Temperature curve optimization control method for three-stage temperature polycondensation of coal tar modification process

Abstract

The present application relates to the field of automatic control technology, and discloses a temperature curve optimization control method for three-stage temperature-variable thermal polycondensation in a coal tar modification process, reaction process data are collected to form a competitive reaction state vector, a polycondensation activity index and a thermal cracking risk equivalent are generated according to the competitive reaction state vector, and a polycondensation progress deviation is determined by the difference between the polycondensation activity index and the thermal cracking risk equivalent; when the deviation indicates that the polycondensation lags behind and the opening of the regulating valve reaches the upper limit, the deviation is used to calculate a pre-compensation time length, the starting judgment temperature of the uniform temperature period is shifted downward to enter the quasi-constant temperature heating mode in advance; in the quasi-constant temperature heating mode, different signs and amplitudes of the heating power increment are given according to the real-time decay rate of the deviation, and the heating power of each sampling period is dynamically adjusted until the net polycondensation difference converges; the present application can actively compensate for the polycondensation lag under the condition of limited heating rate, and avoid the increase of energy consumption caused by the passive extension of the constant temperature time.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology, specifically to a method for optimizing the temperature curve control of three-stage variable-temperature thermal condensation in the coal tar reforming process. Background Technology

[0002] Coal tar upgrading typically employs a three-stage variable-temperature thermopolymerization process. This involves setting different target temperatures and heating rates in the low-temperature induction stage, the medium-temperature main polymerization stage, and the high-temperature deep crosslinking stage to achieve directional polymerization of aromatic compounds and improve the quality of the modified asphalt. In this process, to achieve the ideal polymerization reaction, advanced control methods such as Model Predictive Control (MPC) typically use reaction kinetic models to plan a theoretically optimal heating curve offline. This requires the material temperature within the reactor to rapidly traverse the transition temperature range at a high heating rate at the inflection point where the temperature shifts from the medium-temperature to the high-temperature stage, minimizing the residence time of the material in the active temperature zone of the thermal cracking side reactions.

[0003] In the three-stage variable-temperature thermopolymerization process of coal tar, when the physical heating rate limit of the heating actuator is lower than the heating slope required by the theoretical optimal heating curve, the existing control method lacks an active means to coordinate the mismatch between the theoretical trajectory and the actual execution capability. It can only passively slow down the tracking or extend the isothermal time afterward, which leads to the material being unexpectedly retained in the transition temperature zone where thermal cracking side reactions are active. The deficiency of the polycondensation integral and the aggravation of thermal cracking side reactions are superimposed, ultimately resulting in a decrease in the coking value of the product, an increase in energy consumption, and poor batch stability. Summary of the Invention

[0004] The purpose of this invention is to provide a method for optimizing and controlling the temperature curve of three-stage variable-temperature thermal polycondensation in the coal tar reforming process, so as to solve the problems mentioned above.

[0005] The objective of this invention can be achieved through the following technical solutions: A method for optimizing and controlling the temperature curve of three-stage variable-temperature thermal condensation polymerization in the coal tar upgrading process includes the following steps: S1 collects online spectral data of materials, reactor wall temperature, material center temperature, heating power and regulating valve opening during the coal tar upgrading reaction, forming a competitive reaction state vector characterizing the current reaction state; S2, based on the real-time evolution trend of the competing reaction state vector, generates a polycondensation activity index characterizing the current degree of polycondensation reaction advancement and a thermal decomposition risk equivalent characterizing the degree of accumulation of decomposition side reactions; S3. Compare the real-time difference between the polycondensation activity index and the thermal decomposition risk equivalent, and determine the polymerization progress deviation of the current batch relative to the theoretical optimal trajectory according to the direction and magnitude of the difference deviating from the preset ideal process curve. S4. When the polymerization progress deviation indicates that the actual polycondensation is behind the preset process and the opening of the regulating valve reaches the upper limit, the equivalent reaction time that needs to be compensated in advance is calculated based on the deviation, and the temperature uniformity period of the static constant temperature range is determined. The starting temperature of the temperature uniformity period is shifted downward to enter the quasi-constant temperature heating mode in advance. S5, in the quasi-constant temperature heating mode, dynamically adjusts the sequence of micro-heating power applied in each sampling period according to the real-time decay rate of the polymerization progress deviation, until the cumulative difference between the polycondensation activity index and the thermal decomposition risk equivalent converges to the preset target range, thus completing the online optimization control of the entire temperature change curve.

[0006] As a further aspect of the present invention: the process for generating the polycondensation activity index is as follows: Extract the continuous absorbance change sequence of online spectral data in the preset characteristic wavenumber range from the competitive reaction state vector, and construct a multi-wavelength response trajectory reflecting the degree of aromatic ring condensation in time sequence; Based on the rate of change of absorbance at each wavenumber in the multi-wavelength response trajectory, the wavenumber response values ​​at the same moment are converted into instantaneous condensation contribution values ​​with uniform dimensions according to the catalytic ratio of temperature to reaction rate. The instantaneous condensation contribution value at each moment is accumulated along the reaction process, and the current degree of progress of the condensation reaction is characterized by the proportion of the increment of the accumulated value within the preset time window to the estimated total increment of the whole process. The output is the condensation activity index.

[0007] As a further aspect of the present invention: the process for generating the thermal decomposition risk equivalent is as follows: The real-time difference between the material center temperature and the reactor wall temperature is synchronously extracted from the competing reaction state vector, and the real-time difference is accumulated along the time axis to obtain the accumulated temperature gradient reflecting the degree of local overheating in the reactor. Based on the numerical range of the accumulated temperature gradient, a preset acceleration threshold for the pyrolysis side reaction is matched, and the absorbance shift of the online spectral data in the competitive reaction state vector at the characteristic wavenumber characterizing side chain breakage is used as the pyrolysis induction factor. The pyrolysis inducing factor and the cumulative temperature gradient are weighted together according to a preset correspondence, and the output is the thermal pyrolysis risk equivalent.

