Accurate pyrolysis regulation and control and product analysis system and method for urban biomass solid waste

Abstract

The invention discloses a precise pyrolysis regulation and control and product analysis system and method for urban biomass solid waste, and relates to the technical field of solid waste treatment. Comprising a material characteristic prediction module which is used for carrying out pyrolysis behavior characteristic prediction through a machine learning prediction model after preprocessing urban biomass solid wastes, and obtaining a corresponding pyrolysis behavior characteristic report; the process pyrolysis regulation and control module is used for constructing a feature target vector according to the pyrolysis behavior feature report and a user product quality target, acquiring a corresponding candidate process track through a deep learning prediction model, and adjusting the candidate process track in real time according to product quality data in a pyrolysis reaction treatment process; and the product collection module is used for collecting and analyzing products obtained after the pyrolytic reaction is carried out according to the candidate process tracks. According to the method, the problem of unstable product quality caused by raw material fluctuation in a traditional method can be solved, and the reliability of pyrolysis behavior prediction is improved.

Description

Technical Field

[0001] This invention relates to the field of solid waste treatment technology, specifically to a system and method for precise pyrolysis control and product analysis of urban biomass solid waste. Background Technology

[0002] With the acceleration of urbanization and the continuous improvement of residents' living standards in my country, the amount of municipal solid waste (MSW) generated has continued to rise. According to statistics from the Ministry of Housing and Urban-Rural Development, the amount of urban domestic waste collected nationwide exceeded 270 million tons in 2023, of which biomass components (such as kitchen waste, garden waste, waste paper, and wood materials) accounted for more than 50%. This type of urban biomass solid waste is characterized by high moisture content, complex composition, and easy decomposition. If traditional landfill or incineration methods are used for treatment, it will not only occupy a large amount of land resources, but also easily cause greenhouse gas emissions, leachate pollution, and the generation of toxic byproducts such as dioxins, posing a serious threat to the ecological environment and public health. In recent years, pyrolysis technology has been regarded as an important path to realize the resource utilization of urban biomass waste because it can convert organic matter into high-value-added energy products (such as bio-oil, syngas, and biochar) under anaerobic or hypoxic conditions.

[0003] Chinese invention patent application CN115746886A discloses a multi-source solid waste synergistic thermal treatment method for soil ecological restoration, including the following steps: screening and crushing; pretreatment; drying; pyrolysis: adding the dried solid waste into a pyrolysis device for pyrolysis treatment; high-temperature melting: sending the remaining solid waste after pyrolysis into a gasifier and heating it to a molten state at a temperature of 1400℃-1500℃ with pure oxygen to perform melting treatment, generating tail gas and molten residue; recycling. This invention adopts a method of first oxy-free pyrolysis followed by high-temperature melting to effectively treat solid waste, avoiding various problems such as incomplete treatment, secondary pollution, and serious resource waste caused by existing treatment methods such as sanitary landfill, biocomposting, and incineration. It also fully recovers and reuses the combustible gas, organic liquid, and solid residue generated during the treatment process, which is energy-saving and environmentally friendly, and suitable for large-scale promotion.

[0004] However, the above-mentioned and similar technical solutions still have the following shortcomings: Due to the complex sources and large fluctuations in composition of urban biomass solid waste, its pyrolysis behavior is difficult to predict, resulting in significant fluctuations in the yield distribution (ratio of gas, liquid, and solid phases) and quality (such as calorific value and component purity) of the final product. This severely restricts the consistency and high-value utilization of the product. Furthermore, the means of controlling key process parameters are limited and cannot dynamically respond to specific material characteristics, thus failing to obtain the target product stably and accurately. Summary of the Invention

[0005] The purpose of this invention is to provide a system and method for precise pyrolysis control and product analysis of urban biomass solid waste, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a precise pyrolysis control and product analysis system for urban biomass solid waste, comprising: Material characteristic prediction module: After pretreatment of urban biomass solid waste, the module uses a machine learning prediction model to predict pyrolysis behavior characteristics and obtains a corresponding pyrolysis behavior characteristic report. Process pyrolysis control module: Based on the pyrolysis behavior feature report and user product quality target, construct a feature target vector, and use the feature target vector as input to a deep learning prediction model to output the corresponding candidate process trajectory. At the same time, monitor and acquire product quality data during the pyrolysis reaction process, and adjust the candidate process trajectory in real time by comparing the product quality data with the user product quality target. Product collection module: collects and analyzes the products after pyrolysis reaction according to the candidate process trajectory.

