Process control method and system for waste plastic continuous cracking for oil production with segmented temperature control and rolling optimization

Abstract

This invention belongs to the field of environmental protection technology and discloses a segmented temperature control and rolling optimization process control method and system for continuous pyrolysis of waste plastics to produce oil. The method divides the membrane pyrolysis reactor into A / B / C temperature zones along the material flow direction: zone A is used for rapid film formation and initial pyrolysis of molten material; zone B is used for main pyrolysis and oil production; and zone C is used for deep pyrolysis and oil-gas component tuning. Real-time acquisition of signals such as temperature, pressure, oxygen content, filtration pressure difference, feed rate, and rotation speed is used to construct predictive models for oil yield, carbon residue rate, and pressure difference evolution. The setpoints, back pressure, and condensation intensity of each temperature zone are output using rolling time-domain optimization. When the predicted pressure difference reaches a threshold, a dual-loop non-stop switching is triggered, and the model parameters are written back. Closed-loop control improves the stability of oil yield, reduces the risk of coking and blockage, and enhances safety.

Description

Technical Field

[0001] This invention belongs to the field of environmental protection technology, specifically relating to a process control method and system for segmented temperature control and rolling optimization in the continuous pyrolysis of waste plastics to produce oil. Background Technology

[0002] With the rapid increase in waste plastics, especially polyolefins such as PE / PP, the use of thermal pyrolysis or catalytic pyrolysis to break down polymer chains into combustible fuel gas and recover fuel oil has become one of the important pathways for plastic chemical recycling. To meet industrial demands, existing technologies are constantly evolving from batch reactors to continuous and automated processes. However, continuous pyrolysis processes often have characteristics such as temperature, pressure, oxygen content, and blockage risk. Fluctuations in operating conditions can directly cause significant changes in oil yield, carbon residue rate, and downstream separation and condensation load, thereby affecting the stability and safety of continuous operation.

[0003] For example, Chinese patent CN101074385A discloses a continuous pyrolysis process and equipment for waste plastics. By introducing a solid heat carrier flow that moves in the same direction as the raw material into the pyrolyzer and cooperating with a spiral propulsion mechanism, continuous heating pyrolysis and continuous discharge of pyrolysis gas / solid products are achieved. The background section also points out that traditional intermittent pyrolysis has problems such as slow processing speed, low oil quality, safety risks caused by direct heating, uneven heating during continuous operation, incomplete pyrolysis, or local overheating and charring to form coke residue. It attempts to improve continuous operation by enhancing heat transfer and carrying solid residues out through the heat carrier. While such continuous systems can alleviate uneven heating and coking caused by poor thermal conductivity of plastics to some extent, they still face challenges in industrial settings: ① Fluctuations in raw material composition, impurity content, and melt state lead to frequent changes in reaction load; ② In multi-unit series systems such as pyrolysis-filtration-condensation-compression recovery, temperature, pressure, and flow are coupled, making it difficult to balance oil yield, energy consumption, and stability with single-point or single-loop setpoint control; ③ Coking / blockage usually has a gradual accumulation characteristic, and if only post-event threshold alarms are used for handling, sudden changes in pressure differential / load often occur, leading to load reduction, shutdown for cleaning, or unplanned shutdowns, thus weakening the advantages of continuous systems.

[0004] Therefore, the key problems commonly faced by existing technologies can be summarized as follows: in the multi-unit coupled process of continuous pyrolysis of waste plastics to produce oil, there is a lack of a closed-loop control method and system that can simultaneously consider multi-temperature zone control, multi-variable linkage, comprehensive optimization of oil output and energy consumption, early release of differential pressure / clogging risks, and unified inclusion of safety constraints. Without these capabilities, the equipment often relies on experience-based settings and manual intervention, resulting in large fluctuations in oil output, high energy consumption, easy clogging of filters, a high probability of unplanned shutdowns, and difficulty in maintaining long-term stable operation while ensuring safety boundaries. These problems urgently need to be improved at the process control level to further enhance the stability, economy, and intrinsic safety of continuous pyrolysis oil production units for waste plastics. Summary of the Invention

[0005] To address the aforementioned technical issues, this invention implements segmented temperature control and rolling optimization closed-loop control in the continuous pyrolysis oil production process of waste plastics. Under the premise of meeting safety constraints such as oxygen content, pressure, and pressure difference, it improves the oil yield and reduces the risk of fluctuations and blockage shutdowns, thereby achieving continuous, stable, low-energy-consumption, and safe operation.

[0006] To achieve the above-mentioned technical objectives, the present invention provides the following technical solutions.

[0007] In a first aspect, the present invention provides a process control method for segmented temperature control and rolling optimization in continuous pyrolysis of waste plastics to produce oil. The pyrolysis reactor is sequentially configured with mutually thermally isolated temperature zones A, B, and C along the material flow direction, and each temperature zone is equipped with an independent temperature measurement and heating regulation unit. The method includes:

[0008] S1, collect the temperature and extrusion pressure at multiple points of the waste plastic melting extrusion equipment, the temperature and pressure of the A temperature zone / B temperature zone / C temperature zone, the pressure difference and temperature of the filtration and separation section, the inlet and outlet temperatures and liquid levels of the condensation section, and the online oxygen content of the pressure difference tank;

[0009] S2 establishes a segmented object with melt extrusion temperature, A / B / C temperature zones, and screw speed as the core control parameters. Among them, temperature zone A is the first heating section at the feed end, used to rapidly heat up the 0.3~1.5mm film material and initiate initial pyrolysis and devolatilization; temperature zone B is the second heating section in the middle section, used for main pyrolysis to generate oil and gas; temperature zone C is the third heating section at the discharge end, used for deep pyrolysis and oil and gas component setting, and each of the three temperature zones is independently controlled within the range of 350~450℃.

