Multi-component solid waste compatibility optimization method based on coupling matrix operation and machine learning

Abstract

The present invention discloses a multi-component solid waste compatibility optimization method based on the coupling of matrix operations and machine learning, comprising: obtaining compatibility materials, wherein the compatibility materials include solid waste and kiln working fluid; utilizing a machine learning model nested optimization algorithm to obtain a set of characteristic parameter requirements for the compatibility materials; optimizing the compatibility ratio of the solid waste and kiln working fluid based on the set of characteristic parameter requirements for the compatibility materials to obtain an optimized compatibility ratio of the solid waste and kiln working fluid; obtaining an optimized set of solid waste characteristic parameter requirements based on the optimized solid waste and kiln working fluid compatibility ratio; and completing the solid waste compatibility based on the optimized set of solid waste characteristic parameter requirements and the set of solid waste characteristic parameters. The present invention can ensure the quality of boiler / industrial kiln products while reducing pollutant emissions under normal operating conditions of the boiler / industrial kiln.

Description

Technical Field

[0001] The present invention belongs to the technical field of collaborative solid waste disposal in industrial kilns, and in particular relates to a multi-solid waste matching optimization method based on the coupling of matrix operations and machine learning. Background Art

[0002] With the advancement of urbanization, the emission and disposal of urban solid waste has become increasingly prominent. Urban household waste, municipal sludge, and waste plastics emitted from domestic sources; industrial sludge, oily sludge, and waste tires emitted from urban production sources; and hazardous waste, such as incineration fly ash, hazardous oil waste, and medical waste, are massive, diverse, complex, and widely distributed, posing serious hazards. Their individual disposal is constrained by technology, equipment, and investment. my country's various boilers and industrial kilns (coal-fired boilers, ironmaking blast furnaces, and cement rotary kilns) feature high temperatures of 1,000 to 2,000°C, comprehensive pollutant removal systems, and high-capacity material flows, facilitating their coordinated disposal. This not only addresses the problem of boilers and industrial kilns operating at low loads for extended periods, but also addresses the disposal of large quantities of solid waste.

[0003] However, there are the following difficulties in the process of co-treatment of urban solid waste in industrial kilns: 1) The complexity of multi-source urban solid waste and the differences in the process characteristics of different industrial kilns require that a comprehensive technical solution must be adopted based on the physical and chemical properties of solid waste and the characteristics of pollutants to achieve compatibility with the kiln process system and thermal process; 2) In view of the increased pollution control requirements caused by the co-treatment of urban solid waste in industrial kilns, characteristic pollutants such as dioxins and heavy metals caused by the co-treatment of solid waste must be effectively controlled throughout the entire process.

[0004] Regarding the co-processing of solid waste in industrial kilns, many developed countries began experimental research using cement kilns to treat combustible waste between the 1970s and 1990s. In 1974, the Lawrence Cement Plant in Canada pioneered the use of waste lubricating oil as an alternative fuel in dry-process cement kilns. The treatment process had no adverse effects on the environment or cement quality, demonstrating the feasibility of co-processing solid and hazardous waste in cement kilns. Regarding research on the co-processing of solid waste in industrial kilns in my country, Xiao Haiping et al. found that, without modifying existing cement production equipment, co-processing waste incineration fly ash in cement rotary kilns not only completely decomposes dioxins using the kiln's high temperatures but also effectively solidifies heavy metals. Currently, the main types of waste co-processed by Chinese cement companies include municipal solid waste, sewage sludge, solid waste, and hazardous waste. In recent years, numerous cement companies, including China National Building Materials, Conch, Hongshi, Jinyu Jidong, and Huaxin, have launched co-processing production lines. In 2020 alone, 37 cement kiln co-processing projects for hazardous and solid waste were publicly approved.

[0005] Furthermore, developed countries abroad are rapidly developing the co-treatment of solid waste using pulverized coal furnaces and fluidized bed boilers. These co-processed wastes include biomass (poplar wood, sawdust, straw, orange peel, rice husks, etc.), RDF, sludge, medicinal residues, petroleum coke, peat, plastics, waste tires, and other woody wastes. Currently, China's installed thermal power capacity approaches 1.1×109 kW. Coal-fired power plant boilers have largely undergone ultra-low emission retrofits, and their widespread geographical distribution provides the necessary conditions for co-processing solid waste. Leveraging the efficient power generation systems and centralized pollutant treatment facilities of existing coal-fired power plants, implementing power generation technology coupled with the co-treatment of various solid wastes is one path to achieving low-carbon, clean development for coal-fired power plants. Many power plants in China already utilize sludge treatment facilities, and provinces such as Jiangsu, Guangdong, Shandong, and Zhejiang have already established sludge treatment facilities for power plants. In addition to sludge, numerous engineering trials and application cases have been conducted in power plants around the world using coal-fired boilers to co-process other types of solid waste. Campbell et al. in the UK studied the co-combustion of textile waste and coal in a circulating fluidized bed combustion chamber at different mixing ratios; Leeds University studied the NO x The co-treatment of waste tire powder in the reburning process was studied and found that waste tire powder has better NO x Reburn performance.

