An adaptive parameter control system for ozone deodorization

Abstract

This invention relates to the field of adaptive parameter control technology, specifically disclosing an adaptive parameter control system for ozone deodorization. The system continuously collects data on the ozone inlet concentration, ozone outlet concentration, oil phase feed flow rate, ozone inlet flow rate, pressure difference between the bottom and top of the reaction tower, and reaction temperature within the gas-liquid reaction tower. Based on the collected data, it estimates the gas-liquid exchange activity value using an alternating state recursion and observation correction method. The estimated activity value is compared with the initial activity value to calculate the percentage of activity decay. When a threshold is reached, a feedforward compensation enable signal is generated. A correction coefficient is calculated based on the current activity estimate, the target dissolution efficiency, and the current dissolution efficiency to obtain the corrected ozone inlet flow rate setpoint and adjust the flow rate. The actual ozone inlet flow rate and the newly collected ozone outlet concentration are fed back to the state estimation. This invention avoids reverse adjustment errors caused by mass transfer efficiency decay, thus stabilizing the deodorization effect.

Description

Technical Field

[0001] This invention relates to the field of adaptive parameter control technology, and more specifically to an adaptive parameter control system for ozone deodorization. Background Technology

[0002] In the continuous operation of the ozone deodorization process for recycled base oil from waste mineral oil, when the mass transfer efficiency of the gas-liquid reaction tower gradually decreases due to scaling on the inner wall and blockage of the gas distribution plate, the concentration of undissolved ozone in the exhaust gas actually increases. Existing control methods mistakenly judge the increase in exhaust gas concentration as excessive ozone addition and reduce the inlet air flow in the opposite direction, further exacerbating the deodorization effect. Moreover, there is a lack of real-time online estimation methods for the decline in gas-liquid exchange activity, and the ozone addition cannot be automatically feedforward to compensate for changes in mass transfer status, resulting in long-term instability in the color and odor level of the deodorized oil.

[0003] This invention estimates the gas-liquid exchange activity value online by alternating between state recursion and observation correction. When the activity declines to a threshold, it automatically triggers feedforward compensation, causing the ozone intake flow rate to actively increase as the mass transfer efficiency decreases, thereby avoiding reverse adjustment errors caused by misjudgment. At the same time, through a flow tracking deviation judgment mechanism, the ozone outlet concentration is marked as a valid observation value and fed back to the state estimate only after the flow rate stabilizes, ensuring the convergence of the estimation of the unmeasurable activity value and the reliability of the control system. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive parameter control system for ozone deodorization to solve the problems mentioned above.

[0005] The objective of this invention can be achieved through the following technical solutions: An adaptive parameter control system for ozone deodorization is used in the gas-liquid reaction tower of the ozone deodorization process for waste mineral oil regeneration base oil, executing the following modules in a 15-second control cycle: The data acquisition module collects ozone inlet concentration, ozone outlet concentration, oil phase feed flow rate, ozone inlet flow rate, pressure difference between the bottom and top of the reaction tower, and reaction temperature for the ozone oxidation and deodorization process in the gas-liquid reaction tower. The state estimation module, based on the ozone outlet concentration, pressure difference, and reaction temperature, combined with the current ozone dosage, uses state recursion and observation correction. It takes the gas-liquid exchange activity value estimated in the previous control cycle as a benchmark, and recursively derives the prior activity value for the current cycle based on the current ozone dosage and reaction temperature. It also calculates and predicts the ozone outlet concentration based on the prior activity value and pressure difference. The deviation between the actual collected ozone outlet concentration and the predicted ozone outlet concentration is used to calculate the dynamic correction gain in combination with the current reaction temperature. The dynamic correction gain is then used to correct the prior activity value to obtain the gas-liquid exchange activity estimate for the current cycle. The decay judgment module compares the activity estimate with the initial activity value, calculates the activity decay percentage, and generates a feedforward compensation enable signal when the activity decay percentage reaches or exceeds 20%. The correction calculation module is used to receive the feedforward compensation enable signal, calculate the correction coefficient required for ozone intake flow rate based on the current activity estimate, the preset target dissolution efficiency and the current dissolution efficiency derived from the ozone concentration dissolved in the oil phase, and multiply the current ozone intake flow rate setpoint by the correction coefficient to obtain the corrected ozone intake flow rate setpoint. The flow regulation and feedback module changes the ozone intake flow rate according to the corrected ozone intake flow rate setpoint, and feeds back the changed actual ozone intake flow rate and the newly collected ozone outlet concentration to the state estimation module for state recursion and observation correction in the next cycle.