[0007] Furthermore, the corresponding pyrolysis behavior characteristic report is obtained, including: SA1: Physical pretreatment: crushing, drying and homogenizing urban biomass solid waste to obtain pretreated urban biomass solid waste; SA2: Feature Recognition: The pre-processed urban biomass solid waste is scanned using a near-infrared spectral probe to obtain the spectral signal of the biomass solid waste. The spectral signal of the biomass solid waste and the spectral fingerprint database are used as inputs to a principal component analysis model, and the corresponding feature probability distribution results are output. SA3: Predictive Processing: Based on the characteristic probability distribution results, at least one corresponding material quantitative correction model is determined, and the spectral signal of the biomass solid waste is used as the input of each material quantitative correction model. The corresponding characteristic values ​​are output and obtained. At the same time, the characteristic values ​​are used as the input of the machine learning prediction model and the corresponding prediction index is output. Based on the similarity probability corresponding to the characteristic probability distribution results, the prediction index corresponding to each material quantitative correction model is weighted and fused to obtain the corresponding pyrolysis behavior characteristic prediction report.

[0008] Furthermore, the pre-treated urban biomass solid waste is obtained, including: SA1.1: Crushing process: The urban biomass solid waste to be processed is sequentially transported to the coarse crusher and fine crusher through the conveying mechanism to obtain the final waste fragments; SA1.2: Drying process: The final waste fragments are dried by means of a connected drum dryer and hot air furnace; SA1.3: Homogenization treatment: The dried final waste fragments are laid out in layers inside the homogenization chamber and subjected to shearing motion to obtain pre-treated urban biomass solid waste.

[0009] Furthermore, based on the known spectral data corresponding to multiple different biomass solid wastes, a spectral fingerprint database is constructed. Then, using the principal component analysis model, the spectral signals of the biomass solid wastes are compared with multiple reference spectral signals in the spectral fingerprint database to determine the corresponding similar reference spectral signals and similarity probabilities.

[0010] Furthermore, the candidate process trajectory is adjusted in real time, including: SB1: Trajectory Determination: The feature target vector composed of the pyrolysis behavior feature prediction report and the user product quality target is used as the input of the deep learning prediction model. The output obtains multiple different candidate process trajectories. At the same time, simulation and deduction are performed on all candidate process trajectories, and the corresponding optimal candidate process trajectory is determined based on the simulation and deduction results. SB2: Trajectory Monitoring: Through a distributed control system, the optimal candidate process trajectory is analyzed to obtain the corresponding execution instructions. Based on the execution instructions, different actuators in the pyrolysis reactor are driven. At the same time, during the driving of the actuators, detection data in the pyrolysis reactor, including temperature data, pressure data, and gas composition content data, are collected. SB3: Trajectory Adjustment: The detected data is compared with the process prediction value in the optimal candidate process trajectory to obtain the corresponding real-time data deviation. Simultaneously, the real-time data deviation is compared with a preset trajectory threshold to determine the adjustment state of the candidate process trajectory. Specifically: When the real-time data deviation is greater than the preset trajectory threshold, steps SB1 and SB2 are repeated to adjust the candidate process trajectory until the real-time data deviation is not greater than the preset trajectory threshold; otherwise, the candidate process trajectory is not adjusted.

[0011] Furthermore, the corresponding optimal candidate process trajectories are determined, including: SB1.1: Data Prediction: Based on the set user product quality target, determine the corresponding objective function, and combine the objective function with the final prediction index corresponding to the pyrolysis behavior feature prediction report to obtain the corresponding feature target vector. At the same time, use the feature target vector as the input of the deep learning prediction model and output to obtain multiple different candidate process trajectories. SB1.2: Simulation Determination: Through the set digital twin, each candidate process trajectory is simulated and deduced to obtain the corresponding simulation results. At the same time, the maximum simulation result is determined from the simulation results corresponding to each candidate process trajectory. The candidate process trajectory corresponding to the maximum simulation result is the optimal candidate process trajectory.

[0012] Furthermore, the deep learning prediction model sets constraints during trajectory prediction, including safety constraints, process constraints, environmental constraints, and product quality constraints.

