A method and system for predicting the optimal mix ratio of construction waste recycling

Abstract

The present invention relates to the technical field of building material related data prediction, and specifically to a method and system for predicting the optimal mix ratio of construction waste recycling. The method obtains the target degree by clustering and comprehensively analyzing the differences between the mix ratio vector and the corresponding performance index; adaptively adds new mix ratio vectors according to the target degree to fit the performance curve and segment it, obtains all garbage mix ratio change curves and segments it; obtains the influence contribution of each type of garbage on the performance according to the overall change difference of the mix ratio vector in the performance segment and the correlation between the performance segment and the corresponding garbage mix ratio segment change, and then predicts the performance index of the mix ratio vector in combination with the influence contribution, so as to obtain the optimal mix ratio. The present invention adaptively enriches the fitting data to improve the fitting effect and efficiency of the performance curve, and predicts the performance index of the mix ratio according to the influence contribution of each type of garbage on the performance index, thereby improving the accuracy of the performance index and the optimal mix ratio prediction.

Description

Technical Field

[0001] The invention relates to the technical field of building material related data prediction, and in particular to a method and system for predicting an optimal mix ratio of building waste recycling. Background Art

[0002] With the acceleration of urbanization, the amount of construction waste has increased dramatically. Construction waste recycling technology can use recycled aggregates made from construction waste in construction, reducing resource waste and environmental pollution. However, due to the wide variety of construction waste, the properties of recycled aggregates prepared from different mix ratios of waste vary greatly.

[0003] Technicians usually use experimental methods to determine the optimal mix of various construction wastes, but the experimental methods are cumbersome and time-consuming. Therefore, fitting the mix performance curve is usually used to predict the performance of recycled aggregates prepared with unknown mix ratios, and then obtain the preferred mix ratio corresponding to the best performance; however, since the similarity of the mix ratios only reflects the similarity of the content ratio of each type of waste in the corresponding mix ratio; in the actual preparation process, the performance of the recycled aggregate is nonlinearly related to the mix ratio. Directly predicting the performance of the recycled aggregate based on the similarity of the content of each type of waste ratio may be very different from the performance of the actually prepared recycled aggregate, resulting in low accuracy of the obtained optimal mix ratio, which may cause construction accidents. Summary of the invention

[0004] In order to solve the technical problem of low accuracy of the optimal mix ratio obtained by the prior art, the purpose of the present invention is to provide a method and system for predicting the optimal mix ratio of construction waste recycling. The technical scheme adopted is as follows:

[0005] The present invention provides a method for predicting the optimal mix ratio of construction waste recycling, the method comprising:

[0006] Obtaining a preset number of mix ratio vectors of construction waste in recycled aggregates and performance indicators corresponding to the mix ratio vectors;

[0007] Obtain all clusters of the mix ratio vectors; in each cluster, according to the difference between all the mix ratio vectors and each type of garbage ratio in the cluster center vector, obtain the distribution weight of the corresponding mix ratio vector in the cluster; according to the distribution weight of the mix ratio vector in each cluster and the corresponding performance index, obtain the target degree of the corresponding cluster; according to the target degree, add a new mix ratio vector to each cluster, and fit the performance curve according to all the mix ratio vectors in the cluster and the corresponding performance index;

[0008] According to the change of the tangent slope of the performance curve, all segmentation points are screened and all performance segments are obtained; the garbage ratio curve of each type of garbage is obtained, and the dimension of the mix ratio vector of each type of garbage ratio curve and the performance curve is the same; according to the points in the garbage ratio curve corresponding to the segmentation points, all garbage ratio segments of each type of garbage ratio curve are obtained, and the change correlation between each performance segment and the corresponding garbage ratio segment is obtained; in each performance segment, according to the overall change difference of all mix ratio vectors and the corresponding change correlation, the influence contribution of each type of garbage in the corresponding performance segment to the performance index is obtained;

[0009] Obtain the difference to be predicted of each type of garbage between all mix ratio vectors in the performance curve and the mix ratio vector to be predicted; predict the performance index of the mix ratio vector to be predicted according to the impact contribution of each type of garbage and the difference to be predicted; and select the preferred mix ratio vector in the performance curve to obtain the optimal mix ratio vector.

[0010] Furthermore, the method for obtaining the distribution weight includes:

[0011] In each of the clusters, the absolute value of the first ratio difference between each of the matching ratio vectors and the same category of garbage in the cluster center vector is obtained; in each of the matching ratio vectors, the difference garbage whose absolute value of the first ratio difference is greater than a preset threshold is screened out, and the ratio of the number of the difference garbage in the matching ratio vector to the number of all garbage is obtained;

[0012] Obtain the absolute value mean of the ratio difference between each of the matching ratio vectors and all the same category garbage ratios in the corresponding cluster center vector; multiply the quantity ratio by the corresponding absolute value mean of the ratio difference to obtain the distribution weight corresponding to the matching ratio vector.

[0013] Furthermore, the method for obtaining the target degree includes:

[0014] In each of the clusters, the distribution weight of each of the mix ratio vectors is used as the weight of the corresponding performance indicator to obtain the weighted performance indicator of all the mix ratio vectors in the cluster; the coefficient of variation of the weighted performance indicator is obtained according to the coefficient of variation formula, and the coefficient of variation is the target degree of the cluster.

