaries of the two is not less than 60% for effective coverage. According to this determination criterion, screen the spray response points consistent with the acid mist advancement direction, extract the arrangement order of these nozzles in the advancement direction, record the start and stop times of each point within the acid mist front contact section, and form a spray response trigger sequence. Use the point set with an included angle less than 15° between the response points and the acid mist advancement section direction in this sequence as the direction coincidence point set, classify and count the spatial intervals between each nozzle in this point set and the acid mist advancement path section. The point column with a spacing lower than 120 mm is defined as the synchronization point column. Finally, calibrate all the spray response point columns that meet the conditions as the response synchronization paragraph to obtain the path response synchronization coverage segment value.

[0030] S103: Based on the path response synchronization coverage segment value, align the acid mist movement section and the spray response point sequence of the main and auxiliary paths within the same time window, analyze whether there is a forward or backward shift or starting dislocation in the response rhythm, extract the corresponding state of the rhythm, and obtain the channel linkage structure type label.

[0031] Obtain the acid mist movement segments and spray response point sequences corresponding to the main path and the secondary path within the same time window. Extract the starting time of the spray response and the time annotation of the acid mist front arrival point in each path. Record the time nodes of all response trigger actions on the time axis. Associate the time of each response point with the path segment number it is in. Compare whether the trigger order in the same position segment of the main and secondary paths is staggered. If the response time of a point in the secondary path is earlier than that of the same position segment in the main path by more than 0.8 seconds, it is recorded as a forward response; otherwise, it is a delayed response. Construct a time series difference comparison table for all points with response rhythm differences, and calculate the rhythm offset degree between the main and secondary paths based on the number of rhythm difference counts for different paths. Set the offset degree reference value to 10%. When the proportion of offset quantities is greater than this reference, it is marked as a rhythm inconsistent state. Subsequently, analyze the starting misaligned section in the response rhythm, extract the starting offset range of the first trigger point on the time axis. Set the starting misalignment judgment threshold to 1.5 seconds. If the time difference between the first response points of the two paths is greater than this value, mark the starting misalignment state. In the example, if the time of the first response point in the main path is 12.6 seconds and the time of the first response point in the secondary path is 10.7 seconds, the difference is 1.9 seconds, which exceeds the threshold and is classified into the misaligned section. Combine and organize the misaligned segment numbers and rhythm offset segment numbers to establish a path response rhythm matching relationship table, and mark the four states of synchronous response, forward response, delayed response, and misaligned response respectively to identify the associated pattern of rhythm distribution in the passage. Finally, output the association matrix between each rhythm state and the path segment in the structure as the matching result to obtain the channel linkage structure type label.

[0032] In this embodiment, the specific steps of S2 are as follows: S201: Based on the channel linkage structure type label, obtain the starting point and advancing direction of the droplet front running trajectory, collect the moving direction line segments of the acid mist front within the same path segment, pair the droplet advancing direction and the acid mist moving route, and analyze the angular offset trend of the two direction lines to obtain the direction angle offset value.

[0033] Select a typical channel segment as the analysis object, collect the starting point coordinates of the droplet movement trajectory generated by the spray, record the initial trigger position corresponding to the center point of the nozzle on the channel cross-section, extend the coordinate point along the initial velocity direction of the droplet to form a propulsion vector line, combine the nozzle spray angle range to limit the maximum and minimum offset angles of the propulsion direction, then collect the two position coordinates of the front edge in the acid mist propulsion section at times t1 and t2, connect the two coordinate points to form a line segment in the acid mist movement direction, construct an acid mist direction vector, perform alignment processing on the droplet propulsion direction and the acid mist front direction line segment, extract the included angle θ between the two, and control it within the range of 0° to 90° for direction offset recognition. According to the principle that the included angle is less than 15° when the spray streamline is aligned with the positive direction of the acid mist in industrial operations, set the included angle judgment reference value θ0 as 15°. If the included angle θ is between 0° and θ0, it is calibrated as the direction approaching area. If θ exceeds θ0, it is marked as the direction offset area. For example, if the droplet propulsion direction is the vector (1, 0.2) and the acid mist propulsion direction is the vector (1, 0), according to the vector included angle formula, θ is approximately 11.3°, which is less than the threshold of 15°, and it is judged as being in the same direction and fitting. If multiple spray trigger points all show the characteristic that such offset values are less than θ0, they can be classified as the path of the same direction type. On the contrary, if the included angle value continuously remains above 25° on the sub-path and covers multiple path segment numbers, it indicates that there is a persistent direction deviation in the path. When judging the included angle, summarize the included angle data between each droplet propulsion line and the acid mist front line segment, count the number of paragraphs with an included angle greater than θ0. If the proportion of the statistical result exceeds the set offset threshold σ of 0.3, that is, the total number of paragraphs is n and the offset paragraphs are m, and when m / n > σ, mark that there is a structural offset trend in this path. Finally, combine the distribution state of the included angle offset value and the threshold judgment reference to output the direction consistency result of the path segment and obtain the included angle offset value.

