Hazardous waste matching intelligent optimization method and system

Abstract

This application discloses an intelligent optimization method and system for hazardous waste compatibility, relating to the field of hazardous waste treatment technology. First, a key characteristic information group is determined based on key data sets during the hazardous waste incineration process. When it is determined that any key characteristic information in the key characteristic information group does not meet preset specific conditions, a pre-trained in-situ artificial intelligence computing model is invoked to process the key characteristic information group. Based on the mapping relationship between input data and target performance, a new compatibility formula that meets preset pumpability is predicted. Finally, the new compatibility formula is executed. This application determines whether compatibility formula needs optimization by monitoring key characteristic information during hazardous waste compatibility incineration. When optimization is determined, the in-situ artificial intelligence computing model is invoked to generate a new compatibility formula. This allows for timely adjustment of the compatibility scheme during hazardous waste incineration without frequent manual intervention, significantly improving the incineration efficiency of hazardous waste.

Description

Technical Field

[0001] This application relates to the field of hazardous waste treatment technology, and in particular to a method and system for intelligent optimization of hazardous waste compatibility. Background Technology

[0002] Hazardous waste refers to solid waste (including liquid waste) that needs to be managed in accordance with special regulations because of its hazardous characteristics or potential harmful effects on the ecological environment and human health.

[0003] Incineration is a widely applicable hazardous waste disposal technology, with its core objective being the harmless and reduced-volume treatment of hazardous waste. However, in actual incineration operations, rotary kilns and other incineration systems often face numerous challenges, including: poor adaptability to complex and variable waste compositions, frequent unplanned start-ups due to severe slagging within the furnace, and the risk of exceeding pollutant emission standards. To address these challenges and ensure stable, efficient, and environmentally compliant incineration processes, hazardous waste disposal companies generally need to conduct compatibility analysis before materials are fed into the kiln. This analysis aims to comprehensively analyze the physical and chemical properties of each batch of materials to be incinerated (such as calorific value, volatile matter, sulfur and chlorine content, ash characteristics, and packaging form). Through reasonable material matching, homogenization of the materials entering the kiln in terms of form, calorific value, and chemical composition is achieved, thereby ensuring stable and controllable composition and uniform and balanced combustion, ultimately optimizing the final incineration effect.

[0004] Currently, the commonly used blending methods in hazardous waste disposal are primarily designed to meet the requirements of the final incineration process. This often leads to blending schemes that fail to adequately consider the actual physical state and rheological heat properties of the materials after crushing and mixing. Consequently, the resulting mixtures often exhibit poor homogeneity and flowability. During pumping, the materials tend to have excessively high viscosity, are prone to sedimentation or segregation, and frequently cause high pressure and blockages in the conveying pipelines. To address these issues, on-site personnel must frequently intervene manually. This emergency intervention approach is not only ineffective but also lacks sustainability and severely disrupts the continuous and stable incineration rhythm, ultimately failing to meet the incineration requirements of hazardous waste and resulting in reduced incineration efficiency. Summary of the Invention

[0005] In view of the above problems, this application provides a method and system for intelligent optimization of hazardous waste compatibility to achieve high-efficiency incineration of hazardous waste. The specific solution is as follows:

[0006] The first aspect of this application provides a method for intelligent optimization of hazardous waste compatibility, comprising:

[0007] Acquire key data sets during the hazardous waste incineration process; key data sets include at least emission and combustion data collected from the kiln operating system, pumping pressure and mixing pressure data collected from the automatic control system, and physical state data of the mixture collected from the visual recognition system.

[0008] If the calculation results of emission data and combustion data meet the modified mode conditions in the preset control mode discrimination conditions, the key characteristic information group is calculated based on the key data group; the key characteristics in the key characteristic information group include at least the cohesiveness, flowability and pumpability of the mixture.

[0009] When there are key characteristic information in the key characteristic information group that do not meet the preset characteristic conditions, the pre-trained in-situ artificial intelligence computing model is invoked to process the key characteristic information group and obtain a new compatibility list. The in-situ artificial intelligence computing model is configured to: input the basic dataset and key characteristic information group corresponding to the hazardous waste into the pre-trained compatibility list prediction model included in the in-situ artificial intelligence computing model, and the compatibility list prediction model predicts a new compatibility list that meets the preset pumpability based on the mapping relationship between the input data and the target performance.

[0010] Control the execution of the new pairing order.

[0011] In one possible implementation, when a key characteristic in the key characteristic information group does not meet the preset characteristic conditions, a pre-trained in-situ artificial intelligence computing model is invoked to process the key characteristic information group and obtain a new matching list, including:

[0012] When any one of the key characteristic information in the key characteristic information group—cohesiveness, flowability, and pumpability—does not meet the preset characteristic standard in the preset characteristic conditions, the basic dataset corresponding to the hazardous waste is obtained. The basic dataset includes at least incinerator kiln operation data, hazardous waste inventory data, hazardous waste characteristic information data, real-time data from the pretreatment intelligent control system, and artificial intelligence vision system data. The preset characteristic standards include cohesiveness standard, flowability standard, pumpability standard, combustion emission standard, and a preset reasonable range for calorific value.

[0013] The basic dataset and key feature information are input into the in-situ artificial intelligence computing model to obtain a new pairing list. The in-situ artificial intelligence computing model includes a pre-trained pairing list prediction model and a pre-trained pairing list optimization model. The pairing list prediction model is used to generate an initial pairing list based on the basic dataset and key feature information. The pairing list optimization model is used to adjust the initial pairing list according to a preset iterative adjustment mechanism until a new pairing list that meets the preset pumpability is obtained.

[0014] In one possible implementation, the matching single prediction model is either a machine learning model or a neural network model;

[0015] Matching single optimization model is one of the reinforcement learning model and genetic algorithm model.

[0016] In one possible implementation, the matching order optimization model is used to adjust the initial matching order according to a preset iterative adjustment mechanism until a new matching order that meets the preset pumpability is obtained, including:

[0017] When the iterative adjustment mechanism is a state-action reward mechanism, the process of determining the new pairing list is as follows:

[0018] Determine the state parameters of hazardous waste, including the mixture ratio and incineration process parameters;

[0019] Based on the state parameters, the corresponding reward information for hazardous waste is determined. The reward information includes pumpability score, incineration efficiency and chemical compatibility verification results.

[0020] The adjustment amount of the initial matching list is determined based on the reward information. The action parameters are then corrected using the adjustment amount to obtain a new matching list. The action parameters include the hazardous waste adjustment information in the initial matching list.