[0013] Furthermore, gas samples from the pyrolysis reactor are collected using a high-temperature sampling probe and a heat-traced sampling pipeline. The gas samples are then pretreated by passing them through a primary condenser, a filter, and a dryer. Simultaneously, the pretreated gas samples are analyzed using a gas analyzer to obtain the corresponding gas component content data.

[0014] Furthermore, the products of the pyrolysis reaction following the candidate process trajectory are collected and analyzed, including: SC1: Gas phase product processing: After cooling and separating the high-temperature gas in the pyrolysis reactor, the gas is then subjected to deacidification and filtration processes to obtain pretreated gas, which is then temporarily stored in a gas storage tank. SC2: Liquid product processing: The condensed liquid product is separated into layers using a centrifuge, and the separated liquid products are stored in a storage tank; SC3: Solid product handling: The solid product is cooled to room temperature in a closed cooler using inert gas and then transferred to a biochar silo for temporary storage.

[0015] A method for precise pyrolysis control and product analysis of urban biomass solid waste, using any one of the above-mentioned systems for precise pyrolysis control and product analysis of urban biomass solid waste.

[0016] Compared with the prior art, the beneficial effects of the present invention are: Firstly, this invention uses physical preprocessing combined with near-infrared spectroscopy and principal component analysis models to rapidly identify the characteristics of complex urban biomass solid waste. It also uses a machine learning prediction model to weightedly fuse the quantitative correction results of multiple materials to generate a corresponding pyrolysis behavior characteristic report. This overcomes the problem of unstable product quality caused by raw material fluctuations in traditional methods and improves the reliability of pyrolysis behavior prediction. Secondly, based on the product quality target set by the user, the present invention inputs the feature target vector into the deep learning prediction model to generate multiple sets of candidate process trajectories. At the same time, it combines digital twin technology to simulate and deduce the candidate trajectories, select the optimal process scheme, and monitor the temperature, pressure and gas composition data in the reactor in real time through a distributed control system to dynamically adjust the process parameters, thereby ensuring that the reaction process is always optimized towards the target product. Thirdly, this invention uses high-temperature sampling, gas pretreatment and multi-stage analysis to accurately monitor the gas phase composition and adjust the process trajectory based on real-time data deviation feedback, thereby avoiding the decline in product quality due to fluctuations in operating conditions. Fourthly, the fully automated control of this invention can reduce reliance on human experience and adaptively adjust process parameters through intelligent algorithms, thereby reducing energy consumption and time costs caused by trial and error, and improving overall processing efficiency and economy. Attached Figure Description

[0017] Figure 1 This is a system block diagram of the precise pyrolysis control and product analysis system in this invention; Figure 2 This is a flowchart illustrating the material characteristic prediction method of the present invention; Figure 3 This is a schematic diagram of the process pyrolysis control method in this invention; Figure 4 This is a schematic diagram illustrating the process for determining the optimal candidate process trajectory in this invention. Detailed Implementation

[0018] 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.

[0019] refer to Figure 1This embodiment provides a precise pyrolysis control and product analysis system for urban biomass solid waste. The system includes a material characteristic prediction module, a process pyrolysis control module, and a product collection module. The material characteristic prediction module performs preliminary physical processing on the urban biomass solid waste, including crushing, drying, and homogenization, to obtain pretreated urban biomass solid waste. Simultaneously, a machine learning prediction model is used to predict the pyrolysis behavior characteristics of the pretreated urban biomass solid waste, generating a corresponding pyrolysis behavior characteristic prediction report. The process pyrolysis control module combines the pyrolysis behavior characteristic prediction report obtained by the material characteristic prediction module with the user's product quality target to obtain a corresponding feature target vector. This feature target vector is then used as input to a constructed deep learning prediction model, outputting a corresponding process trajectory, including a coordinated variation scheme for temperature, atmosphere flow rate, and stirring rate. Based on the obtained process trajectory, the pretreated urban biomass solid waste undergoes pyrolysis, and product quality data during the pyrolysis process is monitored in real time. The process trajectory is then adjusted in real time based on the monitored product quality data. It is worth noting that during real-time adjustment of the process trajectory, the priority of the process trajectory adjustment can be specifically set according to the user-defined high-value guidance mode. Specifically, when the high-value guidance mode is to maximize the calorific value of the fuel gas, priority is given to ensuring the stability of temperature and atmosphere. When the high-value guidance mode is to obtain light bio-oil, priority is given to precisely controlling the heating rate and condensation temperature. When the high-value guidance mode is to prepare high specific surface area biochar, priority is given to ensuring the residence time in the high-temperature section and the purity of the atmosphere. The product collection module is used to collect the products (including gaseous products, liquid products, and solid products) after the pyrolysis reaction in the process pyrolysis control module, and to perform relevant analysis and processing on the collected products.