[0006] Blast furnaces are the main facilities in modern ironmaking. In 1995, Germany began the development, research, and application of technologies for the co-treatment of solid waste in blast furnaces. Germany, Japan, and other countries have successively developed technologies for treating electroplating sludge in blast furnaces, detoxification of chromium slag in rotary kilns, and mature technologies for injecting plastics into blast furnaces. Central South University found that the co-treatment of waste incineration fly ash using sintering machines had no negative impact on sintering indicators and slightly improved them. The resulting flue gas could still be discharged in a standardized manner through conventional activated carbon systems. Co-treatment of solid waste in blast furnaces. Babich et al. in Germany studied the reaction kinetics of the co-treatment of waste plastics in blast furnaces and found that the physical properties of waste plastics (grain size, shape, porosity, and specific surface area) had a stronger influence on their conversion than their chemical properties. Currently, domestic enterprises such as Baosteel, Wuhan Iron and Steel, and Shandong Metallurgical have carried out research and engineering practice on various technologies for the co-treatment of solid waste in steel smelting kilns.

[0007] Currently, research on organic solid waste compatibility methods is mostly in the laboratory stage, using experimental research methods. Under different experimental conditions, with different environmental and economic benefits as the goal, the optimal compatibility ratio is tested. Through component and ratio data analysis, most existing studies use a small amount of sludge as an auxiliary material to reduce pollution. This is consistent with the conclusion of Conesa and Juan A that a small amount of sludge can be used as an inhibitor of organic pollutants. However, due to its diverse composition, including kitchen waste, paper, rubber, plastic, metal, textiles, etc., it has a high carbon content and a relatively high O / C atomic ratio, which has great combustion potential. Therefore, most municipal solid waste is used as the primary material.

[0008] Some studies are based on statistics and use machine learning models to predict the compatibility ratio. These calculation methods or experimental studies have not escaped the limitations of the laboratory and are still somewhat different from engineering practice. The specific manifestations are as follows:

[0009] (1) It does not take into account the various parameters of different boiler / industrial kiln operating conditions, such as different combustion temperatures and pollutant emission limits, and different boilers / industrial kilns have different requirements for certain specific resources or energy (coal-fired boilers need to consider calorific value and moisture, cement rotary kilns need to pay attention to the mineral Cao content, and iron blast furnaces need to consider the Fe element content), etc.

[0010] (2) The compatibility only focuses on the compatibility between solid wastes, without considering the boiler / industrial kiln working fluid. Under actual working conditions, the boiler / industrial kiln working fluid is the main substance in the operation of the boiler, but often in experiments, only the compatibility of a small amount of added solid waste is studied. This has resulted in a small proportion of solid waste added in the actual compatibility of the industrial kiln co-treatment of solid waste industry at this stage.

[0011] Therefore, in response to the above difficulties, a boiler / industrial kiln collaborative solid waste disposal method based on mathematical relationships that is closer to actual working conditions is proposed. Summary of the Invention

[0012] To solve the above technical problems, the present invention proposes a multi-solid waste matching optimization method based on the coupling of matrix operations and machine learning, taking into account various parameters in different boiler / industrial kiln operating conditions.

[0013] To achieve the above objectives, the present invention provides a multi-solid waste compatibility optimization method based on the coupling of matrix operations and machine learning, comprising:

[0014] Obtaining initial characteristic parameters and matching materials, wherein the matching materials include solid waste and kiln working fluid;

[0015] Using a machine learning model nested optimization algorithm to obtain a set of characteristic parameter requirements for the compatible materials;

[0016] Obtaining the optimal compatibility ratio of solid waste and kiln working fluid based on the set of characteristic parameter requirements of the compatibility materials, the mathematical relationship between the characteristic parameter requirements of solid waste and the characteristic parameters of kiln working fluid;

[0017] According to the optimal solid waste and kiln working fluid compatibility ratio, an optimal solid waste characteristic parameter requirement set is obtained;

[0018] Based on the optimal solid waste characteristic parameter requirement set and the solid waste characteristic parameter set, solid waste matching is completed.

[0019] Optionally, the initial characteristic parameters include:

[0020] Solid waste characteristic parameter set, solid waste characteristic parameter requirement set and kiln working fluid characteristic parameter requirement set.

[0021] Optionally, obtaining the characteristic parameter requirement set of the compatible materials includes:

[0022] Build an initial machine learning model;

[0023] Optimizing the initial machine learning model to obtain a final machine learning model;

[0024] Based on the final machine learning model and combined with the optimization algorithm, a set of characteristic parameter requirements of the compatible materials is obtained.

[0025] Optionally, the step of obtaining the machine learning model includes:

[0026] Preprocessing the data of the compatible materials, using the preprocessed data to train the initial machine learning model, and then cross-validating the model to obtain the optimal hyperparameters of the algorithm;

[0027] Based on the obtained optimal hyperparameters, the final machine learning model is obtained by combining the input parameters and output performance indicators of the machine learning model.