[0021] In one possible implementation, the matching order optimization model is used to adjust the initial matching order according to a preset iterative adjustment mechanism until a new matching order that meets the preset pumpability is obtained, including:

[0022] When the iterative adjustment mechanism is a population fitness evolution mechanism, the process of determining the new mating list is as follows:

[0023] Multiple matching lists are randomly generated based on the basic dataset and key feature information groups, and the first matching list is composed of multiple matching list sets;

[0024] The chemical compatibility of multiple matching lists is evaluated using a preset security check, and matching lists that do not meet the chemical compatibility standard are eliminated to obtain a second set of matching lists;

[0025] Each pair of pairs in the first pairing set is evaluated using preset key indicators to obtain a fitness score set. The third pairing set is composed of pairing pairs whose scores in the fitness score set reach a preset threshold.

[0026] The proportions of dangerous match combinations in each match set in the third match set are exchanged and recombined to obtain the fourth match set.

[0027] The fourth set of matching pairs is used as the first set of matching pairs. Iteratively, the chemical compatibility of multiple matching pairs is evaluated using a preset safety check. Matches that do not meet the chemical compatibility standard are eliminated, resulting in the second set of matching pairs. Each matching pair in the first set of matching pairs is evaluated using a preset key index to obtain a fitness score set. Matches whose fitness score reaches a preset threshold form the third set of matching pairs. The proportion of dangerous matching combinations in each matching pair in the third set of matching pairs is swapped and recombined to obtain the fourth set of matching pairs. This process continues until a preset iteration stop condition is met. Finally, the matching pair with the highest fitness score in the fourth set of matching pairs is used as the new matching pair.

[0028] In one possible implementation, the in-situ artificial intelligence computing model is a model under the federated learning framework issued by the integrated data center;

[0029] The in-situ AI computing model periodically sends its matching list optimization experience to the integrated data center so that other in-situ AI computing models in the integrated data center can learn from the matching list optimization experience.

[0030] In one possible implementation, after acquiring the key data set from the hazardous waste incineration process, it also includes:

[0031] If the calculation results of the emission data and combustion data collected in the kiln operating system meet the conditions of the feed rhythm adjustment mode in the preset control mode discrimination conditions, the feed recommendation for hazardous waste is determined based on the emission data and combustion data.

[0032] Generate incinerator feed control signals based on feed recommendations;

[0033] The incinerator feed control signal is sent to the incinerator feed system to control the feed rhythm.

[0034] The second aspect of this application provides a hazardous waste compatibility intelligent optimization system, including an inventory and compatibility system, a pretreatment intelligent control system, an SMP system, a visual recognition system, and a kiln operating system;

[0035] The inventory and matching system is used to store hazardous waste inventory data and generate or modify matching lists based on optimization suggestions;

[0036] The SMP system is used for the crushing, mixing and pumping of hazardous waste.

[0037] The visual recognition system is used to collect physical state data of the mixture, and the physical state data includes at least cohesiveness and flowability data;

[0038] The kiln operating system is used to collect emission data and combustion data during the incineration process;

[0039] The pretreatment intelligent control system is communicatively connected to the inventory and matching system, the SMP system, the visual recognition system, and the kiln operating system, respectively, and is used to implement the above-mentioned first aspect or any implementation of the first aspect of the intelligent optimization method for hazardous waste matching.

[0040] A third aspect of this application provides a smart optimization device for hazardous waste compatibility, comprising:

[0041] The acquisition unit is used to acquire key data sets during the hazardous waste incineration process; the key data sets include at least emission data and combustion data collected from the kiln operating system, pumping pressure data and mixing pressure data collected from the automatic control system, and physical state data of the mixture collected from the visual recognition system.

[0042] The calculation unit is used to calculate the key characteristic information group based on the key data group when the calculation results of emission data and combustion data meet the modification mode conditions in the preset control mode discrimination conditions; the key characteristics in the key characteristic information group include at least the cohesiveness, flowability and pumpability of the mixture.

[0043] The processing unit is used to call a pre-trained in-situ artificial intelligence computing model to process the key characteristic information group when the key characteristic information in the key characteristic information group does not meet the preset characteristic conditions, and obtain a new matching order. The in-situ artificial intelligence computing model is configured to: input the basic dataset and key characteristic information group corresponding to the hazardous waste into the pre-trained matching order prediction model included in the in-situ artificial intelligence computing model, and the matching order prediction model predicts a new matching order that meets the preset pumpability based on the mapping relationship between the input data and the target performance of the basic dataset and key characteristic information.

[0044] The execution unit is used to control the execution of new matching orders.

[0045] A fourth aspect of this application provides a hazardous waste compatibility intelligent optimization device, comprising at least one processor and a memory connected to the processor, wherein:

[0046] Memory is used to store computer programs;

[0047] The processor is used to execute computer programs to enable the intelligent optimization device for hazardous waste compatibility to implement the intelligent optimization method for hazardous waste compatibility as described in the first aspect or any implementation thereof.

[0048] The fifth aspect of this application provides a computer storage medium carrying one or more computer programs, which, when executed by an electronic device, enable the electronic device to implement the intelligent optimization method for hazardous waste compatibility as described in the first aspect or any implementation thereof.

[0049] The sixth aspect of this application provides a computer program product, including computer-readable instructions, which, when executed on an electronic device, cause the electronic device to implement the hazardous waste compatibility intelligent optimization method described in the first aspect or any implementation thereof.

[0050] By employing the above technical solution, the intelligent optimization method and system for hazardous waste compatibility provided in this application first determines a key characteristic information group based on key data sets during the hazardous waste incineration process. When it is determined that any key characteristic information in the key characteristic information group does not meet preset specific conditions, a pre-trained in-situ artificial intelligence computing model is invoked to process the key characteristic information group. Based on the mapping relationship between input data and target performance, a new compatibility formula that meets preset pumpability is predicted. Finally, the new compatibility formula is executed. This application determines whether compatibility formula needs optimization by monitoring key characteristic information during the hazardous waste compatibility incineration process. When it is determined that compatibility formula needs optimization, the in-situ artificial intelligence computing model is invoked to generate a new compatibility formula. This allows for timely adjustment of the compatibility scheme during the hazardous waste incineration process without frequent manual intervention, greatly improving the incineration efficiency of hazardous waste. Attached Figure Description

[0051] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0052] Figure 1 A flowchart illustrating the intelligent optimization method for hazardous waste compatibility provided in this application;

[0053] Figure 2 A flowchart illustrating the intelligent optimization method for hazardous waste compatibility provided in this application;

[0054] Figure 3 The control logic diagram of the preprocessing intelligent control system provided in this application;

[0055] Figure 4 A schematic diagram of a hazardous waste compatibility intelligent optimization device provided in this application;

[0056] Figure 5 A schematic diagram of the structure of the intelligent optimization equipment for hazardous waste compatibility provided in this application. Detailed Implementation

[0057] The embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the implementation section of this application is for explaining specific embodiments only and is not intended to limit the scope of this application.