[0020] This embodiment also provides a method for precise pyrolysis control and product analysis of urban biomass solid waste, which uses the aforementioned system for precise pyrolysis control and product analysis of urban biomass solid waste.

[0021] In this embodiment, the real-time raw material characteristics of preprocessed urban biomass solid waste are compared with a spectral fingerprint database to obtain the corresponding feature probability distribution results. Simultaneously, both the real-time raw material characteristics of the preprocessed urban biomass solid waste and the corresponding feature probability distribution results are used as input to a machine learning model, outputting a corresponding pyrolysis behavior feature prediction report. (Reference) Figure 2 This embodiment provides a material characteristic prediction method, which specifically includes the following steps: Step SA1: Physical Pretreatment. This involves the preliminary physical processing of the municipal biomass solid waste to be treated, including crushing, drying, and homogenization, to obtain relatively stable and homogeneous reaction raw materials, i.e., the pretreated municipal biomass solid waste. Details are as follows: Step SA1.1: Crushing Processing. This involves conveying the municipal biomass solid waste to be processed to the feed inlet of a primary crusher via a grab bucket or conveyor belt for initial crushing. In other words, the primary crushing process yields waste fragments smaller than 10-15 cm. It is worth noting that the feed inlet of the primary crusher is equipped with a grating to prevent oversized waste from entering. Specifically, the primary crusher in this embodiment includes a shear crusher and a jaw crusher. The shear crusher uses two counter-rotating rollers with blades to forcibly separate the municipal biomass solid waste into multiple waste fragments through shearing and tearing forces. Simultaneously, the jaw crusher uses fixed and movable jaw plates to crush the municipal biomass solid waste into multiple waste fragments through compression and grinding.

[0022] Furthermore, the discharge port of the coarse crusher is connected to the feed port of the fine crusher to further crush the multiple waste fragments obtained from the coarse crusher, obtaining final waste fragments with uniform particle size and relatively regular shape. Specifically, the fine crusher in this embodiment includes a hammer crusher, which uses multiple rows of hammers hinged on the rotor inside the hammer crusher to repeatedly crush the initial waste fragments obtained from the coarse crusher. Meanwhile, a replaceable screen is provided at the bottom of the fine crusher, and the size of the screen openings can be specifically set according to actual needs, so it is not specifically described in this embodiment. That is to say, the size of the final waste fragments can be specifically screened according to the set screen opening size.

[0023] Step SA1.2: Drying Process. The obtained final waste fragments are transported to the inlet of a rotary drum dryer via a screw conveyor or belt conveyor. Rotating lifting plates or hoisting plates inside the dryer scoop up and scatter the fragments, creating a uniform curtain of material across the drum's cross-section. Simultaneously, a hot air furnace connected to the dryer's outlet supplies high-temperature hot air into the dryer. It's worth noting that in this embodiment, the high-temperature hot air temperature is set between 400-600℃, which can be adjusted according to actual needs. This mixes the high-temperature hot air with the curtain of waste fragments, transferring heat from the hot air to the fragments through convection heat transfer.

[0024] Step SA1.3: Homogenization. The dried waste fragments from Step SA1.2 are transported to a homogenization chamber via a conveyor (e.g., a belt conveyor) to evenly distribute the dried waste fragments inside the chamber. Simultaneously, a twin-helix agitator within the homogenization chamber shears the layered dried waste fragments, resulting in a uniformly mixed batch of dried waste fragments, i.e., pretreated urban biomass solid waste.