[0028] Optionally, based on the final machine learning model and in combination with an optimization algorithm, a characteristic parameter requirement set of the compatible materials is obtained, including:

[0029] According to the boundary conditions of the input parameters, a preset search space is obtained;

[0030] According to the final machine learning model, the optimization algorithm is used with a heuristic search strategy to automatically search in the preset search space, and the characteristic parameter requirement set of the compatible materials is obtained through iteration.

[0031] Optionally, the optimized compatibility ratio includes:

[0032] Constructing a compatibility ratio formula based on the characteristic parameter requirement set of the compatibility materials, the characteristic parameter requirement set of the solid waste, and the characteristic parameter requirement set of the kiln working fluid;

[0033] The maximum value, minimum value of each subset in the solid waste characteristic parameter requirement set and the corresponding subset in the kiln working fluid characteristic parameter requirement set are substituted into the compatibility ratio formula, the solid waste characteristic parameter requirement set and the kiln working fluid characteristic parameter requirement set are optimized to obtain the optimized compatibility ratio.

[0034] Optionally, the compatibility ratio formula is:

[0035] XB+(1-X)C=Z

[0036] X is the proportion of total solid waste in the compatible materials, %; B is the set of solid waste characteristic parameter requirements; C is the set of boiler / industrial kiln working fluid characteristic parameter requirements; Z is the set of compatible material characteristic parameter requirements.

[0037] Optionally, completing the solid waste matching based on the optimized solid waste characteristic parameter requirement set and the solid waste characteristic parameter set includes:

[0038] Obtaining a solid waste compatibility ratio set based on the optimized solid waste characteristic parameter requirement set and the solid waste characteristic parameter set;

[0039] According to the solid waste compatibility ratio set and the maximum and minimum values ​​of the optimized solid waste characteristic parameter requirement set, the allowable solid waste compatibility ratio range is obtained to complete the solid waste compatibility.

[0040] Compared with the prior art, the present invention has the following advantages and technical effects:

[0041] (1) The present invention proposes a multi-solid waste compatibility optimization method based on the coupling of matrix operations and machine learning, which takes into account the various parameter requirements between boiler / industrial kiln working fluids and solid wastes, and is closer to the compatibility in actual working conditions.

[0042] (2) The present invention optimizes the method for the coordinated disposal of multiple solid wastes in industrial kilns, and while complying with national and industry standards, it ensures the quality of boiler / industrial kiln products while reducing pollutant emissions.

[0043] (3) The present invention provides a new formulation idea. In the process of solid waste formulation, it is necessary to combine the original working conditions and consider the calorific value, moisture, corrosive elements, heavy metal content, low melting point components and other resource or energy requirements.

[0044] (4) The present invention aims to improve the utilization rate of solid waste resources, enhance the accuracy of compatibility, and ensure the energy or resource efficiency of boilers / industrial kilns on the basis of ensuring the normal operation of boilers / industrial kilns. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0046] Figure 1 This is a flow chart of a multi-solid waste compatibility optimization method based on the coupling of matrix operations and machine learning according to an embodiment of the present invention;

[0047] Figure 2 This is a flowchart of the machine learning model nesting optimization algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0049] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0050] This embodiment proposes a multi-solid waste compatibility optimization method based on the coupling of matrix operations and machine learning, including the following steps:

[0051] 1. Determine the compatible materials

[0052] Before determining the matching materials, the type of boiler / industrial kiln should be confirmed, and the type of solid waste to be co-processed should be selected according to the boiler / industrial kiln type. For co-processing of solid waste in coal-fired boilers, reference should be made to the requirements of DL / T 1339; for iron-making blast furnaces, reference should be made to GB / T 18512 and GB / T 10322.6; and for cement rotary kilns, reference should be made to DB44 / T 882, HJ 662, and GB / T30760.

[0053] 2. Ideas for Combination Methods

[0054] After confirming the compatible materials, the basic characteristics test of the solid waste in the compatible materials should be carried out to obtain the solid waste characteristic parameter set A;

[0055] Taking the addition of solid waste in existing working conditions as an example, basic characteristic tests were conducted on the solid waste added in multiple examples to obtain the solid waste characteristic parameter requirement range B set;

[0056] The boiler / industrial kiln working fluid characteristic parameter requirements are represented by the various parameters of the boiler / industrial kiln working fluid under normal conditions. Basic characteristic tests are conducted on the boiler / industrial kiln working fluid to obtain a set C of boiler / industrial kiln working fluid parameter requirements;

[0057] The requirements for the characteristic parameters of compatible materials are obtained by using machine learning models, nested optimization algorithms and heuristic searches to obtain the Z set of characteristic parameter requirements for compatible materials;

[0058] Optimize the compatibility ratio of solid waste and kiln working fluid to make the compatible materials meet the characteristic parameter requirements. In order to meet the requirements of all solid waste characteristic parameters B, the compatibility ratio is set flexibly, and the solid waste characteristic parameter requirements are further optimized with the latest compatibility ratio. The specific compatibility ratio between solid wastes can be obtained through the solid waste characteristic parameters and their parameter requirements. The process is as follows: Figure 1 As shown:

[0059] 3. Determine the basic characteristic parameters

[0060] 3.1 Define “Solid Waste Characteristic Parameters” as Set A

[0061] The characteristic parameters of solid waste include the calorific value, moisture content, sulfur content, chlorine content, heavy metal content (lead, zinc, copper, chromium, nickel, manganese, cadmium, arsenic, etc.), low melting point components (potassium oxide, sodium oxide, magnesium oxide, etc.) and other characteristic parameters of solid waste.

