Polyethylene chlorination reaction control system

Abstract

The invention relates to the technical field of supervision and control, in particular to a polyethylene chlorination reaction control system which comprises a hot-pressing time difference extraction module, a viscosity variable-speed recognition module, a chlorine pressure collaborative trend division module, a response centralized recombination module and a stirring amplitude management and control module. According to the method, a multi-source variable data structure is unified through period numbers, cooperative sections with consistent trend directions are constructed, response feature fragments in a continuous period are recognized by combining fluctuation frequency and adjustment amplitude feature extraction, the behavior boundary of each variable in the time dimension is unified, and the recognition precision of multi-variable abnormal evolution is improved; a regulation and control reference is established through trend classification and strength grading, structured expression and controllability analysis of periodic behaviors are realized, the matching between the regulation rhythm and the variable trend is enhanced, and a regulation and control strategy is enabled to have clear boundary recognition, period tracking and behavior intervention capabilities.

Description

Technical Field

[0001] The present invention relates to the technical field of supervisory control, in particular to a polyethylene chlorination reaction control system. Background Art

[0002] The field of supervisory control technology involves real-time monitoring and adjustment of key physical quantities in industrial processes to ensure system operation stability and production efficiency. The core content of this technical field is based on the principle of feedback control, through automatic detection, data acquisition and control strategy execution, to effectively control multiple dynamic variables in complex industrial processes. Supervisory control covers aspects such as signal acquisition, sensor fusion, actuator control, parameter adjustment and control strategy modeling, and is usually used in continuous or batch production scenarios such as chemical industry, metallurgy, energy, and materials. In this field, the systematic integration of various sensor elements, analysis methods and control logic can form a complete closed-loop control chain to ensure the parameter coupling relationship and controllability of reaction states in multi-variable processes.

[0003] The polyethylene chlorination reaction control system monitors reaction conditions such as temperature, pressure, and chlorine delivery during the chlorination modification of polyethylene materials. It uses temperature sensors, pressure sensors, and spectral detection equipment to acquire real-time reaction status data, and utilizes computer-controlled devices to automatically adjust the chlorine flow rate and stirring rate to achieve continuous regulation of reaction conditions. Based on the principles of heat and mass exchange, the system dynamically analyzes the changing trends of reaction parameters and, in conjunction with a gas-liquid distribution model, determines the distribution of the chlorinating agent and calculates the optimal delivery rate of the chlorinating agent. During operation, the system primarily utilizes a multi-point acquisition method based on physical variable measurement, numerical fitting calculations, and proportional-integral regulation to implement quantitative control and error correction of the reaction process, thereby establishing a supervisory control process with coordinated multi-parameter regulation.

[0004] Existing technologies mainly rely on independent measurement and control links of variables, and do not aggregate and analyze the synchronization relationship of multiple types of response data in the cycle time dimension. When the rhythm of variable changes is misplaced or the trend evolution is inconsistent, it is difficult to construct a unified response reference segment, resulting in isolated adjustment basis and difficulty in reflecting the coordinated trend between variables. Especially in working conditions where the interaction between chlorine pressure fluctuations and stirring behavior is obvious, it is impossible to accurately identify the degree of coupling between the control intensity and response changes in continuous cycles, which in turn causes the adjustment strategy to deviate from the actual rhythm of process state changes. For example, when the speed fluctuates frequently without accompanying trend reconstruction and identification, the control action is easily triggered repeatedly, causing the system actuator to be overloaded or the adjustment efficiency to decrease. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a polyethylene chlorination reaction control system.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A polyethylene chlorination reaction control system comprises:

[0007] The hot-pressing time difference extraction module obtains the changing time points of the thermal-sensitive data and pressure data in the polyethylene chlorination process, performs periodic comparison on the time difference between the two, extracts the mutation inflection point period, and generates the hot-pressing time difference mutation result;

[0008] The viscosity change identification module marks the cycle according to the result of the thermal pressing time difference mutation, calls the viscosity recording point data, determines the direction of the cycle viscosity change, extracts and marks the section of unidirectional amplification of the change, and generates a viscosity change concentrated section;

[0009] The chlorine-pressure synergistic trend division module retrieves the chlorination advancement curve and the pressure fluctuation sequence for the viscosity speed change concentrated section period, determines whether the fluctuations are in the same direction in the continuous period, marks the trend synchronization period, and generates a chlorine-pressure synergistic behavior segment set;

[0010] The response concentration reorganization module is based on the period number of the chlorine pressure synergistic behavior segment set, matches the thermal pressure time difference mutation result with the viscosity speed change concentrated segment, extracts the intersection period segment of the three, and generates a multi-source response set;

[0011] The stirring amplitude control module calls the multi-source response set period segment, retrieves the stirring speed change record, identifies the adjustment segment within the period, sets operation restrictions on the adjustment amplitude and duration, and generates a chlorination reaction control result.

[0012] As a further solution of the present invention, the hot pressure time difference mutation result includes the time difference change amplitude, inflection point cycle characteristics, state classification label, and increasing trend sequence; the viscosity speed change concentrated section includes the viscosity change direction, continuous cycle identification, change amplitude characteristics, and concentrated section number; the chlorine pressure collaborative behavior segment set includes the chlorination curve trend, pressure fluctuation trend, trend synchronization cycle, and collaborative relationship label; the multi-source response set includes a cycle number set, trend consistency mark, intersection segment sequence, and multi-source fusion result; the chlorination reaction control result includes the stirring speed change frequency, adjustment amplitude limit value, adjustment duration setting, and stirring operation stability index.

[0013] As a further solution of the present invention, the thermal pressure time difference extraction module includes:

[0014] The direction of cyclic viscosity change refers to the trend direction of the viscosity value continuously rising, falling and remaining stable over time within a thermal pressure mutation cycle;

[0015] The chlorination advancement curve is a time series curve reflecting the trend of chlorine content changing with time during the polyethylene chlorination reaction;

[0016] The pressure fluctuation sequence is a discrete data sequence that records the pressure changes over time during the polymerization reaction, reflecting the dynamic fluctuation characteristics of the pressure;

[0017] The sudden change result of the hot pressing time difference is linked with the viscosity speed change concentrated section to control the stirring amplitude.

[0018] As a further solution of the present invention, the thermal pressure time difference extraction module includes:

[0019] The thermal inflection point identification submodule calculates the temperature increase of adjacent data points based on the collected thermal data, determines whether it exceeds the temperature jump threshold, identifies the mutation point, divides the time period between adjacent mutation points, and generates a thermal cycle segment list;

[0020] The pressure change screening submodule extracts the pressure data of the corresponding time period according to the thermal cycle segment list, calculates the time difference sequence between the pressure mutation moment and the thermal sensitive mutation point, screens the period segments with continuously increasing time differences, and generates a period group with increasing time differences;

[0021] The mutation cycle classification submodule obtains the time difference, mutation amplitude and rate of each segment based on the time difference increment cycle group, performs normalization operation processing, calculates the thermal pressure mutation change of the cycle, classifies the state according to the change range, and generates the thermal pressure time difference mutation result;

[0022] The temperature jump threshold refers to the limit value of the continuous temperature increase required to determine a sudden change in the thermal sensitive data;

[0023] The association between the pressure mutation moment and the thermal-sensitive mutation point is to calculate the time difference between the two, revealing the sequence and linkage relationship between the thermal response and the pressure response.