[0010] S3, Construct a rolling prediction model: with oil yield, carbon slag rate and condensation efficiency as outputs, and the aforementioned manipulation amount, back pressure valve opening and inert gas flow rate as inputs, and set safety constraints: oxygen content ≤0.5%, system pressure controlled by back pressure at -5kPa~0, and pressure difference in the filtration and separation section not exceeding the set threshold.

[0011] S4, solves the constraint optimization in a rolling manner according to the control cycle, and outputs the set values ​​of temperature zone A / B / C, screw speed, feed rate, inert gas replacement / replenishment amount, back pressure valve opening degree and condensation cooling intensity for the next cycle;

[0012] S5, if it is predicted that the pressure difference of the filter separation section will exceed the limit, control the three-way valve to switch to another independent filter separation path to achieve continuous operation without stopping the machine, and model and update the data before and after the switch respectively.

[0013] S6. If any temperature zone exceeds the limit, pressure exceeds the limit, or oxygen content exceeds the limit, then start rapid cold nitrogen charging and lock the extrusion and upstream feed, then write back and correct the rolling prediction model.

[0014] Specifically, temperature sensors and pressure sensors are installed in temperature zones A, B, and C respectively; the temperature setpoint for temperature zone A is 380℃, the temperature setpoint for temperature zone B is 395℃, and the temperature setpoint for temperature zone C is 420℃, and temperature detection, indication, control, alarm, and safety interlock are completed in their respective temperature zones.

[0015] Specifically, the melt extrusion section uses a single-screw extruder and is equipped with 7 temperature measurement points, including 4 temperature measurement points at the electromagnetic heating point, 2 temperature measurement points at the ceramic heating point, and 1 temperature measurement point at the flexible connection heating point. When the temperature at any of the 4 electromagnetic heating points becomes out of control, the melt extrusion section immediately stops and the upstream equipment stops in sequence, and the downstream equipment stops in sequence after a preset delay.

[0016] Specifically, the pyrolysis reactor is equipped with 30 temperature measurement points, 3 pressure measurement points and 1 flow measurement point, with 12 temperature measurement channels set for the reactor and the heating element respectively; when the temperature at any measurement point becomes out of control, the nitrogen switch valve opens and the pyrolysis reactor is shut down, the upstream equipment shuts down in sequence, and the downstream equipment shuts down in sequence after a preset delay.

[0017] Specifically, if any of the three pressure measurement points set in the pyrolysis reactor detects a pressure exceeding the preset upper limit, it is determined to be an overpressure event. When an overpressure event occurs, the interlock action of stopping the extrusion and upstream equipment and opening the pneumatic ball valve is executed, and the entire line is shut down after a preset delay.

[0018] Specifically, the differential pressure tank is equipped with an online oxygen content detection device; before startup, the entire system is purged with nitrogen at least three times through the compressor and nitrogen system, and the oxygen content is controlled at ≤0.5% after purging; during normal operation, the oxygen content of oil and gas is monitored in real time and controlled at ≤0.5%.

[0019] Specifically, the system pressure adopts a back pressure control method. The pipeline from reactor to filter separation to condenser to compressor pre-buffer / differential pressure tank is connected. The compressor is adjusted by the back pressure valve set in front of the differential pressure tank and interlocked, so that the system pressure is controlled within the range of -5kPa to 0 and the pyrolysis reactor is in a slightly positive pressure operation state.

[0020] Specifically, two independent filtration-separation-condensation circuits are set up, with a differential pressure tank in front of the compressor for each circuit; the two circuits are switched by a three-way switch valve at the gas phase outlet of the pyrolysis reactor and a three-way switch valve after the differential pressure tank; the switching includes at least the sequential control of "exiting circuit isolation - entering circuit connection - back pressure recovery".

[0021] Specifically, a differential pressure sensor is installed in the filtration and separation section to monitor the pressure difference ΔP between the top inlet and the bottom of the filtration and separation tower, and compares ΔP with a preset upper limit of differential pressure ΔPlim; when ΔP≥ΔPlim, an alarm is output and the loop switching of S5 is triggered; when the temperature in the filtration and separation section becomes out of control, the cold nitrogen switch is triggered to open and the shutdown interlock is executed.

[0022] Specifically, the set of constraints for the rolling optimization includes at least: upper and lower limits of temperature in temperature zones A / B / C, system pressure range, upper limit of oxygen content, upper limit of filtration and separation pressure difference, and constraint on the rate of change of actuator; the objective function of the rolling optimization includes at least a term for maximizing oil yield, a term for minimizing energy consumption, and a term for minimizing temperature / pressure / pressure difference deviation; and / or, the prediction model is an online updatable model, and after each control cycle, the model parameters are updated online using the actual mixed oil production, cracked residue production, temperature / liquid level changes in the condensation section, and filtration and separation pressure difference data, and the updated model is used to solve the problem in the next rolling cycle.

[0023] The second invention provides a segmented temperature control and rolling optimization process control system for continuous pyrolysis of waste plastics to produce oil, used to implement the method described in the first aspect, including an equipment layer, a control layer and a data storage layer.

[0024] The equipment layer includes at least a melt extrusion device, a membrane pyrolysis reactor with A / B / C temperature zones, a filtration and separation device, a condensation device, a compressor, a differential pressure tank, a cold nitrogen rapid charging device, and a valve actuator.

[0025] The control layer includes at least: a temperature measurement module, a gas flow measurement module, a pressure detection and measurement module, an oxygen content detection module, a waste plastic melting analysis module, a pyrolysis temperature analysis module, an oil and gas condensation efficiency analysis module, a rolling optimization module, a loop switching control module, a safety interlock module, and a production early warning module.

[0026] The data storage layer is used to store process time-series data, optimization setpoints, alarm / interlock events and model parameters, and provides write-back data to the rolling optimization module and the production early warning module.