[0062] Solid waste should be subjected to calorific value testing, industrial analysis, elemental analysis, heavy metal content testing, chemical composition analysis, mineral composition analysis, etc. Each test item should be repeated three times or more, and the average of the single test results for the same solid waste should be taken as the final test result.

[0063] The solid waste characteristic parameters are summarized to form set A, which is specifically expressed as follows:

[0064]

[0065] Where:

[0066] A is a set of solid waste characteristic parameters; the rows in set A represent the unified characteristic parameter data of different solid wastes, and the columns represent the different characteristic parameter data of the same solid. For example, 42 Indicates the specific value of the fourth characteristic parameter of the second type of solid waste, a 43 Indicates the specific situation of the fourth characteristic parameter of concern for the third type of solid waste.

[0067] 3.2 Define “Solid Waste Characteristic Parameter Requirements” as Set B

[0068] The requirements for solid waste characteristic parameters should be consistent with the solid waste characteristic parameters. Taking the solid waste selection and addition ratio in actual engineering applications as an example, characteristic parameter testing is carried out. The specific test items and methods are the same as 3.1.

[0069] The solid waste compatibility ratios under different circumstances in actual engineering applications are multiplied by their different characteristic parameter data to form the solid waste characteristic parameter requirements set B, as follows:

[0070]

[0071] Where:

[0072] B is the set of solid waste characteristic parameter requirements; b1, b2…b n The required range for different characteristic parameters;

[0073] 3.3 Define the C set of “boiler / industrial furnace working fluid characteristic parameter requirements”

[0074] Requirements for boiler / industrial furnace working fluid characteristic parameters: Basic characteristic tests of working fluid materials under normal operating conditions of boilers / industrial furnaces in actual engineering applications shall be conducted. Specific test items and methods are the same as those in 3.1.

[0075] According to the actual application of the project, the boiler / industrial furnace working fluid compatibility ratio and its different basic characteristics test data are multiplied correspondingly to obtain the boiler / industrial furnace working fluid characteristic parameter requirements C set, as follows:

[0076]

[0077] Where:

[0078] C is the set of required parameters of boiler / industrial furnace working fluid characteristics; c1, c2…c n Different characteristic parameter requirements for boiler / industrial furnace working fluids;

[0079] 4. Determine the Z set of characteristic parameters required for compatible materials

[0080] 4.1Z Collection Acquisition Ideas

[0081] Under the condition of meeting the corresponding national standards, the method of nested optimization algorithm of machine learning model is used to obtain the Z set of characteristic parameter requirements of compatible materials. The specific steps are as follows: Figure 2 shown.

[0082] 4.2 Data Preprocessing and Machine Learning Model Construction

[0083] Perform data preprocessing and build machine learning models. This includes the following key steps:

[0084] (1) Data collection: Collect data related to the industrial kiln process, including calorific value, moisture content, corrosive elements, heavy metal content, mineral composition and other input parameters, as well as product concentration as output.

[0085] (2) Data cleaning: Clean the data and process missing values ​​and outliers to ensure the quality and accuracy of the data.

[0086] (3) Feature Engineering: Perform feature engineering, including feature selection, creating interaction terms, or other feature engineering techniques to better utilize input parameters.

[0087] (4) Data standardization: Data standardization or normalization of input parameters to ensure the weight balance between different features.

[0088] (5) Model construction: Three different machine learning models, including artificial neural network (ANN), support regression (SVR), and random forest (RF), were constructed to predict product concentration.

[0089] 4.3 Model Performance Optimization and Selection

[0090] The optimization and selection of model performance includes the following key steps:

[0091] (1) Hyperparameter tuning: Select appropriate hyperparameters such as the number of hidden layers and neuron configuration (for ANN), kernel function and regularization parameter (for SVR), and the number and depth of decision trees (for RF).

[0092] (2) Model training and cross-validation: Each model is trained and cross-validation is used to evaluate model performance to obtain the best combination of hyperparameters and improve model accuracy.

[0093] (3) Performance indicators: Focus on the performance indicators of the model, such as mean mean error (MRE) / root mean square error (RMSE), to ensure that the selected model provides the best product concentration prediction accuracy.

[0094] (4) Optimal model selection: The model that best suits the problem is ultimately selected to achieve the best product concentration prediction accuracy.

[0095] 4.4 Embedded Optimization Algorithms and Heuristic Search

[0096] Using the final model found, the optimization algorithm is embedded to automatically search within a specific area based on the input and output boundary conditions to find the best combination of process conditions. The key processes include the following:

[0097] (1) Optimization algorithm selection: Select an appropriate optimization algorithm to search for the optimal process conditions in the multidimensional parameter space, such as gradient descent, genetic algorithm, particle swarm algorithm or other heuristic algorithms.