[0008] As a further aspect of the present invention: the process for setting the acceleration threshold of the pyrolysis side reaction is as follows: The temperature rise asynchrony between reactor wall temperature and material center temperature during the heating phase is extracted from the competitive reaction state vector, and the slope of the temperature rise asynchrony change over multiple consecutive sampling periods is used as a trend characterization of reactor coking state. By inputting the trend characterization quantity into the rapid response mapping relationship, the critical temperature difference value that significantly aggravates thermal decomposition under the current coking state is matched. Based on the critical temperature difference, and combined with the absorbance baseline drift rate at the characteristic wavenumber characterizing side chain rupture using the online spectral data of the current batch, the value is corrected in real time and output as the acceleration threshold of the cracking side reaction corresponding to the current batch.

[0009] As a further aspect of the present invention: the output process of the polymerization progress deviation is as follows: The difference between the current polycondensation activity index and the thermal decomposition risk equivalent is used as the real-time net polycondensation equivalent. The real-time net polycondensation equivalent is synchronously mapped to the theoretical net polycondensation equivalent at the same moment as defined by the preset ideal process curve, and the instantaneous deviation between the two is calculated. Extract a sequence of instantaneous deviation values ​​along the time axis, including the current moment, for multiple consecutive sampling periods. Based on the continuous increase or decrease trend of the instantaneous deviation values ​​in the sequence, determine whether the current deviation belongs to the condensation lag-dominated type or the cracking advance-dominated type. The judgment result and the final amplitude of the instantaneous deviation value sequence are converted into cumulative lag according to a preset correspondence, and the output is the aggregation progress deviation.

[0010] As a further aspect of the present invention: the initial determination temperature of the temperature equalization period is shifted downward to enter the quasi-constant temperature rise mode earlier, specifically including: Using the cumulative lag included in the polymerization progress deviation as input, and based on the conversion relationship between lag equivalent and compensation time, the pre-compensation time required to catch up with the polycondensation progress at the current heating rate is calculated. Based on the pre-compensation duration and the actual heating rate of the current sampling period, the compensation start temperature point that needs to be triggered in advance is calculated in reverse along the temperature coordinate axis, and the compensation start temperature point is used as the updated judgment temperature. The updated judgment temperature replaces the original starting judgment temperature of the uniform temperature period, and when the material center temperature first reaches the updated judgment temperature, it triggers the entry into the quasi-constant temperature rise mode.

[0011] As a further aspect of the present invention: the calculation process for the pre-compensation duration is as follows: Based on the cumulative lag included in the aggregate schedule deviation, the segment interval where the cumulative lag is located is matched from the preset multi-segment compensation mapping table, and the basic compensation duration corresponding to the segment interval is read. The actual heating rate of the material center temperature in the current sampling period is compared with the preset standard heating rate. Based on the ratio of the two, the basic compensation time is scaled in the same direction to obtain the compensation time after rate correction. Obtain the expected temperature range corresponding to the compensation duration, and determine the overlap between the expected temperature range and the temperature range where the acceleration threshold of the pyrolysis side reaction corresponding to the thermal pyrolysis risk equivalent is located. Based on the overlap, the compensation duration is reduced twice and output as the pre-compensation duration.

[0012] As a further aspect of the present invention: the dynamic adjustment of the applied micro-heating power sequence within each sampling period specifically includes: The decrease in the aggregation progress deviation in each sampling period is used as the real-time decay rate. The real-time decay rates of multiple consecutive sampling periods are arranged in chronological order to form a decay rate sequence. The direction of the end of the decay rate sequence determines whether the current deviation is in the accelerated decay region, the constant decay region, or the decelerated decay region. Based on the different decay regions in which the decay rate sequence is located, the positive and negative signs and amplitude levels of the heating power increment within the current sampling period are assigned respectively to generate the basic power adjustment amount; The basic power adjustment is superimposed with the actual heating power of the previous sampling period to obtain the heating power of the current sampling period, and the heating actuator is controlled according to the heating power.

[0013] The beneficial effects of this invention are: (1) This invention converts temperature setpoint tracking into a dynamic compensation control strategy driven by reaction progress deviation. When the opening of the regulating valve reaches the upper limit and the actual heating rate cannot meet the theoretical requirements, it can actively shift the initial judgment temperature of the uniform temperature period downward and enter the quasi-constant temperature heating mode. It compensates for the deficiency of polycondensation integral by "trading time for space". Compared with the traditional scheme that simply relies on anti-integral saturation or passively extends the constant temperature time, it not only avoids the control failure caused by the heating actuator being in a saturated state for a long time, but also enables the controller to maintain effective control capability under the physical rate limit, and stably converges the cumulative difference between the terminal polycondensation activity index and the thermal decomposition risk equivalent to the target range.

[0014] (2) Based on the real-time decay rate of the polymerization progress deviation in the accelerated, constant or decelerated decay zone, the present invention assigns different signs and amplitude levels to the heating power increment of the current sampling period, so that the adjustment range of the heating power adapts to the state of deviation decay. When the deviation decays at an accelerated rate, the heating power is actively reduced to avoid overshoot. When the deviation decays at a decelerated rate, the heating power is actively increased to maintain the decay momentum. Compared with the conventional PID control method with fixed gain or single rule, it can effectively suppress the frequent oscillation and overshoot of heating power, control the temperature control fluctuation range within 0.5 degrees Celsius per minute, and reduce the mechanical wear and thermal shock caused by the frequent large-scale movement of the actuator. Attached Figure Description

[0015] The invention will now be further described with reference to the accompanying drawings.