[0006] As a further aspect of the present invention: the process of generating the dynamic correction gain is as follows: Based on the currently collected reaction temperature, the temperature influence weight is obtained by interpolation from the preset temperature-gain linear relationship; The absolute value of the deviation between the actual ozone outlet concentration and the predicted ozone outlet concentration is taken, and the absolute value is suppressed according to the preset upper limit of saturation to obtain the deviation adjustment factor. The dynamic correction gain is generated by multiplying the temperature influence weight, the bias adjustment factor, and the positive or negative sign of the bias value.

[0007] As a further aspect of the present invention: the generation of the feedforward compensation enable signal specifically includes: The original percentage of decline is obtained by subtracting the current activity estimate from the initial activity value and then dividing by the initial activity value. The original recession percentage is recursively averaged and filtered with the original recession percentage of the previous two control periods to obtain the filtered recession percentage. The filtered decay percentage is compared with a set threshold. When the filtered decay percentage reaches or exceeds the set threshold for three consecutive cycles, a feedforward compensation enable signal is generated.

[0008] As a further aspect of the present invention: the recursive averaging filtering of the original decay percentage and the original decay percentage of the previous two control periods specifically includes: Extract the original decay percentage of the current control cycle and the original decay percentage of the two immediately preceding control cycles, and calculate the arithmetic mean of the three as a temporary filter value. Calculate the fluctuation range between the original recession percentage of the current period and the original recession percentage of the previous period, and select the corresponding smoothing weight from the preset mapping table according to the numerical range of the fluctuation range. Multiply the temporary filter value by the smoothing weight, and add it to the result of multiplying the filter value of the previous period by 1 and subtracting the smoothing weight to obtain the percentage of decay after filtering in the current period.

[0009] As a further aspect of the present invention: the calculation process of the correction coefficient is as follows: The efficiency compensation ratio is obtained by dividing the preset target dissolution efficiency by the current dissolution efficiency calculated from the dissolved ozone concentration in the oil phase. Divide the current activity estimate by the pre-calibrated initial activity value to obtain the activity retention ratio. Divide the efficiency compensation ratio by the activity retention ratio to obtain the coarse correction coefficient; The coarse correction coefficient and the correction coefficient of the previous control cycle are subjected to a first-order lag filter, and the filter result is used as the correction coefficient output for the current cycle.

[0010] As a further aspect of the present invention: the first-order hysteresis filtering of the coarse correction coefficient and the correction coefficient of the previous control cycle specifically includes: Obtain the final correction coefficient of the output from the previous control cycle; Calculate the difference between the coarse correction coefficient obtained in the current cycle and the final correction coefficient in the previous control cycle; Based on the rate of change of the current percentage of activity decay, the corresponding filter coefficient is found from the preset discrete softening factor table; The difference is multiplied by the filter coefficient and then added to the final correction coefficient of the previous control cycle. The result of the sum is used as the final correction coefficient of the output of the current cycle.

[0011] As a further aspect of the present invention: the preset discrete softening factor specifically includes: The absolute value of the rate of change of the percentage of activity decline is divided into three continuous and non-overlapping numerical bands according to the magnitude of the value: low-speed decline range, medium-speed decline range, and high-speed decline range. A softening factor of 0.9 is assigned to the low-speed descent range, a softening factor of 0.6 is assigned to the medium-speed descent range, and a softening factor of 0.3 is assigned to the high-speed descent range. The mapping relationship between the three numerical bands and the corresponding softening factors is solidified into a discrete table, so that when looking up the table, the softening factor can be directly retrieved as the filtering coefficient according to the interval where the current rate of change is located.

[0012] As a further aspect of the present invention: the state recursion and observation correction for the next cycle specifically includes: The actual ozone intake flow rate after the change is compared with the corrected ozone intake flow rate setpoint, and the flow tracking deviation is calculated. When the absolute value of the deviation is less than the preset deviation tolerance for three consecutive cycles, the newly collected ozone outlet concentration is marked as a valid observation value. The actual ozone intake flow rate, effective observation value, and reaction temperature of the current cycle are packaged and used as input parameters for state recursion in the state estimation module of the next cycle.

[0013] The beneficial effects of this invention are: (1) This invention estimates the gas-liquid exchange activity value in real time by using state recursion and observation correction alternately, and generates a feedforward compensation enable signal when the activity decline reaches the threshold, so that the ozone intake flow rate can be automatically feedforward compensated as the mass transfer efficiency declines, thereby avoiding the control error in traditional control when the ozone concentration in the tail gas increases due to scaling on the inner wall of the gas-liquid reaction tower or blockage of the gas distribution plate, which is mistakenly judged as ozone excess and the intake flow rate is reduced in reverse.