[0027] Specifically, the temperature measurement module is connected to at least 7 temperature measurement points of the melt extrusion equipment, 30 temperature measurement points of the pyrolysis reactor, 15 temperature measurement points of the filtration and separation equipment, and 12 temperature measurement points of the condensation equipment.

[0028] Specifically, the pyrolysis temperature analysis module is used to calculate the target heating power or target temperature setpoint for temperature zones A / B / C based on the film thickness, temperature of the paste-like waste plastic, and material information, and uses the calculation results as the initial value or constraint boundary of the rolling optimization module.

[0029] Specifically, the control layer includes an industrial controller, which includes a processor, a memory, and an industrial communication interface. The industrial communication interface is used to interact with temperature / pressure / differential pressure / oxygen content sensors, as well as heating elements, frequency converters, back pressure valves, three-way valves, cooling actuators, and cold nitrogen switching valves.

[0030] Thirdly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0031] This invention constructs a predictable model using multi-source online signals, and then performs rolling constraint optimization based on this model in each control cycle, while incorporating safety constraints as hard constraints into the solution. The optimized output setpoints are sent to actuators such as temperature zone heating, feeding / speed, back pressure and cooling to form a closed loop, and the predicted values ​​are used to trigger the dual-loop filter separation-condensation non-stop switching in advance. Finally, the process data of the switching and interlocking events are written back for online model correction, thereby realizing adaptive closed-loop process control of "prediction-optimization-execution-writeback".

[0032] Compared to traditional control methods that rely on fixed setpoints and manual intervention or only post-event threshold alarms, this invention achieves dynamic matching between the pyrolysis and condensation sections within the feasible domain through the coordinated adjustment of segmented temperature control and rolling optimization. On the one hand, the linkage optimization of temperature zone settings with feed / speed / back pressure / cooling intensity suppresses increased gasification, coking, and load fluctuations caused by local over-pyrolysis, resulting in an overall increase in oil yield with reduced fluctuations, lower energy consumption, or better unit energy consumption. On the other hand, advance handling based on differential pressure prediction and dual-loop non-stop switching significantly reduce the peak pressure difference of filtration separation and the risk of blockage shutdown, reducing manual intervention and unplanned shutdowns, and improving the stability of continuous operation. At the same time, due to the introduction of setpoint change rate constraints and penalty terms in the optimization, the setpoints of each temperature zone exhibit smooth and small rolling changes, making it easier for the actuator to reach and reducing process disturbances. Ultimately, it balances safety, stability, and economy without relaxing safety boundaries such as oxygen content and pressure. Attached Figure Description

[0033] Figure 1 This is a system structure block diagram of a waste plastic continuous pyrolysis oil production unit.

[0034] Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0035] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of the present invention.

[0036] I. Terminology Explanation

[0037] To ensure a clear and complete description of the technical solution of this invention, the following conventions are used for commonly used terms.

[0038] A / B / C Temperature Zones: Three heating and temperature control zones are set sequentially along the material feeding direction and are thermally isolated from each other; A is the feed-end start-up pyrolysis zone, B is the middle main pyrolysis zone, and C is the discharge-end deep pyrolysis and component setting zone; each temperature zone is equipped with an independent temperature measurement and heating adjustment unit.

[0039] Control cycle : The cycle of rolling optimization calculation and setting value issuance.

[0040] Prediction Time Domain The length of the future time window used for prediction and constraint evaluation in rolling optimization.

[0041] Thin film thickness The thickness of the film formed by the spiral film-forming blade scraping the inner wall of the pyrolysis cylinder.

[0042] Filtration separation pressure difference The pressure difference between the top inlet and the bottom of the filtration separation tower.

[0043] upper limit of pressure difference : Maximum permissible differential pressure threshold.

[0044] Oxygen content The online oxygen content (volume fraction) of the differential pressure tank / oil and gas pipeline is required to meet the following requirements. .

[0045] System pressure The condensing system pressure, regulated by interlocking with the compressor via the back pressure valve, is maintained at... scope.

[0046] Oil yield : Mass fraction of mixed oil produced per unit feed mass (%).

[0047] Condensation efficiency index : An online index used to characterize the strength of condensation, which can be calculated by combining temperature difference and liquid level change rate.

[0048] II. System Structure

[0049] like Figure 1 As shown, the system of the present invention includes a device layer, a control layer and a data storage layer.

[0050] 2.1 Equipment Layer

[0051] The device layer includes at least:

[0052] Melt extrusion equipment: a single-screw extruder with a feeder that heats and melts waste plastics and provides stable feed; typically, it can be configured with 7 temperature measuring points (4 points at the electromagnetic heating point, 2 points at the ceramic heating point, and 1 point at the flexible connection heating point), and is equipped with an extrusion pressure sensor and a motor current sensor.

[0053] Membrane pyrolysis reactor: A geared motor drives the rotating shaft to rotate, which in turn drives the spiral film-forming blades and the spiral cleaning structure; heating elements are arranged on the outside of the pyrolysis cylinder; the reactor is divided into A / B / C temperature zones along the material propulsion direction. Temperature / pressure measuring points can be configured, such as up to 30 temperature measuring points and 3 pressure measuring points, facilitating zoned monitoring and interlocking judgment.

[0054] Filtration and separation equipment: used for the separation and purification of coke powder / droplets / heavy components in oil and gas; can be configured with 15 temperature measuring points (reboiler, reboiler body, separator, filter tank, separation outlet, etc.), inlet / outlet pressure, bottom pressure and differential pressure. Sensor, bottom liquid level sensor.

[0055] Condensation equipment: multi-stage condenser + oil tank; can be configured with 12 temperature measuring points (3 for water inlet, 6 for water return, and 3 for oil outlet) and flow measuring points; oil tank level, pressure and pressure relief valves.