[0058] The embodiments of this application will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that, with technological advancements and the emergence of new scenarios, the technical solutions provided in the embodiments of this application are equally applicable to similar technical problems.

[0059] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such terms are interchangeable where appropriate; this is merely a way of distinguishing objects with the same attributes in the embodiments of this application. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, so that a process, method, system, product, or apparatus that comprises a series of elements is not necessarily limited to those elements, but may include other elements not explicitly listed or inherent to those processes, methods, products, or apparatuses.

[0060] Hazardous waste mainly includes the following two categories: one is waste with clearly defined hazardous characteristics, including but not limited to toxicity, corrosivity, flammability, reactivity, and infectivity; the other is waste that needs to be managed as hazardous waste. Although the above-mentioned hazardous characteristics have not been clearly detected in this type of waste, it cannot be ruled out that it has potential hazards and may have harmful effects on the ecological environment or human health. It needs to be managed in accordance with the requirements for the management of hazardous waste.

[0061] SMP (Shredding, Mixing, Pumping System) is a comprehensive hazardous waste pretreatment system that includes three main stages: crushing, mixing, and pumping. This system can simultaneously handle various hazardous wastes, including solid, liquid, and semi-solid wastes. It is widely used in hazardous waste disposal centers, cement kiln co-processing, and other enterprises, and is mainly used to achieve the key front-end stage of waste homogenization and stable transportation to the incinerator.

[0062] However, the compatibility methods and systems currently widely used in the industry are primarily designed to meet the needs of the final incineration process, prioritizing the chemical compatibility, combustion stability, and restrictions on the introduction of harmful elements (such as chlorine, sulfur, and heavy metals). This approach largely neglects compatibility with the operational requirements of front-end pretreatment systems (especially SMP systems). This disconnect leads to a series of prominent problems and operational bottlenecks: compatibility schemes fail to fully consider the actual physical state and rheological properties of materials after crushing and mixing, resulting in poor homogeneity and flowability of the mixed materials; the materials exhibit excessively high viscosity and are prone to sedimentation or segregation during pumping, frequently causing high pressure and blockages in the conveying pipelines.

[0063] To address these issues, on-site operators are forced to make frequent manual interventions, such as temporarily adjusting material ratios, clearing pipes, or cleaning equipment. This emergency response model, reliant on manual intervention, is not only limited in effectiveness and lacks sustainability, but also severely disrupts the continuous and stable pretreatment and incineration process. The direct consequence is that the material condition after intervention fails to meet the requirements for stable incineration, leading to reduced incineration efficiency, increased energy consumption, greater difficulty in pollutant control, and even affecting kiln stability. Ultimately, the overall hazardous waste disposal efficiency drops significantly due to frequent interruptions and poor results in the pretreatment process, while simultaneously increasing the operational risks and workload for on-site operators.

[0064] In addition, during the treatment of hazardous waste, the integrated hazardous waste pretreatment system suffers from poor operation quality and control effect due to the lack of experience of the kiln operators who are unable to control the feeding rhythm or select control measures effectively.

[0065] To address the aforementioned issues, this application provides a method and system for intelligent optimization of hazardous waste compatibility.

[0066] Optional, see Figure 1 This application provides a flowchart of a method for intelligent optimization of hazardous waste compatibility.

[0067] like Figure 1 As shown, the intelligent optimization method for hazardous waste compatibility includes the following steps:

[0068] Step 101: Obtain key data sets during the hazardous waste incineration process; key data sets include at least emission and combustion data collected from the kiln operating system, pumping pressure and mixing pressure data collected from the automatic control system, and physical state data of the mixture collected from the visual recognition system.

[0069] It should be noted that the intelligent optimization method for hazardous waste compatibility provided in this application is applied to a pretreatment intelligent control system. This pretreatment intelligent control system is designed by this application and is used to optimize the compatibility list. It belongs to the integrated hazardous waste incineration system. In addition to the pretreatment intelligent control system, the integrated hazardous waste incineration system also includes an inventory and compatibility system, an incineration kiln, a kiln operating system, an automatic control system, and a visual recognition system.

[0070] The key data set consists of key datasets collected by subsystems of other hazardous waste incineration integrated systems, including but not limited to emission and combustion data collected by the kiln operating system, pumping pressure and mixing pressure data collected by the automatic control system, and physical state data of the mixture collected by the visual recognition system.

[0071] Specifically, the kiln operating system is mainly used to monitor the combustion process of the mixed materials in the incinerator. During combustion, it collects a large amount of combustion and emission data. This key data is used for real-time monitoring, process adjustment, ensuring compliance with emission standards, and operational safety. Combustion data mainly reflects the thermal working state and combustion efficiency within the kiln. This data is directly related to whether combustion is stable and thorough, and is the core of process control. It includes, but is not limited to, temperature data, pressure data, atmosphere and composition data, and operating parameters. Emission data mainly reflects environmental compliance and pollution control levels. This data is mainly obtained through real-time monitoring by a continuous emission monitoring system installed in front of the chimney. It includes, but is not limited to, the concentration of conventional pollutants, combustion by-products and characteristic pollutants, and physical parameters.

[0072] The automatic control system is the automatic control system of the SMP system, specifically a PLC (Programmable Logic Controller) automatic control system. This system controls the SMP system to complete the normal working cycle. During the control process, it collects pumping pressure data and mixing pressure data of the mixture. These two data are crucial process monitoring parameters that directly reflect the status of material handling and the health of the SMP system. The pumping pressure data represents the resistance of the mixture flowing in the pipeline, and the mixing pressure data represents the difficulty of agitating, cutting, and homogenizing the mixture in the pretreatment tank. It should also be noted that the execution subject of this application, the pretreatment intelligent control system, belongs to the SMP system mentioned above.

[0073] Visual recognition systems are primarily used to identify the physical state parameters of mixtures of hazardous waste. These physical state parameters include, but are not limited to, rheological properties, cohesiveness, flowability data, and process-related state parameters that determine the difficulty of transporting and mixing.

[0074] Optionally, the pretreatment intelligent control system acquires emission and combustion data from the kiln operating system, pumping pressure and mixing pressure data from the automatic control system, and physical state data of the mixture from the visual recognition system. These key datasets collected from the subsystems of the hazardous waste incineration integrated system form a key data set.

[0075] Step 102: If the calculation results of emission data and combustion data meet the modified mode conditions in the preset control mode discrimination conditions, calculate the key characteristic information group based on the key data group; the key characteristics in the key characteristic information group include at least the cohesiveness, flowability and pumpability of the mixture.