[0016] Figure 1 This is a flowchart of the control method of the present invention; Figure 2 This is a flowchart of the process for generating the polycondensation activity index in this invention; Figure 3 This is a flowchart illustrating the process of generating the thermal decomposition risk equivalent in this invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 As shown, this invention provides a method for optimizing and controlling the temperature curve of a three-stage variable-temperature thermal polycondensation process in coal tar upgrading, comprising the following steps: S1 collects online spectral data of materials, reactor wall temperature, material center temperature, heating power and regulating valve opening during the coal tar upgrading reaction, forming a competitive reaction state vector characterizing the current reaction state; S2, based on the real-time evolution trend of the competing reaction state vector, generates a polycondensation activity index characterizing the current degree of polycondensation reaction advancement and a thermal decomposition risk equivalent characterizing the degree of accumulation of decomposition side reactions; S3. Compare the real-time difference between the polycondensation activity index and the thermal decomposition risk equivalent, and determine the polymerization progress deviation of the current batch relative to the theoretical optimal trajectory according to the direction and magnitude of the difference deviating from the preset ideal process curve. S4. When the polymerization progress deviation indicates that the actual polycondensation is behind the preset process and the opening of the regulating valve reaches the upper limit, the equivalent reaction time that needs to be compensated in advance is calculated based on the deviation, and the temperature uniformity period of the static constant temperature range is determined. The starting temperature of the temperature uniformity period is shifted downward to enter the quasi-constant temperature heating mode in advance. S5, in the quasi-constant temperature heating mode, dynamically adjusts the sequence of micro-heating power applied in each sampling period according to the real-time decay rate of the polymerization progress deviation, until the cumulative difference between the polycondensation activity index and the thermal decomposition risk equivalent converges to the preset target range, thus completing the online optimization control of the entire temperature change curve.

[0019] In S1, online spectral data of materials, reactor wall temperature, material center temperature, heating power, and regulating valve opening are collected during the coal tar upgrading reaction to form a competitive reaction state vector characterizing the current reaction state, specifically including: In this embodiment, the modification reaction is carried out in a continuous modification device consisting of a prepolymerization reactor, a polycondensation reaction tower and a vacuum devolatilization tower connected in series. The reactors are connected by a high-temperature asphalt pump and insulated pipelines, and the material flows through each reaction unit in sequence under nitrogen sealing protection.

[0020] Online spectral data acquisition was conducted using three near-infrared spectrometers installed at the outlet pipeline of the prepolymerization reactor, the bottom circulation pipeline of the polycondensation reaction tower, and the bottom outlet pipeline of the vacuum devolatilization tower. At each acquisition point, the liquid material was driven by a circulation pump through the flow cell of the spectrometer. The spectrometer recorded the full-band transmission spectrum of the material within the wavenumber range at a sampling frequency of once per minute, and output the absorbance values ​​corresponding to each wavenumber point in real time as the raw spectral response values ​​characterizing the degree of aromatic ring condensation and side chain breakage in the material.

[0021] Reactor wall temperatures are collected using multiple thermocouples pre-embedded in the outer walls of the prepolymerization reactor, the polycondensation reaction tower, and the vacuum devolatilization tower. Three thermocouples are evenly distributed axially along the outer wall of each reactor. The measuring ends of each thermocouple are tightly fitted to the wall surface and covered with an insulation layer to reduce environmental heat loss. The millivolt signals output by each thermocouple are converted to Celsius temperature values ​​by a temperature transmitter and then uploaded to the data acquisition unit. The arithmetic mean of the measurements taken by the three thermocouples on the same reactor at the same time is taken as the wall temperature of the reactor in the current sampling period.

[0022] The core temperature of the material was collected by inserting sheathed thermocouples into the prepolymerization reactor, polycondensation reaction tower, and vacuum devouring tower. Each sheathed thermocouple extends vertically into the reactor from the top flange, with its measuring end located at the geometric center of the effective volume of each reactor. Each thermocouple is protected by a stainless steel tube. Each thermocouple outputs the core temperature value of the material in the corresponding reactor at the same sampling frequency as the wall temperature thermocouple, with the sampling period uniformly set to one second.

[0023] The heating power is acquired by reading the output signals of the power transmitters of the electric heating elements of the heating jacket of the prepolymer reactor, the heating coil of the polycondensation reaction tower, and the heating system of the vacuum devouring tower. Each power transmitter measures the current and voltage flowing through the corresponding electric heating element in real time, and the product of the two is taken as the instantaneous heating power of each heating unit. This instantaneous value is then transmitted to the data acquisition unit in the form of an analog signal.

[0024] The control valve opening is acquired by reading the valve position feedback signals of each control valve located on the inlet pipe of the heating medium of the prepolymerization reactor, the inlet pipe of the heating medium of the polycondensation reaction tower, and the inlet pipe of the heating medium of the vacuum devouring tower. After receiving the opening command from the controller, each control valve changes the valve core position. Each valve position feedback device converts the actual displacement of its valve core into a standard electrical signal output. The ratio of each signal to the standard signal value corresponding to the fully open valve is the current opening of each control valve.

[0025] At the end of each sampling cycle, the data acquisition unit arranges and combines the online spectral data, reactor wall temperature, material center temperature, heating power, and regulating valve opening of the same reaction unit acquired at the same time in a preset order to generate a competitive reaction state vector for the corresponding reaction unit in that sampling cycle, and stores it in the data buffer for subsequent steps to call. Specifically, this includes: At the end of each sampling cycle, the data acquisition unit combines the five types of data acquired at the same time in the same reaction unit, namely online spectral data, reactor wall temperature, material center temperature, heating power and regulating valve opening, according to the preset arrangement order corresponding to the prepolymerization reactor, polycondensation reaction tower and vacuum devouring tower. The preset arrangement order is to place the characteristic wavenumber absorbance values, reactor wall temperature, material center temperature, heating power, and control valve opening percentage values ​​contained in the online spectral data in sequence. The characteristic wavenumber absorbance values ​​in the online spectral data are arranged in ascending order of wavenumber, and together with the temperature, power, and opening values, they form an array composed of multi-dimensional values. The data acquisition unit marks the timestamp of the sampling period and stores it together with the time sequence number corresponding to the sampling period in the data buffer. This array is the competitive reaction state vector of the reaction unit in the sampling period.