[0056] Non-condensable gas treatment equipment: compressor + gas storage tank; and interlocked with differential pressure tank and back pressure valve.

[0057] Dual-loop configuration: Two independent filtration-separation-condensation systems, each with a differential pressure tank before the compressor; loop switching is achieved through a three-way valve at the reactor gas phase outlet and a three-way valve after the differential pressure tank, ensuring continuous operation without shutting down the machine.

[0058] Cold nitrogen rapid charging device: used for emergency cooling of membrane pyrolysis reactor in case of power failure or temperature runaway, and to quickly stop the pyrolysis reaction; and can be equipped with pressure relief valve and temporary storage tank.

[0059] 2.2 Control Layer and Data Storage Layer

[0060] The control layer adopts a layered architecture with basic closed-loop, upper-level rolling optimization, and safety interlocking.

[0061] Basic closed loop: A / B / C temperature control loop, back pressure valve pressure control loop, cooling water / refrigerant regulation loop, extruder speed and feeding control loop, nitrogen flow control loop, etc.

[0062] Rolling optimization module: periodically reads data, calls the prediction model, solves constraint optimization, and outputs the set values ​​for the next cycle: temperature zone setting, rotation speed, feed rate, nitrogen, back pressure, and cooling intensity.

[0063] Loop switching control module: Based on differential pressure prediction / criteria, the three-way valve is controlled to realize dual-loop switching, and the back pressure valve and compressor are linked during the switching.

[0064] Safety interlock module: triggers cold nitrogen and shutdown interlocks in response to over-temperature, over-pressure, or excessive oxygen content; and provides interlock bypass / reset strategies.

[0065] Data storage layer: Stores raw time-series data, filtered features, optimization solution logs, setpoint issuance records, alarm and interlock events, and model parameter versions; used for traceability and online write-back updates.

[0066] III. Detailed Implementation of the Method

[0067] like Figure 2 As shown, the process control method of the present invention includes: S1 multi-source data acquisition and preprocessing; S2 segmented object establishment; S3 construction of prediction model and safety constraint set; S4 rolling solution of constraint optimization; S5 pressure difference prediction triggering non-stop loop switching; S6 safety interlock and model write-back update.

[0068] 3.1, S1: Multi-source data acquisition, time synchronization and preprocessing

[0069] (1) Data collection object and sampling frequency

[0070] The objects to be collected include at least:

[0071] Melt extrusion section: temperature, extrusion pressure, motor current, feeder speed / opening;

[0072] Pyrolysis reactor: Temperature zones A / B / C A / B / C pressure Reactor total pressure / bottle pressure, rotor / screw speed Heating power ;

[0073] Filtration and separation section: inlet / outlet pressure, bottom pressure, differential pressure Bottom liquid level Temperature of key components;

[0074] Condensation section: Inlet and outlet temperatures of each stage of the condenser Oil outlet temperature, cooling water flow rate Oil tank level Oil tank pressure;

[0075] Differential pressure tank: online oxygen content ,pressure ;

[0076] Actuator status: Back pressure valve opening Three-way valve position, cold nitrogen valve position, etc.

[0077] The sampling frequency can be set according to the project: the basic closed-loop layer usually samples at 1–5 seconds; the rolling optimization layer can perform 1-minute aggregation on the original data (mean, variance, rate of change, etc.) to reduce noise and computational burden.

[0078] (2) Time synchronization and exception handling

[0079] All data collection points use the same timestamp, with a maximum allowable deviation of no more than 1 second;

[0080] Missing value handling: Short-term missing values ​​(<3 sampling periods) are filled with forward padding; long-term missing values ​​trigger model degradation / optimization shutdown strategy, maintaining only the basic closed loop and safety interlock;

[0081] Outlier handling: using 3 Rules or IQR rules remove obvious outliers and retain the original records for retrospective purposes.

[0082] (3) Feature construction

[0083] Construct "time lag and statistical" features for key variables, and for any variable structure:

[0084] moving average Sliding variance , Take 5–15 minutes;

[0085] rate of change ;

[0086] Time Delay Term To reflect thermal inertia and mass transfer hysteresis, Take 5–20 minutes.

[0087] 3.2, S2: Segmentation object creation and temperature zones A / B / C

[0088] (1) The meaning of temperature zones A / B / C and their temperature control boundaries

[0089] Temperature Zone A: Start-up Pyrolysis / Volatile Section: Located near the feed end, the goal is to rapidly heat the film material and initiate initial pyrolysis and volatilization.

[0090] B-zone main cracking section: middle section, responsible for the main cracking and generation of oil and gas;

[0091] C-temperature zone deep cracking and settling section: at the discharge end, it further cracks heavy components and stabilizes the distribution of oil and gas components.

[0092] Thermal isolation between the three temperature zones is achieved through insulation structures or heating partitions; each temperature zone is... Independent adjustment within the range. Settings: The initial values ​​for typical operating conditions are approximately 380℃ for Zone A, approximately 395℃ for Zone B, and approximately 420℃ for Zone C. These can also be adjusted to the nearby range through rolling optimization.

[0093] (2) Membrane formation and cleaning of membrane pyrolysis reactor

[0094] A typical membrane pyrolysis reactor consists of a pyrolysis cylinder equipped with film-forming blades and a spiral cleaning structure inside; and a heating element outside the cylinder. During operation, the beveled edges of the film-forming blades scrape the paste-like waste plastic onto the inner wall of the pyrolysis cylinder to form a thin film, the thickness of which is... Preferred The heating element on the outer wall of the pyrolysis cylinder heats the cylinder to... The membrane instantly decomposes and vaporizes to produce oil and gas, which enters the subsequent filtration and purification through the oil and gas outlet pipe; after decomposition, carbon powder adheres to the inner wall, and the spiral cleaning structure cleans the carbon powder and collects it into the decomposition slag bin, realizing continuous and stable operation.