[0024] As a further solution of the present invention, the viscosity speed change identification module includes:

[0025] The cycle marking submodule marks the time interval number of the mutation segment based on the hot pressing time difference mutation result, maps the corresponding number to the bottom viscosity measurement point data, verifies the cycle attribution according to the measurement point recording time, and generates a cycle viscosity data set;

[0026] The direction judgment submodule calls the periodic viscosity data set, judges the change trend of the viscosity values ​​in adjacent periods, retains sequence segments with consistent change directions, sequentially identifies period segments with continuous unidirectional changes, and generates a set of unidirectional continuation segments;

[0027] The trend annotation submodule extracts the viscosity cumulative change value, cycle length and local fluctuation coefficient based on the one-way continuous segment set, determines the one-way reinforcement characteristics, calculates and obtains the viscosity speed ratio of each segment, maps the ratio interval to the trend type number, and generates the viscosity speed concentrated segment;

[0028] Definition of the change trend: A change trend refers to a pattern in which the viscosity values ​​continuously change in the same direction within adjacent periods;

[0029] Definition of the One-Way Strengthening Feature The one-way strengthening feature refers to a trend phenomenon in which the viscosity changes cumulatively, fluctuates, and lasts for a long period during a continuous one-way viscosity change process.

[0030] As a further solution of the present invention, the chlorine pressure collaborative trend division module includes:

[0031] The data retrieval submodule locates the start and end time points of each cycle segment based on the cycle range marked in the viscosity speed change concentrated section, retrieves the chlorination advancement curve and pressure fluctuation sequence in the corresponding time period, and performs periodic synchronous sorting on the chlorine pressure data to generate a periodic chlorine pressure data set;

[0032] The trend identification submodule calls the periodic chlorine pressure data set, identifies the direction of change of the chlorination curve and pressure fluctuation within the period, compares the signs of the chlorine pressure data, determines whether the fluctuations in consecutive periods are in the same direction, and generates a trend consistent segment sequence;

[0033] The behavior clustering submodule extracts the time range, number of cycles and corresponding chlorine pressure fluctuation characteristic parameters of each trend according to the trend consistent segment sequence, combines them to construct segment features, classifies the segments into the same category according to the trend fitting degree, and generates a chlorine pressure coordinated behavior segment set;

[0034] The definition of the changing direction of the chlorination curve and pressure fluctuation within the cycle refers to the overall trend of the chlorination curve and pressure fluctuation rising, falling and remaining flat within a cycle;

[0035] The definition of the segment characteristics refers to a set of structured information describing a trend segment using trend time range, number of cycles and chlorine pressure fluctuation characteristic parameters;

[0036] The sign comparison is performed by marking the trend signs of the chlorination curve and the pressure fluctuation, and screening out the periodic segments with the same direction between the two to identify the segments with the same trend.

[0037] As a further solution of the present invention, the response centralized reorganization module includes:

[0038] The number alignment submodule matches the hot-pressing time difference mutation results with the corresponding cycles in the viscosity speed change concentrated section based on the cycle numbers identified by the chlorine pressure synergistic behavior segment set, extracts the full set of cycle numbers in the three types of data, establishes a cycle cross-reference table, and generates a multi-source cycle number set;

[0039] The period intersection extraction submodule calls the multi-source period number set, screens period numbers that exist simultaneously in the three sources, locates and extracts the time segments of the intersection period, establishes a list of intersection segments arranged in order, and generates a response intersection period segment;

[0040] The sequence summary submodule integrates the three types of raw data fragments of heat pressure, viscosity, and chlorine pressure in the corresponding segments according to the response intersection period segment, and uniformly numbers and proofreads the time sequence, and generates a multi-source response set according to the continuous time period;

[0041] The period intersection extraction submodule and the sequence summary submodule screen the intersection of period numbers in the three types of sources and locate their time segments, integrate the raw data segments of heat pressure, viscosity and chlorine pressure, and unify the time numbers.

[0042] As a further solution of the present invention, the stirring amplitude control module includes:

[0043] The cycle matching submodule extracts the stirring speed change records within the corresponding cycle segment based on the cycle number marked in the multi-source response set, performs time alignment and cycle number association processing on the speed data, and generates a periodic stirring data set;

[0044] The frequency identification submodule calls the periodic stirring data set, identifies the frequency of changes in stirring speed and direction switching in adjacent cycles, screens the sections with adjustments in continuous cycles, establishes a change feature marker sequence, and generates high-frequency adjustment period segments;

[0045] The amplitude limiting submodule extracts the speed adjustment amplitude, duration and number of changes according to the high-frequency adjustment period, determines whether it exceeds the preset amplitude threshold and time threshold range, divides the control level according to the adjustment intensity, and generates the chlorination reaction control result;

[0046] The time alignment of the speed data and the association of the cycle numbers refers to dividing the time series data of the stirring speed according to the start and end time of each cycle and marking the corresponding cycle number;

[0047] The definition of the segment with high frequency adjustment in the continuous cycle is the time segment in which the stirring speed changes and the number of direction switching times in adjacent cycles;

[0048] The definition of the preset amplitude threshold and time threshold range refers to the setting parameters of the speed change value limit and duration limit used to determine whether the regulation is too strong;

[0049] The specific numerical range of the amplitude threshold is set to be the amplitude of the stirring speed change within a single cycle exceeding ±

[0050] 50RPM is considered as strong throttling behavior.

[0051] Compared with the prior art, the advantages and positive effects of the present invention are:

[0052] In the present invention, the multi-source variable data structure is unified through period numbering, and collaborative segments with consistent trend directions are constructed. The fluctuation frequency and adjustment amplitude feature extraction are combined to identify the response feature segments within continuous periods, unify the behavioral boundaries of each variable in the time dimension, and improve the recognition accuracy of multi-variable abnormal evolution. Through trend classification and intensity grading, a control reference is established to achieve structured expression and controllability analysis of periodic behavior, enhance the matching between the adjustment rhythm and the variable trend, and enable the control strategy to have clear boundary identification, period tracking and behavior intervention capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a system flow chart of the present invention;

[0054] Figure 2 This is a flow chart of the hot pressure time difference extraction module of the present invention;

[0055] Figure 3 This is a flow chart of the viscosity speed change identification module of the present invention;

[0056] Figure 4 This is a flow chart of the chlorine pressure synergy trend division module of the present invention;

[0057] Figure 5 This is a flow chart of the response centralized reorganization module of the present invention;

[0058] Figure 6 This is a flow chart of the stirring amplitude control module of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention 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 invention and are not intended to limit the present invention.

[0060] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0061] See also Figure 1 , a polyethylene chlorination reaction control system includes:

[0062] The hot-pressing time difference extraction module obtains the changing time points of the thermal-sensitive data and pressure data in the polyethylene chlorination process, performs periodic comparison of the time difference between the two, selects the time periods with the same increasing direction of the difference, extracts the mutation inflection point period and performs state classification to generate the hot-pressing time difference mutation result;

[0063] The viscosity change identification module marks the cycle according to the sudden change results of the hot pressing time difference, calls the data of multiple viscosity recording points at the bottom, determines whether the viscosity change direction of adjacent cycles is continuous, extracts the section with unidirectional amplification of the change, marks its trend, and generates the viscosity change concentrated section;

[0064] The chlorine-pressure synergistic trend classification module retrieves the chlorination advancement curve and pressure fluctuation sequence for the period of the viscosity speed concentration segment, determines whether the two fluctuate in the same direction in consecutive cycles, marks the period of synchronized trend and classifies it into the same category, and generates a chlorine-pressure synergistic behavior segment set;

[0065] The response concentration reorganization module is based on the period number of the chlorine pressure synergistic behavior segment set, matches the thermal pressure time difference mutation results with the viscosity speed change concentrated segment, extracts the intersection period of the three, and summarizes them in chronological order to generate a multi-source response set;

[0066] The stirring amplitude control module calls the multi-source response set period segment, retrieves the stirring speed change record, identifies the sections with frequent adjustments in the continuous cycle, sets operation limits for the adjustment amplitude and duration, and generates the chlorination reaction control results.