[0015] Furthermore, the method of adding a new matching ratio vector to the cluster includes:

[0016] The target degree of each cluster is multiplied by a preset second number and then rounded up to obtain the additional number of the mix ratio vectors added to the corresponding cluster in each increase process; in each cluster, the standard deviation of the ratio of each type of garbage in all the mix ratio vectors is obtained; among the preset second number of new mix ratio vectors, the optional mix ratio vectors whose ratios of all types of garbage are less than the corresponding mix ratio standard deviation are screened, and the optional mix ratio vectors corresponding to the increased number are added to each cluster.

[0017] Furthermore, the method for obtaining the performance curve includes:

[0018] All the mix ratio vectors are sorted in ascending order according to the ratio means of all garbage ratios in the mix ratio vectors, and the mix ratio vectors and the corresponding performance indicators are fitted using a multivariate nonlinear regression analysis method to obtain an initial performance curve;

[0019] The optional mix ratio vector is added to each of the clusters according to the adding method, the initial performance curve is updated, and the similarity of the initial performance curve before and after the update is calculated; new optional mix ratio vectors are continuously added to each cluster until a preset cutoff condition is met, and the initial performance curve after the last update is used as the performance curve.

[0020] Furthermore, the method for obtaining the segmentation points includes:

[0021] In the performance curve, the tangent slopes of all data points corresponding to the mix ratio vectors are obtained; the tangent slope ratio of the data point corresponding to the mix ratio vector and the data point corresponding to the next adjacent mix ratio vector is calculated, and the absolute value of the difference between the tangent slope ratio and a preset first normal constant is obtained and normalized to obtain a segmentation point reference value;

[0022] The data point corresponding to the mix ratio vector whose segmentation point reference value is greater than the preset segmentation reference value threshold is used as the segmentation point.

[0023] Furthermore, the method for obtaining the influence contribution of each type of garbage in the performance segment on the performance index includes:

[0024] In each of the performance segments, the cosine similarities between all the mix ratio vectors and the cluster center vector in each cluster are obtained; in each cluster corresponding to the performance segment, the ratio similarity ratio of the cosine similarity of the mix ratio vector and the sum of the cosine similarities of all the mix ratio vectors is obtained, the ratio similarity ratios of all clusters are averaged and then added with a preset second positive constant, to obtain the segment ratio weights of all clusters in the corresponding performance segment;

[0025] The segment ratio weight is multiplied by the change correlation to obtain the contribution of each type of garbage in the performance segment to the performance index.

[0026] Furthermore, the method for predicting the performance index of the mix ratio vector to be predicted includes:

[0027] The influence contribution of each garbage ratio in the performance segment is used as a weight, the weighted average of the differences to be predicted is calculated, and the weighted average is used as the feature similarity between the mix vector to be predicted and the corresponding mix vector; the performance index corresponding to the mix vector with the largest feature similarity is used as the predicted performance index of the mix vector to be predicted.

[0028] Furthermore, the method for obtaining the optimal mix ratio vector includes:

[0029] All mix ratio vectors corresponding to the maximum value of the performance index in the performance curve and the mix ratio vector to be predicted are obtained as the preferred mix ratio vector, and the preferred mix ratio vector corresponding to the lowest sum of all garbage ratios in the preferred mix ratio vector is taken as the optimal mix ratio vector.

[0030] The present invention proposes a system for predicting the optimal mix ratio of construction waste recycling, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any step of the method for predicting the optimal mix ratio of construction waste recycling.

[0031] The present invention has the following beneficial effects:

[0032] The embodiment of the present invention obtains a preset number of mix ratio vectors and corresponding performance indicators, and obtains the target degree of each cluster by clustering comprehensive analysis of the differences in the obtained mix ratio vectors and corresponding performance indicators. The target degree comprehensively reflects the richness of the mix ratio vectors in the cluster by combining the garbage ratio and the performance indicators. Adaptively increasing the mix ratio vectors according to the target degree can improve the fitting effect and efficiency of the performance curve. After fitting the performance curve, considering the nonlinear characteristics of the performance curve, the segmentation points are selected according to the change of the tangent slope of the performance curve and the performance curve is segmented. The segmentation points segment the performance curves with similar change trends, and all garbage ratio change curves are obtained and segmented at the same time, so as to facilitate the analysis of the influence of each type of garbage in each performance segment on the change of the performance indicator. According to the overall change difference of the mix ratio vector in each performance segment and the correlation between the performance segment and the change of the corresponding garbage ratio segment, the influence contribution of each type of garbage ratio vector on the performance indicator is obtained, and then the performance indicator of the mix ratio vector to be predicted is predicted in combination with the influence contribution, so as to avoid the situation where the performance indicator is different only based on the difference of each type of garbage ratio between the mix ratio vectors, and then the optimal mix ratio is accurately obtained. The present invention combines performance indicators to adaptively enrich fitting data to improve the fitting effect and efficiency of the performance curve, and predicts the performance indicators of the mix ratio to be predicted according to the contribution of each type of garbage to the performance indicators to obtain the optimal mix ratio, thereby improving the accuracy of the performance indicator and the optimal mix ratio prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 A method flow chart of a method for predicting the optimal mix ratio of construction waste recycling provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the optimal mix ratio prediction method and system for construction waste recycling proposed by the present invention, its specific implementation method, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0036] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0037] The specific scheme of the method for predicting the optimal mix ratio of construction waste recycling provided by the present invention is described in detail below with reference to the accompanying drawings.