[0095] 3.3, S3: Prediction Model

[0096] The rolling optimization of this invention relies on the predictability of key performance indicators. Therefore, a hybrid prediction framework combining mechanistic constraints and data-driven approaches is employed.

[0097] 3.3.1 Predicting Output and Input Variables

[0098] (1) Predict the output vector

[0099] In each rolling cycle Define the prediction output:

[0100] ;

[0101] in: To predict the oil yield (%); To predict the carbon slag ratio (%); To predict condensation efficiency; To predict the future Filtration separation pressure difference after time; This is the predicted rate of increase in differential pressure.

[0102] (2) Input feature vector

[0103] Define input features:

[0104] ;

[0105] in, For system pressure, This is the feed mass flow rate from the melt extruder to the pyrolysis reactor.

[0106] 3.3.2 Condensation efficiency index

[0107] To avoid relying on oil and gas quality flow rates that are difficult to obtain online, Defined as a normalized combination of the temperature difference term and the liquid level change rate term:

[0108] ;

[0109] in:

[0110] For cooling water temperature rise;

[0111] The rate of increase of the oil tank level;

[0112] , The maximum value within the historical window is used for normalization;

[0113] For weights.

[0114] 3.3.3 Model Structure

[0115] (1) Mechanism constraint layer (level 1)

[0116] The first-level model is used to provide trends and boundaries, with temperature, power, and feed rate as the main drivers, to construct coarse predictions:

[0117] , ; where the mapping function Empirical polynomials or piecewise linear relationships can be adopted. Robust boundaries are provided when data is insufficient or operating conditions change abruptly, preventing data model extrapolation from going out of control.

[0118] (2) Data-driven residual layer (second level)

[0119] The residuals of the second-level model learning the first-level model:

[0120] ;

[0121] ;

[0122] ;

[0123] The final prediction is:

[0124] ;

[0125] ;

[0126] ;

[0127] For the second-level model algorithm, Gradient Boosting Decision Tree (GBDT) can be selected, which is friendly to small samples, nonlinearity, and feature interaction.

[0128] 3.3.4 Dataset Source, Scale, Partitioning, and Label Acquisition

[0129] (1) Data source

[0130] Source: Historical operating data of continuous pilot / production lines (DCS / PLC timing library), laboratory / measurement data;

[0131] Time span: 45 days of stable operation data;

[0132] Original sampling: 1–5 s; Data granularity used for modeling: aggregated by 1 min (mean, variance, rate of change, time lag term).

[0133] (2) Tag acquisition method

[0134] Oil yield Mass balance calculations are performed by combining oil tank metering (mass flow meter or timed weighing) with feed metering (screw scale / loss-in-weight scale);

[0135] carbon slag ratio : Timed / cumulative weighing of pyrolysis slag silos;

[0136] Non-condensable gas ratio: estimated by the compressor flow meter and the gas storage tank;

[0137] Calculated using the above formula based on online temperature and liquid level;

[0138] pressure difference The differential pressure is directly measured by the differential pressure transmitter; the rate of increase in differential pressure. Approximated by difference.

[0139] (3) Sample division

[0140] The training set comprises 70%, the validation set 15%, and the test set 15%; the sets are divided in chronological order to avoid information leakage.

[0141] 3.3.5 Training Methods and Hyperparameters

[0142] Taking GBDT as an example, the reproducible range of hyperparameters is given:

[0143] Number of trees ;

[0144] Maximum depth ;

[0145] Learning rate ;

[0146] Subsampling rate ;

[0147] Minimum leaf sample ;

[0148] Loss function: regression squared error or Huber loss;

[0149] Early stopping strategy: Validation set error is continuous If the wheel does not descend, then stop.

[0150] The evaluation indicators should include at least:

[0151] ;

[0152] ;

[0153] in For the sample size, For the true value, These are predicted values.

[0154] 3.3.6 Online Update

[0155] Online updates employ a lightweight strategy of sliding window retraining and bias correction:

[0156] Every Extract the most recent data in each cycle (e.g., every 60 minutes). Hourly window data (e.g., 24 hours) forms incremental samples;

[0157] The strategy can be either "adding small trees and freezing old trees" or "retraining but limiting the number of trees"; for MLPs, a small learning rate can be used for fine-tuning.

[0158] Force the retention of data before and after the most recent interlocking / circuit switching event as key samples to improve the predictability of abnormal operating conditions.

[0159] 3.4, S4: Objective function, constraints and solution for rolling optimization

[0160] 3.4.1 Issuing setting values

[0161] In the cycle Solve for the setpoint vector of the next period :

[0162]

[0163] 3.4.2 Set of Constraints

[0164] It must include at least the following hard constraints:

[0165] ;

[0166] ;

[0167] ;

[0168] ;

[0169] Rate of change of implementing agency ;

[0170] ;

[0171] ;

[0172] in .

[0173] 3.4.3 Objective Function

[0174] The problem of maximizing oil yield, minimizing energy consumption and fluctuations, and minimizing pressure differential risk can be unified into a minimization problem:

[0175]

[0176] in:

[0177] Energy consumption per unit output (which can be calculated by electricity meter / power meter);

[0178] Weights (can be taken as) To maximize oil yield, Secondly, mitigate the risk of congestion.

[0179] 3.4.4 Solver

[0180] Solver: Sequential quadratic programming (SQP), interior point method, or simplified to quadratic programming (QP) (linearizing / piecewise linearizing the prediction model) can be used.

[0181] Initial value: Use the feasible solution from the previous cycle. Use it as an initial value to reduce convergence time;

[0182] Failure rollback: If the solution times out or is not feasible, rollback to the most recent feasible setpoint, limit, or base PID.

[0183] Safety Bypass: When the safety interlock is triggered, the rolling optimization immediately stops sending, only records and waits for a reset.