[0098] (2) Definition of boundary conditions: Boundary conditions refer to the boundary conditions of the input parameters clearly defined within the characteristic parameter range of the above-mentioned solid waste, boiler / industrial kiln working fluids and matching materials, such as the maximum and minimum values ​​of the input calorific value, moisture, sulfur content, chlorine content, and heavy metal zinc content; the maximum and minimum values ​​of the output heavy metal lead content can limit the search space, ensure that the process conditions are within an acceptable range, and ensure the feasibility of the process implementation.

[0099] (3) Heuristic search: The optimization algorithm uses a heuristic search strategy to automatically search for the best combination of process conditions based on the product concentration and boundary conditions predicted by the model, which is similar to the problem of finding the global optimal solution in a multidimensional space.

[0100] (4) Automatic iteration: The algorithm involves multiple model training and evaluations to continuously improve the proposed process conditions. In each iteration, the algorithm tries different combinations of input parameters, passes them to the trained model, and evaluates the quality of the solution based on the predicted product concentrations and boundary conditions until the best combination of process conditions that meets the optimization goal is found.

[0101] Finally, the Z set of characteristic parameter requirements of compatible materials is obtained through iteration, as follows:

[0102]

[0103] Where:

[0104] Z is the set of required characteristic parameters of compatible materials; z1, z2...z n Requirements for characteristic parameters of compatible materials;

[0105] 5. Calculate the total compatibility ratio range X' after optimization

[0106] The combination material characteristic parameter requirements set Z, solid waste characteristic parameter requirements set B, and boiler / industrial kiln working fluid characteristic parameter requirements set C are calculated according to formula (5):

[0107] XB+(1-X)C=Z (5)

[0108] Where:

[0109] X is the proportion of total solid waste in the compatible materials, %; B is the set of solid waste characteristic parameter requirements; C is the set of boiler / industrial kiln working fluid characteristic parameter requirements; Z is the set of compatible material characteristic parameter requirements;

[0110] B solid waste characteristic parameter set is required to be a range set. To obtain the proportion of total solid waste in the compatible materials in the range X, the maximum and minimum values ​​of each subset in the set are calculated and substituted into the calculation formula (6)(7)(8)(9)(10)

[0111] Xb1+(1-X)c1≥z1 (6)

[0112] Xb2+(1-X)c2≤z2 (7)

[0113] Xb3+(1-X)c3≤z3 (8)

[0114] Xb4+(1-X)C4≤z4 (9)

[0115] Xb n +(1-X)c n ≤z n or Xb n +(1-X)c n ≥z n (10)

[0116] Available

[0117] When b 1max ≥c1≥b 1min hour,

[0118]

[0119] When c1 ≥ b 1max ≥b 1min hour,

[0120]

[0121] When b 1max ≥b 1min When ≥c1,

[0122]

[0123] Similarly

[0124] When b 2 / 3 / 4max ≥c 2 / 3 / 4 ≥b 2 / 3 / 4min hour,

[0125]

[0126] When c 2 / 3 / 4 ≥b 2 / 3 / 4max ≥b 2 / 3 / 4min hour,

[0127]

[0128] When b2 / 3 / 4max ≥b 2 / 3 / 4min ≥c 2 / 3 / 4 hour,

[0129]

[0130] In summary, in order to simultaneously meet all the characteristic parameter requirements of boilers / industrial kilns, the proportion of total solid waste in the matching materials X is intersected under the conditions of the required range of each parameter, and the range of the optimized total matching ratio X' can be obtained;

[0131] 6. Calculation of solid waste characteristic parameter requirements B'

[0132] Substitute the optimized total compatibility ratio X' into (11)

[0133] X'B'+(1-X')C=Z (11)

[0134] Where:

[0135] X' is the total compatibility ratio after optimization, %; B' is the set of solid waste characteristic parameter requirements after optimization; C is the set of boiler / industrial kiln working fluid characteristic parameter requirements; Z is the set of compatible material characteristic parameter requirements;

[0136] Available

[0137]

[0138] In summary, by substituting the range of X' into (12), we can obtain the set of B' ranges of the optimized solid waste characteristic parameter requirements;

[0139] 7. Calculate the allowable compatibility ratio range between solid wastes I

[0140] The mixing relationship formula between solid wastes (13) (14) is used to bring in the optimized solid waste characteristic parameter requirement B' range set, and through the inverse matrix operation (15), the I set can be obtained.

[0141] A·I=B' (13)

[0142] Where:

[0143] I is the solid waste compatibility ratio set, %; A is the solid waste basic characteristic parameter A set; B' is the optimized solid waste characteristic parameter requirement set;

[0144] Expand (13) to get

[0145]

[0146] Where:

[0147] a xyis the basic parameter of the x-characteristic parameter corresponding to the y-type solid waste, %; i n is the proportion of different solid wastes in the matching materials, %; b n 'For different solid waste characteristic parameter requirements;

[0148] I=A - ·B' (15)

[0149] Where:

[0150] I is the solid waste compatibility ratio set; A — is the inverse matrix of the basic characteristic parameters A of solid waste; B' is the required set of characteristic parameters of solid waste after optimization;

[0151] Substituting the B' range set into (15) and taking the maximum and minimum values ​​respectively, the allowable compatibility ratio range I between solid wastes can be obtained.