[0026] Please see Figure 2 or Figure 3 As shown, in S2, based on the real-time evolution trend of the competing reaction state vector, a polycondensation activity index characterizing the current degree of polycondensation reaction advancement and a thermal decomposition risk equivalent characterizing the degree of accumulation of decomposition side reactions are generated, specifically including: The generation process of the polycondensation activity index is carried out as follows: online spectral data in the competitive reaction state vector under the current sampling period is extracted, and continuous absorbance values ​​in the wavenumber range from the reciprocal of 1600 cm² to 1000 cm² are extracted. The absorbance reading of each wavenumber point in the current sampling period is subtracted from the absorbance reading of the same wavenumber point in the previous sampling period to obtain the absorbance change of each wavenumber point. Then, the absorbance change of each wavenumber point is divided by the time interval between the previous sampling period and the current sampling period to obtain the absorbance change rate of each wavenumber point.

[0027] The absorbance change rates at each wavenumber point are arranged in ascending order of wavenumber to obtain the multi-wavelength response trajectory for that sampling period. Based on this, the absorbance change rate corresponding to each wavenumber point in the multi-wavelength response trajectory is multiplied by the temperature-catalytic conversion factor corresponding to that wavenumber point. This temperature-catalytic conversion factor is pre-obtained using the following method: Under laboratory conditions, based on the principle that the reaction rate doubles for every 10 degrees Celsius increase, and using 300 degrees Celsius as the reference temperature, the material center temperature of the current sampling period is subtracted from the reference temperature and divided by 10. The resulting value is used as the exponent, and the power value is calculated with 2 as the base to obtain the relative rate multiple at the current temperature. The preset standard wavenumber weighting coefficient for each wavenumber point is multiplied by this relative rate multiple to obtain the temperature catalytic conversion factor for each wavenumber point.

[0028] The instantaneous polycondensation contribution value for that sampling period is obtained by multiplying the absorbance change rate at each wavenumber point at the same instant by the product of its corresponding temperature catalytic conversion factor. This instantaneous polycondensation contribution value is then added to the total polycondensation contribution accumulated in the previous sampling period to update the total polycondensation contribution. Finally, the sixty instantaneous polycondensation contribution values ​​from the current sampling period and the previous sixty consecutive sampling periods are added together to obtain the increment value within a preset time window. This increment value is divided by the total increment value of the entire process, which has been pre-calibrated in the laboratory, and the resulting ratio is output as a percentage; this ratio is the polycondensation activity index for the current sampling period.

[0029] The generation process of the thermal decomposition risk equivalent is implemented as follows: The reactor wall temperature and material center temperature of the same reaction unit are extracted from the competing reaction state vector of the current sampling period. The reactor wall temperature is subtracted from the material center temperature to obtain the wall-center temperature difference for this sampling period. This temperature difference is then accumulated sequentially along the time axis with the wall-center temperature differences of the previous sampling period and all preceding sampling periods. During accumulation, the temperature difference value of each sampling period is multiplied by the time interval between the current and previous sampling periods before being added to the accumulated temperature gradient of the previous sampling period to obtain the accumulated temperature gradient for the current sampling period. The unit of the accumulated temperature gradient is degrees Celsius multiplied by seconds.

[0030] After obtaining the accumulated temperature gradient of the current sampling period, it is input into the preset cracking side reaction acceleration threshold determination process.

[0031] The determination process is as follows: the current accumulated temperature gradient is compared with a pre-stored threshold for accelerating thermal decomposition side reactions. The specific method for setting this threshold will be detailed later. If the current accumulated temperature gradient does not exceed the threshold, the current thermal decomposition risk is determined to be within the normal range; if it exceeds the threshold, the thermal decomposition risk is determined to be accelerated.

[0032] The absorbance values ​​of each wavenumber point within the reciprocal of 1000 cm² to 800 cm² are extracted from the online spectral data of the competing reaction state vector. The difference between the absorbance values ​​of all wavenumber points within this interval in the current sampling period and the previous sampling period is taken, and the change in this difference from the previous sampling period to the current sampling period is calculated. This change is used as the decomposition induction factor. After obtaining the numerical interval determination result of the accumulated temperature gradient and the value of the decomposition induction factor, the thermal decomposition risk equivalent in the current sampling period is calculated using the following formula: ; in, This represents the equivalent of thermal decomposition risk during the current sampling period. The lysis inducing factor in the current sampling period, This is the dimensionless ratio obtained by dividing the cumulative temperature gradient in the current sampling period by the baseline cumulative temperature gradient, where the baseline cumulative temperature gradient is 3600 degrees Celsius multiplied by seconds. This is the first weighting coefficient, with a value of 0.85. This is the second weighting coefficient, with a value of 0.12. The calculated... The value is output as the equivalent of thermal decomposition risk in the current sampling period.

[0033] The preset process for the acceleration threshold of the cracking side reaction is performed once before each batch of feed enters the reforming reactor and remains unchanged during the reaction of that batch.

[0034] The specific implementation method is as follows: From the competitive reaction state vector of the first 300 sampling periods during the initial heating stage of this batch of raw materials, the reactor wall temperature and material center temperature for each sampling period are extracted. The changes in reactor wall temperature and material center temperature between two adjacent sampling periods are calculated. The change in reactor wall temperature is divided by the change in material center temperature to obtain the temperature rise asynchrony for each sampling period. The temperature rise asynchrony for ten consecutive sampling periods is arranged in chronological order. The last value is subtracted from the first value, and the difference is divided by the total time occupied by the ten sampling periods to obtain the slope of change. This slope of change is used as a trend indicator of the current coking state of the reactor.