[0076] It should be noted that the pre-processing intelligent control system in the specific embodiments provided in this application needs to first determine whether the control mode to be performed next is the modification mode, the feeding rhythm adjustment mode, or other control modes.

[0077] It's easy to understand that the preset control mode discrimination condition is mainly used to determine whether the pretreatment intelligent control system modifies or adjusts the feeding rhythm of the incinerator kiln feeding system. Furthermore, this preset control mode discrimination condition is constructed based on the calculation results of emission and combustion data collected in the kiln operating system.

[0078] Specifically, the judgment principle of the preset control mode is mainly to determine whether the abnormality in the calculation result is due to the inappropriate dosage of each material in the hazardous waste or the inappropriateness of each material in the hazardous waste itself.

[0079] Anomalies caused by inappropriate dosage can manifest as instantaneous fluctuations in combustion data, such as drastic changes in kiln tail temperature, oxygen content, and carbon monoxide concentration within a short period, while the analysis data for hazardous waste remains stable. Another manifestation of inappropriate dosage is short-term exceedances in emission data, such as NO... X The temperature suddenly rose, but based on the combustion data, it was caused by a synchronous increase in the kiln head temperature. This indicates that the total calorific value of the input is too high, and the feeding rate of high-calorific-value waste needs to be reduced.

[0080] The abnormality is caused by unsuitable materials, which manifests as data trends and systematic deviations. For example, after repeated adjustments to the feeding rhythm, certain key parameters continue to slowly develop in one direction.

[0081] The key characteristics in the critical characteristics information group include cohesiveness, flowability, and pumpability. Pumpability refers to the overall ability to be continuously, stably, and smoothly transported in the pumping pipeline of the pumping node in the SMP system. Flowability refers to the ability of a material to flow or deform under its own weight or a small external force, and this characteristic is generally monitored during the feeding process. Cohesiveness refers to the mutual attraction between particles within the material. Moderate cohesiveness helps the material form a homogeneous whole after mixing, effectively preventing the separation of solid particles from the liquid.

[0082] Optionally, the cohesiveness, flowability, and pumpability of the mixture can be obtained by weighting the key characteristics in the key characteristic information group and calculating the key data sets in the key data group.

[0083] Each key characteristic in the key characteristic information group can also be determined separately by the key data group. For example, pumping pressure data is used to determine pumpability, mixing pressure data is used to determine flowability, and physical state data of the mixture obtained from the visual recognition system is used to determine cohesiveness.

[0084] The key characteristics in the aforementioned key characteristic information group also include consistency, moisture content, and bleeding rate. Among them, consistency is a macroscopic empirical indicator describing the softness, hardness, and viscosity of semi-solid materials; moisture content is the mass percentage of water in the material, which is the most basic and sensitive key characteristic; bleeding rate is the tendency and rate at which water (or oil) separates and precipitates from the mixture under static conditions, which is a key indicator for evaluating the stability of the mixture.

[0085] Step 103: When there is key characteristic information in the key characteristic information group that does not meet the preset characteristic conditions, the pre-trained in-situ artificial intelligence computing model is invoked to process the key characteristic information group to obtain a new matching list. The in-situ artificial intelligence computing model is configured to: input the basic dataset and key characteristic information group corresponding to hazardous waste into the pre-trained matching list prediction model included in the in-situ artificial intelligence computing model, and the matching list prediction model predicts the new matching list corresponding to the basic dataset and key characteristic information that meets the preset pumpability based on the pre-set mapping relationship between the input data and the target performance.

[0086] It should be noted that the specific manifestation of a key characteristic information in the key characteristic information group not meeting the preset characteristic conditions is that at least one of the key characteristic information in the key characteristic information group, namely cohesiveness, flowability and pumpability, does not meet the preset characteristic standard in the preset characteristic conditions.

[0087] Correspondingly, the preset characteristic standards in the preset characteristic conditions include cohesiveness standards, flowability standards, pumpability standards, combustion emission standards, and preset reasonable ranges for calorific value.

[0088] The in-situ artificial intelligence computing model is a hybrid architecture built using artificial intelligence algorithms, designed in this application embodiment. It includes a pre-trained match prediction model and a pre-trained match optimization model. The match prediction model can be either a machine learning model or an artificial neural network model, while the match optimization model can be either a reinforcement learning model or a genetic algorithm model.

[0089] This application provides two methods for determining new match lists in specific embodiments. One method is to directly predict the match list using a pre-trained match list prediction model in the in-situ artificial intelligence computing model. The other method is to first predict the initial match list using a pre-trained match list prediction model in the in-situ artificial intelligence computing model, and then adjust the initial match list using a pre-trained match list optimization model in the in-situ artificial intelligence computing model to obtain the new match list.

[0090] Optionally, a basic dataset corresponding to hazardous waste can be obtained, and the basic dataset and key characteristic information group can be input into a pre-trained matching list prediction model in an in-situ artificial intelligence computing model to predict and obtain a new matching list.

[0091] Specifically, the basic dataset includes at least incinerator operation data, hazardous waste inventory data, hazardous waste characteristic information data, real-time data from the pretreatment intelligent control system, and data from the artificial intelligence vision system. This basic data, along with key characteristic information, is input into the compatibility prediction model. The model then calculates a new compatibility formula that meets the preset pumpability based on the pre-established mapping relationship between the input data and the target performance. The compatibility prediction model can be either a machine learning model or a neural network model. For example, the machine learning model can be a supervised machine learning model or an XGBoost model, and the neural network model can be an artificial neural network model.

[0092] Optionally, a basic dataset corresponding to hazardous waste is obtained, and the basic dataset and key characteristic information are input into a pre-trained compatibility prediction model in the in-situ artificial intelligence computing model for prediction to obtain an initial compatibility list; then, the initial compatibility list is input into a pre-trained compatibility optimization model in the in-situ artificial intelligence computing model to adjust the initial compatibility list to obtain a new compatibility list.

[0093] Specifically, the basic dataset includes at least incinerator operation data, hazardous waste inventory data, hazardous waste characteristic information data, real-time data from the pretreatment intelligent control system, and data from the artificial intelligence vision system. This basic data, along with key characteristic information, is input into the batching prediction model. The model uses a pre-established mapping relationship between the input data and target performance to calculate an initial batching that meets the preset pumpability. Then, the initial batching is input into the batching optimization model, where a pre-set iterative adjustment mechanism adjusts the initial batching until a new batching that meets the preset pumpability is obtained.

[0094] The matching single optimization model can be either a reinforcement learning model or a genetic algorithm model. When the matching single optimization model is a reinforcement learning model, the iterative adjustment mechanism is a state-action reward mechanism; when the matching single optimization model is a genetic algorithm model, the iterative adjustment mechanism is a population fitness evolution mechanism.