[0152] Example

[0153] Step 1: Determine the compatible materials

[0154] The matching materials are municipal sludge, waste incineration fly ash and boiler / industrial kiln working fluid. The co-treatment of solid waste in cement kilns should refer to DB44 / T 882, HJ 662 and GB / T 30760. Because waste incineration fly ash contains a large amount of insoluble substances such as Ca, Si, Al, Fe, etc., these components are similar to high-quality lime and can replace the addition of lime; municipal sludge contains a large amount of organic components and a high proportion of combustible components. Its low calorific value on a dry basis is similar to that of lean coal and inferior lignite, which can provide heat for the cement firing process. In summary, waste incineration fly ash and municipal sludge are selected as matching materials.

[0155] Step 2: Determine the type of characteristic parameters

[0156] Considering the impact of various parameters in the matching materials on the normal operation of the boiler / industrial kiln and product quality and the total amount of experimental data, the following characteristic parameter requirements are selected as an example: calorific value, moisture, sulfur content, chlorine content, heavy metal zinc content, and heavy metal lead content.

[0157] Step 3: Determine the set of characteristic parameters

[0158] 1. Determine the set of solid waste characteristic parameters A

[0159] According to the characteristic parameters of calorific value, moisture, sulfur content, chlorine content, heavy metal zinc content, and heavy metal lead content, the calorific value of municipal sludge is derived from literature research and the average value is taken. Waste incineration fly ash is an inorganic substance with a calorific value of 0, but for the convenience of calculation, it is taken as 1. Other data sources are elemental analysis, industrial analysis, and heavy metal content testing of municipal sludge and waste incineration fly ash. The elemental analysis and industrial analysis data were obtained by elemental analyzer (Elemantar: Vario ELcube) and industrial analyzer (5E-MAG6700), respectively. The heavy gold content test data was obtained by inductively coupled plasma mass spectrometry (ICP-MS), as follows:

[0160]

[0161] Where:

[0162] a n1 、a n2 Indicates different types of solid waste, 1 represents municipal sludge, 2 represents waste incineration fly ash; a 1n is the calorific value characteristic parameter of sludge and waste incineration fly ash, kJ / kg; a 2n is the characteristic parameter of moisture content of sludge and waste incineration fly ash, %; a 3n is the characteristic parameter of sulfur content in sludge and waste incineration fly ash, %; a 4n is the characteristic parameter of chlorine content in sludge and waste incineration fly ash, %; a 5n is the characteristic parameter of heavy metal zinc content in sludge and waste incineration fly ash, mg / kg; a 6n is the characteristic parameter of heavy metal lead content in sludge and waste incineration fly ash, mg / kg;

[0163] 2. Determine the solid waste parameter set B

[0164] The solid waste characteristic parameter requirements are obtained by multiplying the municipal sludge and waste incineration fly ash characteristic parameters by the corresponding dosage ratio range to obtain the solid waste characteristic parameter requirements set B, as follows:

[0165]

[0166] Where:

[0167] B is a set of characteristic parameter requirements; b1 is the required range of calorific value in solid waste, kJ / kg; b2 is the required range of moisture content in solid waste, %; b3 is the required range of sulfur content in solid waste, %; b4 is the required range of chlorine content in solid waste, %; b5 is the required range of heavy metal zinc content in solid waste, mg / kg; b6 is the required range of heavy metal lead content in solid waste, mg / kg;

[0168] 3. Determine the boiler / industrial furnace working fluid characteristic parameter requirements C set

[0169] Due to the limited amount of data, coal is used as the main working fluid of the boiler / industrial kiln in this example. In actual projects, comprehensive testing should be carried out in accordance with the requirements of the patent, as follows:

[0170]

[0171] Where: C is the characteristic parameter requirement set; c1 is the calorific value requirement for boiler / industrial kiln working fluid, kJ / kg; c2 is the moisture content requirement for boiler / industrial kiln working fluid, %; c3 is the sulfur content requirement for boiler / industrial kiln working fluid, %; c4 is the chlorine content requirement for boiler / industrial kiln working fluid, %; c5 is the heavy metal zinc content requirement for boiler / industrial kiln working fluid, mg / kg; c6 is the heavy metal lead content requirement for boiler / industrial kiln working fluid, mg / kg;

[0172] Step 4: Determine the Z set of characteristic parameter requirements for compatible materials

[0173] 1. Data preprocessing and machine learning model construction

[0174] Perform data preprocessing and build machine learning models. This includes the following key steps:

[0175] (1) Data collection: Experimental data is used as the data set in this example, which includes input parameters such as calorific value, moisture content, sulfur content, chlorine content, heavy metal zinc content and heavy metal lead content, as well as product concentration as output.

[0176] (2) Data cleaning: Clean the data and process missing values ​​and outliers to ensure the quality and accuracy of the data.