[0034] S202: Based on the included angle offset value, extract the cross-position section of the spray streamline during the acid mist propulsion process, sort out the distribution density of the cross-region within the continuous path length, calculate the ratio of the number of spray points to the number of acid mist propulsion points in the cross-section to obtain the coverage interval of the spray overlap section.

[0035] The specific calculation formula for the ratio of the number of spray points to the number of acid mist propulsion points in the cross-section is: ; Among them, represents the ratio of the number of spray points to the number of acid mist propulsion points in the cross-section, represents the number of spray points in the cross-section, represents the number of acid mist propulsion points in the cross-section, represents the th number of the spray point in the cross-section, represents the The acid mist propulsion point number corresponding to the position of each spray point on the path Represents the maximum value of the acid mist propulsion point numbers within the intersection segment Represents the minimum value of the acid mist propulsion point numbers within the intersection segment Represents the average value of the span of the acid mist propulsion point numbers within the intersection segment

[0036] Assume: Obtained by identifying the spray point numbers through visual flow field tracking ; The number of marked air flow propulsion points within the intersection segment is ; The image processing system extracts and quantifies the spray array numbers to obtain , , , , ; Obtained by tracking based on the aerosol propulsion path numbers , , , , ; Obtained by extracting from the acid mist propulsion point sequence , ; ; Substitute various data for calculation: The numerator part: ; The denominator part: ; The final ratio: ; This result indicates that the ratio of the spray response within the intersection segment in the acid mist propulsion structure is 0.826. This value is used to judge the fitting density degree of the spray streamline and the acid mist propulsion direction, and is an important reference index for judging whether the spray path needs to respond in advance in the subsequent steps, and serves as the source of the quantization index required for generating the coverage interval of the spray overlap segment in the subsequent steps.

[0037] S203: Based on the coverage interval of the spray overlap segment, identify the path fitting area of the spray behavior in the acid mist moving direction in the sub-path, extract the serial number distribution of the continuous fitting segments and the overlapping length of the spray response range, and construct a path contact state index according to the fitting rate and the offset value to obtain the path contact feature offset identifier.

[0038] First, extract the starting positions and response directions of the nozzles at the starting points of each sub-path according to the spray trigger time series, record the nozzle numbers, response starting point coordinates and propulsion vectors, determine the spray propulsion path segments, and then combine the movement trajectories of the acid mist propulsion front to pair the propulsion directions within the same time window, and identify whether the angle between the spray direction and the main acid mist direction is less than the fitting judgment threshold θ1. Set θ1 to 12°, indicating that within this angle, the two directions can be considered the same. If the spray propulsion section of the sub-path satisfies the angle judgment condition in multiple consecutive path segments, then this continuous path segment is used as the spray fitting section; further, count the number sequence of such fitting sections, let the number sequence be from n1 to n6, calculate the intersection section of the spray coverage length and the corresponding acid mist propulsion coverage length for each number section, and define it as the coincidence length Lr; subsequently, compare the ratio of the coincidence length to the total spray propulsion length Lp, calculate the fitting rate η = Lr / Lp. If η is between 0.75 and 1, it is regarded as a good fitting interval. At the same time, refer to the previously generated direction angle offset value θavg, construct the path contact state index S, and define S = η×cos(θavg), which is used to comprehensively represent the response intensity of the sub-path in the fitting section; in the example, if the spray length of a fitting section is 320 mm and the coincidence section length is 256 mm, then η = 0.8, θavg = 10°, cos10° = 0.9848, then S≈0.788, and it is judged as a medium offset type; finally, summarize all the fitting section numbers and the S values of each section, form a path contact index set, sort and number map it, output the state characteristics of the structural contact, and obtain the path contact characteristic offset identifier.

[0039] In this embodiment, the specific steps of S3 are as follows: S301: Based on the path contact characteristic offset identifier, collect the starting time point and duration period in the spray response rhythm of the sub-path, obtain the time starting point and staying section when the acid mist arrives in the same path segment, compare their starting positions in the same time line, and obtain the difference between the response trigger and the air flow arrival order.

[0040] First, extract the spray trigger control records according to the secondary path segment numbers. Select the spray nozzle number and the timestamp in the corresponding trigger control instruction as the start time point. Subsequently, extract the maintenance period of each spray from the spray response duration status records. Suppose the start time corresponding to the number N1 is 12.3 seconds and the duration period is 3.0 seconds. Then the spray period range of this nozzle is from 12.3 seconds to 15.3 seconds. Synchronously, in the acid mist propulsion monitoring data, extract the acid mist propulsion records of the path segment that matches the spatial position of nozzle N1. Select the initial acid mist concentration rising point at the corresponding numbered position in the acid mist sensor as the acid mist arrival start time, set it as 12.8 seconds, and establish the acid mist residence section for the continuously high-concentration section within 12.8 seconds to 14.5 seconds to obtain the acid mist propulsion period. According to the comparison between the spray response start time of 12.3 seconds and the acid mist arrival start time of 12.8 seconds, there is a time difference of 0.5 seconds between the two. Record this difference as a negative value, indicating that the spray arrives and triggers earlier than the acid mist, which is represented as an early response behavior. If this difference is a positive value, it means the spray lags in response. Further, determine whether there is a timing misalignment situation. Establish a synchronization window for the spray timeline and the acid mist propulsion timeline in units of numbered segments. Collect the time position pairs of spray start and acid mist concentration mutation within each window, draw a broken line distribution trend graph, compare and sort the data of multiple groups of secondary path nozzle numbers, label all the time difference distributions, record the sequence of early response or lag response segments, and finally output the response relative position offset results of different numbered segments under the same timeline to obtain the difference between the response trigger and the airflow arrival order.

