Method for determining influence of waste incineration bottom ash on performance of mortar based on hyperspectral data

Abstract

The application provides a kind of based on the influence determination method of waste incineration bottom ash on mortar performance of hyperspectral data, it is related to building material performance analysis technical field, the application is by determining the proportion of influencing component in bottom ash hyperspectral data, to constitute original data, and set gradient determines several groups of experimental groups with slightly different proportions, obtains the characteristic parameter of each experimental group, and constructs bottom ash performance prediction model, in combination with particle swarm optimization algorithm, the bottom ash combination with the highest performance excitation value is selected as the reference group, on this basis, the fitting relationship between the reference group content and compressive strength is established, finally, the theoretical best bottom ash content position and the corresponding maximum compressive strength are determined based on the fitting relationship, and the best content interval is clear. The application realizes the accurate determination of the best bottom ash content interval and the optimization of mortar performance by establishing the bottom ash performance prediction model and combining the content-compressive strength relationship fitting.

Description

Technical Field

[0001] This invention relates to the field of building material performance analysis technology, specifically a method for determining the influence of waste incineration bottom ash on mortar performance based on hyperspectral data. Background Technology

[0002] Alkali-activated mortar has attracted widespread attention because it can utilize industrial waste residue to prepare high-performance, low-carbon, and environmentally friendly building materials. As a potential raw material, the addition of waste incineration bottom ash to alkali-activated mortar may not only change the physical and mechanical properties of the mortar, but also affect its microstructure and durability.

[0003] However, due to the complex physicochemical properties of waste incineration bottom ash and the significant differences in composition between bottom ash from different sources, how to scientifically evaluate its incorporation effect in alkali-activated mortar, determine a reasonable dosage range, and reveal the mechanism of its influence on mortar performance are current technical challenges in research and engineering applications.

[0004] In existing technologies, the utilization of waste incineration bottom ash relies heavily on single-index evaluation, lacking a systematic characterization of the multi-dimensional features of the bottom ash. Furthermore, a quantitative relationship between bottom ash characteristic parameters and alkali-activated mortar performance has not been established, and scientific prediction models are lacking. Moreover, traditional technologies often aim to find a specific optimal bottom ash dosage without fully utilizing intelligent optimization algorithms to optimize the bottom ash mix ratio. The experimental methods lack standardization, resulting in poor data accuracy and repeatability. The impact of multiple factors on mortar performance is not comprehensively considered, and reasonable performance thresholds and optimal dosage ranges are lacking. This leads to low efficiency in bottom ash resource utilization, making it difficult to meet the stability and superior performance requirements of mortar in practical engineering projects.

[0005] Therefore, it is necessary to propose a method for determining the impact of waste incineration bottom ash on mortar performance based on hyperspectral data to solve the aforementioned problem.

[0006] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] The purpose of this invention is to provide a method for determining the influence of waste incineration bottom ash on mortar performance based on hyperspectral data, so as to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data, the specific steps of which include:

[0010] Step 1: Obtain the hyperspectral data of the background gray to be analyzed, and select the reflectance of relevant bands in the hyperspectral image according to the types of influencing components set. Determine the proportion of each influencing component based on the selected reflectance data and form the original data.

[0011] Step 2: Based on the proportion of the original data, determine several experimental groups with different proportions according to the set gradient, determine the total content of active oxides in each experimental group, and obtain the particle size distribution parameters of each experimental group.

[0012] Step 3: Analyze the amorphous phase ratio of each experimental group based on X-ray diffraction technology, and preprocess the total content of active oxides, particle size distribution parameters and amorphous phase ratio of the experimental groups to obtain the characteristic parameters of each experimental group;

[0013] Step 4: Based on the characteristic parameters of each experimental group, establish a performance fitness function to determine the performance excitation value of each experimental group, and construct a background gray performance prediction model. Optimize the background gray performance prediction model based on the particle swarm optimization algorithm, and select the experimental group with the highest performance excitation value as the reference group.

[0014] Step 5: Based on the proportion of the reference group, set several reference groups and mortar dosage schemes, and prepare several concrete specimens according to the dosage schemes. Determine the compressive strength of the concrete specimens and construct a model of the reference group dosage and the compressive strength of the mortar specimens.

[0015] Step 6: Based on the compressive strength model, find the theoretical position of the optimal dosage for the reference group, calculate the corresponding maximum compressive strength, and finally determine the optimal dosage range for the reference group.

[0016] Furthermore, the hyperspectral data of the background gray to be analyzed is obtained, and the reflectance of relevant bands is selected from the hyperspectral image based on the types of influencing components. The method used is as follows:

[0017] The background gray sample to be analyzed was uniformly dried and spread on a uniform and flat substrate. The sample surface was scanned using a hyperspectral imager to obtain a three-dimensional hyperspectral data cube. The data is represented as follows: ,in Represented as spatial coordinates, For wavelength, To obtain the measured spectral intensity, the reflectance of the background gray sample to be analyzed is calculated by comparing it with the spectral response of a pre-obtained standard reflector. The formula used is as follows:

[0018] ;

[0019] in, Indicates spatial location At this point, the wavelength is The reflectance of the sample at that time In spatial location At this point, the wavelength is The intensity of reflected light from the sample at that time Indicates at wavelength The intensity of reflected light from the standard reflector measured at the location;

[0020] X-ray fluorescence spectrometry was used to determine the chemical composition of the background gray, identifying the influencing components in the background gray to be analyzed. Based on the spectral characteristics of the influencing components, the corresponding reflectance band ranges were selected, and the reflectance data of the characteristic bands corresponding to these components were extracted from the hyperspectral data to form a reflectance feature vector. ,in The selected bands are the correlation bands with the components that affect the background gray.

[0021] Furthermore, a mapping model was established, and the proportions of each influencing component were determined based on the selected reflectance data, forming the original data. The method used was as follows:

[0022] A mapping model is constructed based on a neural network. The input layer of the model is used to receive the input reflectance feature vector. Corresponding to the selected The model has reflectivity in each band, and its hidden layer is configured with reflectivity in each band. A number of neurons are used, and a non-linear activation function is employed to process the reflectance feature vector. And extract the feature information from it, the output layer of the model is used to output the first feature in the background gray. The predicted content values ​​of each influencing component;

[0023] Standard base gray samples with known content were collected, and their hyperspectral reflectance data were measured. The reflectance of characteristic bands was extracted, and the corresponding component content was determined by experimental chemical analysis. This data was then used to initialize the model and determine the number of characteristic bands in the input layer. The number of neurons in the hidden layer The weights and biases of the mapping model are initialized, and the mean squared error is used as the loss function, based on the following formula:

[0024] ;

[0025] in, This represents the total value of the loss function, used to measure the magnitude of the error between the model's predictions and the actual values. This indicates the total number of samples with a gray background. This is the index of the gray-based sample, and , For the first The first gray sample The true value of the content of each component is affected. For the first The first sample obtained by model prediction in the gray background sample The predicted content values ​​of each influencing component;

[0026] The weights and biases are iteratively updated using a nonlinear optimization algorithm with the goal of minimizing the loss function. The gradient is calculated and the error is backpropagated to adjust the weights and biases in the mapping model. For new sample data of the background gray to be predicted, the reflectance vector is extracted and input into the trained mapping model to obtain the content of each influencing component in the background gray to be analyzed.