[0095] Specifically, the implementation process of the state-action reward mechanism is as follows: determine the state parameters of the hazardous waste, including but not limited to the mixture ratio and incineration process parameters. Then, determine the reward information corresponding to the hazardous waste based on the state parameters. The reward information is the pumpability score, incineration efficiency, and chemical compatibility verification results determined based on the state parameters. Finally, determine the adjustment amount of the initial formulation based on the reward information, and use the adjustment amount to correct the action parameters. The action parameters include the hazardous waste adjustment information in the initial formulation, and finally obtain the new formulation.

[0096] Specifically, the implementation process of the population fitness evolution mechanism is as follows:

[0097] Multiple matching lists are randomly generated based on the basic dataset and key feature information groups, and the first matching list is composed of multiple matching list sets;

[0098] The chemical compatibility of multiple matching lists is evaluated using a preset security check, and matching lists that do not meet the chemical compatibility standard are eliminated to obtain a second set of matching lists;

[0099] Each pair of pairs in the first pairing set is evaluated using preset key indicators to obtain a fitness score set. The third pairing set is composed of pairing pairs whose scores in the fitness score set reach a preset threshold.

[0100] The proportions of dangerous match combinations in each match set in the third match set are exchanged and recombined to obtain the fourth match set.

[0101] The fourth set of matching singletons is used as the first set of matching singletons. The above process is executed iteratively until the preset iteration stop condition is met. Then, the matching singleton with the highest fitness score in the fourth set of matching singletons is used as the new matching singleton.

[0102] For example, this in-situ artificial intelligence calculation model can also be called an artificial intelligence algorithm calculator, which can automatically generate and adjust the compatibility list of hazardous waste, focusing on the physical properties of the materials, especially pumpability, during the adjustment process. It can also achieve system operation with other existing systems to ensure chemical compatibility.

[0103] This artificial intelligence algorithm calculator is a hybrid framework that integrates predictive modeling and optimization techniques. The predictive modeling technique is implemented in the prediction module of the hybrid framework, and the optimization technique is implemented in the optimization module of the hybrid framework.

[0104] The prediction module uses supervised time machine learning models, such as XGBoost or artificial neural networks, to predict the rheological properties, plastic viscosity, yield stress, and overall pumpability of the mixture based on the characteristics and proportions of hazardous waste and real-time data from each monitoring subsystem throughout the treatment process. Based on the mapping relationship between the input data and the target performance, the model can accurately determine a new mix that meets the preset pumpability.

[0105] The training process of the supervised machine learning model in the prediction module described above can be as follows:

[0106] Historical data is used as the basic training data to establish a mapping relationship between input variables and target performance. Input variables can be basic data and key characteristic information such as the density, chemical composition, and viscosity curve of hazardous waste. Target performance refers to the target performance indicators that need to be achieved by the pre-set intelligent control system for the incineration of hazardous waste. For example, the level of pumpability that hazardous waste needs to achieve during pumping. This model can accurately judge and determine the pumpable slurry that meets the requirements, as well as the corresponding compatibility list of the slurry.

[0107] In practical applications, the implementation logic of supervised machine learning models is as follows: the state classification results of the visual system are encoded as category features and incorporated into the model, and an ensemble method is used to deal with the uncertainty brought about by waste fluctuations.

[0108] The optimization module can specifically employ reinforcement learning models, such as Q-learning or Proximal Policy Optimization. The reinforcement learning model optimizes the mix design based on a state-action reward mechanism to obtain a new mix design. Under this mechanism, the state is the current hazardous waste mixture ratio and process parameters, and the action is to add hazardous waste information according to a specified type or according to the instructed quantity. The reward is jointly determined by the predicted pumpability score, incineration efficiency, and chemical compatibility verification results.

[0109] The optimization module can also employ a genetic algorithm model, which determines new pairings based on a population fitness evolution mechanism. In this mechanism, the population is a potential set of pairings, the evolutionary operations are selection, crossover, and mutation, and the fitness function is based on chemical compatibility verification with pumpability as the core criterion.

[0110] In practical applications, the genetic algorithm model implementation logic enables the optimizer to be directly driven by the initial matching list output from the prediction module, achieving adaptive and constrained efficient search in a huge space of dangerous matching ratios.

[0111] In addition, the various models involved in the in-situ artificial intelligence computing model are all models under the federated learning framework issued by the integrated data center. The in-situ artificial intelligence computing model regularly sends the matching list optimization experience to the integrated data center so that other in-situ artificial intelligence computing models under the integrated data center can learn the matching list optimization experience and achieve experience sharing.

[0112] Specifically, the in-situ AI computing model is connected to the centralized data and analysis center at the SMP vendor's headquarters via the Internet. This center aggregates operational experience from multiple SMP sites, enabling collaborative optimization of the algorithm model.

[0113] The centralized center employs federated learning, enabling geographically distributed sites to collaboratively train models without exchanging sensitive raw data. This protects privacy and complies with hazardous waste management regulations. In the federated learning logic, each site's local model (an in-situ AI computation model, such as a hybrid predictor-optimizer) is trained using its own proprietary data, calculating parameter updates (such as gradients or model weights). These updates are securely aggregated at the central site using secure multi-party computation or differential privacy techniques to form a global model, which is then distributed back to the sites. By leveraging diverse scenarios such as varying waste types across different locations, this mechanism accelerates the process towards autonomy, improving prediction accuracy and optimization efficiency. For example, if a site encounters a new mixture and improves pumping performance as a result, the global model can absorb this knowledge without revealing any specific information about that site. Conversely, if a site experiences a critical failure such as pump pipe blockage, this experience is also incorporated into the global model and distributed to other sites as a preventative measure, allowing all sites in the network to share each other's experiences. During the sharing process, data privacy and security measures such as encryption and anonymization ensure the protection of sensitive information.

[0114] Step 104: Control the execution of the new matching order.

[0115] Finally, the new matching list is sent to the inventory and matching system. After manual confirmation, materials can be loaded according to the new matching list.

[0116] Optional, see Figure 2 This application provides a flowchart illustrating an intelligent optimization method for hazardous waste compatibility.

[0117] like Figure 2 As shown, before controlling the execution of the new matching order, step 201 may be included: sending the new matching order and specific optimization suggestions to the inventory and matching system for relevant verification and manual confirmation, and using the manually confirmed new matching order as the final new matching order.

[0118] Furthermore, step 102 above mentioned determining the next control mode of the pretreatment intelligent control system based on the calculation results of emission and combustion data. The following describes the subsequent processing steps when the calculation results of emission and combustion data collected in the kiln operating system meet the feed rhythm adjustment mode conditions in the preset control mode discrimination criteria.