[0041] S302: Based on the difference between the response trigger and the airflow arrival order, calculate the advance period value when the spray response trigger time in the secondary path is earlier than the acid mist propulsion start point. Align the advance period value with the path set beat sequence, analyze the position of the advance amount in the time structure, and obtain the secondary path response propulsion section number.

[0042] The specific calculation formula for the advance period value when the spray response trigger time in the secondary path is earlier than the acid mist propulsion start point is as follows: ; Wherein, represents the advance period value when the spray response trigger time in the secondary path is earlier than the acid mist propulsion start point, represents the spray response trigger time of path segment represents the acid mist propulsion start time corresponding to path segment , represents the number of adjacent path segments of path segment , represents the path segment number index from 1 to , represents from 1 to of the path segment number index, represents path segment The spatial length, represents the path segment The corresponding beat sequence number, represents the path segment The spray response trigger time in represents the path segment The corresponding starting time for acid mist propulsion.

[0043] Assume: The path segment number is set to A5, and the corresponding spray response trigger time is 104.2 seconds; Parameter is the starting time for acid mist propulsion in path segment i. The moment when the leading gas concentration detected by the high-frequency infrared sensor reaches 0.35 mol per cubic meter is used as the trigger point, and the recorded time is 106.8 seconds; The time difference between the two is calculated as: ; The total number of adjacent path segments is set to 3, which are path segments = B1, B2, B3; The spray response trigger time of path segment B1 is 103.5 seconds, the corresponding starting time for acid mist propulsion is 105.0 seconds, and the path length read from the CAD drawing data is 85 mm, and the spray beat number is 4; The trigger time of path segment B2 is 102.9 seconds, the starting point of propulsion is 105.6 seconds, and the path length mm, and the beat number ; The trigger time of path segment B3 is 104.6 seconds, the starting point of propulsion is 106.4 seconds, and the path length mm, and the beat number .

[0044] Substitute into the calculation terms as follows: The first term: ; ; The product term: -1.5 × 42.5 = -63.75; The second term: ; ; The product term: -2.7 × 40.26 ≈ -108.7; The third term: ; Product term: -1.8 × 40.41 ≈ -72.74; Sum the three terms and divide by 3 to get the average correction term: ; Add the time difference term and the correction term to get: ; This result indicates that there is a significant early behavior of the spray response relative to the acid mist advancement in path segment A5, and the degree of advancement is in the strong offset area in the spatial path and the cycle scheduling structure. The numerical result of the early period is -79.13, which means that the path segment needs to be included in the set of section numbers for the sub-path response advancement for subsequent scheduling judgment and control.

[0045] S303: Based on the section numbers for the sub-path response advancement, delay the main path spray response time to the next-level trigger segment of the current beat sequence, adjust the position and time period of the main path response start point, identify the time overlap area between the front and rear responses, and obtain the sub-path early response scheduling record.

[0046] First, compare the trigger labels of each spray beat in the number sequence with the corresponding beat index values, locate the beat section where the main path spray response is located, move its spray start point to the next number section of the beat where the maximum index value in the sub-path response number is located, perform number conversion on the spray response start time, and mark the position of the new start beat in the spray scheduling control. At the same time, extract the continuous overlap time length between the previous number section of the main path response signal and the last response section of the sub-path, collect the coincidence continuous range value in seconds, and set the number as a n For the last response section of the sub-path with the number a, set its corresponding spray end time as T a , if the delay start time number of the main path spray response is b n , and the corresponding start time is set as T_b, then perform assignment confirmation for the interval between T a and T_b, mark the index set of all overlapping trigger points within this interval, divide the total amount of index points by the length of this section to obtain the overlap density value σ. If σ is greater than 0.6, then mark it as a stable overlap area, record the number set interval, extract the section numbers for the response advancement from each spray trigger section within the sub-path number a n as the reference source for the main path response delay, and finally construct an action chain for the backward synchronous transmission of the sub-path number trigger beats and synchronize it to the scheduling schedule to obtain the sub-path early response scheduling record.

[0047] In this embodiment, the specific steps of S4 are: S401: Based on the scheduling records of the sub-pathway's early response, collect the path extension data of hydrofluoric acid and sulfuric acid in the non-covered section of the spraying area, monitor the moving directions and continuous sections of the two components at the boundary section, classify the changing trend of the position of the advancing line at the edge of the path, and obtain the boundary extension trajectory value of the acidic components.