[0035] The trend characterization is expressed as the rate of change of the ratio of the Celsius difference. After obtaining the trend characterization, it is input into the accelerated response mapping relationship pre-established and stored in the data buffer. The accelerated response mapping relationship is pre-established as follows: under laboratory conditions, the same batch of raw materials is subjected to a reforming reaction at different wall-center temperature differences. The trend characterization value corresponding to the significant intensification of thermal decomposition under each temperature difference condition is recorded. A two-dimensional data table is made up of each temperature difference value and its corresponding trend characterization value, in which each temperature difference value corresponds to a trend characterization value.

[0036] The currently obtained trend indicator value is compared one by one with the trend indicator values ​​in the data table. The two adjacent data points with the smallest difference are found, and the critical temperature difference value corresponding to the trend indicator value is calculated by linear interpolation. Based on the critical temperature difference value, the absorbance values ​​of each wavenumber point in the range from the reciprocal of 1000 square centimeters to the reciprocal of 800 square centimeters within the previous 100 sampling periods are extracted from the competitive reaction state vector of the current batch. The arithmetic mean of the absorbance values ​​of each wavenumber point within the 100 sampling periods is calculated as the baseline value. Then, the absorbance values ​​of each wavenumber point in the same range under the current sampling period are taken, and the ratio of the difference between the current absorbance value and the baseline value to the baseline value is calculated to obtain the absorbance baseline drift rate. The critical temperature difference value is multiplied by one and the absorbance baseline drift rate is subtracted to obtain the temperature difference value after drift rate correction. This temperature difference value is output as the acceleration threshold of the cracking side reaction under the current state of the batch.

[0037] In S3, by comparing the real-time difference between the polycondensation activity index and the thermal decomposition risk equivalent, the polymerization progress deviation of the current batch relative to the theoretical optimal trajectory is determined according to the direction and magnitude of the deviation of the difference from the preset ideal progress curve. The difference between the polycondensation activity index and the thermal decomposition risk equivalent at the same time is taken as the real-time net polycondensation equivalent for the current sampling period.

[0038] The theoretical net polycondensation equivalent value at the same moment is read from the preset ideal process curve. The preset ideal process curve is a reference trajectory curve calculated and stored in advance before the reaction of each batch of raw materials begins, based on the initial polycondensation activity index calibration value and the initial thermal decomposition risk equivalent calibration value of that batch of raw materials, and according to the kinetic evolution law of the polycondensation reaction under the target temperature curve. The curve is plotted with time as the abscissa and net polycondensation equivalent value as the ordinate. Each time point on the curve corresponds to a theoretical net polycondensation equivalent value. The theoretical net polycondensation equivalent value at that moment is subtracted from the real-time net polycondensation equivalent value, and the absolute value of the difference is taken as the instantaneous deviation value for that sampling period.

[0039] After obtaining the instantaneous deviation value of the current sampling period, the data acquisition unit retrieves the instantaneous deviation values ​​of the last thirty sampling periods stored in the data buffer. Together with the instantaneous deviation value of the current sampling period, a total of 31 values ​​are arranged into an instantaneous deviation value sequence in the order of the sampling periods from first to last.

[0040] In the instantaneous deviation value sequence, each pair of adjacent values ​​is compared one by one, and the number of times the subsequent value is greater than the preceding value and the number of times the subsequent value is less than the preceding value are counted. If the number of times the subsequent value is greater than the preceding value exceeds the number of times the subsequent value is less than the preceding value, then the sequence is determined to have an overall upward trend; otherwise, the sequence is determined to have an overall downward trend.

[0041] After determining the overall trend of the sequence, calculate the difference between the arithmetic mean of the last five instantaneous deviation values ​​in the sequence and the arithmetic mean of the first five instantaneous deviation values ​​in the sequence. If the difference is greater than zero and the sequence as a whole shows an upward trend, the current deviation type is determined to be the cleavage-leading type; if the difference is less than zero and the sequence as a whole shows a downward trend, the current deviation type is determined to be the condensation-lag type; if neither of the above conditions is met, the current deviation type is determined to be the intermediate type.

[0042] After obtaining the determination result of the deviation type, the determination result and the terminal amplitude of the instantaneous deviation value sequence are converted into the cumulative lag: the last instantaneous deviation value in the instantaneous deviation value sequence is taken as the terminal amplitude, and the 26th to 30th instantaneous deviation values ​​in the sequence are taken as the arithmetic mean of the five instantaneous deviation values ​​as the recent mean.

[0043] If the deviation type is condensation lag-dominated, then the final amplitude is added to the recent average and multiplied by a conversion factor of 0.8, and the resulting product is taken as the cumulative lag. If the deviation type is the predominantly leading type of splitting, then the final amplitude is added to the recent average and multiplied by the conversion factor of 0.4, and the resulting product is used as the cumulative lag. If the deviation type is intermediate, the final amplitude is added to the recent average and then multiplied by a conversion factor of 0.6. The resulting product is the cumulative lag. The cumulative lag calculated for the above three deviation types is output in seconds, and this output value is the aggregation progress deviation in the current sampling period.

[0044] The calculated aggregation progress deviation is stored in the data cache for subsequent steps to retrieve sequentially according to the sampling period. The preset ideal progress curve and the values ​​of conversion factors 0.8, 0.4, and 0.6 are all pre-stored in the data cache and can be updated once before each batch of raw materials is fed in, based on the initial calibration value of that batch of raw materials.

[0045] In S4, when the polymerization progress deviation indicates that the actual polycondensation lags behind the preset process and the regulating valve opening reaches the upper limit, the equivalent reaction time requiring pre-compensation is calculated based on the deviation, and the temperature uniformity period of the static isothermal zone is determined. The starting temperature of the temperature uniformity period is then shifted downward to advance the entry into the quasi-isothermal heating mode. Specifically, this includes: The process of shifting the initial temperature during the equalization period downward to enter the quasi-constant temperature rise mode in advance is implemented as follows: Under the current sampling period, determine whether the polymerization progress deviation is greater than zero and whether the opening of the regulating valve reaches more than 95%.