[0014] (2) The present invention compares the actual ozone intake flow rate with the corrected set value, calculates the flow tracking deviation, and only when the absolute value of the deviation is less than the preset deviation tolerance for three consecutive cycles, the newly collected ozone outlet concentration is marked as a valid observation value and fed back to the state estimation module. This avoids the interference of measurement noise introduced by the transient flow regulation with state recursion and observation correction, and ensures the stable convergence of state estimation. Attached Figure Description

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

[0016] Figure 1 This is a system block diagram of the present 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, the present invention is an adaptive parameter control system for ozone deodorization, comprising: The data acquisition module collects ozone inlet concentration, ozone outlet concentration, oil phase feed flow rate, ozone inlet flow rate, pressure difference between the bottom and top of the reaction tower, and reaction temperature during continuous operation of the ozone oxidation deodorization process in the gas-liquid reaction tower. The state estimation module, based on the collected ozone outlet concentration, pressure difference, and reaction temperature, combined with the current ozone dosage, uses a method of alternating state recursion and observation correction to simultaneously estimate the gas-liquid exchange activity value and the oil phase dissolved ozone concentration, thus obtaining the activity estimate value. The decay judgment module compares the activity estimate with the pre-calibrated initial activity value, calculates the activity decay percentage, and generates a feedforward compensation enable signal when the activity decay percentage reaches or exceeds a set threshold. The correction calculation module is used to receive the feedforward compensation enable signal, calculate the correction coefficient required for ozone intake flow rate based on the current activity estimate, the preset target dissolution efficiency and the current dissolution efficiency derived from the ozone concentration dissolved in the oil phase, and obtain the corrected ozone intake flow rate setpoint. The flow regulation and feedback module changes the ozone intake flow rate according to the corrected ozone intake flow rate setpoint, and feeds back the changed actual ozone intake flow rate and the newly collected ozone outlet concentration to the state estimation module for state recursion and observation correction in the next cycle.

[0019] In the data acquisition module, for the ozone oxidation deodorization process within the gas-liquid reaction tower, the ozone inlet concentration, ozone outlet concentration, oil phase feed flow rate, ozone inlet gas flow rate, pressure difference between the bottom and top of the reaction tower, and reaction temperature are collected during continuous operation. Specifically, this includes: In the ozone deodorization process for recycled base oil from waste mineral oil, the ozone oxidation deodorization process within the gas-liquid reaction tower requires real-time acquisition of multiple parameters during continuous operation. Specifically, an ultraviolet absorption ozone concentration analyzer is installed on the ozone generator outlet pipeline to continuously measure the ozone inlet concentration in milligrams per liter. Another ozone concentration analyzer of the same type is installed on the tail gas emission pipeline at the top of the gas-liquid reaction tower to continuously measure the ozone outlet concentration. A mass flow meter is installed on the oil phase feed pipeline to continuously measure the oil phase feed flow rate in cubic meters per hour. A thermal gas mass flow meter is installed on the ozone inlet pipeline to continuously measure the ozone inlet flow rate in standard cubic meters per hour. Pressure transmitters are installed at the bottom of the gas-liquid reaction tower near the gas distribution plate and at the top of the gas phase space, respectively. The pressure difference between the two is used to obtain the pressure difference between the bottom and top of the reaction tower, in kilopascals. A platinum resistance temperature sensor is installed in the oil phase region in the middle of the reaction tower to continuously measure the reaction temperature in degrees Celsius. The sampling frequency of all the above sensors is set to 1 Hz. The collected signals are transmitted to the controller through the analog input channel. The controller performs an arithmetic average of the collected instantaneous values ​​every 15 seconds as a control cycle to obtain the mean value of each parameter within that cycle, which is used for subsequent state estimation.

[0020] In the state estimation module, based on the collected ozone outlet concentration, pressure difference, and reaction temperature, and combined with the current ozone dosage, a method of alternating state recursion and observation correction is used to simultaneously estimate the gas-liquid exchange activity value and the oil-phase dissolved ozone concentration, obtaining the activity estimate value, specifically including: The gas-liquid exchange activity value estimated in the previous control cycle is used as the baseline value. The gas-liquid exchange activity value characterizes the ease with which ozone is transferred from the gas phase to the oil phase, and its value ranges from 0 to 0.2 milligrams per second. At the start of the current control cycle, the controller reads the gas-liquid exchange activity estimate stored in the previous cycle, and simultaneously acquires the ozone dosage and reaction temperature collected in the current cycle. The ozone dosage is calculated by multiplying the ozone inlet concentration by the ozone inlet flow rate, in milligrams per second. The prior activity value for the current cycle is calculated according to a preset recursive rule: the prior activity value equals the gas-liquid exchange activity estimate from the previous cycle minus a decay term. This decay term is obtained by multiplying the gas-liquid exchange activity estimate from the previous cycle by a constant, which is then multiplied by the control cycle duration. This constant has a value of 0.001.