[0048] First, extract the previously generated sub-pathway response data, including the time difference between the spraying trigger time point and the initial movement time of the acid mist, the corresponding path number, and the relative order in the time series. Determine whether the response scheduling is in an early state by comparing the time differences. If the time difference ΔT satisfies ΔT > 0.5 s, it is recorded as a valid early response and incorporated into subsequent extension analysis. Subsequently, for the section that is not fully covered in the spraying area, determine its spatial boundary range. The specific method is to perform regional grid division on the spraying flow density distribution map and record the coordinate boundary points of the uncovered area between the nozzles, and then establish the boundary path line of the uncovered section. Collect the movement path data of hydrofluoric acid and sulfuric acid in this area. The required data includes the distribution of gas concentration along the boundary path, the flow velocity change curve, and the time series evolution trajectory. The path extension data of hydrofluoric acid and sulfuric acid is obtained through high-speed photography combined with image tracking. Locate the leading edge points once every 5 frames of images and calculate their cumulative displacement along the path direction. Set the effective length threshold L t of 50 mm. If the extension length exceeds L t in multiple consecutive time frames, it is judged that boundary extension has occurred. At the same time, monitor the moving directions and continuous sections of the two acidic components at the boundary section. The moving direction is calculated by the angle between the tangent direction of the boundary path and the gas front advancing vector. When the angle θ satisfies 0° ≤ θ ≤ 15°, it is regarded as co-directional movement, and 15° < θ ≤ 90° is offset propulsion. The determination of the continuous section is based on the path length continuously maintained in the same moving direction. If the continuous co-directional propulsion length is greater than 30 mm, it is classified as a continuous section. The changing trend of the position of the advancing line is evaluated by the lateral offset distance Δx of the leading edge advancing points at consecutive moments. When Δx shows a monotonically increasing or decreasing change and its fluctuation range is less than 5 mm, it is classified as stable propulsion. If Δx fluctuates by more than 10 mm between any two adjacent moments, it is judged as severe offset. Finally, based on the cross-analysis of the aforementioned path extension data, moving direction determination, continuous section identification, and advancing line trend classification, obtain the boundary extension trajectory value of the acidic components.

[0049] S402: Based on the boundary extension trajectory value of the acidic components, compare the response areas that are not covered by droplets in the same passage section in the current and the previous cycle, screen out the sequence of position sections that are not covered in two consecutive cycles, and extract the continuous uncovered sections to obtain the periodic spraying fault section index.

[0050] First, a spatio-temporal comparison is made between the current cycle and the unresponsive area in the acid mist spray reaction path of the previous cycle to identify the passage numbers where the acid mist is not covered by droplets in the two cycles. By sequentially calling the droplet density distribution and acid mist residue indicators in the response matrix, path nodes with droplet action intensity lower than the average boundary extension intensity are screened to form an initial set of unresponsive areas. Further, the passage numbers of each path segment in this set are sorted, and a continuous sequence of adjacent numbers with uncovered paths in the middle is extracted. The time domain position of the corresponding section of this sequence is marked in combination with the response cycle time period, and then it is overlapped and compared with the time slice of the reaction propulsion path in the acid mist movement trajectory to mark the actual continuous unresponsive time span. Furthermore, the continuous fault segments in adjacent time slices are merged to construct the periodic spray missing block. Finally, the main and secondary path boundary sequences where droplets do not form a response are classified according to the passage number, and the continuous segments of the overlapping response missing between the current cycle and the previous cycle are output to obtain the periodic spray fault section index.

[0051] S403: Based on the periodic spray fault section index, extract the distribution of residual components of hydrofluoric acid and sulfuric acid in the edge area of the spray response surface, and process the spatial offset relationship between the component migration direction at the boundary and the blank section of droplet contact to identify the path segment area where the acid mist does not form a reaction, thereby obtaining the component-triggered response boundary section.

[0052] First, segment the spray response frequency data of the fault section according to the spray operation time axis, and judge based on whether the number of spray starts per unit time is less than the set reference value. The reference value is set to no less than 3 times of spray starts within every 60 seconds. If it is lower than this value, it is marked as the fault section. Then, with these fault sections as the center, an edge area extending 5 cm around is used as the analysis area range. The thermal imaging map of the residual concentrations of hydrofluoric acid and sulfuric acid in the edge area is obtained through high-resolution infrared imaging. The images are collected once every 2 seconds for each frame. A total of 300 consecutive images within 10 minutes are selected. For each frame, the pixel gray value is calibrated to the concentration value. The conversion coefficient is set to correspond to 0.012 mol / m³ per 1 gray unit with reference to the laboratory spectral analysis results. When the gray value of a certain area remains higher than 200 units for more than 60 seconds in consecutive images and there is no spray trace, it is determined as a high-concentration residue area. Subsequently, the coordinates of these high-concentration residue areas are classified, the center point coordinates of this area in each frame are extracted, and the displacement amounts in the x and y directions in consecutive frames are calculated, so as to obtain the component migration direction. If the change angle of the direction accumulates more than 45° within 10 frames, it is defined as a sharp turn, otherwise it is regarded as stable advancement. Then, these migration directions are spatially compared with the droplet distribution trajectories in the spray operation image. The blank section contacted by the droplets is determined as the area where the temperature drop amplitude in the thermal imaging map of the droplet distribution does not exceed 1.5°C. If the area of this area is greater than 50 cm² and remains stable for more than 30 seconds, it is determined as an effective blank section. The offset calculation is carried out for the spatial relationship between this blank section and the component migration direction. The offset amount is based on the shortest straight-line distance between the center points of the two. If the offset distance is greater than 5 cm and lasts for more than three time periods, it is marked as a significant offset area. Finally, the path sections with high concentrations of hydrofluoric acid and sulfuric acid residues, spatial offsets with the droplet blank interval, and stable moving directions are integrated and statistically analyzed to identify the path section areas where the acid mist does not form a reaction, and finally the component trigger response boundary section is obtained.