[0027] The content of all influencing components in the base gray to be analyzed is summarized to form the raw data of the component proportions of the base gray sample, which serves as the basis for subsequent analysis.

[0028] Furthermore, a gradient was established to determine several experimental groups with different proportions, the total content of active oxides in each experimental group was determined, and the particle size distribution parameters of each experimental group were obtained. The method used was as follows:

[0029] Assume the original data for the proportion of base gray components is: ,in Indicates the first The original proportions of each influencing component, and Choose the allowable step size for each component and define the gradient vector: ,in, Indicates the first The adjustment step size of each influencing component, This is an index of the components that influence the background gray, and ;

[0030] Select the number of adjustment steps This represents the number of levels of variation of each base gray component around the original proportion. A gradient variation is applied to each component to generate several candidate proportion values. The formula used is:

[0031] ;

[0032] in, Indicates the base gray level The influencing component is in the first Candidate proportion values ​​under each adjustment level To adjust the index of the hierarchy, and The range of values ​​is arrive , Indicates rounding down, and is limited to... ;

[0033] All experimental groups are generated by permuting and combining the candidate values ​​of all components;

[0034] The formula used to calculate the total content of reactive oxides in each experimental group is as follows:

[0035] ;

[0036] in, This indicates the total content of active oxides in each group of base gray. , , , These represent the mass percentages of silicon dioxide, aluminum oxide, calcium oxide, and iron oxide in each group of base gray.

[0037] The particle size distribution parameters of each group of base ash are obtained, including fineness, uniformity, and specific surface area. The particle size distribution data is automatically measured and output based on a laser particle size analyzer program. Commonly used characteristic particle sizes include... , and , The particle size corresponding to a 10% cumulative distribution indicates the fine-grained segment. This represents the particle size corresponding to the 50% cumulative distribution, indicating the median particle size. It represents the particle size corresponding to 90% cumulative distribution, indicating the coarse-grained segment;

[0038] Define the particle size of the base gray as The arithmetic mean particle size is used to characterize the fineness of each group of base ash particles, based on the following formula:

[0039] ;

[0040] in, This indicates the fineness of each group of base coats, specifically the average diameter of the particles in each group of base coats. For each group of base gray, the first The median particle size across the particle size range For each group of base gray, the first Volume fraction of each particle size range;

[0041] The uniformity of each base coat is defined based on the degree of concentration of particle size distribution, using the following formula:

[0042] ;

[0043] in, This indicates the uniformity of the base coat in each group;

[0044] Assuming the particle shape is spherical, based on the density of each batch of base ash obtained from a laser particle size analyzer and the calculated fineness of each batch of base ash, the specific surface area of ​​each batch of base ash particles is defined using the formula for calculating the specific surface area of ​​a sphere. The formula used is as follows:

[0045] ;

[0046] in, This represents the specific surface area per unit mass of particles in each group of base ash. The density of each group of base gray.

[0047] Furthermore, the proportion of amorphous phase in the base ash of each experimental group was analyzed, and the total content of active oxides, particle size distribution parameters, and proportion of amorphous phase in each experimental group were preprocessed to obtain characteristic parameters for each group. The method used was as follows:

[0048] The mineral crystalline phase content in each set of base ash was determined using X-ray diffraction, and the proportion of amorphous phases was calculated using the internal standard method. The formula used is as follows:

[0049] ;

[0050] in, This indicates the proportion of amorphous phase in each group of base gray. It is the integral of the diffraction peak intensity of the mineral crystal phase for each group of background gray. It is the total diffraction intensity of the mineral crystalline phase content in each group of base gray;

[0051] The total content of active oxides, particle size distribution parameters, and proportion of amorphous phase in each experimental group were normalized and preprocessed, and all data were converted to [database name missing]. The formula used to determine this is:

[0052] ;

[0053] in, This represents the normalized value of each experimental group on a certain parameter after normalization. This represents the actual measured value of each experimental group on a certain parameter. This indicates the minimum value of this parameter among all base gray combinations. This indicates the maximum value of the parameter across all background gray combinations;

[0054] The total content of reactive oxides in the experimental group after normalization was determined. Particle size distribution parameters , , and the proportion of amorphous phases These serve as characteristic parameters for each experimental group.

[0055] Furthermore, a performance fitness function was established to calculate the performance excitation value of the base ash in each experimental group, and a base ash performance prediction model was constructed. The model was optimized based on the particle swarm optimization algorithm, using the following method:

[0056] Based on the reaction kinetics formula and particle size effect, a performance fitness function is constructed to obtain the performance activation value of each group of base ash. The formula used is as follows:

[0057] ;

[0058] in, This indicates the performance activation value of each group of base gray. The activation energy of each group of base ash. The gas constant is The reaction temperature. , , These are the reaction sensitivity coefficients for the total content of active oxides, the proportion of amorphous phase, and the particle size distribution parameter, respectively.

[0059] The total content of active oxides, particle size distribution parameters, and amorphous phase ratio of each group of base grays are combined into a feature vector. A base gray performance prediction model is constructed based on a neural network learning structure. The gray performance prediction model includes an input layer, a hidden layer, and an output layer. The input layer is responsible for receiving the feature vector. The hidden layer is an intermediate layer in the neural network between the input and output layers, responsible for feature extraction and information processing of the input feature vector. A nonlinear activation function is used in the hidden layer to capture complex patterns. The output layer generates the final prediction result of the model, that is, the performance excitation value of each group of base grays. ;

[0060] The particle swarm optimization algorithm is used to search for optimal combinations of background gray parameters, thereby optimizing the background gray performance prediction model. A particle swarm is initialized, with each particle representing a set of background gray parameter combinations. The position and velocity of the particles are set, where position represents a combination of weights and thresholds in the prediction model, and velocity defines the particle's direction and stride in the search space. At the beginning of each iteration, a set of weights and thresholds is randomly generated as the initial positions of the particles. The initial positions of the particles are generated based on a normal distribution, ensuring that the particles are uniformly distributed throughout the search space. The velocity and position of the particles are updated using the following formula:

[0061] ;

[0062] ;

[0063] in, Indicates the first Individual particles The speed of time, Indicates the first Individual particles The position of time - For inertial weights, , It is a learning factor. , It is a random number. Indicates the first The optimal position of each particle Indicates the globally optimal position. For the first Individual particles Location at any moment For the first Individual particles The speed of time;

[0064] Preset performance activation threshold The process of repeatedly evaluating loss and updating particle position and velocity is repeated, and the performance excitation value of the current background gray combination is calculated in each iteration. The performance excitation value calculated in each iteration is compared with... If the performance excitation value of the current background gray combination exceeds the preset threshold, the particle corresponding to the current background gray combination is defined as the optimal particle. The position of the found optimal particle will be used to update the weights and thresholds of the model. By retraining the model, we can ensure that its prediction on the model reaches the best performance.