[0038] See also Figure 1 , which shows a method flow chart of a method for predicting the optimal mix ratio of construction waste recycling provided by an embodiment of the present invention, the method comprising the following steps:

[0039] Step S1, obtaining a preset number of mix ratio vectors of construction waste in recycled aggregate and performance indicators corresponding to the mix ratio vectors.

[0040] In order to predict the optimal mix ratio of construction waste recycling to prepare recycled aggregate with the best performance, the embodiment of the present invention first randomly obtains the mix ratio of each type of construction waste in a preset number of recycled aggregates, and obtains the performance index of the recycled aggregate corresponding to each mix ratio through an experimental method, and then analyzes the influence of each type of garbage content in each mix ratio on the performance of the recycled aggregate, so as to accurately predict the performance of the recycled aggregate prepared by the corresponding mix ratio, so as to obtain the optimal mix ratio. The performance indicators include key indicators such as compressive strength, water absorption, adhesion and particle size distribution, which are of great significance for evaluating the quality of recycled aggregate. In one embodiment of the present invention, the preset number is 50, and the implementer can set it according to the actual situation.

[0041] In one embodiment of the present invention, in order to facilitate the analysis of mix data, the proportion content of all types of garbage in each mix is ​​constructed into a mix vector, and each vector element in the mix vector represents the proportion content of different construction waste, where the construction waste includes masonry materials, ceramic materials, and wood materials, and the performance indicators of the recycled aggregate corresponding to each mix vector are obtained according to the experimental method.

[0042] In other embodiments of the present invention, the types of construction waste and vector dimensions in the mix ratio vector can be selected and set according to the actual needs of the implementer.

[0043] Step S2, obtaining all clusters of mix ratio vectors; in each cluster, obtaining the distribution weight of the corresponding mix ratio vector in the belonging cluster according to the difference between all mix ratio vectors and each garbage ratio in the cluster center vector; obtaining the target degree of the corresponding cluster according to the distribution weight of the mix ratio vector in each cluster and the corresponding performance index; adding a new mix ratio vector to each cluster according to the target degree, and fitting the performance curve according to all the mix ratio vectors in the cluster and the corresponding performance index.

[0044] The present invention obtains the influence of the recycled aggregate mix ratio on the performance by fitting the curve of the performance index changing with the corresponding mix ratio, and then predicts and obtains the optimal mix ratio. In order to reduce the number of experiments to obtain the performance index of the recycled aggregate corresponding to each mix ratio, so as to improve the efficiency of obtaining the optimal mix ratio, the present invention first analyzes the approximate distribution of the few mix ratios obtained and the known recycled aggregate performance indexes, and then adaptively adds the mix ratios and performance indexes as fitting data according to the distribution, so as to fit a more accurate performance curve, which is convenient for subsequent prediction of the optimal mix ratio; and considering that the clustering algorithm can more intuitively reflect the differences between the data and the approximate distribution of the data, it is convenient for subsequent adaptive enrichment of the fitting data to improve the fitting effect of the performance curve. Therefore, the embodiment of the present invention first obtains all clusters of the mix ratio vector.

[0045] In one embodiment of the present invention, a DBSCAN algorithm is specifically used to cluster a preset number of mix ratio vectors to obtain all clusters and the cluster center vector corresponding to each cluster. The size of each cluster reflects the number of similar mix ratio vectors and the approximate distribution of all mix ratio vectors. The DBSCAN algorithm is a prior art and will not be described in detail here.

[0046] Taking into account that the difference between the cluster sample data and the cluster center data in each cluster not only reflects the similarity between the data, but also reflects the overall distribution of all samples in the cluster, in order to adaptively enrich the fitting data used to fit the performance curve, in each cluster, the embodiment of the present invention obtains the distribution weight of the corresponding mix ratio vector in the belonging cluster according to the difference between all mix ratio vectors and each garbage ratio in the cluster center vector.

[0047] Preferably, in one embodiment of the present invention, the method for obtaining the weights includes: in each cluster, obtaining the absolute value of the first ratio difference between each mix ratio vector and the same category of garbage in the cluster center vector; in each mix ratio vector, screening out the difference garbage whose absolute value of the first ratio difference is greater than a preset threshold, and obtaining the quantity ratio of the number of difference garbage in the corresponding mix ratio vector to the number of all garbage; obtaining the mean of the absolute value of the ratio difference between each mix ratio vector and all the same category of garbage in the corresponding cluster center vector; multiplying the quantity ratio by the mean of the absolute value of the corresponding ratio difference to obtain the distribution weight of the corresponding mix ratio vector. The calculation formula of the distribution weight is:

[0048]

[0049] in, is the nth cluster in the lth cluster l The distribution weight of the mix ratio vector, n l is the serial number of the matching ratio vector in the lth cluster, U is the number of all garbage in the matching ratio vector, is the nth cluster in the lth clusterl The number of difference garbage whose absolute value of the difference between the first ratio in the matching ratio vector and the cluster center vector is greater than the preset threshold, is the nth cluster in the lth cluster l The absolute value mean of the difference between the ratio of the matching ratio vector and the corresponding cluster center vector of all garbage of the same category. In one embodiment of the present invention, the preset threshold is 0.2, and the implementer can set it according to the actual situation.