[0025] Step SA2: Feature Recognition. This involves scanning and identifying the pre-processed urban biomass solid waste obtained in Step SA1.3 using a near-infrared spectral probe to acquire the corresponding original biomass solid waste spectral signal. Simultaneously, the original biomass solid waste spectral signal undergoes spectral preprocessing, including but not limited to baseline drift elimination and vector normalization, to obtain the pre-processed biomass solid waste spectral signal. In other words, the pre-processed biomass solid waste spectral signal only includes spectral signals related to the chemical molecular structure of the material. It is worth noting that the spectral preprocessing methods in this embodiment are all conventional techniques, and therefore are not specifically described in this embodiment.

[0026] Furthermore, based on known spectral data corresponding to multiple different biomass solid wastes, a corresponding spectral fingerprint database is constructed. Simultaneously, both the constructed spectral fingerprint database and the preprocessed biomass solid waste spectral signals are used as inputs to a principal component analysis (PCA)-based model. Through PCA, the preprocessed biomass solid waste spectral signals are compared with each reference spectral signal in the spectral fingerprint database to determine the similar reference spectral signals corresponding to the preprocessed biomass solid waste spectral signals and their corresponding similarity probabilities. In other words, through PCA, the characteristic probability distribution results of the corresponding preprocessed biomass solid waste spectral signals are obtained, including the corresponding similar reference spectral signals and their corresponding similarity probabilities.

[0027] Step SA3: Predictive Processing. Based on the feature probability distribution obtained in Step SA2, the material category corresponding to each similarity probability is determined. Simultaneously, based on the determined material category, the corresponding material quantitative correction model (e.g., partial least squares regression model) is selected. In this embodiment, the preprocessed biomass solid waste spectral signal obtained in Step SA2 is used as the input to each selected material quantitative correction model, and the output is the characteristic value corresponding to each material quantitative correction model.

[0028] Furthermore, the characteristic values ​​corresponding to the material quantitative correction model are used as input to the set machine learning prediction model, and the output is the corresponding prediction index, which is the prediction index corresponding to the material quantitative correction model. Specifically, based on the similarity probability corresponding to each material category, the prediction indices corresponding to each material category are weighted and fused to determine the final prediction index. In other words, based on the obtained final prediction index, a corresponding pyrolysis behavior characteristic prediction report is constructed.

[0029] In this embodiment, a constructed deep learning prediction model is used to combine the pyrolysis behavior feature prediction report and the user's product quality target to obtain the corresponding process trajectory. Based on the comparison between the product quality data obtained through real-time monitoring and the user's product quality target, the obtained process trajectory is adjusted in real time. (Reference) Figure 3 and Figure 4 This embodiment provides a method for controlling process pyrolysis, which specifically includes the following steps: Step SB1: Trajectory Determination. The feature target vector, composed of the pyrolysis behavior feature prediction report and the user product quality target, is used as input to the deep learning prediction model. The output yields multiple candidate process trajectories. Then, using a constructed digital twin, each candidate process trajectory is simulated and analyzed to determine the optimal candidate process trajectory. Details are as follows: Step SB1.1: Data Prediction. This involves setting user product quality goals through the user interface and determining the corresponding objective function based on these goals. In other words, at least one user product quality goal is weighted and combined to construct the corresponding objective function. Simultaneously, based on the pyrolysis behavior feature prediction report obtained in Step SA3, the corresponding final prediction index is determined. The determined final prediction index and the constructed objective function are then concatenated to obtain the corresponding feature target vector.

[0030] Furthermore, the obtained feature target vectors are used as input to the constructed deep learning prediction model, and the output yields multiple corresponding candidate process trajectories. It is worth noting that during trajectory prediction using the deep learning prediction model, corresponding constraints can be set according to actual needs, including safety constraints, process constraints, environmental constraints, and product quality constraints. In other words, based on different constraints and corresponding feature target vectors, multiple different candidate process trajectories can be obtained. Each candidate process trajectory includes a scheme for the coordinated variation of temperature, atmosphere flow rate, and stirring rate.

[0031] Step SB1.2: Simulation Determination. This involves using a digital twin to simulate and extrapolate each different candidate process trajectory obtained in Step SB1.1. Specifically, based on the user's product quality objectives, the pyrolysis behavior characteristic prediction report, and each candidate process trajectory, a digital twin is set up, and a simulation (virtual pyrolysis simulation) is performed to obtain simulation results for each candidate process trajectory, including the final product yield, calorific value, and composition. Simultaneously, the final product yield, calorific value, and composition corresponding to each simulation result are weighted and fused to obtain the corresponding simulation result data. It is worth noting that only data fusion is performed during the weighted fusion process; dimensional fusion is not involved.