[0046] If both of the above conditions are met simultaneously, it is determined that the current actual polymerization progress is lagging behind the preset process and the heating actuator is close to saturation output. At this time, the temperature downward migration procedure is triggered. After triggering this procedure, the cumulative lag amount contained in the polymerization progress deviation is extracted. This cumulative lag amount is in seconds.

[0047] The calculation process for the pre-compensation duration is carried out as follows: using the cumulative lag as input, the multi-segment compensation mapping table stored in the data buffer is retrieved.

[0048] The mapping table is constructed as follows: Under laboratory conditions, using the same batch of raw materials and the same heating rate, the cumulative lag is tested in the ranges of 0 to 30 seconds, 30 to 60 seconds, 60 to 120 seconds, and above 120 seconds. The basic compensation time required to catch up with the polycondensation progress is then determined, and a two-dimensional data table is created by mapping the endpoints of each interval to the corresponding basic compensation time. During the current sampling period, the current cumulative lag is compared with each endpoint in the mapping table to determine the segmented interval in which the cumulative lag falls, and the basic compensation time is read from the record corresponding to that segmented interval.

[0049] For example: when the cumulative lag is in the range of 60 to 120 seconds, the corresponding basic compensation duration is 90 seconds; when it is in the range of 30 to 60 seconds, the corresponding basic compensation duration is 45 seconds; when it is in the range of 0 to 30 seconds, the corresponding basic compensation duration is 20 seconds; and when it is above 120 seconds, the corresponding basic compensation duration is 180 seconds.

[0050] After obtaining the basic compensation duration, the material center temperature values ​​of the most recent ten sampling cycles within the current sampling period are retrieved. The temperature value of the tenth sampling cycle is subtracted from the temperature value of the first sampling cycle, and the resulting temperature difference is divided by the total time occupied by the ten sampling cycles to obtain the actual heating rate in the current sampling period, expressed in degrees Celsius per minute. Simultaneously, the standard heating rate, pre-stored in the data buffer, is retrieved. This standard heating rate is the target value of the average heating rate of this batch of raw materials within the third heating interval defined in the preset ideal process curve, and is set to 2.5 degrees Celsius per minute. The current actual heating rate is divided by the standard heating rate to obtain the rate ratio. This rate ratio is multiplied by the basic compensation duration, and the product is used as the compensation duration after rate correction.

[0051] After obtaining the compensation time after rate correction, this compensation time is multiplied by the actual heating rate in the current sampling period to obtain the expected temperature range value. The material center temperature in the current sampling period is added to this expected temperature range value to obtain the upper limit temperature of the expected temperature range. The material center temperature in the current sampling period is used as the lower limit temperature of the expected temperature range, thus forming the expected temperature range.

[0052] The acceleration threshold for the current batch's pyrolysis side reaction is retrieved. A preset temperature difference bandwidth value is subtracted from this threshold to obtain the lower limit, and the same value is added to it to obtain the upper limit, forming the acceleration threshold temperature range. This temperature difference bandwidth value is preset to five degrees Celsius. The overlap length between the expected temperature range and the acceleration threshold temperature range on the temperature axis is calculated. This overlap length is divided by the total length of the expected temperature range to obtain the overlap ratio. If the overlap ratio is greater than 0.3, the rate-corrected compensation time is multiplied by 0.7 to obtain the second-reduced compensation time. If the overlap ratio is less than or equal to 0.3, no reduction is performed, and the rate-corrected compensation time is directly output as the pre-compensation time.

[0053] After obtaining the pre-compensation duration, the compensation start temperature point is calculated in reverse along the temperature coordinate axis based on the pre-compensation duration and the actual heating rate under the current sampling period.

[0054] The specific method is as follows: Multiply the pre-compensation duration by the actual heating rate to obtain the temperature value that needs to be crossed for compensation. Subtract this temperature value from the original constant temperature start judgment temperature of the third high-temperature zone. The resulting temperature value is used as the compensation start temperature point, and this compensation start temperature point is used as the updated judgment temperature. The updated judgment temperature is output to the controller, replacing the original temperature of the uniform temperature period start judgment stored in the controller.

[0055] The controller continuously monitors the material's center temperature. When the material's center temperature reaches the updated judgment temperature for the first time, it triggers the entry into a quasi-constant temperature heating mode. This quasi-constant temperature heating mode refers to a heating state in which the material's center temperature rises slowly at the actual heating rate and is allowed to fluctuate within a range of 0.5 degrees Celsius per minute.

[0056] The endpoints of each interval in the above multi-segment compensation mapping table, as well as the corresponding basic compensation duration, standard heating rate, temperature difference bandwidth (5.0 degrees Celsius), overlap ratio (0.3), and reduction factor (0.7), are all pre-stored in the data cache area and are calibrated once before each batch of raw materials is fed in, based on the initial calibration value of that batch of raw materials.

[0057] In S5, in the quasi-isothermal heating mode, the sequence of micro-heating power applied in each sampling period is dynamically adjusted according to the real-time decay rate of the polymerization progress deviation, until the cumulative difference between the polycondensation activity index and the thermal decomposition risk equivalent converges to the preset target range, thus completing the online optimization control of the entire temperature change curve, specifically including: The dynamic adjustment process of the heating power sequence is implemented as follows: During the operation of the quasi-constant temperature heating mode, at the beginning of each sampling period, the polymerization progress deviation of the current sampling period and the polymerization progress deviation of the previous sampling period are read. The polymerization progress deviation of the previous sampling period is subtracted from the polymerization progress deviation of the current sampling period. The difference is used as the reduction of the polymerization progress deviation in the current sampling period. This reduction is the real-time decay rate of the sampling period.

[0058] Arrange the real-time decay rate of the current sampling period and the real-time decay rate of the previous 20 consecutive sampling periods in chronological order to form a decay rate sequence containing 21 real-time decay rate values.