[0119] First, feed recommendations for hazardous waste are determined based on emission and combustion data. Then, a feed control signal for the incinerator is generated based on the feed recommendations and sent to the incinerator feed system to control the feed rhythm.

[0120] The above process involves adjusting the feeding rhythm of the incinerator to meet the operational requirements of the incinerator.

[0121] In summary, the pretreatment intelligent control system provided in this application embodiment can realize the judgment and selection of the above control mode. Specifically, it determines whether to adjust the feeding rhythm based on the calculation results of emission data and combustion data, thereby automatically realizing the adjustment and control of the feeding rhythm, effectively avoiding the problem of insufficient operator experience, and improving the quality level of kiln operation.

[0122] In summary, this application provides an intelligent optimization method for hazardous waste compatibility. First, it determines a set of key characteristic information based on key data sets during the hazardous waste incineration process. When it is determined that any key characteristic information in this set does not meet preset specific conditions, a pre-trained in-situ artificial intelligence computing model is invoked to process the key characteristic information set. Based on the mapping relationship between input data and target performance, a new compatibility plan that meets preset pumpability is predicted. Finally, the new compatibility plan is executed. This application determines whether compatibility plan optimization is needed by monitoring key characteristic information during hazardous waste compatibility incineration. When optimization is determined, the in-situ artificial intelligence computing model is invoked to generate a new compatibility plan. This allows for timely adjustment of the compatibility scheme during hazardous waste incineration without frequent manual intervention, significantly improving the incineration efficiency of hazardous waste.

[0123] For example, see Figure 3 The control logic diagram of the preprocessing intelligent control system provided in this application.

[0124] like Figure 3 As shown, the pretreatment intelligent control system obtains emission and combustion data from the kiln operating system, pumping pressure and mixing pressure data from the PLC automatic control system of the SMP system, and physical state data of the mixture from the vision recognition system.

[0125] Then, the pumpability of the mixture during the pumping process in the SMP system is evaluated using pumping pressure data. The flowability of the mixture is evaluated using both mixing pressure data and physical state data. The cohesiveness of the mixture is evaluated using the physical state data. When the current pumpability, flowability, and cohesiveness of the mixture do not meet the preset characteristic conditions, the modification requirement is calculated based on the above characteristic data and the real-time data of the mixture. This modification requirement triggers the execution of the modification plan. The modification plan reads material data from the inventory and compatibility system and selects materials that match the modification plan in conjunction with the pre-processing intelligent control system and the material selection characteristic database. At the same time, it predicts whether the modification plan meets the modification requirement. Specifically, the modification requirement can be whether the modification plan meets the preset characteristic conditions. If it does, the modification plan is qualified, and a feeding suggestion is generated. This suggestion is sent to the inventory and compatibility system. After compatibility calculation, specifically including the calculation of compatibility, thermal stability, and harmful elements, an optimization suggestion is determined. After manual verification and confirmation, the initial compatibility sheet is modified to determine the optimized compatibility sheet, and the material is fed according to the optimized compatibility sheet.

[0126] In the above process, the pretreatment intelligent control system calculates whether a modification operation needs to be triggered or whether the feeding rhythm needs to be changed based on the emission data and combustion data of the kiln operating system.

[0127] Among these measures, determining the chemical compatibility, calorific value stability, and harmful element content of the mixture in the modification scheme is to meet the requirements of incineration treatment; ensuring pumpability is to ensure that the pumping distance and pumping resistance are within a reasonable range, thereby effectively avoiding pumping pipeline blockage; controlling the feeding rhythm is to meet the kiln operation requirements; ensuring the cohesiveness of the mixture in the modification scheme is to ensure that the material does not segregate and that the mixing is homogenized; and ensuring the flowability of the mixture is to ensure continuous production, controllable equipment operation, and stable system operation.

[0128] This application embodiment also provides a hazardous waste compatibility intelligent optimization system, which includes an inventory and compatibility system, a pretreatment intelligent control system, an SMP system, a visual recognition system, and a kiln operating system;

[0129] The inventory and matching system is used to store hazardous waste inventory data and generate or modify matching lists based on optimization suggestions;

[0130] SMP systems are used for the crushing, mixing, and pumping of hazardous waste.

[0131] The visual recognition system is used to collect physical state data of the mixture, which includes at least cohesiveness and flowability data.

[0132] The kiln operating system is used to collect emission and combustion data during the incineration process;

[0133] The pretreatment intelligent control system is connected to the inventory and compatibility system, the SMP system, the vision recognition system, and the kiln operating system to implement any of the hazardous waste compatibility intelligent optimization methods provided in the embodiments of this application.

[0134] This application also provides a hazardous waste compatibility intelligent optimization device.

[0135] Please see Figure 4 , Figure 4 This is a schematic diagram of a hazardous waste compatibility intelligent optimization device provided in this application. Figure 4 As shown, the device includes:

[0136] The system comprises an acquisition unit 10, a calculation unit 20, a processing unit 30, and an execution unit 40; wherein:

[0137] The acquisition unit 10 is used to acquire key data sets during the hazardous waste incineration process; the key data sets include at least emission data and combustion data collected in the kiln operating system, pumping pressure data and mixing pressure data collected in the automatic control system, and physical state data of the mixture collected in the visual recognition system.

[0138] The calculation unit 20 is used to calculate a key characteristic information group based on the key data group when the calculation results of emission data and combustion data meet the modification mode conditions in the preset control mode discrimination conditions; the key characteristics in the key characteristic information group include at least the cohesiveness, flowability and pumpability of the mixture.

[0139] Processing unit 30 is used to call a pre-trained in-situ artificial intelligence computing model to process the key characteristic information group when the key characteristic information in the key characteristic information group does not meet the preset characteristic conditions, and obtain a new matching order; the in-situ artificial intelligence computing model is configured to: input the basic dataset and key characteristic information group corresponding to hazardous waste into the pre-trained matching order prediction model included in the in-situ artificial intelligence computing model, and the matching order prediction model predicts a new matching order that meets the preset pumpability corresponding to the basic dataset and key characteristic information based on the pre-set mapping relationship between the input data and the target performance;

[0140] Execution unit 40 is used to control the execution of new matching orders.

[0141] In one embodiment, the processing unit 30 is specifically used for:

[0142] When any one of the key characteristic information in the key characteristic information group—cohesiveness, flowability, and pumpability—does not meet the preset characteristic standard in the preset characteristic conditions, the basic dataset corresponding to the hazardous waste is obtained. The basic dataset includes at least incinerator kiln operation data, hazardous waste inventory data, hazardous waste characteristic information data, real-time data from the pretreatment intelligent control system, and artificial intelligence vision system data. The preset characteristic standards include cohesiveness standard, flowability standard, pumpability standard, combustion emission standard, and a preset reasonable range for calorific value.