[0067] The results of hot pressure time difference mutation include the time difference change amplitude, inflection point cycle characteristics, state classification label, and increasing trend sequence. The viscosity change concentrated segment includes the viscosity change direction, continuous cycle identification, change amplitude characteristics, and concentrated segment number. The chlorine pressure synergistic behavior segment set includes the chlorination curve trend, pressure fluctuation trend, trend synchronization cycle, and synergistic relationship labeling. The multi-source response set includes the cycle number set, trend consistency mark, intersection segment sequence, and multi-source fusion results. The chlorination reaction control results include the stirring speed change frequency, adjustment amplitude limit value, adjustment duration setting, and stirring operation stability index.

[0068] See also Figure 2 , the thermal pressure time difference extraction module includes:

[0069] The thermal inflection point identification submodule calculates the temperature increase of adjacent data points based on the collected thermal data, determines whether it exceeds the temperature jump threshold, identifies the mutation point, divides the time period between adjacent mutation points, and generates a thermal cycle segment list;

[0070] The thermal inflection point identification submodule first continuously acquires raw data from the temperature sensor based on a preset sampling frequency, fixed at once per second. The system then generates a time-stamped temperature sequence in real time. For example, suppose the monitoring system continuously collects temperature data of 36.2°C, 36.5°C, 38.1°C, 39.8°C, 41.1°C, 42.9°C, 43.2°C, 43.3°C, 43.3°C, and 43.2°C within the first 10 seconds of a thermal process. Starting from the second data point, the module calculates the temperature increase from the previous data point to determine whether it exceeds the set transition threshold, which can be set to 1.5°C. If the temperature change is less than the threshold, the point is skipped. If it is greater than or equal to the threshold, the current point is marked as a thermal inflection point. In the above sequence, the temperature increases by 1.6 degrees Celsius from the second to the third second, 1.7 degrees Celsius from the third to the fourth second, and 1.3 degrees Celsius from the fourth to the fifth second. Since both of these changes exceed the trip threshold, the third and fourth seconds are identified as thermally sensitive mutation points, and the system records them in the mutation point list. After identifying all mutation points, the system sorts them according to their time tags and further divides the time segments between these mutation points. Each time segment starts with the previous mutation point and ends with the next mutation point. The time elapsed between these two points is calculated and recorded as a thermal cycle segment. For example, if the third and fifth seconds are two mutation points, the time range of this cycle segment is from the third to the fifth second, with a total duration of two seconds. This process is repeated for all adjacent mutation point pairs, forming a list of thermal cycle segments. Each cycle segment in this list contains three fields: start time, end time, and duration, which are used for subsequent pressure data correlation analysis. Each cycle segment can be used to track and analyze sudden changes in thermal reactions. This segmentation allows the system to break down complex, continuous thermal data into multiple event-related sudden changes, constructing targeted temperature behavior data segments. This processing method is suitable for typical scenarios such as high-temperature process monitoring, motor thermal protection, and environmental anomaly monitoring.

[0071] The pressure change screening submodule extracts the pressure data of the corresponding time period according to the thermal cycle segment list, calculates the time difference sequence between the pressure mutation moment and the thermal sensitive mutation point, screens the period segments with continuously increasing time differences, and generates a period group with increasing time differences;

[0072] After receiving the list of thermal cycle segments generated by the thermal inflection point identification module, the pressure change screening submodule first parses the time range of each cycle segment in the list and extracts the pressure data for the corresponding time period. In practice, the system reads a pressure data sequence consisting of pressure values ​​collected with the same timestamp, sampled once per second, consistent with the temperature data. For example, if a cycle segment starts at the tenth second and ends at the fifteenth second, the system retrieves all pressure data points between the tenth and fifteenth seconds, forming a continuous pressure subsequence. Next, the module performs a mutation determination on the data within this subsequence, calculating the difference between each pressure data point and the previous one. If the magnitude of the change exceeds a preset pressure jump threshold (e.g., 2.0 kPa), the point is recorded as a pressure jump point. After determining the pressure jump point, the system calculates the time difference between the timestamp of that point and the timestamp of the first thermal jump point in the current cycle segment to determine the time interval between the thermal response and the pressure response. This interval is defined as the thermal-pressure time difference for that cycle segment. The module sequentially stores the time difference of each cycle segment in a time difference sequence and then checks whether the sequence contains a monotonically increasing characteristic. The judgment process sequentially compares whether the current time difference is greater than the time difference of the previous cycle. If all time difference values ​​show a sequentially increasing trend, the thermal cycle group is identified as a group with increasing time difference. For example, in a test system, four cycle segments were sequentially obtained, with pressure mutation points detected at 1.0 seconds, 1.3 seconds, 1.7 seconds, and 2.4 seconds after the occurrence of the thermal sensitivity mutation point. This sequence constitutes a continuously increasing group. The system applies this judgment logic to all detected cycle groups and classifies those that meet the increasing time difference rule for subsequent thermal and pressure characteristic calculation and state identification. This screening process effectively eliminates cycle segments with no significant linkage characteristics in the thermal and pressure responses, focusing on key cycles with clear response characteristics of heating followed by pressure, thereby improving the targetedness and efficiency of the entire analysis system. This process is applicable to thermal and pressure response modeling scenarios for various dynamic conditions, such as fluid transport monitoring, reactor temperature control, and geothermal energy transfer.

[0073] The mutation cycle classification submodule obtains the time difference, mutation amplitude and rate of each segment based on the time difference increasing cycle group, and performs normalization operation processing using the formula:

[0074]

[0075] The thermal pressure mutation change of the period is obtained by calculation, and the state is classified according to the change interval to generate the thermal pressure time difference mutation result;

[0076] Among them, D represents the thermal pressure mutation change, d represents the thermal pressure time difference within the cycle, a represents the thermal pressure mutation amplitude, and v represents the mutation rate. is the mean rate, T is the average duration of the cycle;

[0077] The thermal pressure mutation variation is an indicator that measures the strength of the mutation correlation between temperature changes and pressure responses within a certain cycle of the polyethylene chlorination process. It comprehensively considers factors such as the time difference between the temperature mutation point and the pressure mutation point, the amplitude and rate of the pressure mutation, and the average duration of the entire cycle. After normalization, it reflects the significance and abnormality of the thermal pressure response within the cycle.

[0078] Dimensional normalization process of formula parameters:

[0079] The heat pressure time difference d, mutation amplitude a, rate v and average rate The cycle duration T is scaled according to the maximum value and statistical characteristic value and converted into dimensionless;

[0080] The mutation cycle classification submodule obtains the hot pressure time difference d, mutation amplitude a, mutation rate v, and rate mean within each period based on the time difference increasing cycle group. And the average duration of the cycle T, and after normalization, enter the following formula:

[0081]

[0082] The parameters are explained as follows: d represents the time difference between the pressure mutation point and the thermal mutation point in the current cycle, in seconds; a represents the pressure change amplitude in the current cycle, in kPa; v represents the rate at which the pressure mutation occurs (the pressure change divided by the duration of the change). is the average value of the pressure mutation rate of all cycle segments; T represents the length of the cycle segment in seconds;

[0083] Take sample 1 data and assume its parameters are:

[0084] d = 4.2s, obtained by the difference between the pressure mutation point 208.2s and the thermal sensitivity mutation point 204.0s;

[0085] a=8.5kPa, obtained by increasing the pressure from 102.3 to 110.8;

[0086] v = 2.1 kPa / s, which is derived from the fact that an 8.5 kPa change takes 4.0 seconds.