[0053] In this embodiment, the specific steps of S5 are as follows: S501: Based on the component trigger response boundary section, collect the moving direction, advancement sequence, and flow velocity stable section of the acid mist in the corresponding path section, sort out the advancement time sequence of the starting position and the front section of the acid mist in the path on the time axis, monitor the coincidence distribution of the acid mist movement behavior and the response section, and obtain the acid mist moving direction sorting set.

[0054] First, extract the spatial coordinate information and acid mist initial diffusion time tags of each path segment in the boundary section. Divide the path segments into a numbered sequence in order, and deploy infrared imaging devices and high-frequency concentration probes inside each numbered area to collect the real-time change trajectory of acid mist concentration. The imaging data is frame-sampled at 1-second intervals and the coordinate points are recorded. Subsequently, perform pixel gray-scale analysis on the acid mist front concentration area in the collected images. The advancing direction is obtained through the straight-line connection direction of the centroid positions of the high-concentration areas in three consecutive frames of images. The angle θ between each advancing direction vector and the main axis direction of the path needs to be calculated and classified. If θ is less than 30°, it is defined as consistent advancement. If it is between 30° and 60°, it is regarded as biased advancement. If the included angle exceeds 60°, it is regarded as reverse fluctuation advancement. At the same time, extract the advancement order according to the sequence in the time series, and sort each path number according to the first arrival time of the front. If there are multiple path points reaching at the same time, sort the priority according to the abscissa from small to large. For the acid mist flow rate in each numbered area, divide the set sampling point distance by the time difference when the front concentration reaches the threshold to calculate the advancement rate within a single path segment. For example, if the distance between two monitoring points in path A is 0.4 meters and the time difference between the front and rear arrivals is 6 seconds, the flow rate is 0.067 m / s. Screen the stability of the flow rates in all path segments. If the variation range of the flow rate within a unit time does not exceed ±10%, it is regarded as a stable flow rate segment, otherwise it is a fluctuating segment. Then, perform coordinate backtracking on the starting position of the acid mist, and extract the pixel centroid point where the high-concentration patch first appears as the starting position. Pair the coordinate points of all starting points and front areas in the path on the time axis, and generate an advancement time sequence map. Subsequently, compare the spatial coincidence degree between the advancement path and the component trigger response boundary section. If the coincidence range between the path advancement point and the boundary section coordinates is within ±2 cm and the coincidence time interval does not exceed 3 seconds, it is determined as an effective coincidence. Finally, integrate all the path information that meets the advancement direction, advancement order, stable flow rate, clear advancement start and end positions, and spatio-temporal coincidence with the response section to obtain the acid mist movement direction sorting set.

[0055] S502: Based on the acid mist movement direction sorting set, obtain the start time and corresponding node numbers of the spray response points, compare the advancement rhythm of the acid mist in the path segment with the trigger period of the spray response points, mark the sequence of the reaction nodes, and obtain the response node time sequence structure table.

[0056] First, extract the start time and the front arrival time recorded in the acid mist propulsion sequence of each path segment to form the start and end time pairs of acid mist propulsion. At the same time, derive the start time of the spray response point and the corresponding node number from the spray control record. The node numbers are obtained by numbering in the spatial order of the nozzle layout. Each numbered point corresponds to a spatial coordinate value and its trigger time tag, and they are arranged in ascending order along the time axis in the spray response record. Compare and match the acid mist path segment numbers and the spray node numbers on the spatial coordinate axis. If the distance between the end point of the acid mist front propulsion path and the center coordinate of the spray response point is less than 5 cm, it is considered that the path segment and the response node match. Then, compare the acid mist propulsion time and the spray start time in each pair of matching items. If the start of acid mist propulsion is earlier than the start time of the spray response point, mark this node as a late response. If the time difference between the two is less than 1 second, mark it as a synchronous response. If the spray response time is more than 1 second earlier than the start of acid mist propulsion, record it as an early response. Through the above comparison process, obtain the relative position status of each response node on the time axis. Arrange all response nodes in the order of path segments, generate the timing tag Ti for each node, and make it correspond one by one with the acid mist path propulsion sorting sequence. At the same time, construct a timing matching matrix. In this matrix, use the path segment number as the row and the response node number as the column, and fill the elements with response status identification values, where the synchronous response is 1, the early response is 2, and the late response is 3. Traverse the rows and columns of the response matrix to screen out all response type distribution rules, and further export the response node time series data after arranging them in ascending order according to the path segment number. Finally, integrate the node number, start time, acid mist propulsion start time, response status, and path number into a set of structured data sets to obtain the response node timing structure table.