[0021] The ozone outlet concentration is predicted based on the prior activity value and the currently collected pressure difference between the bottom and top of the reaction tower. The calculation process for predicting the ozone outlet concentration is as follows: First, calculate the product of the pressure difference and the prior activity value. Then, multiply this product by a preset proportionality coefficient. This proportionality coefficient is obtained through offline calibration. The calibration method involves collecting multiple sets of pressure difference and measured ozone outlet concentration data when the mass transfer efficiency of the gas-liquid reaction tower is in its initial state, and obtaining the coefficient value through linear regression. This coefficient is set to 0.35. The result of the above product and proportionality coefficient calculation is used as the predicted ozone outlet concentration, in milligrams per liter.

[0022] The controller compares the actual ozone outlet concentration collected in this cycle with the predicted ozone outlet concentration to obtain the deviation value. This deviation value is equal to the actual ozone outlet concentration minus the predicted ozone outlet concentration, and the result can be positive or negative.

[0023] The dynamic correction gain is calculated based on the currently acquired reaction temperature. This dynamic correction gain corrects for the error between the prior activity value and the true value. The specific calculation process is as follows: First, based on the current reaction temperature, the temperature influence weight is obtained by interpolation from the temperature-gain linear relationship pre-stored in the controller. This linear relationship is experimentally determined: at reaction temperatures of 30°C, 50°C, and 80°C, the optimal gain value that makes the state estimation converge is tested, and the temperature and gain value are fitted to a straight line. The equation of the line is gain equal to 0.01 multiplied by temperature minus 0.1. For the current temperature, this linear equation is substituted to calculate the temperature influence weight.

[0024] The absolute value of the deviation is taken and compared with a preset saturation upper limit, which is set to 1.5 mg / L. If the absolute value of the deviation is less than or equal to the saturation upper limit, the deviation adjustment factor is equal to the absolute value of the deviation; if the absolute value of the deviation is greater than the saturation upper limit, the deviation adjustment factor is equal to the saturation upper limit. Finally, the temperature influence weight is multiplied by the deviation adjustment factor, and then multiplied by the sign of the deviation value (i.e., the deviation value divided by the absolute value of the deviation to get a positive or negative 1). The result of multiplying these three factors is the dynamic correction gain.

[0025] The prior activity value is corrected using the aforementioned dynamic correction gain to obtain the gas-liquid exchange activity estimate for the current cycle. The correction process is as follows: first, the product of the dynamic correction gain and the deviation value is calculated; then, this product is added to the prior activity value, and the result is output as the gas-liquid exchange activity estimate for the current cycle. This estimate is stored in the controller for recursive calculation in the next control cycle and for subsequent degradation judgment steps. Through the alternating execution of the above recursion and correction, an updated gas-liquid exchange activity estimate can be obtained for each control cycle.

[0026] In the decay judgment module, the estimated activity value is compared with the pre-calibrated initial activity value to calculate the activity decay percentage. When the activity decay percentage reaches or exceeds a set threshold, a feedforward compensation enable signal is generated, specifically including: The controller compares the estimated gas-liquid exchange activity obtained in the current cycle with the pre-calibrated initial activity value to calculate the initial degradation percentage. The initial activity value is calibrated as follows: after the gas-liquid reactor has completed cleaning and is put into operation for the first time, with mass transfer efficiency at its optimal state, the estimated gas-liquid exchange activity is continuously collected for 10 control cycles, and its arithmetic mean is calculated as the initial activity value, in units of one second. The initial degradation percentage is calculated as follows: first, the initial activity value is subtracted from the current activity estimate to obtain the activity decrease; then, the activity decrease is divided by the initial activity value, and the result is multiplied by 100% and expressed as a percentage.

[0027] For example: if the initial activity value is 0.085 per second and the current activity estimate is 0.068 per second, then the activity decrease is 0.017 per second. Dividing by 0.085 per second gives 0.2, and multiplying by 100% gives the original decay percentage of 20%.

[0028] To prevent false triggering caused by single-cycle measurement noise or instantaneous fluctuations, the controller performs a recursive average filter on the original decay percentage to obtain the filtered decay percentage. The specific filtering process is as follows: The controller internally stores the original decay percentage values ​​from the three most recent control cycles: the original decay percentage of the current cycle, the original decay percentage of the previous control cycle, and the original decay percentage of the two control cycles prior. These three values ​​are first retrieved, and their arithmetic mean is calculated as a temporary filter value.