[0050] To further help understand how to obtain differential garbage, the present invention uses an example to illustrate how to obtain differential garbage. Specifically, first, obtain the nth cluster in the lth cluster. l The absolute value of the difference between the first ratio in the ratio vector and the cluster center vector, where the nth l The mix ratio vector is (a nl , b nl , c nl ), the corresponding cluster center vector o l for (a ol , b ol , c ol ), a, b, c are the symbols of each type of garbage in the mix ratio vector, then the nth l The absolute value of the first ratio difference between the ratio vector and the cluster center vector of the a-th type of garbage is |a nl ―a ol |, similarly obtain |b nl ―b ox |with|c nl ―c ol |, in the nth l In a mix ratio vector, the type of garbage corresponding to the absolute value of the first mix ratio difference of each type of garbage being greater than a preset threshold is regarded as the difference garbage in the mix ratio vector.

[0051] In the calculation formula of distribution weight, the quantity ratio of the number of difference garbage in the mix ratio vector to the number of all garbage reflects the proportion of garbage types with similar proportions in the mix ratio vector and all mix ratio vectors in the corresponding cluster. The fewer the number of garbage types with similar proportions, the lower the similarity of all mix ratio vectors in the cluster, and the worse the effect of the subsequent fitting performance curve. Multiplying the quantity ratio by the mean of the absolute value of the corresponding ratio difference reflects the average difference level of the proportion content of each type of garbage. The larger the average difference, the lower the similarity of all mix ratio vectors in the cluster. When fitting the performance curve subsequently, more abundant fitting data needs to be added to the cluster.

[0052] Since the distribution weight only reflects the difference in the proportion of various types of garbage in each mix ratio vector in the cluster, and the performance of recycled aggregates with similar mix ratios may vary greatly, adding fitting data based on the difference in garbage content alone cannot obtain more targeted and rich fitting data; therefore, when analyzing the fitting performance curve, the distribution difference of performance indicators in the cluster is also combined to improve the fitting effect of the performance curve. Based on this, the embodiment of the present invention obtains the target degree of the corresponding cluster according to the distribution weight of the mix ratio vector in each cluster and the corresponding performance indicator.

[0053] Preferably, in one embodiment of the present invention, the method for obtaining the target degree includes, in each cluster, obtaining the weighted performance index of the performance index of the corresponding mix ratio vector according to the distribution weight of each mix ratio vector, and obtaining the coefficient of variation of the weighted performance index according to the coefficient of variation formula, wherein the coefficient of variation is the target degree of the corresponding cluster. Since the coefficient of variation formula is a prior art, it will not be described in detail here, and only a calculation formula for obtaining the target degree according to the coefficient of variation formula is briefly described in one embodiment of the present invention. The calculation formula for the target degree includes:

[0054]

[0055] Among them, α l is the target degree of the lth cluster, N l is the number of matching ratio vectors in the lth cluster, n l is the serial number of the matching ratio vector in the lth cluster, is the nth cluster in the lth cluster l The distribution weights of the mix ratio vectors, is the nth cluster in the lth cluster l The performance index of a mix ratio vector, norm() represents a standard normalization function, sigmoid() represents a normalization function, and in other embodiments of the present invention, the implementer may adopt other normalization methods.

[0056] In the calculation formula of the target degree, the coefficient of variation is used to obtain the nth l The distribution of the weighted average performance indicators of the mix ratio vectors reflects the overall difference level of all performance indicators and various garbage ratio contents in the cluster. The larger the target degree, the lower the richness of the mix ratio data in the corresponding cluster, and the worse the subsequent fitting effect. Therefore, more abundant data should be added to the cluster to improve the fitting effect of the performance curve.

[0057] The target degree of each cluster reflects the richness of the mix ratio vector in the corresponding cluster based on the garbage ratio and performance index. Adding mix ratio vectors to the cluster according to the target degree can more accurately improve the data richness of clusters with relatively discrete garbage ratio content and performance index distribution and large differences, thereby fitting the performance curve more accurately. Therefore, the embodiment of the present invention first adds mix ratio vectors to the corresponding cluster according to the target degree, and obtains the performance curve according to the mix ratio vector fitting.