[0177] (3) Feature Engineering: Perform feature engineering, including feature selection, creating interaction terms, or other feature engineering techniques to better utilize input parameters.

[0178] (4) Data standardization: Data standardization or normalization of input parameters to ensure the weight balance between different features.

[0179] (5) Model construction: Three different machine learning models, including artificial neural network (ANN), support regression (SVR), and random forest (RF), were constructed to predict product concentration.

[0180] 2. Model performance optimization and selection

[0181] The optimization and selection of model performance include the following key steps:

[0182] (1) Hyperparameter tuning: Select appropriate hyperparameters such as the number of hidden layers and neuron configuration (for ANN), kernel function and regularization parameter (for SVR), and the number and depth of decision trees (for RF).

[0183] (2) Model training and cross-validation: Each model is trained and cross-validation is used to evaluate model performance to obtain the best combination of hyperparameters and improve model accuracy.

[0184] (3) Performance indicators: Focus on the performance indicators of the model, such as mean mean error (MRE) / root mean square error (RMSE), to ensure that the selected model provides the best product concentration prediction accuracy.

[0185] (4) Optimal model selection: Finally, the model that best fits the problem is selected to achieve the best product concentration prediction accuracy. Random forest (RF) is the best choice for this model.

[0186] 3. Embedding optimization algorithms and heuristic search

[0187] Using the found RF optimal model (taken as examples in this model), we embed the optimization algorithm and automatically search within a specific area based on the input and output boundary conditions to find the best combination of process conditions. This includes the following key processes:

[0188] (1) Optimization algorithm selection: Select an appropriate optimization algorithm to search for the optimal process conditions in the multidimensional parameter space, such as gradient descent, genetic algorithm, particle swarm algorithm or other heuristic algorithms.

[0189] (2) Boundary condition definition: Clearly define the boundary conditions of the input parameters, limit the search space, ensure that the process conditions are within an acceptable range, and ensure the feasibility of process implementation.

[0190] (3) Heuristic search: The optimization algorithm uses a heuristic search strategy to automatically search for the best combination of process conditions based on the product concentration and boundary conditions predicted by the model, which is similar to the problem of finding the global optimal solution in a multidimensional space.

[0191] (4) Automatic iteration: The algorithm includes multiple model training and evaluation steps to continuously improve the proposed process conditions. In each iteration, the algorithm tries different combinations of input parameters, passes them to the trained RF model, and evaluates the quality of the solution based on the predicted product concentrations and boundary conditions until the best combination of process conditions that meets the optimization goal is found.

[0192] Finally, the Z set of characteristic parameter requirements of compatible materials is obtained, as follows:

[0193]

[0194] In the formula: Z is the set of characteristic parameter requirements of compatible materials; z1 is the calorific value requirement of compatible materials, kJ / kg; z2 is the moisture content requirement of compatible materials, %; z3 is the sulfur content requirement of compatible materials, %; z4 is the chlorine content requirement of compatible materials, %; z5 is the heavy metal zinc content requirement of compatible materials, mg / kg; z6 is the heavy metal lead content requirement of compatible materials, mg / kg;

[0195] Step 5: Calculate the optimized total compatibility ratio X'

[0196] According to the linear relationship between the compatibility materials, solid waste, and boiler / industrial kiln working fluids, by inputting specific values ​​and optimizing X, the optimized total compatibility ratio X' is obtained. The specific calculation is as follows:

[0197] XB+(1-X)C=Z

[0198] Where:

[0199] X is the proportion of total solid waste in the compatible materials, %; B is the set of solid waste characteristic parameter requirements; C is the set of boiler / industrial kiln working fluid characteristic parameter requirements; Z is the set of compatible material characteristic parameter requirements;

[0200] Expand the above formula and substitute the corresponding values ​​for calculation

[0201] Xb1+(1-X)c1≥z1

[0202] Xb2+(1-X)c2≤z2

[0203] Xb3+(1-X)c3≤z3

[0204] Xb4+(1-X)c4≤z4

[0205] Xb5+(1-X)c5≤z5)

[0206] Xb6+(1-X)c6≤z6

[0207] Substitute numerical values ​​and calculate the solution

[0208]

[0209]

[0210]

[0211]

[0212]

[0213]

[0214] In summary, in order to simultaneously meet the requirements of all characteristic parameters in boilers / industrial kilns, the proportion of total solid waste in the matching materials X is intersected under the conditions of the required range of each parameter to obtain the optimized total matching ratio X′.