[0065] Based on the optimized performance prediction model for the base gray, the experimental group with the highest performance excitation value was selected as the reference group.

[0066] Furthermore, several reference groups and mortar dosage schemes were established, and several concrete specimens were prepared according to the dosage schemes. The compressive strength of the concrete specimens was determined, and a model of the compressive strength of the reference group dosages and mortar specimens was constructed. The method used was as follows:

[0067] Based on the particle swarm optimization-optimized base gray performance prediction model, an experimental group with the best performance was selected as the reference group for dosing design, and different mass percentage dosings were designed. Series, determining the doping gradient range in The admixture dosage is expressed as the proportion of the base mortar mass to the total mortar mass. Each component is accurately weighed according to the ratio, mixed thoroughly, and cubic mortar specimens are prepared according to the standard. The specimens are cured to the specified age, and the compressive strength of the mortar specimens under each admixture gradient is determined. ;

[0068] Using the dosage of each reference group as the model input and the compressive strength of the corresponding mortar specimens as the model output, a compressive strength model is constructed. Based on the chemical principle of alkali-activated mortar with base cement, and the trend of the compressive strength of the mortar specimens first increasing and then decreasing with the dosage, a quadratic polynomial is used to fit the nonlinear change of compressive strength with dosage. The formula used is as follows:

[0069] ;

[0070] in, This indicates the compressive strength at different dosage levels after curing. Indicates the dosage of the reference group. , and These are the fitting parameters;

[0071] Sorting the doping levels of multiple reference groups measured in the experiment Compressive strength of corresponding mortar specimens A set of fitting equations was constructed to minimize the error between the true and fitted values ​​of the compressive strength of the mortar specimens under the gradient admixture of each reference group. The formula used is as follows:

[0072] ;

[0073] in, Indicates the first The square of the compressive strength error of the mortar specimens. For the first The true value of the compressive strength of the mortar specimens. It refers to the number of mortar specimens. This serves as an index for mortar specimens, and ;

[0074] Based on the fitted equation system, respectively... , and Find the partial derivatives and set them equal to zero to obtain a set of normal equations. Solve the normal equations to obtain the fitting parameters. , and The value of the fitting parameter , and Substitute the values ​​into a quadratic polynomial to obtain the fitting relationship between the amount of base lime and the compressive strength of the mortar specimen.

[0075] Furthermore, the theoretical position of the optimal base ash content is identified, and the corresponding maximum compressive strength is calculated. Finally, the optimal base ash content range is determined using the following formula:

[0076] Based on the fitted relationship between the base cement content and the compressive strength of mortar specimens, the theoretical position of the optimal base cement content is determined. This theoretical position is the maximum point of the fitted relationship, which is defined as the theoretical position of the optimal base cement content. , will obtain Substitute the values ​​into the fitted formula to calculate the maximum compressive strength of multiple mortar specimens. Set the compressive strength threshold coefficient Find the satisfying The optimal doping range, making Solving for two real roots yields two roots. and This is the optimal dosage range. .

[0077] Compared with the prior art, the beneficial effects of the present invention are:

[0078] This invention establishes a quantitative relationship between bottom ash characteristic parameters and alkali-activated mortar performance by systematically characterizing multi-dimensional feature parameters such as the main chemical composition, particle size distribution, and amorphous phase ratio of waste incineration bottom ash. This significantly improves the scientificity and accuracy of bottom ash performance evaluation. Furthermore, by utilizing performance fitness functions and neural network prediction models, combined with particle swarm optimization algorithms, the invention intelligently optimizes the combination of bottom ash parameters, realizing full-process data analysis and precise screening of bottom ash components and proportions. This overcomes the problems of traditional methods, such as one-sided evaluation, strong reliance on experience, and limited optimization means.

[0079] Furthermore, this invention establishes a fitting relationship between the base ash content and the mortar compressive strength by designing mortar specimens with gradient base ash content, and scientifically determines the optimal base ash content range and its corresponding maximum compressive strength using mathematical methods. This effectively solves the problems of traditional base ash content design lacking theoretical guidance and ignoring the fluctuations in experiments. Attached Figure Description

[0080] Figure 1 This is a schematic diagram of the overall method flow of the present invention. Detailed Implementation

[0081] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0082] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0083] Example:

[0084] Please see Figure 1 A method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data, the specific steps of which include:

[0085] Step 1: Obtain the hyperspectral data of the background gray to be analyzed, and select the reflectance of relevant bands in the hyperspectral image according to the types of influencing components set. Determine the proportion of each influencing component based on the selected reflectance data and form the original data.

[0086] In a specific embodiment of the present invention, we first grind the waste incineration bottom ash sample and control the bottom ash particle size to 75. The purpose of this measure is to reduce the impact of particle size on chemical composition analysis, improve the accuracy and repeatability of X-ray fluorescence spectrometry results, and avoid signal attenuation due to excessively large particle size. Based on the measurement results, the main chemical components of the bottom ash are silicon dioxide, aluminum oxide, calcium oxide, and iron oxide. According to the chemical analysis data of actual waste incineration bottom ash, the upper and lower limits of the content of these four oxides are determined: silicon dioxide content is between 30% and 50%, aluminum oxide content is between 10% and 20%, calcium oxide content is between 15% and 35%, and iron oxide content is between 5% and 15%.

[0087] It should be noted that different chemical components in the background gray exhibit characteristic absorption peaks in the near-infrared and short-wave infrared ranges. For example, the most prominent chemical component in the background gray, silica, has a dominant characteristic absorption band around 2200 nm, alumina around 2300 nm, calcium oxide around 2330-2500 nm, and iron oxide around 500-600 nm. These major chemical components can be detected in the hyperspectral image of the background gray because the minerals they constitute have unique and identifiable spectral absorption characteristics. Hyperspectral imaging technology can capture these characteristics, thereby enabling spectral-based component analysis, which is the basis for the feasibility of this scheme.

[0088] Furthermore, the hyperspectral data of the background gray to be analyzed is obtained, and the reflectance of relevant bands is selected from the hyperspectral image based on the types of influencing components. The method used is as follows:

[0089] The background gray sample to be analyzed was uniformly dried and spread on a uniform and flat substrate. The sample surface was scanned using a hyperspectral imager to obtain a three-dimensional hyperspectral data cube. The data is represented as follows: ,in Represented as spatial coordinates, For wavelength, To obtain the measured spectral intensity, the reflectance of the background gray sample to be analyzed is calculated by comparing it with the spectral response of a pre-obtained standard reflector. The formula used is:

[0090] ;

[0091] in, Indicates spatial location At this point, the wavelength is The reflectance of the sample at that time In spatial location At this point, the wavelength is The intensity of reflected light from the sample at that time Indicates at wavelength The intensity of reflected light from the standard reflector measured at the location;

[0092] X-ray fluorescence spectrometry was used to determine the chemical composition of the background gray, identifying the influencing components in the background gray to be analyzed. Based on the spectral characteristics of the influencing components, the corresponding reflectance band ranges were selected, and the reflectance data of the characteristic bands corresponding to these components were extracted from the hyperspectral data to form a reflectance feature vector. ,in The selected bands are the correlation bands with the components that affect the background gray.