[0029] For example: if the original decline percentages for the three cycles are 20%, 18% and 22% respectively, then the sum of the three is 60%, which is divided by 3 to get a temporary filter value of 20%.

[0030] Calculate the fluctuation range between the original decline percentage of the current period and the original decline percentage of the previous control period. The fluctuation range is equal to the absolute value of the difference between the two, divided by the average of the two, and then multiplied by 100%. The result is expressed as a percentage. The controller internally uses a pre-defined mapping table that divides the fluctuation range into three consecutive and non-overlapping numerical intervals: fluctuation range less than 5% corresponds to a smoothing weight of 0.8, fluctuation range between 5% and 15% corresponds to a smoothing weight of 0.5, and fluctuation range greater than 15% corresponds to a smoothing weight of 0.2. Based on the interval in which the currently calculated fluctuation range falls, the corresponding smoothing weight is selected from the mapping table.

[0031] Obtain the filtered decay percentage calculated in the previous control cycle, multiply the temporary filter value by the smoothing weight, and add it to the result of multiplying the filter value of the previous cycle by one minus the smoothing weight. The sum is used as the filtered decay percentage for the current control cycle.

[0032] For example: if the temporary filter value is 20%, the smoothing weight is 0.8, and the filter value of the previous period is 18%, then the percentage decline after filtering in this period is equal to 20% multiplied by 0.8 plus 18% multiplied by 0.2, that is, 16% plus 3.6% equals 19.6%.

[0033] The controller compares the percentage of filtered degradation with a preset threshold. The threshold is predetermined based on process requirements, ranging from 15% to 25%; in this embodiment, it is set to 20%. The controller records whether the percentage of filtered degradation reaches or exceeds this threshold for each control cycle and monitors this for three consecutive control cycles. When the percentage of filtered degradation reaches or exceeds the threshold for three consecutive control cycles, the controller generates a feedforward compensation enable signal. This signal is a Boolean value; a value of 1 indicates that the enable is active, and a value of 0 indicates that it is inactive. The feedforward compensation enable signal is transmitted to the subsequent correction calculation step to trigger the compensation calculation of the ozone intake flow rate. If the percentage of filtered degradation is lower than the threshold in any of the three consecutive cycles, the enable signal remains at 0, and no compensation action is triggered.

[0034] In the correction calculation module, the feedforward compensation enable signal is received. Based on the current activity estimate, the preset target dissolution efficiency, and the current dissolution efficiency derived from the dissolved ozone concentration in the oil phase, the correction coefficient required for the ozone inlet flow rate is calculated to obtain the corrected ozone inlet flow rate setpoint. Specifically, this includes: After receiving the feedforward compensation enable signal, the controller initiates the correction coefficient calculation process when the enable signal value is 1. The controller calculates the efficiency compensation ratio by reading the preset target dissolution efficiency and the current dissolution efficiency calculated from the dissolved ozone concentration in the oil phase. The target dissolution efficiency is a fixed value in the process design, which is 85% in this embodiment. The current dissolution efficiency is calculated by dividing the dissolved ozone concentration in the oil phase obtained in the state estimation step by the ozone inlet concentration, and then multiplying the quotient by 100%. The result is expressed as a percentage.

[0035] For example: if the dissolved ozone concentration in the oil phase is 85 mg / L and the ozone inlet concentration is 100 mg / L, then the current dissolution efficiency is 85%. The efficiency compensation ratio is equal to the target dissolution efficiency divided by the current dissolution efficiency, that is, 85% divided by 85% equals 1.0.

[0036] The controller calculates the activity retention ratio by dividing the estimated gas-liquid exchange activity obtained in the current cycle by a pre-calibrated initial activity value. The activity retention ratio reflects the remaining proportion of mass transfer performance and its value ranges from 0 to 1. For example, if the current activity estimate is 0.068 sec and the initial activity value is 0.085 sec, then the activity retention ratio is 0.8.

[0037] The controller obtains the coarse correction coefficients. The calculation process for the coarse correction coefficients follows the mathematical formula: ; in, Indicates the coarse correction factor; This represents the efficiency compensation ratio, which is dimensionless. This represents the activity retention ratio. Taking the aforementioned values ​​as an example, if the efficiency compensation ratio is 1.0 and the activity retention ratio is 0.8, then the coarse correction factor is 1.25. This coarse correction factor is used to compensate for insufficient effective ozone dissolution caused by a decrease in mass transfer efficiency.