[0087] is the average rate of the current sample group;

[0088] T = 12.0s, the length of the current thermal cycle;

[0089] The calculation is as follows:

[0090] 1. Calculate the product:

[0091] d·a=4.2×8.5=35.7;

[0092] 2. Calculate the absolute value of the rate difference and the square root of the period:

[0093]

[0094] 3. Combine the denominators and complete the main formula:

[0095]

[0096] The results show that the comprehensive change of thermal stress mutation in sample 1 is 10.02. If the state classification threshold is set as follows:

[0097] D<6: stable;

[0098] 6≤D<9: transitional type;

[0099] D≥9: mutant type;

[0100] According to the calculation results, sample 1 can be classified as a "mutation type" period segment. This classification will form a thermal pressure time difference mutation result set in the subsequent output stage and be used for subsequent decision-making of the system response logic. The benefit of the formula is that by combining d, a, v, The five parameters T eliminate the interference of cycle length and rate differences on mutation identification, making the mutation type identification more unified and distinctive.

[0101] See also Figure 3 , the viscosity speed change identification module includes:

[0102] The cycle marking submodule marks the time interval number of the mutation segment based on the thermal pressure time difference mutation result, maps the corresponding number to the bottom viscosity measurement point data, verifies the cycle attribution according to the measurement point recording time, and generates a cycle viscosity data set;

[0103] First, the system numbers the thermal pressure mutation segments output in the previous stage according to time. Assuming that the mutation segments identified in a certain acquisition are from 105 seconds to 135 seconds, 150 seconds to 190 seconds, and 205 seconds to 235 seconds, the system numbers them P1, P2, and P3 in sequence. A period identification table is established for each number. Next, the original viscosity data in the complete time series is extracted from the bottom viscosity acquisition sensor. For example, a certain measuring point is collected at intervals of 0.5 seconds, and the sample is obtained as follows:

[0104] t=106.0s, η=185.3mPa·s;

[0105] t=152.5s, η=188.1mPa·s;

[0106] For each data point, the system extracts the timestamp of each sampling record and compares it with the time interval of the above mutation segment one by one, that is, to determine whether the time value meets the interval inequality condition

[0107] t≥t start ∧t≤t end ;

[0108] If the conditions are met, it will be marked as periodic data with the corresponding number. For example, if the timestamp 106.0 seconds is within the interval P1, it belongs to the P1 periodic segment, and the time 152.5 seconds belongs to P2. And so on. The entire measurement point data set is traversed, and each data point is marked with the period number to which it belongs. The final data structure is {number, timestamp, viscosity value}, for example, {P1, 106.0s, 185.3mPa·s}, {P2, 152.5s, 188.1mPa·s}. The measurement point matching and number attribution operations in all segments are repeated. After the system is sorted out, a matching set of period numbers and viscosity data is formed to construct a periodic viscosity data set for subsequent trend extraction and processing.

[0109] The direction judgment submodule calls the periodic viscosity data set, judges the change trend of the viscosity values ​​in adjacent periods, retains the sequence segments with consistent change directions, sequentially identifies the period segments with continuous unidirectional changes, and generates a set of unidirectional continuation segments;

[0110] First, all viscosity sampling values ​​in each period are aggregated according to the period number, and the mean viscosity value in the period is calculated. For example, the sample viscosity sequence in period P1 is {185.3, 185.9, 186.0, 185.5}, ​​then its mean is The period P2 is {188.1, 187.6, 188.3, ​​188.0}, and the mean is Similarly, obtain the viscosity mean sequence of multiple consecutive period segments, perform difference calculation on the viscosity mean values ​​of two adjacent periods, and determine their change direction. The difference calculation formula is: If the result is greater than zero, it is a positive change; if it is less than zero, it is a negative change; if it is equal to zero, it is marked as no change. In this embodiment, The direction marker is "positive", the system records the trend direction of the change in the entire cycle number set, and obtains the direction sequence such as {+, +, -, +, -}, and then filters out the continuous segment with the same direction, for example, {+, +} constitutes a continuous positive segment, and {-} is marked as a reverse segment alone. During the direction marking process, once the direction change is detected, the current segment is truncated, and a new segment recording is started. Finally, the system identifies multiple continuous unidirectional change cycle segments, and records their start and end points according to the original cycle number. For example, segment S1 corresponding to cycle {P1, P2} is a continuous positive segment, and segment S2 is {P3} which is a reverse segment. The system organizes these continuous unidirectional cycle segments into a unidirectional continuation segment set, which is used as the input basis for subsequent viscosity trend marking.

[0111] The trend marking submodule extracts the viscosity cumulative change value, cycle length and local fluctuation coefficient according to the unidirectional continuation segment set, judges the unidirectional reinforcement feature, and uses the formula:

[0112]

[0113] The operation obtains the viscosity speed ratio value of each segment, maps the ratio interval to the trend type number, and generates a viscosity speed set of segments;

[0114] Wherein, w represents the viscosity speed ratio value, m i represents the viscosity increment of the i-th cycle, z j represents the viscosity fluctuation value in the j-th cycle, q represents the total number of cycles in the segment, h represents the viscosity fluctuation amplitude set in the cycle, and k represents the number of cycles in the current unidirectional continuation segment.

[0115] The viscosity speed ratio value is used to describe whether the unidirectional change trend of viscosity in the continuous cycle has the enhancement feature. It combines the increment, fluctuation amplitude and cycle number of the viscosity change of each cycle, calculates a ratio value, and is used to judge the stability and trend intensity of the viscosity change in the segment, so as to identify different types of change modes such as rapid enhancement, stable enhancement or fluctuation repetition;

[0116] The trend marking submodule receives the unidirectional continuation segment set generated in the previous step, executes the calculation process for each cycle in each segment, and obtains the viscosity speed ratio value w, and the calculation formula is:

[0117]

[0118] The parameters are explained as follows: k is the number of cycles contained in the continuation segment, m i represents the viscosity difference between the beginning and the end of the i-th cycle, that is, the viscosity increment, z jIt represents the sum of the absolute deviations of all sampled viscosities from the mean value in the jth period, q represents the total number of periods in the current segment, h represents the maximum viscosity fluctuation amplitude set within each period, and max(h) represents its maximum value. The following takes segment S1 as an example. S1 contains three periods: P1, P2, and P3. The sampling data are as follows:

[0119] P1 initial and final viscosity: 185.3→186.0, m1=0.7, viscosity sampling values ​​{185.3, 185.9, 186.0, 185.5}, ​​mean 185.675, deviation z1=|185.3-185.675|+|185.9-185.675|+|186.0-185.675|+|185.5-185.675|=0.375+0.225+0.325+0.175=1.1, maximum fluctuation range h1=0.7

[0120] P2 initial and final viscosity: 188.1→188.0, m2=-0.1, viscosity sampling values ​​{188.1, 187.6, 188.3, ​​188.0}, mean 188.0, deviation z2=0.1+0.4+0.3+0=0.8, fluctuation range h2=0.7

[0121] P3 initial and final viscosity: 187.0→186.5, m3=-0.5, sampling values ​​{187.0, 186.6, 186.4, 186.5}, mean 186.625, deviation z3=0.375+0.025+0.225+0.125=0.75, fluctuation range h3=0.6