[0057] S503: Based on the response node timing structure table, connect the sorted gas-liquid reaction nodes in the path segment in the order of the start rhythm, extract the path fragments and time series of the complete response chain, and combine and associate the path structures to obtain the acid mist treatment response control sequence.

[0058] First, read each group of node numbers, start times, path segment numbers, and response status identifications in the table, arrange them in ascending order according to the start time from early to late to form a start rhythm sequence, and then construct a spatial connection relationship based on the path segment number to judge whether any two adjacent nodes are spatially continuous. If the straight-line distance between the coordinates corresponding to the two node numbers does not exceed 15 cm and they are in the same acid mist propulsion direction, it is considered that they have physical continuity. Classify and number these physically continuous response nodes to form a logical chain number group. Then, re-mark the sequence numbers of the nodes within each logical chain, and assign values from the starting point to the end point as L1, L2, L3...L n, then extract the response time of each node and the path segment number to generate a time series, record the interval value ΔTᵢ between each beat, and determine whether it is a stable response rhythm. If the continuous change of ΔTᵢ does not exceed ±20%, it is regarded as a stable rhythm. If the change exceeds ±50%, it is recorded as a rhythm fluctuation segment. Then, based on the stable rhythm segment, extract the complete response chain. Each response chain needs to include the starting point response time, the end response time, the starting and ending numbers of the path segment, and the intermediate transition path sequence. After determining the complete response chain, further cross-compare the path number information shared between all response chains, extract the response chains with overlapping segments or adjacent segments for combination and association. If the distance between the end nodes of two response chains and the starting node number of another chain is less than 10 cm and the time difference is less than 3 seconds, it is determined that they can be connected. After connection, generate a new response combination chain, mark the order and time interval of the combination chain again, and record its total path length and total time length to form a complete path response process. Then, output the time series and path segment number corresponding to each combination chain in a unified numbering manner as the main sequence, and at the same time mark the rhythm state and physical connection relationship between each segment. Finally, integrate all the path segments and their time series data obtained through sorting, matching, screening, and connection to obtain the acid mist treatment response control sequence.

[0059] Please refer to Figure 2 , the present invention also provides a graphite waste acid gas treatment system based on artificial intelligence. The system is used to implement the above-mentioned graphite waste acid gas treatment method based on artificial intelligence. The system includes: The path recognition module obtains the gas propulsion section positions of the main path and the secondary path in the graphite acid mist pipeline, compares the spatial coincidence area between the acid mist propulsion area and the spray starting point, analyzes the synchronous relationship between the staggered length and the coincidence time period, and obtains the channel linkage structure type label; The contact recognition module, based on the channel linkage structure type label, measures the angle between the running track of the leading edge of the liquid droplet and the moving direction of the acid mist, records the continuous distribution section of the cross-sectional area where the atomized spray streamline intersects with the acid mist front, and matches the flow direction consistent section formed in the secondary channel to obtain the path contact characteristic offset identifier; The response scheduling module, based on the path contact characteristic offset identifier, compares the start time of the spray rhythm of the secondary path with the acid mist propulsion time period in terms of time sequence, checks whether there is a rhythm coincidence section in the front and back order, advances the spray response of the secondary path to the response node before the acid mist propulsion, and delays the spray rhythm of the main path to the next-level beat node to obtain the early response scheduling record of the secondary path; The boundary recognition module, based on the early response scheduling record of the secondary path, analyzes the path propulsion change of hydrofluoric acid and sulfuric acid outside the spray coverage, tracks the reaction edge path where no droplet coverage is formed in two consecutive cycles, and identifies the spatial area where the acid mist does not enter the neutralization reaction to obtain the component trigger response boundary section; The path control module triggers and responds to the boundary section based on the components, sorts out the chronological order between the advancing direction of the acid mist in the path section and the atomization response point, and strings together the start beats of all reaction nodes as the rhythm main line for the output control of the path behavior, thereby obtaining the acid mist treatment response control sequence.