[0143] The basic dataset and key feature information are input into the in-situ artificial intelligence computing model to obtain a new pairing list. The in-situ artificial intelligence computing model includes a pre-trained pairing list prediction model and a pre-trained pairing list optimization model. The pairing list prediction model is used to generate an initial pairing list based on the basic dataset and key feature information. The pairing list optimization model is used to adjust the initial pairing list according to a preset iterative adjustment mechanism until a new pairing list that meets the preset pumpability is obtained.

[0144] In one embodiment, the matching single prediction model in the processing unit 30 is one of a machine learning model and a neural network model;

[0145] Matching single prediction models are a type of machine learning model and neural network model.

[0146] In one embodiment, the processing unit 30 is specifically used for:

[0147] When the iterative adjustment mechanism is a state-action reward mechanism, the process of determining the new pairing list is as follows:

[0148] Determine the state parameters of hazardous waste, including the mixture ratio and incineration process parameters;

[0149] Based on the state parameters, the corresponding reward information for hazardous waste is determined. The reward information includes pumpability score, incineration efficiency and chemical compatibility verification results.

[0150] The adjustment amount of the initial matching list is determined based on the reward information. The action parameters are then corrected using the adjustment amount to obtain a new matching list. The action parameters include the hazardous waste adjustment information in the initial matching list.

[0151] In one embodiment, the processing unit 30 is specifically used for:

[0152] When the iterative adjustment mechanism is a population fitness evolution mechanism, the process of determining the new mating list is as follows:

[0153] Multiple matching lists are randomly generated based on the basic dataset and key feature information groups, and the first matching list is composed of multiple matching list sets;

[0154] The chemical compatibility of multiple matching lists is evaluated using a preset security check, and matching lists that do not meet the chemical compatibility standard are eliminated to obtain a second set of matching lists;

[0155] Each pair of pairs in the first pairing set is evaluated using preset key indicators to obtain a fitness score set. The third pairing set is composed of pairing pairs whose scores in the fitness score set reach a preset threshold.

[0156] The proportions of dangerous match combinations in each match set in the third match set are exchanged and recombined to obtain the fourth match set.

[0157] The fourth set of matching pairs is used as the first set of matching pairs. Iteratively, the chemical compatibility of multiple matching pairs is evaluated using a preset safety check. Matches that do not meet the chemical compatibility standard are eliminated, resulting in the second set of matching pairs. Each matching pair in the first set of matching pairs is evaluated using a preset key index to obtain a fitness score set. Matches whose fitness score reaches a preset threshold form the third set of matching pairs. The proportion of dangerous matching combinations in each matching pair in the third set of matching pairs is swapped and recombined to obtain the fourth set of matching pairs. This process continues until a preset iteration stop condition is met. Finally, the matching pair with the highest fitness score in the fourth set of matching pairs is used as the new matching pair.

[0158] In one embodiment, the in-situ artificial intelligence computing model in the processing unit 30 is a model under the federated learning framework issued by the integrated data center;

[0159] The in-situ AI computing model periodically sends its matching list optimization experience to the integrated data center so that other in-situ AI computing models in the integrated data center can learn from the matching list optimization experience.

[0160] In one embodiment, the hazardous waste compatibility intelligent optimization device further includes a feeding rhythm adjustment unit;

[0161] This feed rhythm adjustment unit is specifically used for:

[0162] If the calculation results of the emission data and combustion data collected in the kiln operating system meet the conditions of the feed rhythm adjustment mode in the preset control mode discrimination conditions, the feed recommendation for hazardous waste is determined based on the emission data and combustion data.

[0163] Generate incinerator feed control signals based on feed recommendations;

[0164] The feed control signal is sent to the incinerator feed system to control the feed rhythm.

[0165] This application also provides an intelligent optimization device for hazardous waste compatibility in its embodiments. (Reference) Figure 5The diagram illustrates a structural schematic suitable for implementing the intelligent optimization device for hazardous waste compatibility provided in this application. The intelligent optimization device for hazardous waste compatibility in the embodiments of this application may include, but is not limited to, fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 5 The hazardous waste compatibility intelligent optimization device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0166] like Figure 5 As shown, the intelligent optimization device for hazardous waste compatibility may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. When the intelligent optimization device for hazardous waste compatibility is powered on, the RAM 603 also stores various programs and data required for the operation of the device. The processing unit 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0167] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 608 including, for example, memory cards, hard drives, etc.; and communication devices 609. Communication device 609 allows the hazardous waste compatibility intelligent optimization device to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 5 A smart optimization device for hazardous waste compatibility is shown, featuring various apparatuses; however, it should be understood that implementation or possession of all shown apparatuses is not required. Additional or fewer apparatuses may be implemented alternatively.

[0168] This application also provides a computer storage medium that carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any of the hazardous waste compatibility intelligent optimization methods provided in this application.

[0169] This application also provides a computer program product including computer-readable instructions, which, when executed on an electronic device, cause the electronic device to implement any of the hazardous waste compatibility intelligent optimization methods provided in this application.

[0170] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.

[0171] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, training equipment, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0172] In the above embodiments, the implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, in the form of a computer program product.

[0173] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, training device, or data center to another website, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

Claims

1. A hazardous waste matching intelligent optimization method, characterized in that, Applications in pretreatment intelligent control systems include: Acquire key data sets during the hazardous waste incineration process; the key data sets include at least emission data and combustion data collected from the kiln operating system, pumping pressure data and mixing pressure data collected from the automatic control system, and physical state data of the mixture collected from the visual recognition system. If the calculation results of the emission data and combustion data meet the modification mode conditions in the preset control mode discrimination conditions, a key characteristic information group is calculated based on the key data group; the key characteristics in the key characteristic information group include at least the cohesiveness, flowability and pumpability of the mixture. When any key characteristic information in the key characteristic information group does not meet the preset characteristic conditions, a pre-trained in-situ artificial intelligence computing model is invoked to process the key characteristic information group to obtain a new compatibility list. The in-situ artificial intelligence computing model is configured to input the basic dataset corresponding to the hazardous waste and the key characteristic information group into a pre-trained compatibility list prediction model included in the in-situ artificial intelligence computing model. The compatibility list prediction model predicts the new compatibility list corresponding to the basic dataset and the key characteristic information that meets the preset pumpability based on the pre-set mapping relationship between the input data and the target performance. Control the execution of the new matching order.