[0215] X′∈[0, 0.2611]

[0216] Step 6: Calculate the optimized solid waste characteristic parameter requirement B′

[0217] Substitute the optimized compatibility ratio X′ into the following formula

[0218] X′B′+(1-X′)C=Z

[0219] Substitute the corresponding numerical calculation to obtain the optimized solid waste requirement B′. The specific results are as follows

[0220]

[0221] Step 7: Calculate the allowable compatibility ratio range of solid waste I

[0222] The solid waste characteristic parameter set A is known, and the optimized solid waste characteristic parameter requirement set B′ is known. Substitute the A and B′ sets into their compatibility ratio relationship formula to calculate the allowable compatibility ratio range set I between solid wastes, as follows:

[0223] A·I=B′

[0224] Where: I is the solid waste compatibility ratio set, %; A is the solid waste basic characteristic parameter A set; B' is the optimized solid waste characteristic parameter requirement set;

[0225] Substitute the known conditions and expand to obtain

[0226]

[0227] Where: i1 is the allowable addition ratio of municipal sludge; i2 is the allowable addition ratio of fly ash from waste incineration;

[0228] Take the maximum and minimum values ​​in the B′ set respectively, calculate the range of i1 and i2, and obtain the allowable proportion range of solid waste addition

[0229] i1∈[0.0036, 0.2282]

[0230] i2∈[0.0025, 0.2125]

[0231] This compatibility ratio method can provide guidance for the coordinated disposal of solid waste by boilers / industrial kilns. While ensuring the normal operation of the boilers / industrial kilns, it can both ensure product quality and reduce pollutant emissions. This compatibility method offers a new direction for research on solid waste compatibility. The invention provides a new approach to solid waste compatibility. The solid waste compatibility process must be combined with existing operating conditions, taking into account calorific value, moisture content, corrosive elements and heavy metal content, low-melting-point components, and other resource or energy requirements.

[0232] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A multi-component solid waste compatibility optimization method based on the coupling of matrix operations and machine learning, characterized in that: include: Obtaining initial characteristic parameters and matching materials, wherein the matching materials include solid waste and kiln working fluid; Using a machine learning model nested optimization algorithm to obtain a set of characteristic parameter requirements for the compatible materials; Obtaining the optimal compatibility ratio of solid waste and kiln working fluid based on the set of characteristic parameter requirements of the compatibility materials, the mathematical relationship between the characteristic parameter requirements of solid waste and the characteristic parameters of kiln working fluid; The optimized compatibility ratio includes: Constructing a compatibility ratio formula based on the characteristic parameter requirement set of the compatibility materials, the characteristic parameter requirement set of the solid waste, and the characteristic parameter requirement set of the kiln working fluid; Substituting the maximum value and minimum value of each subset in the solid waste characteristic parameter requirement set and the corresponding subset in the kiln working fluid characteristic parameter requirement set into the compatibility ratio formula, optimizing the solid waste characteristic parameter requirement set and the kiln working fluid characteristic parameter requirement set to obtain the optimized compatibility ratio; The compatibility ratio formula is: XB+(1-X)C=Z X is the proportion of total solid waste in the compatible materials, %; B is the set of solid waste characteristic parameter requirements; C is the set of boiler / industrial kiln working fluid characteristic parameter requirements; Z is the set of compatible material characteristic parameter requirements; According to the optimal solid waste and kiln working fluid compatibility ratio, an optimal solid waste characteristic parameter requirement set is obtained; Based on the optimal solid waste characteristic parameter requirement set and the solid waste characteristic parameter set, solid waste matching is completed.

2. The multi-solid waste compatibility optimization method based on the coupling of matrix operation and machine learning according to claim 1 is characterized in that: The initial characteristic parameters include: Solid waste characteristic parameter set, solid waste characteristic parameter requirement set and kiln working fluid characteristic parameter requirement set.

3. The multi-solid waste compatibility optimization method based on the coupling of matrix operation and machine learning according to claim 1 is characterized in that: The characteristic parameter requirements for obtaining the compatible materials include: Build an initial machine learning model; Optimizing the initial machine learning model to obtain a final machine learning model; Based on the final machine learning model and combined with the optimization algorithm, a set of characteristic parameter requirements of the compatible materials is obtained.

4. The multi-solid waste compatibility optimization method based on the coupling of matrix operation and machine learning according to claim 3 is characterized in that: The steps of obtaining the machine learning model include: Preprocessing the data of the compatible materials, using the preprocessed data to train the initial machine learning model, and then cross-validating the model to obtain the optimal hyperparameters of the algorithm; Based on the obtained optimal hyperparameters, the final machine learning model is obtained by combining the input parameters and output performance indicators of the machine learning model.

5. The multi-solid waste compatibility optimization method based on the coupling of matrix operation and machine learning according to claim 3 is characterized in that: Based on the final machine learning model and combined with the optimization algorithm, the characteristic parameter requirement set of the compatible materials is obtained, including: According to the boundary conditions of the input parameters, a preset search space is obtained; According to the final machine learning model, the optimization algorithm is used with a heuristic search strategy to automatically search in the preset search space, and the characteristic parameter requirement set of the compatible materials is obtained through iteration.

6. The multi-component solid waste compatibility optimization method based on the coupling of matrix operation and machine learning according to claim 1 is characterized in that: Based on the optimal solid waste characteristic parameter requirement set and the solid waste characteristic parameter set, completing solid waste matching includes: Obtaining a solid waste compatibility ratio set based on the optimal solid waste characteristic parameter requirement set and the solid waste characteristic parameter set; According to the solid waste compatibility ratio set and the maximum and minimum values ​​of the optimal solid waste characteristic parameter requirement set, the allowable solid waste compatibility ratio range is obtained to complete the solid waste compatibility.

Images