[0122] Substitute the data into the formula:

[0123] Find the square of the sum of the viscosity increases:

[0124] (∑m i ) 2 =(0.7-0.1-0.5) 2 =0.1 2 =0.01;

[0125] Calculate the total deviation and the absolute value:

[0126] |∑z j |=|1.1+0.8+0.75|=2.65;

[0127] Number of cycles: q = 3, maximum fluctuation amplitude: max(h) = 0.7, the denominator is:

[0128]

[0129] The final calculated ratio is:

[0130]

[0131] The results show that the viscosity change direction is continuous but fluctuates significantly, and the viscosity speed ratio is negative. The ratio judgment criteria are set as follows:

[0132] w>10: stable enhanced type;

[0133] 0<w≤10: slowly increasing type;

[0134] -5≤w≤0: Fluctuation type;

[0135] w<-5: rapid reversal type;

[0136] The calculated result, w≈-1.086, classifies this section as "fluctuation type," which is used to output the viscosity-shifting concentrated section number and type. The formula's innovation lies in the introduction of a trend-enhancing term based on the square of the viscosity increase, and the inclusion of fluctuations, total number of cycles, and local maximum amplitude as balancing factors. This allows the result to comprehensively reflect the trend and stability of viscosity changes.

[0137] See also Figure 4 , the chlorine pressure synergistic trend division module includes;

[0138] The data retrieval submodule locates the start and end time points of each cycle segment based on the cycle range marked in the viscosity speed concentration section, retrieves the chlorination advancement curve and pressure fluctuation sequence within the corresponding period, and synchronizes the chlorine pressure data to generate a periodic chlorine pressure data set;

[0139] First, the original viscosity time series data needs to be extracted. During the data analysis process, all viscosity measurement values ​​are sorted in chronological order according to the acquisition time to form a complete sequence, and the viscosity difference between adjacent time nodes is calculated. Then, the viscosity change rate sequence is obtained. The unit of this sequence is the viscosity change value per minute. For example, when the viscosity is 42mPa·s at 10:00 and 48mPa·s at 10:01, the change rate is 6mPa·s / min. The system divides the segments according to the duration of each change rate. When the continuous change rate is greater than 5mPa·s / min, the segment is preliminarily identified as the starting point of the speed change segment. If the rate continues for more than 3 minutes, the starting point is confirmed to be valid and marked as the beginning of a cycle segment until the change rate drops to less than 2

[0140] mPa·s / min and lasts for two minutes, the cycle segment is considered to have ended, and the segment is finally regarded as a valid cycle segment; when setting the above rate threshold, reference is made to the typical data of viscosity fluctuations in historical wellbore operations, in which the general background change rate is concentrated within 2mPa·s / min, while the fluctuation peak caused by the operation can usually reach 5-7mPa·s / min. Therefore, 5mPa·s / min is set as the threshold for judging the starting point of the speed change segment, and 2mPa·s / min is set as the end judgment value, and the time thresholds are set to 3 minutes and 2 minutes respectively to filter out short-term pulse disturbances; after obtaining the cycle segment, the chlorination advancement curve data and pressure fluctuation sequence are called according to the start and end time windows of the cycle segment. The specific calling operation is carried out by matching the conductivity monitoring data within the time range (as the chlorination concentration The chlorination advancement curve data is usually sampled every 30 seconds, and the pressure fluctuation sequence is sampled every 10 seconds. Therefore, time interpolation and alignment processing is required. The chlorination data are interpolated to a group of 10 seconds to make it consistent with the time axis of the pressure fluctuation data. Then, they are paired one by one according to the timestamps to generate a synchronous data point array. Each data point contains time, chlorination concentration and pressure value. Finally, all data points are combined to form a periodic chlorine pressure data set. For example, the data point sequence extracted in the segment from 12:00:00 to 12:10:00 can be: the first point is the chlorination concentration of 1010μS / cm and the pressure of 5.3MPa at 12:00:10, and the second point is the chlorination concentration of 1035μS / cm and the pressure of 5.5MPa at 12:00:20. This is recorded point by point until the end of the cycle.

[0141] The trend identification submodule calls the periodic chlorine pressure data set to identify the direction of change of the chlorination curve and pressure fluctuation in each period, compares the signs of the chlorine pressure data to determine whether it fluctuates in the same direction in consecutive periods, and generates a sequence of trend-consistent segments;

[0142] First, extract the chlorination concentration sequence and pressure value sequence of all sampling points in each cycle, and calculate the overall change direction of each cycle respectively, that is, by judging the signs of the first and last values ​​in the cycle, if the last value is greater than the first value, the corresponding change direction of the cycle is recorded as "+1", indicating an upward trend, if the last value is less than the first value, it is recorded as "-1", and if they are equal, it is recorded as "0"; for example, if the chlorination concentration rises from 950μS / cm to 1170μS / cm in a certain cycle, the chlorination trend is recorded as +1; if the pressure rises from 5.1MPa to 5.4MPa, the pressure trend is also recorded as +1; then compare the signs of the chlorination trend and the pressure trend, if the same signs are the same, it means the fluctuation is in the same direction, otherwise it is recorded as inconsistent trends; in multiple consecutive cycles, the system continuously compares the trend marks of each cycle, and if a certain period of continuous fluctuations is not consistent, the system will automatically compare the trend marks of each cycle, and if the last value is greater than the first value, the system will automatically compare the trend marks of each cycle, and if the last value is less ... If all periodic trends within a consecutive cycle have the same sign, the segment is recorded as a trend-consistent segment sequence. For example, if the chlorination and pressure trends are both +1 for four consecutive cycles, it constitutes a continuous trend-consistent sequence, and the duration is the total time of the four cycles. For the "equal" situation in trend judgment, if the numerical difference is less than the set slight fluctuation threshold, it is considered a stable trend. The threshold is set as the chlorination concentration change not exceeding ±10μS / cm and the pressure change not exceeding ±0.05MPa, which is determined based on the background fluctuation range in historical data. If it exceeds the threshold, it is considered that the trend has a clear direction. Through this cycle-by-cycle trend comparison method, it is possible to clearly identify the time segments in which the chlorination concentration and pressure show consistent fluctuations, thereby providing structured trend data for subsequent behavioral analysis.

[0143] The behavior clustering submodule extracts the time range, number of cycles and corresponding chlorine pressure fluctuation characteristic parameters of each trend segment based on the sequence of trend-consistent segments, combines them to construct segment feature expressions, and classifies segments into the same category based on trend similarity to generate a chlorine pressure collaborative behavior segment set;

[0144] The following features are extracted for each segment with consistent trend: the start and end time of the segment, the number of cycles, the average chlorination concentration change amplitude, the average pressure change amplitude, the mean and variance of the change rate in each cycle. First, the number of cycles is calculated, that is, the number of consecutive cycles with the same trend in the segment is counted. For example, if a segment with consistent trend contains 5 cycles, the number of cycles is 5; the start and end time is determined by the starting point of the first cycle and the end point of the last cycle; the average chlorination change amplitude is calculated by calculating the average of the difference between the final and initial chlorination values ​​in all cycles. For example, if the changes in the 5 cycles are 210, 180, 230, 220, and 190 μS / cm, respectively, the average is 206 μS / cm; the average pressure change amplitude is calculated in the same way; the mean of the change rate is calculated by dividing the difference between the data points within each cycle by the rate value of time, and then averaging it. The variance is calculated by the square difference method of the rate value; taking the chlorination concentration as an example, if the 5 data points within a cycle are The chlorination values ​​are 1100, 1140, 1170, 1200, and 1230 μS / cm, respectively, with a 10-second interval. The average chlorination rate for this period is (1230–1100) / 40 seconds = 3.25 μS / cm / s, which converts to 195 μS / cm / min. The characteristic rate for this period is obtained by averaging these values ​​over multiple periods. All eigenvalues ​​are combined to form a characteristic expression vector for the segment. Similarity is then determined between the vectors of all segments with consistent trends. If the differences in average rate, amplitude, and number of cycles for multiple segments are within a preset tolerance (chlorination amplitude difference less than 30 μS / cm, pressure difference less than 0.2 MPa, and cycle number difference no more than 2), they are classified as belonging to the same collaborative behavior category. Finally, similar segments are combined to generate a chlorine-pressure collaborative behavior segment set, providing a data foundation for state identification or anomaly detection in the next stage.