2. The intelligent optimization method for hazardous waste compatibility according to claim 1, characterized in that, When a key characteristic in the key characteristic information group does not meet the preset characteristic conditions, a pre-trained in-situ artificial intelligence computing model is invoked to process the key characteristic information group to obtain a new matching list, including: When any one of the key characteristic information in the key characteristic information group—cohesiveness, flowability, and pumpability—does not meet the preset characteristic standard in the preset characteristic conditions, a basic dataset corresponding to the hazardous waste is obtained; the basic dataset includes at least incinerator kiln operation data, hazardous waste inventory data, hazardous waste characteristic information data, real-time data of the pretreatment intelligent control system, and artificial intelligence vision system data; the preset characteristic standards include cohesiveness standards, flowability standards, pumpability standards, combustion emission standards, and a preset reasonable range of calorific value; The basic dataset and the key characteristic information group are input into the in-situ artificial intelligence computing model to obtain the new matching list; the in-situ artificial intelligence computing model includes a pre-trained matching list prediction model and a pre-trained matching list optimization model; the matching list prediction model is used to generate an initial matching list based on the basic dataset and the key characteristic information group; the matching list optimization model is used to adjust the initial matching list according to a preset iterative adjustment mechanism until a new matching list that meets the preset pumpability is obtained.

3. The intelligent optimization method for hazardous waste compatibility according to claim 2, characterized in that, The matching prediction model is one of a machine learning model and a neural network model; The matching single optimization model is one of the reinforcement learning model and the genetic algorithm model.

4. The intelligent optimization method for hazardous waste compatibility according to claim 2, characterized in that, The matching order optimization model is used to adjust the initial matching order according to a preset iterative adjustment mechanism until a new matching order that meets the preset pumpability is obtained, including: When the iterative adjustment mechanism is a state-action reward mechanism, the process for determining the new pairing list is as follows: Determine the state parameters of hazardous waste, including the mixture ratio and incineration process parameters during the incineration process; The reward information corresponding to the hazardous waste is determined based on the state parameters. The reward information includes pumpability score, incineration efficiency and chemical compatibility verification results. The adjustment amount of the initial matching list is determined based on the reward information, and the action parameters are corrected using the adjustment amount to obtain the new matching list; the action parameters include the hazardous waste adjustment information in the initial matching list.

5. The intelligent optimization method for hazardous waste compatibility according to claim 2, characterized in that, The matching order optimization model is used to adjust the initial matching order according to a preset iterative adjustment mechanism until a new matching order that meets the preset pumpability is obtained, including: When the iterative adjustment mechanism is a population fitness evolution mechanism, the process for determining the new mating list is as follows: Multiple matching lists are randomly generated based on the basic dataset and key characteristic information groups, and the multiple matching lists form the first matching list set; The chemical compatibility of the multiple matching lists is evaluated using a preset security check, and matching lists that do not meet the chemical compatibility standard are eliminated to obtain a second set of matching lists; Each matching unit in the first matching unit set is evaluated using preset key indicators to obtain a fitness score set. The matching units in the fitness score set that reach a preset threshold form a third matching unit set. The proportions of dangerous matching combinations in each matching set in the third matching set are exchanged and recombined to obtain the fourth matching set. The fourth set of matching pairs is used as the first set of matching pairs. The chemical compatibility of the multiple matching pairs is evaluated using a preset security check. Matching pairs that do not meet the chemical compatibility standard are eliminated to obtain the second set of matching pairs. Each matching pair in the first set of matching pairs is evaluated using a preset key index check to obtain a fitness score set. Matching pairs whose fitness score reaches a preset threshold form the third set of matching pairs. The proportion of dangerous matching combinations in each matching pair in the third set of matching pairs is exchanged and recombined to obtain the fourth set of matching pairs. This process continues until a preset iteration stop condition is met. The matching pair with the highest fitness score in the fourth set of matching pairs is then used as the new matching pair.

6. The intelligent optimization method for hazardous waste compatibility according to claim 1, characterized in that, The in-situ artificial intelligence computing model is a model under the federated learning framework issued by the integrated data center; The in-situ artificial intelligence computing model periodically sends the matching list optimization experience to the integrated data center, so that other in-situ artificial intelligence computing models in the integrated data center can learn the matching list optimization experience.

7. The intelligent optimization method for hazardous waste compatibility according to claim 1, characterized in that, After obtaining the key data set in the hazardous waste incineration process, it also includes: If the calculation results of the emission data and combustion data collected in the kiln operating system meet the conditions of the feed rhythm adjustment mode in the preset control mode discrimination conditions, the feed recommendation for hazardous waste is determined based on the emission data and combustion data. Generate incinerator feed control signals based on the feed recommendations; The incinerator feed control signal is sent to the incinerator feed system to control the feed rhythm.

8. A hazardous waste compatibility intelligent optimization system, characterized in that, This includes an inventory and matching system, a pre-treatment intelligent control system, an SMP system, a vision recognition system, and a kiln operating system; The inventory and matching system is used to store hazardous waste inventory data and generate or modify matching lists based on optimization suggestions; The SMP system is used for the crushing, mixing and pumping of hazardous waste. The visual recognition system is used to collect physical state data of the mixture, and the physical state data includes at least cohesiveness and flowability data; The kiln operating system is used to collect emission data and combustion data during the incineration process; The pretreatment intelligent control system is communicatively connected to the inventory and matching system, the SMP system, the visual recognition system, and the kiln operating system, respectively, and is used to implement the intelligent optimization method for hazardous waste matching as described in any one of claims 1 to 7.

9. A hazardous waste compatibility intelligent optimization device, characterized in that, include: The acquisition unit is used to acquire key data sets during the hazardous waste incineration process. The key data set includes at least emission and combustion data collected from the kiln operating system, pumping pressure and mixing pressure data collected from the automatic control system, and physical state data of the mixture collected from the visual recognition system. The calculation unit is used to calculate a key characteristic information group based on the key data group when the calculation results of the emission data and combustion data meet the modification mode conditions in the preset control mode discrimination conditions; the key characteristics in the key characteristic information group include at least the cohesiveness, flowability and pumpability of the mixture. The processing unit is configured to, when there is key characteristic information in the key characteristic information group that does not meet preset characteristic conditions, invoke a pre-trained in-situ artificial intelligence computing model to process the key characteristic information group and obtain a new matching order; the in-situ artificial intelligence computing model is configured to: input the basic dataset corresponding to hazardous waste and the key characteristic information group into the pre-trained matching order prediction model included in the in-situ artificial intelligence computing model, and the matching order prediction model predicts the new matching order corresponding to the basic dataset and the key characteristic information that meets the preset pumpability based on the pre-set mapping relationship between the input data and the target performance; An execution unit is used to control the execution of the new matching order.

10. A hazardous waste compatibility intelligent optimization device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the hazardous waste compatibility intelligent optimization device can implement the hazardous waste compatibility intelligent optimization method as described in any one of claims 1 to 7.

Images