[0058] Preferably, in one embodiment of the present invention, considering that the larger the target degree is, the lower the richness of the data in the corresponding cluster is, and the more appropriate data should be added to the cluster; considering that while adding the mix ratio data to enrich the data in the cluster, adding the mix ratio vector with too large difference to the cluster will affect the quality of the cluster, thereby causing the target degree of the cluster to be abnormal and reducing the fitting effect. Based on this, the method of adding the mix ratio vector to the cluster includes multiplying the target degree of each cluster by the preset second number and rounding up to obtain the increase number of mix ratio vectors added to the corresponding cluster during each increase process; in each cluster, obtaining the ratio standard deviation of each type of garbage ratio in all the mix ratio vectors; screening the optional mix ratio vectors whose ratios of all types of garbage are less than the corresponding ratio standard deviation in the preset second number of new mix ratio vectors, and adding the corresponding increase number of optional mix ratio vectors to each cluster. In one embodiment of the present invention, the preset second number is 50, and the implementer can set it according to the actual situation.

[0059] It should be noted that when adding a new mix vector to a cluster, the performance index corresponding to the mix vector must be obtained at the same time in order to enrich the fitting data and improve the fitting effect of the subsequent performance curve.

[0060] After obtaining the target degree of each cluster, a corresponding number of new mix vectors are added to the cluster in a more targeted manner according to the target degree, which can improve the fitting efficiency and effect; then, a performance curve is fitted according to all the mix vectors and corresponding performance indicators in the cluster, so as to subsequently predict the optimal mix. In one embodiment of the present invention, the performance curve is fitted with the mix vector as the horizontal coordinate and the performance indicator corresponding to the mix vector as the horizontal coordinate.

[0061] Preferably, in one embodiment of the present invention, the method for obtaining the performance curve includes: sorting all the mix ratio vectors in ascending order according to the mean value of the ratio of all garbage ratios in the mix ratio vector, fitting the mix ratio vector and the corresponding performance index by using the method of multivariate linear regression analysis, and obtaining the initial performance curve; adding optional mix ratio vectors to each cluster according to the addition method, updating the initial performance curve, and calculating the similarity of the initial performance curve before and after the update; continuously adding new optional mix ratio vectors to each cluster until the preset cutoff condition is met, and the initial performance curve after the last update is used as the performance curve. In the embodiment of the present invention, the mix ratio vectors are sorted in ascending order according to the mean value of the ratio, and the multivariate nonlinear regression analysis method can accurately describe the relationship between the performance and the ratio of various garbage; by continuously adding suitable optional mix ratio vectors to the cluster, the fitting effect is improved; considering that continuously adding mix ratio vectors to the cluster can provide a large amount of rich fitting data, but cannot reduce the number of experiments to improve the fitting efficiency, it is necessary to set the corresponding cutoff condition; since there is no obvious difference in the fitting curve before and after adding the mix ratio vector, the curve characteristics will not change significantly if the mix ratio vector is continued to be added, then it is considered that the ideal fitting effect has been achieved. Due to the inconsistent length of the initial performance curves before and after the update, the DTW algorithm is specifically used in the embodiment of the present invention to obtain the similarity of the two initial performance curves before and after the update, and the obtained similarity is normalized using an exponential function. When the similarity is less than the preset similarity threshold, the addition of the mix ratio vector is stopped and the initial performance curve is stopped. In one embodiment of the present invention, the preset similarity threshold is set to 0.65, and the implementer can set it according to the actual situation. Multiple nonlinear regression analysis and DTW algorithm are prior arts and will not be repeated here.

[0062] Step S3, according to the change of the tangent slope of the performance curve, all segmentation points are screened and all performance segments are obtained; the garbage ratio curve of each type of garbage is obtained, and the dimension of the mix ratio vector of each garbage ratio curve and the performance curve is the same; according to the points in the garbage ratio curve corresponding to the segmentation points, all garbage ratio segments of each garbage ratio curve are obtained, and the change correlation between each performance segment and the corresponding garbage ratio segment is obtained; in each performance segment, according to the overall change difference of all mix ratio vectors and the corresponding change correlation, the contribution of each type of garbage in the corresponding performance segment to the performance index is obtained.

[0063] The performance indicators of recycled aggregates and various construction waste mix ratios do not show a simple linear relationship. There may be huge differences in the performance indicators corresponding to similar mix ratios, and there is a trend that the performance indicators first increase and then decrease with the mix ratio vector. In order to analyze and obtain the changes in performance indicators, the embodiment of the present invention first screens all segmentation points and obtains all performance segments according to the changes in the tangent slope of the performance curve, and then segments the performance curve according to the changes in the performance curve, and then obtains the impact of each garbage ratio change on the performance indicator change within the segments with similar change trends.

[0064] Preferably, in one embodiment of the present invention, considering that the change in the slope of adjacent data points in the curve reflects the change trend of the curve, when the change trend of the curve segment before and after a certain data point is greatly different, the point is considered to be the inflection point of the curve, and segmentation can be performed at the inflection point; based on this, the method for obtaining the segmentation point includes, in the performance curve, obtaining the tangent slope of all data points corresponding to the mix vector; calculating the tangent slope ratio of the data point corresponding to the mix vector and the data point corresponding to the next adjacent mix vector, obtaining the absolute value of the difference between the tangent slope ratio and the preset first normal number, and then normalizing to obtain the segmentation point reference value; the data point corresponding to the mix vector whose segmentation point reference value is greater than the preset segmentation point reference value threshold is used as the segmentation point. The calculation formula for the segmentation point reference value is:

[0065]

[0066] In the formula, γ nl represents the segment point reference value of the data point corresponding to the nlth mix ratio vector, k nl represents the tangent slope of the nlth mix ratio vector corresponding to the data point, k nl+1 It represents the tangent slope of the data point corresponding to the nl+1th mix ratio vector, exp() is an exponential function with the natural constant e as the base, and b is a preset first normal number. In one embodiment of the present invention, the preset first normal number b is 1, and the implementer can set it according to actual needs.