[0145] See also Figure 5 , the response centralized reorganization module includes:

[0146] The number alignment submodule matches the cycle numbers identified by the chlorine-pressure synergistic behavior segment set with the corresponding cycles in the viscosity-variable concentrated segment, extracts the full set of cycle numbers from the three types of data, establishes a cycle cross-reference table, and generates a multi-source cycle number set.

[0147] First, all cycle numbers are extracted from the collaborative behavior segment set. The process reads the identification field of each collaborative behavior segment, and uses the "cycle number" as the main index item to form a number list. For example, the numbers extracted from the collaborative behavior segment are: 4, 5, 6, 8, 10. The system records them in sequence and proceeds to the next step of processing; then reads the thermal pressure time difference mutation result data, and the system groups the mutation records in the data set by cycle number. Each record includes the cycle number and the mutation start time, mutation end time and mutation amplitude. The thermal pressure mutation cycle number is extracted and formed into a set. For example The hot pressure number set is: 3, 4, 5, 9, 10; compare the collaborative behavior segment number set with the thermal pressure mutation set, and the system performs a number comparison operation one by one. If the number exists in both sets, such as numbers 4 and 5, it is recorded as a valid alignment candidate; then read the viscosity speed concentrated segment number set, each data contains the time window of viscosity mutation and the cycle number to which it belongs. The system also extracts the cycle number field to form a set. For example, if the viscosity set is: 1, 2, 5, 6, 10, perform a third comparison to find the number that exists in both the chlorine pressure collaborative behavior segment and the thermal pressure mutation segment. , the cycle number of the viscosity speed change section, the screening rule is the number that appears in all three sets. If numbers 5 and 10 meet the conditions, then numbers 5 and 10 are valid cross-cycle numbers; to ensure that the cycle data with aligned numbers have time consistency, it is necessary to check whether the difference range of the time periods corresponding to the cycle numbers in the three sources is within the threshold limit. The time error tolerance is set to 30 seconds. This value comes from the maximum range statistics of the deviation of the same cycle boundary division by different data sources in previous experiments. If it exceeds 30 seconds, it is invalid alignment. If a number is 10:00 to 10 in chlorine pressure :08, the hot pressing is 10:01 to 10:09, and the viscosity is 10:00 to 10:07, and the maximum difference between the three is 60 seconds, then the number will not be included in the final set; finally, the system integrates the full set of period numbers from the three data sources to generate a full set number set, removes duplicates and arranges them in ascending order, and establishes a cross-reference table by comparing the period numbers and time periods in different sources. Each reference entry records the period number, the start and end points of the corresponding time segments of the three sources, the time difference range, and the matching result status, thereby generating a period number set with a unified structure of multi-source numbers.

[0148] The cycle intersection extraction submodule calls the multi-source cycle number set, filters the cycle numbers that exist in the three sources at the same time, locates and extracts the time segments of the intersection cycle, establishes a list of intersection segments arranged in order, and generates the response intersection cycle segment;

[0149] First, read the full set of numbers generated by the previous module and the source records corresponding to the numbers, and perform a classification check on each number item to determine whether it exists in all three types of sources; this step traverses the full set number list and performs an existence check on each number. If a number exists in the chlorine pressure collaborative behavior segment number set and also appears in the thermal pressure mutation result and viscosity speed change result numbers, then the number is considered to meet the intersection extraction condition; for example, the number set is: 3, 4, 5, 6, 7, 8, 9, 10, of which the chlorine pressure numbers are: 4, 5, 6, 10, the thermal pressure numbers are: 5, 6, 8, 10, and the viscosity numbers are: 5, 6, 10, then numbers 5, 6, and 10 are period numbers that meet the intersection condition; after the system confirms the intersection number, it extracts the corresponding start time and end time. The specific operation is to index the start and end time fields from the three types of data sources according to the period number to form a time list. For example, number 5 is from 11:00 to 11:08, hot pressing is 11:01 to 11:09, and viscosity is 11:00 to 11:07. The earliest starting point and the latest ending point of the three time periods are taken to determine the final intersection segment as 11:00 to 11:09. The total length of the time period is 9 minutes. If the length is less than the preset cycle minimum limit of 5 minutes, the number is discarded; at the same time, the coverage status of the three types of data in the segment is recorded. If a number exists in the three types of number sets, but the original measurement data is missing in the actual data segment, it is marked as an incomplete intersection. The system sorts the time periods of all valid intersection period numbers in ascending order by start time and attaches a sequential number to each intersection period segment. For example, the intersection period segments are numbered 1, 2, and 3, and the corresponding numbers are 5, 6, and 10. The number, start and end time, and data source consistency status of each segment are recorded to form a list of intersection segments arranged in chronological order. Finally, a list of data segment numbers that meet the three-source synchronization is output for subsequent data integration module calls.

[0150] The sequence summary submodule integrates the three types of raw data fragments of heat pressure, viscosity, and chlorine pressure in the corresponding segments according to the response intersection period, and uniformly numbers and proofreads the time sequence to generate a multi-source response set according to the continuous time period;

[0151] The sequence aggregation sub-module receives the response intersection period segment, and sequentially processes according to each intersection period segment number. Each period segment corresponds to a unified time window. The system performs three types of data aggregation and integration operations on the data in the time window, and extracts three types of original data segments of hot pressure, viscosity and chlorine pressure. The hot pressure data is derived from the wellbore pressure sensor, with a sampling interval of 20 seconds, and the data format is a pair of time stamp and pressure value, such as 11:00:00-1.85MPa, 11:00:20-1.88MPa, and so on. The viscosity data has a sampling interval of 60 seconds, and the data item is a pair of time point and viscosity value, such as 11:00:00-72mPa·s, 11:01:00-75mPa·s. The chlorine pressure data has a sampling interval of 10 seconds, and the data item is a pair of time point and chlorine concentration value, such as 11:00:00-1150μS / cm, 11:00:10-1170μS / cm. The data points between the start and end times of all intersection period segments are extracted in turn. The system takes the 10-second interval of the chlorine pressure data as the reference time axis, and performs interpolation processing on the hot pressure and viscosity data. The interpolation method is linear value estimation, that is, the intermediate value of two adjacent original sampling points is inserted as the estimated value in the missing time point. For example, the hot pressure at 11:00:00 and 11:00:20 is 1.85 and 1.88 MPa, respectively, and the inserted hot pressure value at 11:00:10 is 1.865 MPa. After the unified time axis is established, the estimated or original values of the three types of data at the same time point are integrated into a single record. Each record item is recorded with a unified number, and the number is numbered from the start point of the intersection period segment every 10 seconds, such as the first data point with number 1, corresponding to time 11:00:00, number 2 for 11:00:10, and so on. Finally, each response intersection period segment forms a response data set that is unified in time sequence and aligned in data source. Each record contains three numerical fields of hot pressure, viscosity and chlorine pressure. The system saves the multi-source response set number list of each period segment in a structure for subsequent use by the state identification module.