[0184] 3.5, S5: Differential Pressure Prediction Trigger

[0185] 3.5.1 Criteria for Differential Pressure Prediction

[0186] Relying on Post-event triggering can easily lead to short-term overruns and downtime risks. This invention introduces predictive criteria, such as:

[0187] ;

[0188] in, As a pre-trigger threshold, such as setting To predict the lead time, such as 10-20 minutes.

[0189] 3.5.2 Switching Sequence Control

[0190] To ensure pressure continuity and prevent backflow, the switching can be performed in the following order.

[0191] Entering switching preparation state: Freeze temperature zone set change rate, back pressure valve enters pressure stabilization mode;

[0192] Open the critical valves of the backup circuit to the standby position and confirm that the differential pressure tank pressure is within the allowable range;

[0193] Switch the reactor gas phase outlet three-way valve to the standby circuit;

[0194] Synchronously switch the three-way valve after the differential pressure tank to establish a complete circuit;

[0195] Adjust the back pressure valve / compressor to make Stay ;

[0196] The original circuit is isolated, allowing for online maintenance / replacement of filter components;

[0197] Record switch timestamp The switching event is then written to the database for model updates and anomaly tracking.

[0198] 3.5.3 Segmented Modeling Updates Before and After Switching

[0199] Since switching alters the pressure drop characteristics and condensation load, this invention divides the data into two segments:

[0200] Switch front end ;

[0201] Switch to the back end ;

[0202] The pressure difference growth rate model or the residual model are updated respectively to enable the prediction to converge quickly on the new loop and avoid misjudgment caused by the reuse of the model.

[0203] 3.6, S6: Over-temperature / over-pressure / over-oxygen interlocking and cold nitrogen emergency response, and its model write-back

[0204] 3.6.1 Interlock Triggering Conditions

[0205] Temperatures in any temperature zone are excessively high;

[0206] Pressure exceeded: , , Any pressure exceeding the upper limit is considered overpressure;

[0207] Online oxygen content exceeds limit: ;

[0208] The differential pressure tank experienced excessive and uncontrolled pressure.

[0209] The temperature of the filtration and separation device ran out of control.

[0210] 3.6.2 Interlocking Action Sequence

[0211] Start rapid charging with cold nitrogen to urgently cool the membrane pyrolysis reactor;

[0212] Interlocking the extrusion and upstream feed to inhibit the further progress of the reaction;

[0213] Open necessary pressure relief / venting paths to prevent environmental emissions;

[0214] The downstream equipment is shut down according to the preset delay sequence to avoid abnormal liquid level / pressure in the condensation section;

[0215] Full event log: trigger cause, trigger point, temperature / pressure trajectory, action execution time, and recovery conditions.

[0216] 3.6.3 Interlocking data write-back for model correction

[0217] The data before and after interlocking will be used as samples of extreme operating conditions for focused updates:

[0218] Dynamic response model from temperature runaway to cooling;

[0219] The relationship between differential pressure tank pressure and back pressure valve / compressor linkage;

[0220] Evolution of oxygen content anomalies and a model of nitrogen replacement efficiency.

[0221] This allows subsequent rolling optimizations to identify potential risks earlier and take lenient interventions, reducing the probability of interlocking.

[0222] IV. Specific Application Examples

[0223] The following example uses a 10 t / d pilot-scale PE / PP waste plastic continuous pyrolysis oil production unit, with a 12-hour pilot test. The tests used the same batch of raw materials, the same equipment, and the same safety thresholds, differing only in the control strategy.

[0224] 4.1 Experimental conditions and evaluation indicators

[0225] (1) Raw materials and operating conditions

[0226] Raw materials: PE / PP mixed waste plastics, moisture content <1%, impurities <2%, particle size 10–20 mm;

[0227] Feed rate: ;

[0228] Thin film thickness: ;

[0229] Initial temperature zone settings: A 380℃, B 395℃, C 420℃;

[0230] Oxygen content constraint: Nitrogen purging three times before starting the engine;

[0231] System pressure constraints: ;

[0232] Differential pressure threshold: Pre-trigger ;

[0233] Scrolling optimization parameters: Predicting the time domain Forecast lead time .

[0234] (2) Evaluation indicators

[0235] Average oil yield and standard deviation;

[0236] Carbon slag ratio, non-condensable gas ratio;

[0237] Unit power consumption (kWh / t);

[0238] System pressure standard deviation (reflects pressure stabilization capability);

[0239] Number of times exceeding limits / interlocks within 24 hours;

[0240] Peak pressure difference during filtration and separation versus the number of manual interventions.

[0241] 4.2 Comparison of Control Methods: Example 1 and Comparative Example 1

[0242] 4.2.1 Example 1: The present invention features segmented temperature control, rolling optimization, differential pressure prediction, and non-stop switching.

[0243] Control method: Continuous optimization periodic output of A / B / C temperature zone settings, feed rate, rotation speed, nitrogen, back pressure, and cooling intensity; when... In advance, switch to the backup filter separation-condensation loop and update the model in segments based on the data before and after the switch.

[0244] 4.2.2 Comparative Example 1: Traditional fixed settings, manual intervention

[0245] Control method: Temperature zones A / B / C are fixed; the operator adjusts the feeding and cooling intensity based on experience; the filtration and separation operates in a single loop. achieve The system will then be shut down, manually cleaned, or subjected to a short-term load reduction; that is, the rolling optimization and predictive switching control method of this invention will not be used.

[0246] 4.2.3 Test Comparison Results

[0247] The hourly process data from the 12-hour pilot test are shown in Table 1.

[0248] Table 1. Comparison of 12-hour pilot tests

[0249]

[0250] The key indicators of Example 1 and Comparative Example 1 are shown in Table 2.