[0032] Furthermore, based on the simulation results data corresponding to each candidate process trajectory, the maximum simulation result data is determined. In other words, the candidate process trajectory corresponding to the maximum simulation result data is the determined optimal candidate process trajectory.

[0033] Step SB2: Trajectory Monitoring. This involves analyzing the optimal candidate process trajectory determined in Step SB1.2 using a distributed control system to identify the execution commands for different actuators (such as temperature controllers, flow meters, and variable frequency motors) within the pyrolysis reactor. Specifically, the pretreated urban biomass solid waste obtained in Step SA1.3 is transferred to the pyrolysis reactor. Based on the identified execution commands, the corresponding actuators within the reactor are driven (e.g., controlling the temperature control loop based on temperature commands, and controlling the flow control loop based on flow commands), thereby performing the pyrolysis reaction on the pretreated urban biomass solid waste.

[0034] Furthermore, during the pyrolysis reaction process, a sensor network installed inside the pyrolysis reactor collects corresponding detection data, including temperature data, pressure data, and gas composition data. Specifically, multiple armored thermocouples and pressure transmitters are installed inside the pyrolysis reactor to detect and acquire corresponding internal temperature and pressure data. In this embodiment, armored thermocouples are installed in the upper, middle, and lower axial sections and the center and edge radially inside the pyrolysis reactor through thermocouple sheath interfaces on the outer shell of the reactor to detect and acquire temperature data at each detection point, and obtain the corresponding temperature gradient data based on the temperature data at each detection point. Simultaneously, a pressure tap is provided at the top of the pyrolysis reactor, and the pressure tap is connected to the pressure transmitter via a pressure tapping pipe to monitor and acquire corresponding pressure data. It is worth noting that the pressure tapping pipe in this embodiment is equipped with a condenser pipe or an isolation pipe to prevent high-temperature or corrosive media from directly damaging the pressure transmitter.

[0035] Furthermore, a high-temperature sampling probe is installed on the gas outlet pipe of the pyrolysis reactor to collect corresponding gas samples. It is worth noting that the high-temperature sampling probe in this embodiment is equipped with a sintered metal filter to remove dust and tar particles. Simultaneously, the sampling pipeline of the high-temperature sampling probe is equipped with heat tracing and insulation, with the corresponding temperature set at 150-200℃, to prevent condensation of high-boiling-point tar and water vapor in the sampling pipeline, which could cause blockage and compositional distortion.

[0036] Furthermore, the sampled gas is pretreated before being analyzed to determine its components. Specifically, the sampled gas is transferred to a primary condenser to cool it to its dew point temperature, condensing and removing moisture and heavy tar. Simultaneously, the cooled gas sample is transferred to a filter (e.g., a ceramic or polyester fiber filter) to remove fine particulate matter. The filtered gas sample is then further transferred to a membrane dryer or vortex cooler for drying. The dried gas sample is then transferred through a pressure regulating valve and a rotor flow meter to a gas analyzer (e.g., an online mass spectrometer) for gas analysis, thereby obtaining the corresponding gas components and their respective volume concentration percentages.

[0037] Step SB3: Trajectory Adjustment. This involves determining the real-time data deviation between the detected data and the predicted process value from the optimal candidate process trajectory identified in Step SB1.2, based on the detection data (temperature, pressure, and gas composition data) obtained in Step SB2. Simultaneously, the obtained real-time data deviation is compared with a preset trajectory threshold (which can be set according to actual data requirements, and is not specifically described in this embodiment). Based on the comparison result, the adjustment status of the corresponding candidate process trajectory is determined. Specifically: If the deviation of the acquired real-time data exceeds the preset trajectory threshold, steps SB1 and SB2 are repeated to adjust the corresponding candidate process trajectory until the deviation of the acquired real-time data does not exceed the preset trajectory threshold. Conversely, if the deviation of the acquired real-time data does not exceed the preset trajectory threshold, the corresponding candidate process trajectory is not adjusted and continues to operate in its current state.