[0152] Please refer to Figure 6 The stirring amplitude control module includes:

[0153] The period matching sub-module extracts the stirring speed change records in the corresponding period segment based on the period number marked in the multi-source response set, performs time alignment and period number association processing on the speed data, and generates a period stirring data set.

[0154] The period matching sub-module is based on the annotated period number in the multi-source response set, such as the period number 1-5, and in the execution process, first determines the time stamp boundary corresponding to the period number, for example, the starting time corresponding to period 1 is 10s, and the ending time is 70s, extracts all stirring speed record data in the interval from the original record, and samples to form a time sequence with 1s as the time interval, and extracts the stirring speed record in the time period of period 2 (such as 70s-140s), period 3 (140s-210s), etc. Each data point in the extracted time sequence includes a time point and a corresponding stirring speed value, for example, in period 1, the time point 15s corresponds to the stirring speed of 300r / min, and the time point 16s corresponds to the stirring speed of 305r / min, and so on, to form a period stirring speed sequence containing 60 data points. Next, align the time sequences of each period segment on the unified time axis, take the period starting time as the reference, normalize all time points, normalize 10s to 70s to 0s to 60s, and update the stirring speed value of each point simultaneously. Then, bind the period number and the corresponding time sequence to form a mapping relationship, so that each period stirring data contains period number, normalized time, and stirring speed value three fields, for example, the data segment numbered 1 contains normalized time points 0-60s, and the corresponding stirring speed values are 300, 305, 310, etc. The data segment numbered 2 is normalized to 0-70s, and the recorded stirring speed changes are 280, 285, 290, etc. Finally, the period stirring data set is formed by aggregating, and each record in the data set is bound by time alignment and number, which guarantees the continuity and consistency of the period feature extraction in subsequent analysis.

[0155] The frequency identification sub-module calls the period stirring data set, identifies the change frequency and direction switching of the stirring speed in adjacent periods, filters the segments with adjustments in consecutive periods, establishes a variable feature label sequence, and generates a high-frequency adjustment period segment.

[0156] The frequency identification submodule analyzes the change frequency and direction switching of the stirring speed between adjacent periods after calling the periodic stirring data set. In the execution process, first, the speed change value between every two adjacent time points in the period section is calculated to determine the change direction, for example, in the normalized time points 0s to 1s of period 1, the speed increases from 300 r / min to 305 r / min, the direction is upward, and if it decreases from 305 r / min to 295 r / min in 1s to 2s, the direction is downward. On this basis, the number of direction switches in each period is counted, and the speed regulation frequency is judged. If the number of direction switches in a period section exceeds 4 times, it is considered that the period section is a high-frequency regulation section, for example, in period 3, there are 5 direction switches in the 60s period, and in period 4, there are 3 direction switches. Period 3 is a high-frequency regulation section, and period 4 is not a high-frequency section. The frequency threshold needs to be set in the frequency judgment, which is set to 4 times. The reason is that the speed adjustment frequency in the normal stirring process does not exceed once every 15s, that is, the number of switches in the 60s period should be ≤4 times, and more than that is high-frequency behavior. This threshold is obtained by statistical analysis of 100 normal running data samples. More than 90% of the period sections in the sample have a switch number of 3 times or less. Therefore, 4 times is used as the threshold. By executing the above judgment on period 1 to period 5, it can be obtained that period 1 is 3 times, period 2 is 4 times, period 3 is 5 times, period 4 is 2 times, and period 5 is 6 times. Period 3 and 5 are marked as high-frequency regulation period sections, and the final generated variable feature label sequence is {period 1: 0, period 2: 0, period 3: 1, period 4: 0, period 5: 1}, where "1" represents a high-frequency regulation section, and "0" represents a non-high-frequency section. Form a high-frequency regulation period data list for subsequent module calls.

[0157] The amplitude limiting submodule extracts the speed regulation amplitude, duration and change number according to the high-frequency regulation period section, judges whether it exceeds the preset amplitude threshold and time threshold range, divides the regulation level according to the adjustment intensity, and generates the chlorination reaction control result;

[0158] The amplitude limitation submodule extracts the speed adjustment amplitude, adjustment duration and number of adjustments in the high-frequency adjustment cycle segment, and compares them with the preset threshold item by item. The adjustment amplitude refers to the difference between the maximum speed and the minimum speed, the adjustment duration refers to the total time from the first change to the last change, and the number of adjustments is the total number of times the speed change direction is switched. In the specific implementation, take cycle 3 as an example. The maximum speed in this cycle is 340r / min, the minimum is 260r / min, the amplitude is 80r / min, the first change occurs in the 5th second, the last change is in the 55th second, the duration is 50s, and the direction is switched 5 times in total. It is compared with the threshold in turn and the amplitude threshold is set to 70r / min. This value comes from the maximum acceptable adjustment amplitude during the stable stirring process of the industrial reactor, which is generally maintained between 60 and 75r / min. / min is the median value as the boundary, the duration threshold is set to 150s, and the adjustment number threshold is 4 times. These thresholds are set by referring to the maximum tolerance values ​​of each parameter in 200 process monitoring samples. In cycle 3, the amplitude is 80r / min>70r / min, the duration is 50s<150s, and the number of adjustments is 5 times>4 times. Therefore, both conditions are met and it is classified as the medium-intensity regulation level; in cycle 5, the maximum amplitude is 100r / min, the duration is 240s, and the number of adjustments is 6 times, all of which are greater than the threshold and it is classified as the high-intensity regulation level; in cycle 2, the amplitude is 50r / min<70r / min, the time is 120s<150s, and the number of adjustments is 3 times<4 times. All three items are within the limit and it is classified as the no-significant regulation level. Finally, all judgment results are divided into intensity according to the comparison results of the three parameters with the threshold and a chlorination reaction control result data set is generated.

[0159] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A polyethylene chlorination reaction control system, characterized in that: The system comprises: The hot-pressing time difference extraction module obtains the changing time points of the thermal-sensitive data and pressure data in the polyethylene chlorination process, performs periodic comparison on the time difference between the two, extracts the mutation inflection point period, and generates the hot-pressing time difference mutation result; The viscosity change identification module marks the cycle according to the result of the thermal pressing time difference mutation, calls the viscosity recording point data, determines the direction of the cycle viscosity change, extracts and marks the section of unidirectional amplification of the change, and generates a viscosity change concentrated section; The chlorine-pressure synergistic trend division module retrieves the chlorination advancement curve and the pressure fluctuation sequence for the viscosity speed change concentrated section period, determines whether the fluctuations are in the same direction in the continuous period, marks the trend synchronization period, and generates a chlorine-pressure synergistic behavior segment set; The response concentration reorganization module is based on the period number of the chlorine pressure synergistic behavior segment set, matches the thermal pressure time difference mutation result with the viscosity speed change concentrated segment, extracts the intersection period segment of the three, and generates a multi-source response set; The stirring amplitude control module calls the multi-source response set period segment, retrieves the stirring speed change record, identifies the adjustment segment within the period, sets operation restrictions on the adjustment amplitude and duration, and generates a chlorination reaction control result.