[0038] To avoid control oscillations caused by drastic changes in the correction coefficient, the controller performs a first-order hysteresis filter on the coarse correction coefficient. The filtering process is as follows: the controller internally stores the final correction coefficient output from the previous control cycle, denoted as... Calculate the difference between the current period's coarse correction factor and the final correction factor. The filter coefficient is determined based on the rate of change of the current percentage of activity decay. The rate of change of the percentage of activity degradation is defined as: the percentage degradation after filtering in the current cycle minus the percentage degradation after filtering in the previous cycle, divided by the control cycle duration (15 seconds), expressed as a percentage per second. The controller uses the absolute value of this rate of change to look up the corresponding filter coefficient from a preset discrete softening factor table. The filter coefficients range from 0 to 1; the smaller the value, the stronger the filtering effect. Calculate the final correction coefficient for this cycle's output using the following mathematical formula. The calculation expression is: ; For example: If the final correction coefficient of the previous cycle was 1.10, the current coarse correction coefficient is 1.25, the difference is 0.15, and the filter coefficient obtained from the table is 0.6, then the final correction coefficient of this cycle is equal to 1.10 plus 0.6 multiplied by 0.15, that is, 1.10 plus 0.09 equals 1.19. This final correction coefficient is used as the correction coefficient for the output of this cycle.

[0039] The discrete softening factor table is established as follows: The absolute value of the rate of change of the percentage of activity decay is divided into three continuous and non-overlapping numerical bands. The first numerical band is the low-speed decline range, corresponding to an absolute value of the rate of change less than 0.5 percentage points per second; the second numerical band is the medium-speed decline range, corresponding to an absolute value of the rate of change greater than or equal to 0.5 percentage points per second and less than 1.5 percentage points per second; the third numerical band is the high-speed decline range, corresponding to an absolute value of the rate of change greater than or equal to 1.5 percentage points per second. Next, a softening factor is assigned to each numerical band: 0.9 for the low-speed decline range, 0.6 for the medium-speed decline range, and 0.3 for the high-speed decline range. This allocation principle is based on experimental testing: when mass transfer efficiency decays slowly, the correction coefficient should not change abruptly, so a larger softening factor of 0.9 is used to maintain smoothness; during medium-speed decay, a medium softening factor of 0.6 is used to balance response speed and stability; during high-speed decay, rapid compensation is required, so a smaller softening factor of 0.3 is used to allow the correction coefficient to track the coarse correction coefficient more quickly. The above mapping relationship is permanently stored in the controller's read-only memory, forming a discrete table. During operation, the controller first calculates the absolute value of the rate of change of the current percentage of activity decay, determines which numerical range it falls into, and then extracts the corresponding softening factor as the filtering coefficient. .

[0040] After obtaining the final correction coefficient, the controller multiplies the ozone intake flow rate setpoint for the current cycle by this coefficient to obtain the corrected ozone intake flow rate setpoint. This corrected setpoint is then transmitted to the flow regulation stage to adjust the ozone intake flow rate. Simultaneously, the final correction coefficient is stored and used as the final correction coefficient for the previous cycle in the next control cycle. Through this method, the controller achieves adaptive compensation for the ozone intake flow rate, maintaining the actual dissolution efficiency near the target value.

[0041] In the flow regulation and feedback module, the ozone intake flow rate is adjusted according to the corrected ozone intake flow rate setpoint, and the adjusted actual ozone intake flow rate and the newly collected ozone outlet concentration are fed back to the state estimation module for state recursion and observation correction in the next cycle. Specifically, this includes: The controller sends the corrected ozone intake flow rate setpoint to the thermal gas mass flow controller on the ozone intake pipeline. This flow controller automatically adjusts the opening of the regulating valve according to the setpoint, so that the actual ozone intake flow rate follows the setpoint. Simultaneously, the controller continuously collects the actual ozone intake flow rate value measured by the thermal gas mass flow meter.

[0042] The controller calculates the flow tracking deviation. The flow tracking deviation equals the actual ozone intake flow rate minus the corrected ozone intake flow rate setpoint. The controller performs this calculation at the end of each control cycle (cycle length is 15 seconds).

[0043] For example: if the corrected setting is 3.0 standard cubic meters per hour, and the actual measured value is 2.94 standard cubic meters per hour, then the flow tracking deviation is -0.06 standard cubic meters per hour. The absolute value of the deviation is 0.06 standard cubic meters per hour. The controller stores this deviation value and its absolute value in an internal register for subsequent judgment.