[0038] In this embodiment, a corresponding process knowledge base is constructed based on the pyrolysis chemical reaction principle and historical data. Based on the obtained real-time data deviations, relevant variables are determined within the constructed process knowledge base. In other words, the corresponding objective function is reset based on the determined relevant variables. This involves returning to step SB1 and repeating steps SB1 and SB2 to adjust the determined optimal candidate process trajectory.

[0039] In this embodiment, the product collection module collects and analyzes the products after the pyrolysis reaction according to the candidate process trajectory in step SB3 to determine the corresponding product collection, specifically including the following steps: Step SC1: Gas-phase product processing. This involves cooling and separating the high-temperature gas from the pyrolysis reactor using a quench tower or condenser. Simultaneously, the cooled and separated gas undergoes deacidification treatment using desulfurization and denitrification towers to remove acidic gases. Then, tar droplets and solid particles are removed from the deacidified gas using a tar trap (e.g., an electrostatic precipitator or a high-efficiency fiber filter) and a dust filter. Finally, the obtained gas is temporarily stored in a gas holder (e.g., a wet or dry gas holder). It is worth noting that during gas transfer, flow meters are used for metering, and the corresponding calorific value is monitored in real time using a calorific value meter.

[0040] Step SC2: Liquid Phase Product Processing. The liquid product condensed from the quench tower or condenser in Step SC1 is transferred to a liquid collection tank and then separated into layers using a centrifuge. Specifically, the upper layer of the separated liquid product is light bio-oil, and the lower layer is process wastewater. Simultaneously, through a stratified drain valve, the light bio-oil and process wastewater are directed to separate liquid storage tanks.

[0041] Step SC3: Solid Product Processing. This involves discharging the solid product (biochar) from the bottom of the pyrolysis reactor using a rotary valve or screw conveyor and transferring it to a closed cooler (such as a screw cooler or jacketed water cooler) for cooling to room temperature under inert gas protection. The cooled biochar is then temporarily stored in a biochar silo.

[0042] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended embodiments and their equivalents.

Claims

1. A precise pyrolysis control and product analysis system for urban biomass solid waste, characterized in that, Including: Material characteristic prediction module: After pretreatment of urban biomass solid waste, the module uses a machine learning prediction model to predict pyrolysis behavior characteristics and obtains a corresponding pyrolysis behavior characteristic report. Process pyrolysis control module: Based on the pyrolysis behavior feature report and user product quality target, construct a feature target vector, and use the feature target vector as input to a deep learning prediction model to output the corresponding candidate process trajectory. At the same time, monitor and acquire product quality data during the pyrolysis reaction process, and adjust the candidate process trajectory in real time by comparing the product quality data with the user product quality target. Product collection module: collects and analyzes the products after pyrolysis reaction according to the candidate process trajectory.

2. The precise pyrolysis control and product analysis system for urban biomass solid waste according to claim 1, characterized in that, The corresponding pyrolysis behavior characteristic report is obtained, including: SA1: Physical pretreatment: crushing, drying and homogenizing urban biomass solid waste to obtain pretreated urban biomass solid waste; SA2: Feature Recognition: The pre-processed urban biomass solid waste is scanned using a near-infrared spectral probe to obtain the spectral signal of the biomass solid waste. The spectral signal of the biomass solid waste and the spectral fingerprint database are used as inputs to a principal component analysis model, and the corresponding feature probability distribution results are output. SA3: Predictive Processing: Based on the characteristic probability distribution results, at least one corresponding material quantitative correction model is determined, and the spectral signal of the biomass solid waste is used as the input of each material quantitative correction model. The corresponding characteristic values ​​are output and obtained. At the same time, the characteristic values ​​are used as the input of the machine learning prediction model and the corresponding prediction index is output. Based on the similarity probability corresponding to the characteristic probability distribution results, the prediction index corresponding to each material quantitative correction model is weighted and fused to obtain the corresponding pyrolysis behavior characteristic prediction report.

3. The precise pyrolysis control and product analysis system for urban biomass solid waste according to claim 2, characterized in that, The pretreated urban biomass solid waste obtained includes: SA1.1: Crushing process: The urban biomass solid waste to be processed is sequentially transported to the coarse crusher and fine crusher through the conveying mechanism to obtain the final waste fragments; SA1.2: Drying process: The final waste fragments are dried by means of a connected drum dryer and hot air furnace; SA1.3: Homogenization treatment: The dried final waste fragments are laid out in layers inside the homogenization chamber and subjected to shearing motion to obtain pre-treated urban biomass solid waste.