[0067] In the calculation formula of the segment point reference value, the tangent slope ratio is used to represent the relative change trend of the curves of the data points corresponding to the adjacent mix ratio vectors. After subtracting the preset first normal number, the change difference is more intuitively reflected. The absolute value of the difference between the tangent slope ratio and the preset first normal number is negatively correlated and mapped to an exponential function for normalization to facilitate subsequent application.

[0068] According to the above segmentation point reference value calculation formula, the segmentation point reference values ​​of the data points corresponding to all the mix ratio vector data in the performance curve are obtained, and the data points whose segmentation point reference values ​​are greater than the preset segmentation reference values ​​are used as segmentation points. In one embodiment of the present invention, the preset segmentation point reference value is 0.45. When the segmentation point reference value of a data point is greater than 0.45, it is considered that the curve changes drastically, and the data points corresponding to the mix ratio vector are used as segmentation points to segment the performance curve.

[0069] In order to analyze the influence of the garbage ratio of each type of garbage on the performance, after fitting the performance curve reflecting the relationship between the performance and the mix ratio vector, the embodiment of the present invention further obtains the garbage ratio curve of each type of garbage according to the fitting order of all the mix ratio vectors in the performance curve and the garbage ratio of each type of garbage in the mix ratio vector, and the dimensions of each type of garbage ratio curve and the mix ratio vector of the performance curve are the same. In one embodiment of the present invention, each type of garbage ratio in the mix ratio vector in the performance curve is used as the ordinate, and the sorting sequence number of the corresponding data point of the mix ratio vector is used as the abscissa to construct the garbage ratio curve of the corresponding type of garbage.

[0070] Taking into account the nonlinearity of the performance curve, and taking into account that the dimensions of the mix ratio vector of each type of garbage ratio curve and the performance curve are the same, in order to more intuitively and accurately analyze the impact of each type of garbage in the mix ratio vector on the performance, the embodiment of the present invention obtains all garbage ratio segments of each type of garbage ratio curve according to the segmentation point and the corresponding point in the garbage ratio curve, and then obtains the impact of each type of garbage on the performance index according to the change of the ratio of each type of garbage in each segment and the change of the corresponding performance segment; after obtaining all garbage ratio segments of all types of garbage, obtain the change correlation between each performance segment and the corresponding garbage ratio segment of each type of garbage, and further more intuitively compare the impact of a certain type of garbage on the performance change. In one embodiment of the present invention, the Pearson correlation coefficient between the performance segment and the corresponding comparison ratio segment of each type of garbage is obtained, and the absolute value of the Pearson correlation coefficient is used as the change correlation between each performance segment and the corresponding garbage ratio segment of each type of garbage. The calculation of the Pearson correlation coefficient is a prior art and will not be repeated here.

[0071] In other embodiments of the present invention, the change correlation may also be obtained through other technical means such as Chebyshev distance, distance correlation analysis, etc., which are not limited here.

[0072] In each performance segment, the influence of a certain type of garbage on the performance is obtained only based on the change of a certain type of garbage ratio, which is susceptible to the influence of other types of garbage; considering that when the overall change of all mix ratio vectors is small, if the performance index changes with the change of a certain type of garbage ratio in the mix ratio vector, it means that the influence of this type of garbage ratio on the performance index is large, anyway, the influence is small; when the overall mix ratio vector changes little, the correlation between the change of a certain type of garbage in all mix ratio vectors and the change of performance is analyzed separately, so as to focus on analyzing the influence of each type of garbage on the performance index. Therefore, in each performance segment, according to the overall change difference of all mix ratio vectors and the corresponding change correlation, the embodiment of the present invention obtains the contribution of each type of garbage to the performance index in the corresponding performance segment.

[0073] Preferably, in one embodiment of the present invention, considering the similarity between the mix ratio vectors in all clusters and the corresponding cluster center vector can help analyze the overall changes in the ratios of all types of garbage, and the correlation between the performance segment and the corresponding ratio segment of each type of garbage can focus on the impact of each type of garbage ratio change on the performance index. Based on this, the method for obtaining the impact contribution of each type of garbage ratio in the performance segment on the performance index includes: in each performance segment, obtaining the cosine similarity between all mix ratio vectors in each cluster and the cluster center vector; in each cluster of the corresponding performance segment, obtaining the ratio similarity ratio of the cosine similarity of the mix ratio vector and the sum of the cosine similarities of all mix ratio vectors, averaging the ratio similarity ratios of all clusters and adding a preset second normal constant to obtain the segment ratio weights of all clusters in the corresponding performance segment; multiplying the segment ratio weight by the change correlation to obtain the impact contribution of each type of garbage ratio in the performance segment on the performance index. The calculation formula for the impact contribution is expressed as:

[0074]

[0075] in, is the impact contribution of the ath garbage in the jth performance segment, M j is the number of data points corresponding to the mix ratio vector in the jth performance segment, m j is the serial number of the data point corresponding to the mix ratio vector in the jth performance segment, is the mth performance segment of the jth performance segment j The cosine similarity between the matching ratio vector and the corresponding cluster center vector, S is the mth performance segment of the jth performance segment. j The total number of match ratio vectors in the cluster to which the match ratio vector belongs, d s is the mth performance segment of the jth performance segment j The cosine similarity between the sth matching ratio vector in the cluster to which the corresponding data point belongs and the corresponding cluster center vector, R aIt represents the correlation between the j-th performance segment and the j-th garbage distribution segment of the a-th garbage.

[0076] In the calculation formula of the impact contribution, the segmented mix weight reflects the overall change of all mix ratio vectors in the performance segment; the change correlation reflects the linear correlation between the performance index and various garbage ratios in the corresponding segment. The greater the change correlation, the greater the linear correlation between the change of the performance segment and the change of the garbage ratio segment of the a-th garbage, that is, the greater the possibility that the performance index in the performance segment changes with the change of the a-th garbage ratio. If the segmented mix weight is larger at this time, it means that the overall change difference of all mix ratio vectors in the performance segment is smaller, and the change of other types of garbage ratios has less interference with the change of the index affecting the analysis of the change of the a-th garbage ratio, then the corresponding a-th garbage ratio has a greater impact contribution on the performance index; if the segmented mix weight is smaller at this time, it means that the overall change difference of all mix ratio vectors in the performance segment is larger, and the change of other types of garbage ratios is more likely to interfere with the change of the index affecting the analysis of the change of the i-th garbage ratio, then the impact contribution needs to be reduced.

[0077] Step S4, obtaining the difference to be predicted of each type of garbage between all mix ratio vectors in the performance curve and the mix ratio vector to be predicted; predicting the performance index of the mix ratio vector to be predicted according to the impact contribution of each type of garbage and the difference to be predicted; screening the preferred mix ratio vector in the performance curve to obtain the optimal mix ratio vector.

[0078] The optimal mix vector of recycled aggregate should have the best performance. In order to predict the optimal mix vector, it is necessary to obtain the performance indicators of all the mix vectors to be tested, and then screen out the optimal mix according to the performance indicators. In order to improve the fitting efficiency, the present invention predicts the performance indicators of the mix vector to be predicted according to the performance curve; considering that the performance indicators corresponding to similar mix vectors are more likely, the embodiment of the present invention first obtains the difference to be predicted for each type of garbage between all the mix vectors in the performance curve and the mix vector to be predicted. The method for obtaining the difference to be predicted for each type of garbage is the same as the method for obtaining the first ratio difference of each type of garbage ratio in step S2, which will not be repeated here.

[0079] Considering the characteristics of the performance curve, the performance indicators corresponding to similar mix ratio vectors may be greatly different. The accuracy of the performance indicators obtained only based on the similarity between mix ratio vectors is not high, and the impact contribution reflects the impact of each garbage ratio on the performance indicator in the corresponding performance segment. Therefore, the embodiment of the present invention predicts the performance indicators of the mix ratio vector to be predicted based on the impact contribution of each garbage and the difference to be predicted.

[0080] Preferably, in one embodiment of the present invention, considering that the difference in the same type of garbage ratio between mix vectors reflects the similarity of the mix, the greater the possibility that similar mix vectors correspond to similar performance indicators; and considering the nonlinear characteristics of the performance curve and the different degrees of influence of each garbage ratio on the performance, there are certain differences in the performance indicators predicted by obtaining the similarity between mix vectors based only on the difference in the same type of garbage ratio between mix vectors. Based on this, the prediction method of the performance indicator of the mix vector to be predicted includes taking the influence contribution of each garbage ratio in the performance segment as a weight, calculating the weighted average of the differences to be predicted, and taking the weighted average as the characteristic similarity between the mix vector to be predicted and the corresponding mix vector; taking the performance indicator corresponding to the mix vector with the largest characteristic similarity as the predicted performance indicator of the mix vector to be predicted.

[0081] Since one performance index may correspond to multiple mix vectors, after obtaining the performance indexes of all the mix vectors to be predicted, the embodiment of the present invention selects all the preferred mix vectors to obtain the optimal mix vector.

[0082] Preferably, in one embodiment of the present invention, the method for obtaining the optimal mix ratio vector includes: obtaining all mix ratio vectors corresponding to the maximum values ​​of performance indicators in the performance curve and the mix ratio vector to be predicted as preferred mix ratio vectors, and taking the preferred mix ratio vector corresponding to the lowest sum of all garbage ratios in the preferred mix ratio vector as the optimal mix ratio vector.

[0083] Taking into account the cost factors while taking into account the best performance, the preferred mix vector with the lowest content of various types of garbage in the mix vector is used as the optimal mix vector to assist professional technicians in preparing recycled aggregates.