[0059] Calculate the arithmetic mean of the last five values ​​in the decay rate sequence, then calculate the arithmetic mean of the first five values ​​in the decay rate sequence. Subtract the arithmetic mean of the first five values ​​from the arithmetic mean of the last five values ​​to obtain the terminal trend difference.

[0060] If the difference in the direction of the endpoint is greater than 0.02, the current deviation is determined to be in the acceleration and decay zone; if the difference in the direction of the endpoint is less than -0.02, the current deviation is determined to be in the deceleration and decay zone. If the difference in the direction of the endpoint is between -0.02 and 0.02, then the current deviation is determined to be in the constant rate decay zone.

[0061] After determining the type of the current attenuation zone, the heating power increment within the current sampling period is assigned a corresponding positive or negative sign and amplitude level based on the determination result.

[0062] The specific assignment method is as follows: if the current deviation is in the accelerated decay zone, the sign of the heating power increment is set to negative, and the amplitude level is set to level two, corresponding to an amplitude of 2.0 kW; if the current deviation is in the constant decay zone, the sign of the heating power increment is set to zero, and the amplitude level is set to level zero, corresponding to an amplitude of zero kW; if the current deviation is in the deceleration zone, the sign of the heating power increment is set to positive, and the amplitude level is set to level three, corresponding to an amplitude of 3.0 kW.

[0063] After completing the above settings, the set sign and the corresponding amplitude are combined to generate the base power adjustment amount under the current sampling period. The base power adjustment amount is in kilowatts. A positive value indicates an increase in heating power, a negative value indicates a decrease in heating power, and a zero value indicates that the current heating power remains unchanged.

[0064] After obtaining the base power adjustment, the actual heating power value of the previous sampling period is retrieved, and this actual heating power value is added to the base power adjustment in the current sampling period to obtain the heating power command value for the current sampling period.

[0065] The heating power command value is compared with the maximum allowable heating power of the heating actuator. If the command value is greater than the maximum allowable heating power, the command value is limited to the maximum allowable heating power value before being output. If the command value is less than the minimum maintaining power of the heating actuator, the command value is raised to the minimum maintaining power value before being output. If the command value is between the two, the command value is output directly. The output heating power command value is transmitted to the heating actuator, which controls the heating actuator to output the corresponding heating power according to the command value.

[0066] The 2.0 kW, 3.0 kW, and 0 kW values ​​corresponding to the aforementioned amplitude levels, the ±0.02 threshold for the end-point direction difference, and the 21 sampling period lengths of the decay rate sequence are all pre-stored in the data buffer. A one-time calibration is performed before each batch of raw materials is fed in, based on the rated power of the heating actuator and process requirements. The above process is repeated in each sampling period until the difference between the polycondensation activity index and the thermal decomposition risk equivalent in the same sampling period is within the preset target range for three consecutive sampling periods. At this point, the power adjustment process is terminated, completing the online optimization control of the entire temperature curve.

[0067] The working principle of this invention is as follows: Online spectral data of materials in each reaction unit during the modification reaction process, reactor wall temperature, material center temperature, heating power, and regulating valve opening are collected and combined into a competitive reaction state vector. Based on the real-time evolution trend of this state vector, a polycondensation activity index and a thermal decomposition risk equivalent are generated. The direction and magnitude of the deviation of the difference between these two values ​​from the preset ideal process curve determine the polymerization progress deviation. When the deviation indicates that polycondensation is lagging and the regulating valve opening reaches its upper limit, a pre-compensation time is output after matching the basic compensation time from a multi-segment compensation mapping table based on the deviation, corrected by the actual heating rate, and determined by the overlap with the temperature range of the acceleration threshold of the decomposition side reaction. Based on this, the initial determination temperature of the isothermal period is shifted downwards to enter the quasi-constant temperature heating mode earlier. In this mode, different signs and amplitude levels of the heating power increment are assigned according to the acceleration, constant speed, or deceleration decay zone to which the real-time decay rate of the polymerization progress deviation belongs, dynamically adjusting the heating power sequence of each sampling cycle until the difference between the polycondensation activity index and the thermal decomposition risk equivalent converges to the preset target range, thereby achieving online optimization control of the entire temperature change curve.

[0068] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for optimizing and controlling the temperature curve of three-stage variable-temperature thermal polycondensation in the coal tar upgrading process, characterized in that, Includes the following steps: S1 collects online spectral data of materials, reactor wall temperature, material center temperature, heating power and regulating valve opening during the coal tar upgrading reaction, forming a competitive reaction state vector characterizing the current reaction state; S2, based on the real-time evolution trend of the competing reaction state vector, generates a polycondensation activity index characterizing the current degree of polycondensation reaction advancement and a thermal decomposition risk equivalent characterizing the degree of accumulation of decomposition side reactions. The generation process of the polycondensation activity index is as follows: Extract the continuous absorbance change sequence of online spectral data in the preset characteristic wavenumber range from the competitive reaction state vector, and construct a multi-wavelength response trajectory reflecting the degree of aromatic ring condensation in time sequence; Based on the rate of change of absorbance at each wavenumber in the multi-wavelength response trajectory, the wavenumber response values ​​at the same moment are converted into instantaneous condensation contribution values ​​with uniform dimensions according to the catalytic ratio of temperature to reaction rate. The instantaneous condensation contribution value at each moment is accumulated along the reaction process, and the current degree of progress of the condensation reaction is characterized by the proportion of the increment of the accumulated value within the preset time window to the estimated total increment of the whole process. The output is the condensation activity index. The process for generating the thermal decomposition risk equivalent is as follows: The real-time difference between the material center temperature and the reactor wall temperature is synchronously extracted from the competing reaction state vector. The real-time difference of each sampling period is multiplied by the time interval between the current sampling period and the previous sampling period and then accumulated along the time axis to obtain the cumulative temperature gradient reflecting the degree of local overheating in the reactor. Based on the numerical range of the accumulated temperature gradient, a preset acceleration threshold for the pyrolysis side reaction is matched, and the absorbance shift of the online spectral data in the competitive reaction state vector at the characteristic wavenumber characterizing side chain breakage is used as the pyrolysis induction factor. The pyrolysis inducing factor and the cumulative temperature gradient are weighted together according to a preset correspondence, and the output is the thermal pyrolysis risk equivalent. S3. Compare the real-time difference between the polycondensation activity index and the thermal decomposition risk equivalent, and determine the polymerization progress deviation of the current batch relative to the theoretical optimal trajectory according to the direction and magnitude of the difference deviating from the preset ideal process curve. S4. When the polymerization progress deviation indicates that the actual polycondensation is behind the preset process and the opening of the regulating valve reaches the upper limit, the equivalent reaction time that needs to be compensated in advance is calculated based on the deviation, and the temperature uniformity period of the static constant temperature range is determined. The starting temperature of the temperature uniformity period is shifted downward to enter the quasi-constant temperature heating mode in advance. S5, in the quasi-constant temperature heating mode, dynamically adjusts the sequence of micro-heating power applied in each sampling period according to the real-time decay rate of the polymerization progress deviation, until the numerical difference between the polycondensation activity index and the thermal decomposition risk equivalent in the same sampling period is within the preset target range for three consecutive sampling periods, thus completing the online optimization control of the entire temperature change curve.