[0060] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for treating graphite waste acid gas based on artificial intelligence, characterized in that: It includes the following steps: S1: Identify the gas movement segments of the main and secondary paths in the graphite acid mist passage, compare the synchronous movement ranges of the acid mist propulsion area and the atomization spray starting point, analyze the staggered lengths that appear synchronously between the two segments, process the response matching relationship between the main and secondary paths within the same time window, and obtain the channel linkage structure type label; S2: Based on the channel linkage structure type label, measure the angular deviation between the running trajectory of the droplet front edge and the forward movement route of the acid mist, match the flow overlap segment formed by the spraying behavior and the acid mist direction in the secondary path, and obtain the path contact feature offset identifier; S3: Based on the path contact feature offset identifier, compare the spraying trigger rhythm of the secondary path with the acid mist arrival time period, check the overlap relationship between the front and back sequences, advance the spraying response of the secondary path before the acid mist movement action, and delay the spraying of the main path to the next time beat position to obtain the secondary path early response scheduling record; S4: Based on the secondary path early response scheduling record, extend the analysis of the path changes of hydrofluoric acid and sulfuric acid in the non-fully covered segment of the spraying area, track the edge of the reaction zone where no droplet coverage is formed in two consecutive cycles, and locate the diffusion area where the acid mist does not enter the reaction state to obtain the component trigger response boundary section; S5: Based on the component trigger response boundary section, sort out the movement direction of the acid mist on the path segment in the diffusion area where it has not entered the reaction state and the sequence of the first and subsequent actions of the spraying response points, and connect the start rhythms between the gas-liquid reaction nodes to obtain the acid mist treatment response control sequence.

2. The method for treating graphite waste acid gas based on artificial intelligence according to claim 1, characterized in that: The channel linkage structure type label includes the main path gas movement segment, the secondary path gas movement segment, the staggered length, and the response matching relationship; the path contact feature offset identifier includes the angular deviation, the cross-sectional distribution, and the flow overlap segment; the secondary path early response scheduling record includes the spraying trigger rhythm, the acid mist arrival time period, the start advancement position, and the time beat adjustment; the component trigger response boundary section includes the path change area, the reaction zone edge, and the diffusion area; the acid mist treatment response control sequence includes the movement direction, the response point sequence, the start rhythm, and the path relationship.

3. The method for treating graphite waste acid gas based on artificial intelligence according to claim 1, wherein: The specific steps of S1 are as follows: S101: Obtain the acid mist propulsion segment information in the main and secondary paths of the graphite acid mist passage, collect the starting point position, flow range, and direction annotation of the acid mist propulsion in the differential paths, compare the time coverage difference and position offset of the acid mist movement starting areas in the two path segments, and obtain the path propulsion starting point position difference value; S102: Based on the path propulsion starting point position difference value, locate the atomization spray starting point positions of the main and secondary paths, extract the spatial contact area between the spraying response surface and the acid mist front, and screen the spraying response point sequence that coincides with the acid mist movement direction to obtain the path response synchronous coverage segment value; S103: Based on the path response synchronous coverage segment value, align the acid mist movement segments and the spraying response point sequences of the main and secondary paths within the same time window, analyze whether there is a front-back offset or starting misalignment in the response rhythm, and extract the rhythm corresponding state to obtain the channel linkage structure type label.

4. The method for treating graphite waste acid gas based on artificial intelligence according to claim 1, characterized in that: The specific steps of S2 are as follows: S201: based on the channel linkage structure type label, the starting point and propulsion direction of the droplet front edge running trajectory are obtained, the moving direction line segment of the acid mist front in the same path segment is collected, the droplet propulsion direction and the acid mist moving route are matched, and the angle deviation trend of the two direction lines is analyzed to obtain the direction angle deviation value; S202: based on the direction angle offset value, extract the intersection position section of the spray streamline in the process of acid mist advancement, sort out the distribution density of the intersection area within the continuous path length, calculate the ratio of the number of spray points in the intersection section to the number of acid mist advancement points, and obtain the spray overlap section coverage interval; S203: Based on the coverage interval of the spray overlapping segment, identify the path fitting area of the spray behavior in the secondary path in the direction of acid mist movement, extract the overlapping length of the serial number distribution of the continuous fitting segment and the spray response range, construct the path contact state index according to the fitting rate and the offset value, and obtain the path contact feature offset mark.

5. The method for treating graphite waste acid gas based on artificial intelligence according to claim 4, characterized in that: The specific calculation formula for the ratio of the number of spray points to the number of acid mist propulsion points in the intersection section is: ； Among them, represents the ratio of the number of spray points to the number of acid mist propulsion points in the intersection section, represents the number of spray points in the intersection section, represents the number of acid mist propulsion points in the intersection section, represents the number of the th spray point in the intersection section, represents the number of the acid mist propulsion point corresponding to the th spray point at the corresponding position on the path, represents the maximum value of the numbers of the acid mist propulsion points in the intersection section, represents the average value of the spans of the numbers of the acid mist propulsion points in the intersection section.