[0093] Furthermore, a mapping model was established, and the proportions of each influencing component were determined based on the selected reflectance data, forming the original data. The method used was as follows:

[0094] A mapping model is constructed based on a neural network. The input layer of the model is used to receive the input reflectance feature vector. Corresponding to the selected The model has reflectivity in each band, and its hidden layer is configured with reflectivity in each band. A number of neurons are used, and a non-linear activation function is employed to process the reflectance feature vector. And extract the feature information from it, the output layer of the model is used to output the first feature in the background gray. The predicted content values ​​of each influencing component;

[0095] Standard base gray samples with known content were collected, and their hyperspectral reflectance data were measured. The reflectance of characteristic bands was extracted, and the corresponding component content was determined by experimental chemical analysis. This data was then used to initialize the model and determine the number of characteristic bands in the input layer. The number of neurons in the hidden layer The weights and biases of the mapping model are initialized, and the mean squared error is used as the loss function, based on the following formula:

[0096] ;

[0097] in, This represents the total value of the loss function, used to measure the magnitude of the error between the model's predictions and the actual values. This indicates the total number of samples with a gray background. This is the index of the gray-based sample, and , For the first The first gray sample The true value of the content of each component is affected. For the first The first sample obtained by model prediction in the gray background sample The predicted content values ​​of each influencing component;

[0098] The weights and biases are iteratively updated using a nonlinear optimization algorithm with the goal of minimizing the loss function. The gradient is calculated and the error is backpropagated to adjust the weights and biases in the mapping model. For new sample data of the background gray to be predicted, the reflectance vector is extracted and input into the trained mapping model to obtain the content of each influencing component in the background gray to be analyzed.

[0099] The content of all influencing components in the base gray to be analyzed is summarized to form the raw data of the component proportions of the base gray sample, which serves as the basis for subsequent analysis.

[0100] Step 2: Based on the proportion of the original data, determine several experimental groups with different proportions according to the set gradient, determine the total content of active oxides in each experimental group, and obtain the particle size distribution parameters of each experimental group.

[0101] In a specific embodiment of the present invention, we obtain the particle size distribution parameters of each group of base ash using a laser particle size analyzer, and perform comprehensive preprocessing of the characteristic parameters of the base ash in combination with the total content of active oxides. This process not only systematically reveals the physical particle size characteristics and chemical amorphous structure characteristics of the base ash, but also provides a scientific basis for quantifying its potential reactivity and hydration performance. By integrating the three key parameters of particle size distribution, chemical composition and amorphous phase content, we can more accurately reflect the comprehensive performance of the base ash.

[0102] It should be noted that, in order to facilitate random sampling, the content of each component can be divided into several discrete points within a set range, for example, with a step size of 1% or 2%, to ensure the diversity of random proportions and facilitate calculation. For each influencing component, a value is randomly generated. The specific method is as follows: generate random floating-point numbers or integers within the content range of each component, and use a computer random function to generate a value between the ranges of each component. Generate a sufficient number of values ​​according to the designed experimental scale to serve as the base gray sample proportioning scheme for subsequent experiments.

[0103] Furthermore, a gradient was established to determine several experimental groups with different proportions, the total content of active oxides in each experimental group was determined, and the particle size distribution parameters of each experimental group were obtained. The method used was as follows:

[0104] Assume the original data for the proportion of base gray components is: ,in Indicates the first The original proportions of each influencing component, and Choose the allowable step size for each component and define the gradient vector: ,in, Indicates the first The adjustment step size of each influencing component, This is an index of the components that influence the background gray, and ;

[0105] Select the number of adjustment steps This represents the number of levels of variation of each base gray component around the original proportion. A gradient variation is applied to each component to generate several candidate proportion values. The formula used is:

[0106] ;

[0107] in, Indicates the base gray level The influencing component is in the first Candidate proportion values ​​under each adjustment level To adjust the index of the hierarchy, and The range of values ​​is arrive , Indicates rounding down, and is limited to... ;

[0108] In the above formula, the specific logic for generating candidate proportional values ​​is as follows: For example, when When, it means that the component content can be taken as follows based on the original ratio: , , In this way, each influencing component has There are a total of 100 candidate ratio values. There are 10 components, so the theoretical number of combinations of all components is 100. However, due to The number of combinations will be very large. In actual experiments, strategies are often adopted to reduce the number of combinations. We adopt the method of adjusting only the key components and keeping the other components fixed. The key components here include silicon dioxide, aluminum oxide, calcium oxide and iron oxide in the base gray.

[0109] All experimental groups are generated by permuting and combining the candidate values ​​of all components;

[0110] The formula used to calculate the total content of reactive oxides in each experimental group is as follows:

[0111] ;

[0112] in, This indicates the total content of active oxides in each group of base gray. , , , These represent the mass percentages of silicon dioxide, aluminum oxide, calcium oxide, and iron oxide in each group of base gray.

[0113] The particle size distribution parameters of each group of base ash are obtained, including fineness, uniformity, and specific surface area. The particle size distribution data is automatically measured and output based on a laser particle size analyzer program. Commonly used characteristic particle sizes include... , and , The particle size corresponding to a 10% cumulative distribution indicates the fine-grained segment. This represents the particle size corresponding to the 50% cumulative distribution, indicating the median particle size. It represents the particle size corresponding to 90% cumulative distribution, indicating the coarse-grained segment;

[0114] Define the particle size of the base gray as The arithmetic mean particle size is used to characterize the fineness of each group of base ash particles, based on the following formula:

[0115] ;

[0116] in, This indicates the fineness of each group of base coats, specifically the average diameter of the particles in each group of base coats. For each group of base gray, the first The median particle size across the particle size range For each group of base gray, the first The volume fraction of each particle size interval; in the above formula, the median particle size of each interval is averaged and weighted with its corresponding volume fraction. This comprehensively reflects the average particle size level of the entire sample. Using volume weighting instead of a simple arithmetic mean more accurately reflects the distribution of actual material particles, especially when the particle volume distribution is uneven, avoiding the unreasonable influence of extreme values ​​on the results. The fineness of each base ash... The bigger the better, because The larger the particle size, the finer the particles of the base mortar, which can accelerate the reaction rate, shorten the curing time, and significantly improve the early and long-term mechanical properties of the mortar.