2. The method for optimizing and controlling the temperature curve of the three-stage variable-temperature thermal polycondensation process in the coal tar upgrading process according to claim 1, characterized in that, The process for setting the acceleration threshold of the pyrolysis side reaction is as follows: The temperature rise asynchrony between reactor wall temperature and material center temperature during the heating phase is extracted from the competitive reaction state vector, and the slope of the temperature rise asynchrony change over multiple consecutive sampling periods is used as a trend characterization of reactor coking state. By inputting the trend characterization value into the accelerated response mapping relationship, the critical temperature difference value that significantly aggravates thermal decomposition under the current coking state is matched. Based on the critical temperature difference, and combined with the absorbance baseline drift rate at the characteristic wavenumber characterizing side chain rupture using the online spectral data of the current batch, the value is corrected in real time and output as the acceleration threshold of the cracking side reaction corresponding to the current batch.

3. The method for optimizing and controlling the temperature curve of the three-stage variable-temperature thermal polycondensation process in the coal tar upgrading process according to claim 1, characterized in that, The process for outputting the polymerization progress deviation is as follows: The difference between the current polycondensation activity index and the thermal decomposition risk equivalent is used as the real-time net polycondensation equivalent. The real-time net polycondensation equivalent is synchronously mapped to the theoretical net polycondensation equivalent at the same moment as defined by the preset ideal process curve, and the instantaneous deviation between the two is calculated. Extract a sequence of instantaneous deviation values ​​along the time axis, including the current moment, for multiple consecutive sampling periods. Based on the continuous increase or decrease trend of the instantaneous deviation values ​​in the sequence, determine whether the current deviation belongs to the condensation lag-dominated type or the cracking advance-dominated type. The judgment result and the final amplitude of the instantaneous deviation value sequence are converted into cumulative lag according to a preset correspondence, and the output is the aggregation progress deviation.

4. The method for optimizing and controlling the temperature curve of the three-stage variable-temperature thermal polycondensation process in the coal tar upgrading process according to claim 1, characterized in that, The initial temperature for the temperature equalization period is shifted downwards to advance the entry into the quasi-constant temperature rise mode, specifically including: Using the cumulative lag included in the polymerization progress deviation as input, and based on the conversion relationship between lag equivalent and compensation time, the pre-compensation time required to catch up with the polycondensation progress at the current heating rate is calculated. Based on the pre-compensation duration and the actual heating rate of the current sampling period, the compensation start temperature point that needs to be triggered in advance is calculated in reverse along the temperature coordinate axis, and the compensation start temperature point is used as the updated judgment temperature. The updated judgment temperature replaces the original starting judgment temperature of the uniform temperature period, and when the material center temperature first reaches the updated judgment temperature, it triggers the entry into the quasi-constant temperature rise mode.

5. The method for optimizing and controlling the temperature curve of the three-stage variable-temperature thermal polycondensation process in the coal tar upgrading process according to claim 4, characterized in that, The calculation process for the pre-compensation duration is as follows: Based on the cumulative lag included in the aggregate schedule deviation, the segment interval where the cumulative lag is located is matched from the preset multi-segment compensation mapping table, and the basic compensation duration corresponding to the segment interval is read. The actual heating rate of the material center temperature in the current sampling period is compared with the preset standard heating rate. The basic compensation time is scaled based on the ratio of the actual heating rate to the standard heating rate to obtain the compensation time after rate correction. Obtain the expected temperature range corresponding to the compensation duration, and determine the overlap between the expected temperature range and the temperature range where the acceleration threshold of the pyrolysis side reaction corresponding to the thermal pyrolysis risk equivalent is located. Based on the overlap, the compensation duration is reduced twice and output as the pre-compensation duration.

6. The method for optimizing and controlling the temperature curve of the three-stage variable-temperature thermal polycondensation process in the coal tar upgrading process according to claim 1, characterized in that, The dynamic adjustment of the applied micro-heating power sequence within each sampling period specifically includes: The decrease in the aggregation progress deviation in each sampling period is used as the real-time decay rate. The real-time decay rates of multiple consecutive sampling periods are arranged in chronological order to form a decay rate sequence. The direction of the end of the decay rate sequence determines whether the current deviation is in the accelerated decay region, the constant decay region, or the decelerated decay region. Based on the different decay regions in which the decay rate sequence is located, the positive and negative signs and amplitude levels of the heating power increment in the current sampling period are assigned respectively to generate the basic power adjustment amount; The basic power adjustment is superimposed with the actual heating power of the previous sampling period to obtain the heating power of the current sampling period, and the heating actuator is controlled according to the heating power.

Images