4. The precise pyrolysis control and product analysis system for urban biomass solid waste according to claim 2, characterized in that, Based on the known spectral data corresponding to multiple different biomass solid wastes, a spectral fingerprint database is constructed. Then, using the principal component analysis model, the spectral signals of the biomass solid wastes are compared with multiple reference spectral signals in the spectral fingerprint database to determine the corresponding similar reference spectral signals and similarity probabilities.

5. The precise pyrolysis control and product analysis system for urban biomass solid waste according to claim 1, characterized in that, Real-time adjustment of the candidate process trajectory includes: SB1: Trajectory Determination: The feature target vector composed of the pyrolysis behavior feature prediction report and the user product quality target is used as the input of the deep learning prediction model. The output obtains multiple different candidate process trajectories. At the same time, simulation and deduction are performed on all candidate process trajectories, and the corresponding optimal candidate process trajectory is determined based on the simulation and deduction results. SB2: Trajectory Monitoring: Through a distributed control system, the optimal candidate process trajectory is analyzed to obtain the corresponding execution instructions. Based on the execution instructions, different actuators in the pyrolysis reactor are driven. At the same time, during the driving of the actuators, detection data in the pyrolysis reactor, including temperature data, pressure data, and gas composition content data, are collected. SB3: Trajectory Adjustment: The detected data is compared with the process prediction value in the optimal candidate process trajectory to obtain the corresponding real-time data deviation. Simultaneously, the real-time data deviation is compared with a preset trajectory threshold to determine the adjustment state of the candidate process trajectory. Specifically: When the real-time data deviation is greater than the preset trajectory threshold, steps SB1 and SB2 are repeated to adjust the candidate process trajectory until the real-time data deviation is not greater than the preset trajectory threshold. Conversely, the candidate process trajectory will not be adjusted.

6. The precise pyrolysis control and product analysis system for urban biomass solid waste according to claim 5, characterized in that, The corresponding optimal candidate process trajectory was determined, including: SB1.1: Data Prediction: Based on the set user product quality target, determine the corresponding objective function, and combine the objective function with the final prediction index corresponding to the pyrolysis behavior feature prediction report to obtain the corresponding feature target vector. At the same time, use the feature target vector as the input of the deep learning prediction model and output to obtain multiple different candidate process trajectories. SB1.2: Simulation Determination: Through the set digital twin, each candidate process trajectory is simulated and deduced to obtain the corresponding simulation results. At the same time, the maximum simulation result is determined from the simulation results corresponding to each candidate process trajectory. The candidate process trajectory corresponding to the maximum simulation result is the optimal candidate process trajectory.

7. The precise pyrolysis control and product analysis system for urban biomass solid waste according to claim 5, characterized in that, The deep learning prediction model sets constraints during trajectory prediction, including safety constraints, process constraints, environmental constraints, and product quality constraints.

8. The precise pyrolysis control and product analysis system for urban biomass solid waste according to claim 5, characterized in that, Gas samples from the pyrolysis reactor are collected using a high-temperature sampling probe and a heat-traced sampling pipeline. The gas samples are then pretreated by passing them through a primary condenser, a filter, and a dryer. Simultaneously, the pretreated gas samples are analyzed using a gas analyzer to obtain the corresponding gas component content data.

9. The precise pyrolysis control and product analysis system for urban biomass solid waste according to claim 1, characterized in that, The products of the pyrolysis reaction following the candidate process trajectory are collected and analyzed, including: SC1: Gas phase product processing: After cooling and separating the high-temperature gas in the pyrolysis reactor, the gas is then subjected to deacidification and filtration processes to obtain pretreated gas, which is then temporarily stored in a gas storage tank. SC2: Liquid product processing: The condensed liquid product is separated into layers using a centrifuge, and the separated liquid products are stored in a storage tank; SC3: Solid product handling: The solid product is cooled to room temperature in a closed cooler using inert gas and then transferred to a biochar silo for temporary storage.

10. A method for precise pyrolysis control and product analysis of urban biomass solid waste, characterized in that, The system for precise pyrolysis control and product analysis of urban biomass solid waste, as described in any one of claims 1-9, was used.

Images