[0117] The uniformity of each base coat is defined based on the degree of concentration of particle size distribution, using the following formula:

[0118] ;

[0119] in, This indicates the uniformity of each group of base coats; in the above formula, uniformity... It reflects the width of the particle size distribution. The closer the value is to 1, the better the uniformity, the more concentrated the particle distribution, and the more uniform the particle size. The base mortar with uniform particle size distribution can ensure that the particles in the mortar or concrete are arranged more regularly, reduce the gaps between large and fine particles, and improve the density.

[0120] Assuming the particle shape is spherical, based on the density of each batch of base ash obtained from a laser particle size analyzer and the calculated fineness of each batch of base ash, the specific surface area of ​​each batch of base ash particles is defined using the formula for calculating the specific surface area of ​​a sphere. The formula used is as follows:

[0121] ;

[0122] in, This represents the specific surface area per unit mass of particles in each group of base ash. The density of each base layer is given; in the above formula, it is assumed that the particles are spherical, and the formula is derived based on the ratio of surface area to volume of particles per unit volume. The specific surface area per unit volume of particles is further divided by the density to obtain the specific surface area per unit mass. The larger the better, which means that the total surface area per unit mass of the material is larger. A larger surface area provides more reaction interfaces, which is conducive to the occurrence of chemical reactions or physical adsorption.

[0123] It should be noted that the reason for calculating the total active oxide content of each group of base mortar is that a higher active oxide content indicates a stronger potential pozzolanic reaction performance of the base mortar, which is beneficial to improving the strength and durability of the mortar. Furthermore, the active oxide content index allows for the selection of stable and suitable base mortar mix ratios, reducing the randomness of experiments. We chose fineness, uniformity, and specific surface area as particle size distribution parameters for each group of base mortar because fineness reflects the average particle size and is a direct indicator of the overall particle size distribution of the base mortar; the finer the base mortar particles, the larger the surface area. Higher reactivity generally indicates more complete hydration, promoting the formation of cementitious substances and improving mortar performance. Uniformity reflects the breadth of the base ash particle size distribution, indicating the concentration or dispersion of particle size. Specific surface area, the total surface area per unit mass of particles, directly relates to the contact area and reaction rate between particles and the alkali activator. Therefore, obtaining these particle size distribution parameters can comprehensively and scientifically characterize the base ash particle size properties, providing a solid foundation for subsequent material performance analysis and optimization design. We selected three characteristic particle sizes for calculating fineness, uniformity, and specific surface area. This indicates that the particle size in the sample is smaller than The particles comprise 10% of the total volume, representing the distribution range of the finest particles in the base gray sample and reflecting the characteristics of the fine-grained segment. It can be used to evaluate the content and fineness of the finest particles in the base ash. For active materials, fine particles often have higher reactivity and have an important impact on subsequent reaction performance. This indicates that 50% of the particles in the sample have a particle size less than or equal to [the specified value]. , It can be understood as the "medium particle size" of the sample, which is the median of the particle distribution and reflects the size of the "main" particles in the particle size distribution of the base gray sample. This indicates that 90% of the particles in the sample have a particle size less than or equal to 100%. This represents the particle size characteristics of the coarse-grained segment. Used to characterize the proportion and particle size of coarse particles in the base gray sample. The presence of coarse particles can affect the material's filling properties, distribution uniformity, and final performance.

[0124] Step 3: Analyze the amorphous phase ratio of each experimental group based on X-ray diffraction technology, and preprocess the total content of active oxides, particle size distribution parameters and amorphous phase ratio of the experimental groups to obtain the characteristic parameters of each experimental group.

[0125] In a specific embodiment of the present invention, the amorphous phase ratio in the base gray is used as one of the characteristic parameters because the higher the amorphous phase ratio of the base gray, the better its potential activity and subsequent reaction performance. Using it as a characteristic parameter helps the model to more accurately predict and distinguish the performance excitation values ​​of base grays in different experimental groups. Since the dimensions of each characteristic parameter are different and the numerical ranges vary greatly, we have performed normalization processing. All features are mapped to a unified range, and the normalized data helps the optimization algorithm converge faster and avoid getting trapped in local optima or oscillations.

[0126] Furthermore, the proportion of amorphous phase in the base ash of each experimental group was analyzed, and the total content of active oxides, particle size distribution parameters, and proportion of amorphous phase in each experimental group were preprocessed to obtain characteristic parameters for each group. The method used was as follows:

[0127] The mineral crystalline phase content in each set of base ash was determined using X-ray diffraction, and the proportion of amorphous phases was calculated using the internal standard method. The formula used is as follows:

[0128] ;

[0129] in, This indicates the proportion of amorphous phase in each group of base gray. It is the integral of the diffraction peak intensity of the mineral crystal phase for each group of background gray. It is the total diffraction intensity of the mineral crystalline phase content in each group of base gray. In the above formula, since the mineral crystalline phase in the base gray will produce diffraction, that is, all the diffraction peaks in the sample come from the crystalline phase, the total diffraction intensity can be regarded as the sum of the peak intensities of the crystalline phase. The method of indirectly calculating the proportion of amorphous phase by "1 minus the proportion of crystalline phase" is a commonly used and reasonable method in the quantitative analysis of amorphous phase.

[0130] The total content of active oxides, particle size distribution parameters, and proportion of amorphous phase in each experimental group were normalized and preprocessed, and all data were converted to [database name missing]. The formula used to determine this is:

[0131] ;

[0132] in, This represents the normalized value of each experimental group on a certain parameter after normalization. This represents the actual measured value of each experimental group on a certain parameter. This indicates the minimum value of this parameter among all base gray combinations. This indicates the maximum value of the parameter across all background gray combinations;

[0133] The total content of reactive oxides in the experimental group after normalization was determined. Particle size distribution parameters , , and the proportion of amorphous phases These serve as characteristic parameters for each experimental group.

[0134] Step 4: Based on the characteristic parameters of each experimental group, establish a performance fitness function to determine the performance excitation value of each experimental group, and construct a background gray performance prediction model. Optimize the background gray performance prediction model based on the particle swarm optimization algorithm, and select the experimental group with the highest performance excitation value as the reference group.

[0135] In a specific embodiment of this invention, we establish a fitness function to comprehensively transform the key indicators of the background gray into a single numerical indicator—the performance excitation value—and provide an evaluation standard for the optimization algorithm. The particle swarm optimization algorithm is a global optimization method based on swarm intelligence, which requires evaluating the quality of each "solution" through the fitness function. The fitness function serves as the objective function, guiding the exploration and utilization of the search space during the optimization process. By calculating the performance excitation value of each combination, the particle swarm optimization algorithm can identify the background gray combination with better performance and gradually approach the optimal solution.