2. The polyethylene chlorination reaction control system according to claim 1, characterized in that: The hot pressure time difference mutation result includes the time difference change amplitude, inflection point cycle characteristics, state classification label, and increasing trend sequence; the viscosity speed change concentrated section includes the viscosity change direction, continuous cycle identification, change amplitude characteristics, and concentrated section number; the chlorine pressure collaborative behavior segment set includes the chlorination curve trend, pressure fluctuation trend, trend synchronization cycle, and collaborative relationship label; the multi-source response set includes the cycle number set, trend consistency mark, intersection segment sequence, and multi-source fusion result; the chlorination reaction control result includes the stirring speed change frequency, adjustment amplitude limit value, adjustment duration setting, and stirring operation stability index.

3. The polyethylene chlorination reaction control system according to claim 1, characterized in that: The thermal pressure time difference extraction module includes: The direction of cyclic viscosity change refers to the trend direction of the viscosity value continuously rising, falling and remaining stable over time within a thermal pressure mutation cycle; The chlorination advancement curve is a time series curve reflecting the trend of chlorine content changing with time during the polyethylene chlorination reaction; The pressure fluctuation sequence is a discrete data sequence that records the pressure changes over time during the polymerization reaction, reflecting the dynamic fluctuation characteristics of the pressure; The sudden change result of the hot pressing time difference is linked with the viscosity speed change concentrated section to control the stirring amplitude.

4. The polyethylene chlorination reaction control system according to claim 1, characterized in that: The thermal pressure time difference extraction module includes: The thermal inflection point identification submodule calculates the temperature increase of adjacent data points based on the collected thermal data, determines whether it exceeds the temperature jump threshold, identifies the mutation point, divides the time period between adjacent mutation points, and generates a thermal cycle segment list; The pressure change screening submodule extracts the pressure data of the corresponding time period according to the thermal cycle segment list, calculates the time difference sequence between the pressure mutation moment and the thermal sensitive mutation point, screens the period segments with continuously increasing time differences, and generates a period group with increasing time differences; The mutation cycle classification submodule obtains the time difference, mutation amplitude and rate of each segment based on the time difference increment cycle group, performs normalization operation processing, calculates the thermal pressure mutation change of the cycle, classifies the state according to the change range, and generates the thermal pressure time difference mutation result; The temperature jump threshold refers to the limit value of the continuous temperature increase required to determine a sudden change in the thermal sensitive data; The association between the pressure mutation moment and the thermal-sensitive mutation point is to calculate the time difference between the two, revealing the sequence and linkage relationship between the thermal response and the pressure response.

5. The polyethylene chlorination reaction control system according to claim 4, characterized in that: The viscosity speed change identification module includes: The cycle marking submodule marks the time interval number of the mutation segment based on the hot pressing time difference mutation result, maps the corresponding number to the bottom viscosity measurement point data, verifies the cycle attribution according to the measurement point recording time, and generates a cycle viscosity data set; The direction judgment submodule calls the periodic viscosity data set, judges the change trend of the viscosity values ​​in adjacent periods, retains sequence segments with consistent change directions, sequentially identifies period segments with continuous unidirectional changes, and generates a set of unidirectional continuation segments; The trend annotation submodule extracts the viscosity cumulative change value, cycle length and local fluctuation coefficient based on the one-way continuous segment set, determines the one-way reinforcement characteristics, calculates and obtains the viscosity speed ratio of each segment, maps the ratio interval to the trend type number, and generates the viscosity speed concentrated segment; Definition of the change trend: A change trend refers to a pattern in which the viscosity values ​​continuously change in the same direction within adjacent periods; Definition of the One-Way Strengthening Feature The one-way strengthening feature refers to a trend phenomenon in which the viscosity changes cumulatively, fluctuates, and lasts for a long period during a continuous one-way viscosity change process.

6. The polyethylene chlorination reaction control system according to claim 5, characterized in that: The chlorine pressure collaborative trend division module includes: The data retrieval submodule locates the start and end time points of each cycle segment based on the cycle range marked in the viscosity speed change concentrated section, retrieves the chlorination advancement curve and pressure fluctuation sequence in the corresponding time period, and performs periodic synchronous sorting on the chlorine pressure data to generate a periodic chlorine pressure data set; The trend identification submodule calls the periodic chlorine pressure data set, identifies the direction of change of the chlorination curve and pressure fluctuation within the period, compares the signs of the chlorine pressure data, determines whether the fluctuations in consecutive periods are in the same direction, and generates a trend consistent segment sequence; The behavior clustering submodule extracts the time range, number of cycles and corresponding chlorine pressure fluctuation characteristic parameters of each trend according to the trend consistent segment sequence, combines them to construct segment features, classifies the segments into the same category according to the trend fitting degree, and generates a chlorine pressure coordinated behavior segment set; The definition of the changing direction of the chlorination curve and pressure fluctuation within the cycle refers to the overall trend of the chlorination curve and pressure fluctuation rising, falling and remaining flat within a cycle; The definition of the segment characteristics refers to a set of structured information describing a trend segment using trend time range, number of cycles and chlorine pressure fluctuation characteristic parameters; The sign comparison is performed by marking the trend signs of the chlorination curve and the pressure fluctuation, and screening the periodic segments with the same direction between the two.

7. The polyethylene chlorination reaction control system according to claim 6, characterized in that: The response centralized reorganization module includes: The number alignment submodule matches the hot-pressing time difference mutation results with the corresponding cycles in the viscosity speed change concentrated section based on the cycle numbers identified by the chlorine pressure synergistic behavior segment set, extracts the full set of cycle numbers in the three types of data, establishes a cycle cross-reference table, and generates a multi-source cycle number set; The period intersection extraction submodule calls the multi-source period number set, screens period numbers that exist simultaneously in the three sources, locates and extracts the time segments of the intersection period, establishes a list of intersection segments arranged in order, and generates a response intersection period segment; The sequence summary submodule integrates the three types of raw data fragments of heat pressure, viscosity, and chlorine pressure in the corresponding segments according to the response intersection period segment, and uniformly numbers and proofreads the time sequence, and generates a multi-source response set according to the continuous time period; The period intersection extraction submodule and the sequence summary submodule screen the intersection of period numbers in the three types of sources and locate their time segments, integrate the raw data segments of heat pressure, viscosity and chlorine pressure, and unify the time numbers.

8. The polyethylene chlorination reaction control system according to claim 7, characterized in that: The stirring amplitude control module includes: The cycle matching submodule extracts the stirring speed change records within the corresponding cycle segment based on the cycle number marked in the multi-source response set, performs time alignment and cycle number association processing on the speed data, and generates a periodic stirring data set; The frequency identification submodule calls the periodic stirring data set, identifies the frequency of changes in stirring speed and direction switching in adjacent cycles, screens the sections with adjustments in continuous cycles, establishes a change feature marker sequence, and generates high-frequency adjustment period segments; The amplitude limiting submodule extracts the speed adjustment amplitude, duration and number of changes according to the high-frequency adjustment period, determines whether it exceeds the preset amplitude threshold and time threshold range, divides the control level according to the adjustment intensity, and generates the chlorination reaction control result; The time alignment of the speed data and the association of the cycle numbers refers to dividing the time series data of the stirring speed according to the start and end time of each cycle and marking the corresponding cycle number; The definition of the segment with high frequency adjustment in the continuous cycle is the time segment in which the stirring speed changes and the number of direction switching times in adjacent cycles; The definition of the preset amplitude threshold and time threshold range refers to the setting parameters of the speed change value limit and duration limit used to determine whether the regulation is too strong; The specific numerical range of the amplitude threshold is set to be the amplitude of the stirring speed change exceeding ±50RPM is considered as strong regulation behavior.

Images