[0251] Table 2 Summary of Key Indicators for Comparative Example 1 and Example 1

[0252]

[0253] As can be seen from Tables 1 and 2, in the comparative example 1 of the traditional control method, the average oil yield is about 72% and fluctuates greatly. After the filter pressure difference increases, a significant peak appears, triggering manual cleaning and causing process disturbance. In the embodiment 1 of the present invention, the average oil yield is 75% and the fluctuation range is extremely small. The change range of filter pressure difference is significantly smaller, no over-limit events occur, and the process is continuous.

[0254] As can be seen from Tables 1 and 2, after adopting the segmented temperature control and rolling optimization control of the present invention, the system can achieve coordinated adjustment through rolling optimization without relaxing safety constraints such as oxygen content, system pressure, and filtration separation pressure difference. , , Key control parameters such as cooling intensity are used to ensure that the pyrolysis and condensation sections operate in a matched manner within the feasible region, thereby effectively suppressing unstable factors such as increased gasification and coking caused by local over-pyrolysis. This is reflected in an overall increase in oil yield and a significant convergence of fluctuations, as shown in Table 1. Meanwhile, based on... The differential pressure prediction criterion, combined with dual-loop switching sequence control, can complete the non-stop switching before the differential pressure approaches the threshold, avoiding the risk of differential pressure peak and shutdown due to blockage, and significantly reducing the probability of manual intervention and unplanned shutdown, as shown in Table 2.

[0255] 4.3 Comparison of Switching Methods

[0256] 4.3.1 Example 2: Adaptability under raw material fluctuation disturbances

[0257] Based on Example 1, the PP proportion in the raw material was increased from 40% to 60% at the 6th hour (simulating incoming material fluctuations), and maintained... and constant.

[0258] Without online updates: the model's prediction of the oil yield of new feedstocks will produce systematic biases, and rolling optimization will tend to set the temperature zone too high, leading to an increase in the gasification rate;

[0259] When using online updates: after approximately 2-3 rolling cycles, the model bias converges significantly, and the optimized output will... Set the temperature down by 2–5°C and make minor adjustments. and ,make If the oil yield remains within the target range, it will return to a stable level.

[0260] 4.3.2 Comparative Example 2: Switching using only the absolute pressure difference threshold

[0261] Control method: When Only switch the loop; after switching, the old model is still used, and no segmented updates are performed before and after the switch.

[0262] Results: Switching trigger lag. It is more likely to produce "peaks"; after switching, due to changes in pressure drop characteristics, the model becomes inaccurate for a short time, causing the set value to oscillate and the oil output rate to fluctuate more; more manual confirmation and intervention are required.

[0263] Statistics two hours after the disturbance: the oil yield decreased by approximately 1.2 percentage points in Comparative Example 2 and by approximately 0.4 percentage points in Example 2, followed by a rapid recovery, and the rate of increase in filtration differential pressure was also observed. The price decreased by about 15%, and the coking trend was suppressed.

[0264] In summary, by adopting the technical solution of this invention, the overall oil yield is improved and the fluctuation is significantly reduced without relaxing safety constraints such as oxygen content, system pressure, and differential pressure. At the same time, by predicting differential pressure and switching between dual loops without stopping the machine, the differential pressure risk is addressed in advance, avoiding differential pressure peaks and shutdowns / cleaning caused by blockages, thereby significantly reducing manual intervention and unplanned shutdowns.

[0265] The foregoing description of embodiments of the present invention, through which those skilled in the art are able to implement or use the present invention, will be readily apparent to those skilled in the art. Various modifications to these embodiments will be readily apparent to those skilled in the art. The general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novelty disclosed herein.

Claims

1. A process control method for segmented temperature control and rolling optimization in continuous pyrolysis of waste plastics to produce oil, characterized in that, The pyrolysis reactor is sequentially configured with thermally isolated temperature zones A, B, and C along the material flow direction, and each temperature zone is equipped with an independent temperature measurement and heating control unit; the method includes: S1, collect the temperature and extrusion pressure at multiple points of the waste plastic melting extrusion equipment, the temperature and pressure of the A temperature zone / B temperature zone / C temperature zone, the pressure difference and temperature of the filtration and separation section, the inlet and outlet temperatures and liquid levels of the condensation section, and the online oxygen content of the pressure difference tank; S2 establishes a segmented object with melt extrusion temperature, A / B / C temperature zones, and screw speed as the core control parameters. Among them, temperature zone A is the first heating section at the feed end, used to rapidly heat up the 0.3-1.5mm film material and initiate initial pyrolysis and devolatilization; temperature zone B is the second heating section in the middle section, used for main pyrolysis to generate oil and gas; temperature zone C is the third heating section at the discharge end, used for deep pyrolysis and oil and gas component setting, and the three temperature zones are independently controlled within the range of 350-450℃. S3, Construct a rolling prediction model: with oil yield, carbon slag rate and condensation efficiency as outputs, and the aforementioned manipulation amount, back pressure valve opening and inert gas flow rate as inputs, and set safety constraints: oxygen content ≤0.5%, system pressure controlled by back pressure at -5kPa~0, and pressure difference in the filtration and separation section not exceeding the set threshold. S4, solves the constraint optimization in a rolling manner according to the control cycle, and outputs the set values ​​of temperature zone A / B / C, screw speed, feed rate, inert gas replacement / replenishment amount, back pressure valve opening degree and condensation cooling intensity for the next cycle; S5, if it is predicted that the pressure difference of the filter separation section will exceed the limit, control the three-way valve to switch to another independent filter separation path to achieve continuous operation without stopping the machine, and model and update the data before and after the switch respectively. S6. If any temperature zone exceeds the limit, pressure exceeds the limit, or oxygen content exceeds the limit, then start rapid cold nitrogen charging and lock the extrusion and upstream feed, then write back and correct the rolling prediction model.