[0044] The controller determines whether flow tracking is stable. A preset deviation tolerance is determined based on flow measurement accuracy and process requirements; in this embodiment, the deviation tolerance is 0.1 standard cubic meters per hour. The controller continuously monitors for three control cycles. At the end of each cycle, it calculates the absolute value of the flow tracking deviation for that cycle and compares it with the deviation tolerance. When the absolute value of the deviation for three consecutive control cycles is less than 0.1 standard cubic meters per hour, the controller determines that the current flow tracking has entered a stable state. Under this condition, the controller marks the newly acquired ozone outlet concentration for that control cycle as a valid observation. If the absolute value of the deviation for any control cycle is greater than or equal to 0.1 standard cubic meters per hour, the flow tracking is unstable, and the newly acquired ozone outlet concentration is not marked as a valid observation but rather as an invalid observation and is not used. This judgment mechanism ensures that only ozone outlet concentration data acquired during periods of stable flow can be used for state recursion, avoiding interference from measurement noise introduced by flow fluctuations that could affect the accuracy of state estimation.

[0045] The controller packages the actual ozone intake flow rate, the ozone outlet concentration marked as a valid observation, and the currently acquired reaction temperature into a set of input parameters for the state recursion and observation correction steps of the next control cycle. Specifically, the controller allocates three contiguous memory locations to write the actual ozone intake flow rate (in standard cubic meters per hour), the valid observation value (in milligrams per liter), and the reaction temperature (in degrees Celsius), respectively. If the ozone outlet concentration is marked as an invalid observation, the controller does not write it into the packaged parameters but instead uses the valid observation value from the previous cycle, or sets the position of that observation value to an empty flag. In the state recursion of the next cycle, the controller will skip the observation correction step and only perform state recursion. This processing method ensures that the state estimation can still operate stably during the transient flow regulation period and will not diverge due to short-term flow fluctuations. The packaged input parameters are stored in the controller, awaiting use by the state estimation in the next control cycle. Through the above feedback loop, the actual ozone intake flow rate change information and effective ozone outlet concentration data are continuously fed back to the state estimation step, forming a complete closed-loop adaptive control.

[0046] The working principle of this invention is as follows: In the ozone deodorization process of waste mineral oil regeneration base oil, for the ozone oxidation deodorization process in the gas-liquid reaction tower, the ozone inlet concentration, ozone outlet concentration, oil phase feed flow rate, ozone inlet flow rate, pressure difference between the bottom and top of the reaction tower, and reaction temperature are continuously collected. The parameters are arithmetically averaged every 15 seconds as a control cycle. Based on the collected ozone outlet concentration, pressure difference, and reaction temperature, combined with the current ozone dosage, a method of alternating state recursion and observation correction is used to simultaneously estimate the gas-liquid exchange activity value and the dissolved ozone concentration in the oil phase, obtaining an activity estimate. The activity estimate is then compared with a pre-calibrated... The initial activity value is compared, and the percentage of activity decay is calculated. When the percentage of activity decay reaches or exceeds a set threshold, a feedforward compensation enable signal is generated. After receiving the enable signal, the correction coefficient required for ozone intake flow rate is calculated based on the current activity estimate, the preset target dissolution efficiency, and the current dissolution efficiency derived from the ozone concentration dissolved in the oil phase. The corrected ozone intake flow rate setpoint is obtained. The ozone intake flow rate is changed according to the setpoint, and the changed actual ozone intake flow rate and the newly collected ozone outlet concentration are fed back to the state estimation step for state recursion and observation correction in the next cycle, forming a closed-loop adaptive control.

[0047] 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. An adaptive parameter control system for ozone deodorization, characterized in that, The gas-liquid reaction tower used in the ozone deodorization process for waste mineral oil regeneration base oil executes the following modules in a 15-second control cycle: The data acquisition module collects ozone inlet concentration, ozone outlet concentration, oil phase feed flow rate, ozone inlet flow rate, pressure difference between the bottom and top of the reaction tower, and reaction temperature for the ozone oxidation and deodorization process in the gas-liquid reaction tower. The state estimation module, based on the ozone outlet concentration, pressure difference, and reaction temperature, and combined with the current ozone dosage, employs state recursion and observation correction. Using the gas-liquid exchange activity value estimated in the previous control cycle as a benchmark, it recursively derives the prior activity value for the current cycle based on the current ozone dosage and reaction temperature. Furthermore, it calculates and predicts the ozone outlet concentration based on the prior activity value and pressure difference. The deviation between the actual collected ozone outlet concentration and the predicted ozone outlet concentration, combined with the current reaction temperature, is used to calculate a dynamic correction gain. This dynamic correction gain is then used to correct the prior activity value, yielding the estimated gas-liquid exchange activity value for the current cycle. The generation process of the dynamic correction gain is as follows: Based on the currently collected reaction temperature, the temperature influence weight is obtained by interpolation from the preset temperature-gain linear relationship; The absolute value of the deviation between the actual ozone outlet concentration and the predicted ozone outlet concentration is taken, and the absolute value is suppressed according to the preset upper limit of saturation to obtain the deviation adjustment factor. The dynamic correction gain is generated by multiplying the temperature influence weight, the bias adjustment factor, and the positive or negative sign of the bias value. The decay judgment module compares the activity estimate with the initial activity value, calculates the activity decay percentage, and generates a feedforward compensation enable signal when the activity decay percentage reaches or exceeds 20%. The correction factor calculation module receives the feedforward compensation enable signal, calculates the correction factor required for the ozone inlet flow rate based on the current activity estimate, the preset target dissolution efficiency, and the current dissolution efficiency derived from the dissolved ozone concentration in the oil phase, and multiplies the current ozone inlet flow rate setpoint by the correction factor to obtain the corrected ozone inlet flow rate setpoint. The calculation process of the correction factor is as follows: The efficiency compensation ratio is obtained by dividing the preset target dissolution efficiency by the current dissolution efficiency calculated from the dissolved ozone concentration in the oil phase. Divide the current activity estimate by the pre-calibrated initial activity value to obtain the activity retention ratio. Divide the efficiency compensation ratio by the activity retention ratio to obtain the coarse correction coefficient; The coarse correction coefficient and the correction coefficient of the previous control cycle are subjected to a first-order lag filter, and the filtering result is used as the correction coefficient output for the current cycle. The flow regulation and feedback module changes the ozone intake flow rate according to the corrected ozone intake flow rate setpoint, and feeds back the changed actual ozone intake flow rate and the newly collected ozone outlet concentration to the state estimation module for state recursion and observation correction in the next cycle.