[0136] Furthermore, a performance fitness function was established to calculate the performance excitation value of the base ash in each experimental group, and a base ash performance prediction model was constructed. The model was optimized based on the particle swarm optimization algorithm, using the following method:

[0137] Based on the reaction kinetics formula and particle size effect, a performance fitness function is constructed to obtain the performance activation value of each group of base ash. The formula used is as follows:

[0138] ;

[0139] in, This indicates the performance activation value of each group of base gray. The activation energy of each group of base ash. The gas constant is The reaction temperature. , , These represent the reaction sensitivity coefficients for the total content of active oxides, the proportion of amorphous phase, and the particle size distribution parameter, respectively. In the above formulas, the performance fitness function is constructed based on the Arrhenius reaction kinetics formula and the BHS model. The total content of active oxides in the base gray is... The higher the concentration, the greater the reaction potential, and the higher the amorphous phase ratio. The larger the value, the stronger the amorphous structure advantage of the base gray, resulting in higher activity and reactivity. These two factors are related to the Arrhenius term. The combination reflects the effects of temperature and activation energy on reaction rate or activity, conforming to the basic principles of chemical reaction kinetics, and is reflected in the denominator. The BHS model describes the particle size effect, showing that greater fineness, uniformity, and specific surface area generally increase the reactivity of the base ash. A more uniform particle size distribution facilitates effective particle contact and reaction, reduces the obstruction of reaction interface exposure by large particles, and improves overall activity. When fineness, uniformity, and specific surface area are greater, the value of the denominator in the expression increases, and the reciprocal of the overall denominator decreases, leading to a higher performance activation value. The larger the size, the finer the texture, the higher the uniformity, and the larger the specific surface area. A higher value indicates a higher performance activation value for the base gray, signifying greater reactivity and utilization efficiency, and a higher reaction sensitivity coefficient. , , Size ratio set The purpose of this setting is to reduce the number of parameters by unifying the response sensitivity coefficient, avoid parameter overfitting, and improve the generalization ability of the model among different base gray samples. Furthermore, the content of active oxides, the proportion of amorphous phase, and the particle size parameter are all key factors affecting the activity of base gray. Giving them the same weight helps to fully reflect their combined effects.

[0140] The total content of active oxides, particle size distribution parameters, and amorphous phase ratio of each group of base grays are combined into a feature vector. A base gray performance prediction model is constructed based on a neural network learning structure. The gray performance prediction model includes an input layer, a hidden layer, and an output layer. The input layer is responsible for receiving the feature vector. The hidden layer is an intermediate layer in the neural network between the input and output layers, responsible for feature extraction and information processing of the input feature vector. A nonlinear activation function is used in the hidden layer to capture complex patterns. The output layer generates the final prediction result of the model, that is, the performance excitation value of each group of base grays. ;

[0141] The particle swarm optimization algorithm is used to search for optimal combinations of background gray parameters, thereby optimizing the background gray performance prediction model. A particle swarm is initialized, with each particle representing a set of background gray parameter combinations. The position and velocity of the particles are set, where position represents a combination of weights and thresholds in the prediction model, and velocity defines the particle's direction and stride in the search space. At the beginning of each iteration, a set of weights and thresholds is randomly generated as the initial positions of the particles. The initial positions of the particles are generated based on a normal distribution, ensuring that the particles are uniformly distributed throughout the search space. The velocity and position of the particles are updated using the following formula:

[0142] ;

[0143] ;

[0144] in, Indicates the first Individual particles The speed of time, Indicates the first Individual particles The position of time - For inertial weights, , It is a learning factor. , It is a random number. Indicates the first The optimal position of each particle Indicates the globally optimal position. For the first Individual particles Location at any moment For the first Individual particles The speed of time;

[0145] Preset performance activation threshold The process of repeatedly evaluating loss and updating particle position and velocity is repeated, and the performance excitation value of the current background gray combination is calculated in each iteration. The performance excitation value calculated in each iteration is compared with... If the performance excitation value of the current background gray combination exceeds the preset threshold, the particle corresponding to the current background gray combination is defined as the optimal particle. The position of the found optimal particle will be used to update the weights and thresholds of the model. By retraining the model, we can ensure that its prediction on the model reaches the best performance.

[0146] Based on the optimized performance prediction model for the base gray, the experimental group with the highest performance excitation value was selected as the reference group.

[0147] Step 5: Based on the proportion of the reference group, set several reference groups and mortar dosage schemes, and prepare several concrete specimens according to the dosage schemes. Determine the compressive strength of the concrete specimens and construct a model of the reference group dosage and the compressive strength of the mortar specimens.

[0148] In a specific embodiment of the present invention, we designed the dosage gradient range to be 0% to 30%. The optimizable base ash was added to the alkali-activated mortar in the following order: 0%, 5%, 10%, 15%, 20%, 25%, and 30%. The effect of different dosages of base ash on the mechanical properties of mortar is often nonlinear. The reason why we chose a quadratic polynomial as the fitting relationship for the compressive strength model is that the compressive strength of mortar often shows a trend of first rising and then falling with the change of base ash dosage, presenting a "parabolic" shape. The quadratic polynomial can fit this kind of single-peak nonlinear trend very well. Compared with higher-order polynomials, it avoids the problems of overfitting and overly complex models.

[0149] Furthermore, several reference groups and mortar dosage schemes were established, and several concrete specimens were prepared according to the dosage schemes. The compressive strength of the concrete specimens was determined, and a model of the compressive strength of the reference group dosages and mortar specimens was constructed. The method used was as follows:

[0150] Based on the particle swarm optimization-optimized base gray performance prediction model, an experimental group with the best performance was selected as the reference group for dosing design, and different mass percentage dosings were designed. Series, determining the doping gradient range in The admixture dosage is expressed as the proportion of the base mortar mass to the total mortar mass. Each component is accurately weighed according to the ratio, mixed thoroughly, and cubic mortar specimens are prepared according to the standard. The specimens are cured to the specified age, and the compressive strength of the mortar specimens under each admixture gradient is determined. ;

[0151] Using the dosage of each reference group as the model input and the compressive strength of the corresponding mortar specimens as the model output, a compressive strength model is constructed. Based on the chemical principle of alkali-activated mortar with base cement, and the trend of the compressive strength of the mortar specimens first increasing and then decreasing with the dosage, a quadratic polynomial is used to fit the nonlinear change of compressive strength with dosage. The formula used is as follows:

[0152] ;

[0153] in, This indicates the compressive strength at different dosage levels after curing. This indicates the amount of base ash added. , and These are the fitting parameters; in the above formula, for The reason for adding the absolute value and the negative sign is that the quadratic polynomial must first satisfy the condition that its "opening" is downward, and that it follows a trend of first rising and then falling. The negative sign is used because it prevents errors in subsequent calculations of the fitting parameters. Non-positive numbers appear.

[0154] Sorting the doping levels of multiple reference groups measured in the experiment Compressive strength of corresponding mortar specimens A set of fitting equations was constructed to minimize the error between the true and fitted values ​​of the compressive strength of the mortar specimens under the gradient admixture of each reference group. The formula used is as follows:

[0155] ;

[0156] in, Indicates the first The square of the compressive strength error of the mortar specimens. For the first The true value of the compressive strength of the mortar specimens. It refers to the number of mortar specimens. This serves as an index for mortar specimens, and ;

[0157] Based on the fitted equation system, respectively... , and Find the partial derivatives and set them equal to zero to obtain a set of normal equations. Solve the normal equations to obtain the fitting parameters. , and The value of the fitting parameter , and Substitute the values ​​into a quadratic polynomial to obtain the fitting relationship between the amount of base lime and the compressive strength of the mortar specimen.