2. The process control method as described in claim 1, characterized in that, Temperature and pressure sensors are respectively installed in temperature zones A, B, and C; the temperature setpoint for zone A is 380℃, for zone B it is 395℃, and for zone C it is 420℃, and each zone performs temperature detection, indication, control, alarm, and safety interlock within its respective zone; and / or, the melt extrusion section uses a single-screw extruder with 7 temperature measurement points, including 4 temperature measurement points at the electromagnetic heating point, 2 temperature measurement points at the ceramic heating point, and 1 temperature measurement point at the flexible connection heating point; when the temperature at any of the 4 electromagnetic heating points becomes out of control, the melt extrusion section immediately stops, and the upstream equipment stops sequentially, and the downstream equipment stops sequentially after a preset delay.

3. The process control method as described in claim 1, characterized in that, The pyrolysis reactor is equipped with 30 temperature measurement points, 3 pressure measurement points, and 1 flow measurement point. The reactor and heating element are each equipped with 12 temperature measurement channels. When the temperature at any measurement point becomes out of control, the nitrogen switch valve opens and the pyrolysis reactor shuts down. The upstream equipment shuts down sequentially, and the downstream equipment shuts down sequentially after a preset delay. And / or, if any of the 3 pressure measurement points on the pyrolysis reactor detects a pressure exceeding a preset upper limit, it is determined to be an overpressure event. When an overpressure event occurs, the interlock action of stopping the extrusion and upstream equipment and opening the pneumatic ball valve is executed, and the entire line shuts down after a preset delay.

4. The process control method as described in claim 1, characterized in that, The differential pressure tank mentioned in step S1 is equipped with an online oxygen content detection device; before startup, the entire system is purged with nitrogen at least three times through the compressor and nitrogen system, and the oxygen content is controlled at ≤0.5% after purging; during normal operation, the oxygen content of the oil and gas is monitored in real time and controlled at ≤0.5%; And / or, the system pressure adopts a back pressure control method, with the reactor-filtration-condenser-compressor pre-buffer / differential pressure tank pipeline connected, and the compressor is adjusted by the back pressure valve set in front of the differential pressure tank in an interlocking manner, so that the system pressure is controlled within the range of -5kPa to 0 and the pyrolysis reactor is in a slightly positive pressure operating state.

5. The process control method as described in claim 1, characterized in that, Two independent filtration-condensation loops are set up, with a differential pressure tank installed before the compressor in each loop. Switching between the two loops is achieved through a three-way valve at the gas phase outlet of the pyrolysis reactor and a three-way valve after the differential pressure tank. The switching includes at least the sequential control of "exiting loop isolation - entering loop connection - back pressure recovery". And / or, a differential pressure sensor is installed in the filtration section to monitor the pressure difference ΔP between the top inlet and bottom of the filtration tower, and compares ΔP with a preset upper limit of differential pressure ΔPlim. When ΔP ≥ ΔPlim, an alarm is output and the loop switching in S5 is triggered. When the temperature in the filtration and separation section becomes out of control, the cold nitrogen switch is triggered to open and a shutdown interlock is executed.

6. The process control method as described in claim 1, characterized in that, The set of constraints for the rolling optimization in step S4 includes at least: upper and lower limits of temperature in temperature zones A / B / C, system pressure range, upper limit of oxygen content, upper limit of filtration separation pressure difference, and constraint on the rate of change of actuator; the objective function of the rolling optimization includes at least a term for maximizing oil yield, a term for minimizing energy consumption, and a term for minimizing temperature / pressure / pressure difference deviation; and / or, the prediction model is an online updatable model, and after each control cycle, the model parameters are updated online using the actual mixed oil production, cracked residue production, temperature / liquid level changes in the condensation section, and filtration separation pressure difference data, and the updated model is used to solve the problem in the next rolling cycle.

7. A segmented temperature control and rolling optimization process control system for continuous pyrolysis of waste plastics to produce oil, used to implement the process control method according to any one of claims 1-6, characterized in that, It includes the device layer, control layer, and data storage layer; The equipment layer includes at least a melt extrusion device, a membrane pyrolysis reactor with A / B / C temperature zones, a filtration and separation device, a condensation device, a compressor, a differential pressure tank, a cold nitrogen rapid charging device, and a valve actuator. The control layer includes at least: a temperature measurement module, a gas flow measurement module, a pressure detection and measurement module, an oxygen content detection module, a waste plastic melting analysis module, a pyrolysis temperature analysis module, an oil and gas condensation efficiency analysis module, a rolling optimization module, a loop switching control module, a safety interlock module, and a production early warning module. The data storage layer is used to store process time-series data, optimization setpoints, alarm / interlock events and model parameters, and provides write-back data to the rolling optimization module and the production early warning module.

8. The process control system as described in claim 7, characterized in that, The temperature measurement module is connected to at least 7 temperature measurement points of the melt extrusion equipment, 30 temperature measurement points of the pyrolysis reactor, 15 temperature measurement points of the filtration and separation equipment, and 12 temperature measurement points of the condensation equipment; and / or, the thermal pyrolysis temperature analysis module is used to calculate the target heating power or target temperature setpoint of temperature zones A / B / C based on film thickness, temperature of paste-like waste plastic, and material information, and uses the calculation results as the initial value or constraint boundary of the rolling optimization module.

9. The process control system as described in claim 7, characterized in that, The control layer includes an industrial controller, which includes a processor, a memory, and an industrial communication interface. The industrial communication interface is used to exchange data with temperature / pressure / differential pressure / oxygen content sensors, as well as heating elements, frequency converters, back pressure valves, three-way valves, cooling actuators, and cold nitrogen switching valves.

10. A computer-readable storage medium having a computer program stored thereon, the computer program implementing the process control method according to any one of claims 1-6 when executed by a processor.

Images