2. The adaptive parameter control system for ozone deodorization according to claim 1, characterized in that, The generation of the feedforward compensation enable signal specifically includes: The original percentage of decline is obtained by subtracting the current activity estimate from the initial activity value and then dividing by the initial activity value. The original recession percentage is recursively averaged and filtered with the original recession percentage of the previous two control periods to obtain the filtered recession percentage. The filtered decay percentage is compared with a set threshold. When the filtered decay percentage reaches or exceeds the set threshold for three consecutive cycles, a feedforward compensation enable signal is generated.

3. The adaptive parameter control system for ozone deodorization according to claim 2, characterized in that, The recursive averaging filter of the original recession percentage and its original recession percentage from the previous two control periods specifically includes: Extract the original decay percentage of the current control cycle and the original decay percentage of the two immediately preceding control cycles, and calculate the arithmetic mean of the three as a temporary filter value. Calculate the fluctuation range between the original recession percentage of the current period and the original recession percentage of the previous period, and select the corresponding smoothing weight from the preset mapping table according to the numerical range of the fluctuation range. Multiply the temporary filter value by the smoothing weight, and add it to the result of multiplying the filter value of the previous period by 1 and subtracting the smoothing weight to obtain the percentage of decay after filtering in the current period.

4. The adaptive parameter control system for ozone deodorization according to claim 1, characterized in that, The step of performing a first-order hysteresis filter on the coarse correction coefficient and the correction coefficient of the previous control cycle specifically includes: Obtain the final correction coefficient of the output from the previous control cycle; Calculate the difference between the coarse correction coefficient obtained in the current cycle and the final correction coefficient in the previous control cycle; Based on the rate of change of the current percentage of activity decay, the corresponding filter coefficient is found from the preset discrete softening factor table; The difference is multiplied by the filter coefficient and then added to the final correction coefficient of the previous control cycle. The result of the sum is used as the final correction coefficient of the output of the current cycle.

5. The adaptive parameter control system for ozone deodorization according to claim 4, characterized in that, The preset discrete softening factor specifically includes: The absolute value of the rate of change of the percentage of activity decline is divided into three continuous and non-overlapping numerical bands according to the magnitude of the value: low-speed decline range, medium-speed decline range, and high-speed decline range. A softening factor of 0.9 is assigned to the low-speed descent range, a softening factor of 0.6 is assigned to the medium-speed descent range, and a softening factor of 0.3 is assigned to the high-speed descent range. The mapping relationship between the three numerical bands and the corresponding softening factors is solidified into a discrete table, so that when looking up the table, the softening factor can be directly retrieved as the filtering coefficient according to the interval where the current rate of change is located.

6. The adaptive parameter control system for ozone deodorization according to claim 1, characterized in that, The state recursion and observation correction for the next cycle specifically include: The actual ozone intake flow rate after the change is compared with the corrected ozone intake flow rate setpoint, and the flow tracking deviation is calculated. When the absolute value of the deviation is less than the preset deviation tolerance for three consecutive cycles, the newly collected ozone outlet concentration is marked as a valid observation value. The actual ozone intake flow rate, effective observation value, and reaction temperature of the current cycle are packaged and used as input parameters for state recursion in the state estimation module of the next cycle.

Images