[0158] Step 6: Based on the compressive strength model, find the theoretical position of the optimal dosage for the reference group, calculate the corresponding maximum compressive strength, and finally determine the optimal dosage range for the reference group.

[0159] In a specific embodiment of the present invention, the theoretical position of the optimal base ash content can be obtained based on the obtained fitting relationship, and then substituted into the fitting formula to obtain the maximum compressive strength. To obtain the optimal content range, a compressive strength threshold coefficient is set here. , The value is determined by obtaining the mean and standard deviation of compressive strength through multiple sets of repeated tests, and then determining the performance fluctuation range in combination with the confidence interval to guide the selection of the threshold. The maximum strength is subtracted by twice the standard deviation to obtain the minimum allowable strength value, and the ratio of the minimum strength to the maximum strength is defined as the compressive strength threshold coefficient.

[0160] It should be noted that the reason we sought the optimal dosage range of the base mortar rather than the optimal dosage value is that, in actual production and construction, the physicochemical properties, particle size distribution, and the ratio of cement and activator of the base mortar will fluctuate and be uncertain. These fluctuations will cause the actual dosage to deviate from the theoretical optimal value. A single optimal value is difficult to guarantee stable and optimal performance. By determining the dosage range, we can resist the impact of batch differences in materials and process fluctuations, ensuring continuous and stable performance. Selecting any value within the dosage range can ensure that the mortar performance is in excellent condition, reducing risks. Furthermore, based on the dosage range, it can be flexibly adjusted in combination with other proportioning parameters to meet the performance requirements of different projects.

[0161] Furthermore, the theoretical position of the optimal base ash content is identified, and the corresponding maximum compressive strength is calculated. Finally, the optimal base ash content range is determined using the following formula:

[0162] Based on the fitted relationship between the base cement content and the compressive strength of mortar specimens, the theoretical position of the optimal base cement content is determined. This theoretical position is the maximum point of the fitted relationship, which is defined as the theoretical position of the optimal base cement content. , will obtain Substitute the values ​​into the fitted formula to calculate the maximum compressive strength of multiple mortar specimens. Set the compressive strength threshold coefficient Find the satisfying The optimal doping range, making Solving for two real roots yields two roots. and This is the optimal dosage range. .

[0163] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0164] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0165] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0166] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data, characterized in that, The specific steps include: Step 1: Obtain the hyperspectral data of the background gray to be analyzed, and select the reflectance of relevant bands in the hyperspectral image according to the types of influencing components set. Determine the proportion of each influencing component based on the selected reflectance data and form the original data. Step 2: Based on the proportion of the original data, determine several experimental groups with different proportions according to the set gradient, determine the total content of active oxides in each experimental group, and obtain the particle size distribution parameters of each experimental group. Step 3: Analyze the amorphous phase ratio of each experimental group based on X-ray diffraction technology, and preprocess the total content of active oxides, particle size distribution parameters and amorphous phase ratio of the experimental groups to obtain the characteristic parameters of each experimental group; Step 4: Based on the characteristic parameters of each experimental group, establish a performance fitness function to determine the performance excitation value of each experimental group, and construct a background gray performance prediction model. Optimize the background gray performance prediction model using a particle swarm optimization algorithm, and select the experimental group with the highest performance excitation value as the reference group. The method used to establish the performance fitness function and construct the background gray performance prediction model is as follows: Based on the Arrhenius reaction kinetics formula and the BHS model, a performance fitness function is constructed to obtain the performance excitation value of each group of bottom ash. The total active oxide content, particle size distribution parameters and amorphous phase ratio of each group of bottom ash are combined into a feature vector. The feature vector of each experimental group is used as the input of the model and the performance excitation value of each experimental group is used as the output of the model to construct a bottom ash performance prediction model. Step 5: Based on the proportion of the reference group, set several reference groups and mortar dosage schemes, and prepare several concrete specimens according to the dosage schemes. Determine the compressive strength of the concrete specimens and construct a model of the reference group dosage and the compressive strength of the mortar specimens. Step 6: Based on the compressive strength model, find the theoretical position of the optimal dosage for the reference group, calculate the corresponding maximum compressive strength, and finally determine the optimal dosage range for the reference group.

2. The method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data according to claim 1, characterized in that, The hyperspectral data of the background gray to be analyzed is obtained, and the reflectance of relevant bands is selected from the hyperspectral image based on the types of influencing components. The method used is as follows: The background gray sample to be analyzed was uniformly dried and spread on a smooth, even substrate. The sample surface was then scanned using a hyperspectral imager to obtain a three-dimensional hyperspectral data cube, represented as follows: ,in Represented as spatial coordinates, For wavelength, The measured spectral intensity is compared with the spectral response of a pre-acquired standard reflector to calculate the reflectance of the background gray sample to be analyzed. The chemical composition of the background gray was determined by X-ray fluorescence spectrometry to identify the influencing components in the background gray to be analyzed. Based on the spectral characteristics of the influencing components, the corresponding reflectance band ranges were selected, and the reflectance data of the characteristic bands corresponding to these components were extracted from the hyperspectral data to form a reflectance feature vector. The reflectance feature vector contains the bands related to the influencing components of the background gray.

3. The method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data according to claim 2, characterized in that, A mapping model was established, and the proportions of each influencing component were determined based on the selected reflectance data, forming the original data. The method used was as follows: A mapping model is constructed based on a neural network. The input layer of the model receives the input reflectance feature vector, corresponding to the selected... The model has reflectivity in each band, and its hidden layer is configured with reflectivity in each band. The model uses a non-linear activation function to process the reflectivity feature vector and extract its feature information. The output layer of the model is used to output the first neuron in the background gray. The predicted content values ​​of each influencing component; Standard base gray samples with known content were collected, and their hyperspectral reflectance data were measured. The reflectance of characteristic bands was extracted, and the corresponding component content was determined by experimental chemical analysis. This data was then used to initialize the model and determine the number of characteristic bands in the input layer. The number of neurons in the hidden layer Initialize the weights and biases of the mapping model and use the mean squared error as the loss function; The weights and biases are iteratively updated using a nonlinear optimization algorithm with the goal of minimizing the loss function. The gradient is calculated and the error is backpropagated to adjust the weights and biases in the mapping model. For new sample data of the background gray to be predicted, the reflectance vector is extracted and input into the trained mapping model to obtain the content of each influencing component in the background gray to be analyzed. The content of all influencing components in the base gray to be analyzed is summarized to form the raw data of the component proportions of the base gray sample, which serves as the basis for subsequent analysis.