6. The method for treating graphite waste acid gas based on artificial intelligence according to claim 1, characterized in that: The specific steps of S3 are: S301: Based on the path contact feature offset mark, the start time point and duration period in the secondary path spray response rhythm are collected, the time starting point and the stay section of the acid mist arrival in the same path segment are obtained, and the starting positions of the two in the same timeline are compared to obtain the difference between the response trigger and the airflow arrival sequence; S302: Based on the difference between the response trigger and the airflow arrival sequence, the advance time period value of the spray response trigger time in the secondary path is calculated to be earlier than the acid mist advancement starting point, the advance time period value is aligned with the path setting beat sequence, the position of the advance amount in the time structure is analyzed, and the secondary path response advancement section number is obtained; S303: Based on the secondary path response advancement section number, the main path spray response time is postponed to the next level trigger segment of the current beat sequence, the position and time period of the main path response start point are adjusted, the time overlap area between the previous and next responses is identified, and the secondary path advance response scheduling record is obtained.

7. The method for treating graphite waste acid gas based on artificial intelligence according to claim 6, wherein: The numerical calculation formula for the advance period in which the spray response trigger time in the secondary path is earlier than the acid mist advancement starting point is specifically: ； Among them, represents the sub-path The numerical value of the advance period when the spray response trigger time in is earlier than the acid mist propulsion starting point, represents the path segment The spray response trigger time of represents the path segment The corresponding acid mist propulsion starting point time, represents the path segment The number of adjacent path segments, represents from 1 to The path segment number index of represents the path segment The spatial length of represents the path segment The corresponding beat sequence number, represents the path segment The spray response trigger time in represents the path segment The corresponding acid mist propulsion starting point time.

8. The method for treating graphite waste acid gas based on artificial intelligence according to claim 1, characterized in that: The specific steps of S4 are: S401: Based on the advance response scheduling record of the secondary passage, the path extension data of fluoric acid and sulfuric acid in the uncovered section of the spraying area are collected, the moving direction and duration of the two components in the boundary section are monitored, the position change trend of the advancement line at the edge of the path is classified, and the boundary extension trajectory value of the acidic component is obtained; S402: Based on the boundary extension trajectory value of the acidic component, the response area of the same passage section in the current and previous cycles that is not covered by droplets is compared, the position segment sequence that is not covered in two consecutive cycles is screened out, and the continuous uncovered segments are extracted to obtain the periodic spray fault segment index; S403: Based on the periodic spray fault segment index, the residual component distribution of fluoric acid and sulfuric acid in the edge area of the spray response surface is extracted, the spatial offset relationship between the component migration direction at the boundary of the treatment and the blank section of the droplet contact is compared, the path segment area where the acid mist does not form a reaction is identified, and the component triggered response boundary segment is obtained.

9. The method for treating graphite waste acid gas based on artificial intelligence according to claim 1, characterized in that: The specific steps of S5 are: S501: Based on the component-triggered response boundary section, collect the moving direction, advancing sequence, and flow velocity stable section of the acid mist within the corresponding path segment, organize the starting position of the acid mist in the path and the advancing time sequence of the front section on the time axis, monitor the coincidence distribution of the acid mist movement behavior and the response section, and obtain the acid mist moving direction sorting set; S502: Based on the acid mist moving direction sorting set, obtain the start time of the spray response point and the corresponding node number, compare the advancing rhythm of the acid mist within the path segment with the triggering period of the spray response point, mark the time sequence priority of the reaction nodes, and obtain the response node time sequence structure table; S503: Based on the response node time sequence structure table, connect the sorted gas-liquid reaction nodes in the path segment in the order of the start rhythm, extract the path segment and time series of the complete response chain, and combine and associate the path structures to obtain the acid mist treatment response control sequence.

10. The graphite waste acid gas treatment system based on artificial intelligence is characterized in that: The system is used to implement the method for treating graphite waste acid gas based on artificial intelligence according to any one of claims 1-9. The system includes: The path recognition module obtains the gas advancing section positions of the main path and the secondary path in the graphite acid mist pipeline, compares the spatial coincidence area between the acid mist advancing area and the spray starting point, analyzes the synchronous relationship between the staggered length and the coincidence time period, and obtains the channel linkage structure type label; The contact recognition module, based on the channel linkage structure type label, measures the angle between the running trajectory of the droplet front edge and the moving direction of the acid mist, records the continuous distribution section of the atomized spray streamline and the acid mist front intersection area, and matches the flow direction consistent section formed in the secondary channel to obtain the path contact feature offset identifier; The response scheduling module, based on the path contact feature offset identifier, conducts a time sequence comparison between the start of the spray rhythm of the secondary path and the acid mist advancing time period, checks whether there is a rhythm coincidence section in the front and back order, advances the spray response of the secondary path to the response node before the acid mist advances, and delays the spray rhythm of the main path to the next-level beat node to obtain the secondary path early response scheduling record; The boundary recognition module, based on the secondary path early response scheduling record, analyzes the path advancement changes of hydrofluoric acid and sulfuric acid outside the spray coverage, tracks the reaction edge path where no droplet coverage is formed in two consecutive cycles, and identifies the spatial area where the acid mist does not enter the neutralization reaction to obtain the component-triggered response boundary section; The path control module, based on the component-triggered response boundary section, sorts out the time sequence between the advancing direction of the acid mist in the path segment and the atomized response point, connects the start beats of all reaction nodes as the rhythm main line of the path behavior output control, and obtains the acid mist treatment response control sequence.

Images