4. The method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data according to claim 3, characterized in that, A gradient was established to determine several experimental groups with different proportions. The total content of active oxides in each experimental group was determined, and the particle size distribution parameters of each group were obtained. The method used was as follows: Based on the original data of the proportions of the base gray components, the allowable variation step size for each component is selected, and a gradient vector is defined, with the number of adjustment steps chosen. This is used to represent the number of levels of change of each background gray component around the original ratio. Gradient changes are made to each component to generate several candidate ratio values. By arranging and combining the candidate values ​​of all components, all experimental groups are generated. The total content of active oxides in each experimental group was calculated based on the mass percentage of silica, alumina, calcium oxide and iron oxide in the base ash of each experimental group. The particle size distribution parameters of each group of base ash are obtained, including fineness, uniformity, and specific surface area. The particle size distribution data is automatically measured and output based on a laser particle size analyzer program. Commonly used characteristic particle sizes include... , and , The particle size corresponding to a 10% cumulative distribution indicates the fine-grained segment. This represents the particle size corresponding to the 50% cumulative distribution, indicating the median particle size. This represents the particle size corresponding to 90% cumulative distribution, indicating the coarse-grained segment. The bottom gray particle size is defined as... The arithmetic mean particle size is used to characterize the fineness of each group of base ash, and the uniformity of each group of base ash is defined by the concentration of particle size distribution. Assuming the particle shape is spherical, the specific surface area of ​​each group of bottom ash particles is defined by using the formula for calculating the specific surface area of ​​a sphere, based on the density of each group of bottom ash particles obtained from the laser particle size analyzer and the fineness of each group of bottom ash particles calculated.

5. The method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data according to claim 4, characterized in that, The proportion of amorphous phase in the base ash of each experimental group was analyzed, and the total content of active oxides, particle size distribution parameters, and proportion of amorphous phase in each experimental group were preprocessed to obtain characteristic parameters for each group. The method used was as follows: The mineral crystalline phase content in each group of base ash was determined using X-ray diffraction. The proportion of amorphous phase was calculated using the internal standard method. The total content of active oxides, particle size distribution parameters, and the proportion of amorphous phase in each experimental group were normalized before being converted to [a more accurate translation]. Between these parameters, the total content of active oxides, particle size distribution parameters, and the proportion of amorphous phase in the experimental groups after normalization were used as characteristic parameters for each experimental group.

6. The method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data according to claim 5, characterized in that, A performance fitness function was established to calculate the performance excitation value of the base ash in each experimental group, and a base ash performance prediction model was constructed. The method used was as follows: A gray performance prediction model is constructed based on a neural network learning structure. The gray performance prediction model includes an input layer, a hidden layer, and an output layer. The input layer is responsible for receiving feature vectors. The hidden layer is an intermediate layer in the neural network between the input layer and the output layer. It is responsible for feature extraction and information processing of the input feature vectors. Furthermore, a non-linear activation function is used in the hidden layer to capture complex patterns. The output layer generates the final prediction result of the model, namely the performance excitation value of each set of gray.

7. The method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data according to claim 6, characterized in that, The performance prediction model for the background gray area was optimized using the particle swarm optimization algorithm. The experimental group with the highest performance excitation value was selected as the reference group. The method used was as follows: The particle swarm optimization algorithm is used to search for optimizable combinations of background gray parameters, thereby optimizing the background gray performance prediction model. A particle swarm is initialized, with each particle representing a set of background gray parameter combinations. The position and velocity of the particles are set, where the position represents a combination of weights and thresholds of the prediction model, and the velocity defines the direction and stride of the particle's movement in the search space. At the beginning of each iteration, a set of weights and thresholds is randomly generated as the initial position of the particles. The initial position of the particles is generated based on a normal distribution, and the particles are uniformly distributed throughout the search space. The velocity and position of the particles are then updated. The process of evaluating loss and updating particle position and velocity is repeated by setting a preset performance excitation threshold. In each iteration, the performance excitation value of the current background gray combination is calculated and compared with the performance excitation threshold. If the performance excitation value of the current background gray combination exceeds the preset threshold, the particle corresponding to the current background gray combination is defined as the optimal particle. The position of the found optimal particle will be used to update the weights and threshold of the model. By retraining the model, it is ensured that its prediction on the model reaches the best performance. Based on the optimized performance prediction model for the base gray, the experimental group with the highest performance excitation value was selected as the reference group.

8. The method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data according to claim 7, characterized in that, Several reference groups and mortar admixture schemes were established, and several concrete specimens were prepared according to the admixture schemes. The compressive strength of the concrete specimens was determined, and a model of the compressive strength of the reference group admixtures and mortar specimens was constructed. The method used was as follows: Based on the particle swarm optimization-optimized base gray performance prediction model, an experimental group with the best performance was selected as the reference group for dosing design, and different mass percentage dosings were designed. Series, determining the doping gradient range in The admixture dosage is expressed as the proportion of the base ash mass to the total mortar mass. Each component is accurately weighed according to the ratio, mixed evenly, and cubic mortar specimens are prepared according to the standard. The specimens are cured to the age specified in the experiment, and the compressive strength of the mortar specimens under each admixture gradient is determined. Using the dosage of each reference group as the input of the model and the compressive strength of the mortar specimens corresponding to each reference group as the output of the model, a compressive strength model is constructed. Based on the chemical principle of alkali-activated mortar with the addition of base cement, the compressive strength of the mortar specimens first increases and then decreases with the dosage. A quadratic polynomial is used to fit the nonlinear change of compressive strength with dosage. By compiling the compressive strength of the corresponding mortar specimens and the dosage of multiple reference groups measured in the experiment, a set of fitting equations is constructed to minimize the error between the true value and the fitted value of the compressive strength of the mortar specimens under the gradient dosage of each reference group. Based on the fitted equations, partial derivatives of the fitted parameters are calculated and set to zero to obtain a normal equation system. The values ​​of the fitted parameters are obtained by solving the normal equation system. Finally, the fitted parameters are substituted into a quadratic polynomial to obtain the fitted relationship between the base cement content and the compressive strength of the mortar specimen.

9. The method for determining the influence of waste incineration bottom ash on mortar properties based on hyperspectral data according to claim 8, characterized in that, The theoretical location of the optimal base ash content is determined, and the corresponding maximum compressive strength is calculated. Finally, the optimal base ash content range is determined using the following formula: Based on the fitting relationship between the base lime content and the compressive strength of mortar specimens, the theoretical position of the optimal base lime content is found. The theoretical position is the maximum point of the fitting relationship. The maximum point of the fitting relationship is defined as the theoretical position of the optimal base lime content. The abscissa of the obtained theoretical position of the optimal base lime content on the fitting curve is substituted into the fitting relationship to calculate the maximum compressive strength in multiple groups of mortar specimens. A compressive strength threshold coefficient is set to screen out the base lime content range in the fitting curve where the compressive strength is not lower than the threshold limit. An equation is established between the fitting relationship and the maximum compressive strength under the compressive strength threshold coefficient. Finally, the equation is solved to obtain two real roots. These two roots constitute the optimal content range that satisfies the compressive strength threshold condition.

Images