[0084] In summary, the embodiment of the present invention obtains a preset number of mix vectors and corresponding performance indicators, and obtains the target degree by clustering and comprehensively analyzing the differences between the obtained mix vectors and corresponding performance indicators; adaptively adds new mix vectors according to the target degree to fit the performance curve; segments the performance curve according to the change of the tangent slope of the performance curve, obtains all garbage mix change curves and segments them; obtains the influence contribution of each type of garbage on the performance according to the overall change difference of the mix vector in each performance segment and the correlation between the performance segment and the corresponding garbage mix segment change, and then predicts the performance indicators of the mix vector to be predicted in combination with the influence contribution, so as to obtain the optimal mix. The present invention combines performance indicators to adaptively enrich fitting data to improve the fitting effect and efficiency of the performance curve, and predicts the performance indicators of the mix to be predicted according to the influence contribution of each type of garbage on the performance indicators, thereby improving the prediction accuracy of the performance indicators and the accuracy of the optimal mix.

[0085] The present invention also proposes a system for predicting the optimal mix ratio of construction waste recycling, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any step of a method for predicting the optimal mix ratio of construction waste recycling.

[0086] A construction waste regeneration performance curve fitting method and system embodiment:

[0087] With the acceleration of urbanization, the amount of construction waste has increased dramatically. Construction waste recycling technology can put recycled aggregates made from construction waste into construction, reducing resource waste and environmental pollution. However, due to the wide variety of construction waste, the properties of recycled aggregates prepared from waste with different mix ratios vary greatly. At present, technicians usually obtain mix ratios randomly and fit performance curves through continuous experiments. The randomness of the mix ratio and the cumbersome and time-consuming experimental fitting process result in low fitting effect and efficiency. The present invention provides a construction waste recycling performance curve fitting method, the method comprising:

[0088] Step S1, obtaining a preset number of mix ratio vectors of construction waste in recycled aggregate and performance indicators corresponding to the mix ratio vectors.

[0089] Step S2, obtaining all clusters of mix ratio vectors; in each cluster, obtaining the distribution weight of the corresponding mix ratio vector in the belonging cluster according to the difference between all mix ratio vectors and each garbage ratio in the cluster center vector; obtaining the target degree of the corresponding cluster according to the distribution weight of the mix ratio vector in each cluster and the corresponding performance index; adding a new mix ratio vector to each cluster according to the target degree, and fitting the performance curve according to all the mix ratio vectors in the cluster and the corresponding performance index.

[0090] Among them, step S1 and step S2 have been described in detail in the above-mentioned embodiment of the method for predicting the optimal mix ratio of construction waste recycling, and will not be repeated here.

[0091] The present invention also proposes a construction waste regeneration performance curve fitting system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any step of a construction waste regeneration performance curve fitting method.

[0092] The embodiment of the present invention obtains a preset number of mix ratio vectors and corresponding performance indicators, and obtains a target degree by clustering and comprehensively analyzing the differences between the obtained mix ratio vectors and corresponding performance indicators; the target degree comprehensively reflects the richness of the mix ratio vectors in the cluster by combining the garbage ratio and the performance indicators, and adaptively increasing the mix ratio vectors according to the target degree can improve the fitting effect and efficiency of the performance curve. The present invention adaptively enriches the fitting data by clustering and comprehensively analyzing the differences between the mix ratio vectors and the corresponding performance indicators, thereby improving the fitting effect and efficiency of the performance curve.

[0093] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0094] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for predicting the optimal mix ratio of construction waste recycling, characterized in that: The method comprises: Obtaining a preset number of mix ratio vectors of construction waste in recycled aggregates and performance indicators corresponding to the mix ratio vectors; Obtain all clusters of the mix ratio vectors; in each cluster, according to the difference between all the mix ratio vectors and each type of garbage ratio in the cluster center vector, obtain the distribution weight of the corresponding mix ratio vector in the cluster; according to the distribution weight of the mix ratio vector in each cluster and the corresponding performance index, obtain the target degree of the corresponding cluster; according to the target degree, add a new mix ratio vector to each cluster, and fit the performance curve according to all the mix ratio vectors in the cluster and the corresponding performance index; According to the change of the tangent slope of the performance curve, all segmentation points are screened and all performance segments are obtained; the garbage ratio curve of each type of garbage is obtained, and the dimension of the mix ratio vector of each type of garbage ratio curve and the performance curve is the same; according to the points in the garbage ratio curve corresponding to the segmentation points, all garbage ratio segments of each type of garbage ratio curve are obtained, and the change correlation between each performance segment and the corresponding garbage ratio segment is obtained; in each performance segment, according to the overall change difference of all mix ratio vectors and the corresponding change correlation, the influence contribution of each type of garbage in the corresponding performance segment to the performance index is obtained; Obtain the difference to be predicted of each type of garbage between all mix ratio vectors in the performance curve and the mix ratio vector to be predicted; predict the performance index of the mix ratio vector to be predicted according to the impact contribution of each type of garbage and the difference to be predicted; and select the preferred mix ratio vector in the performance curve to obtain the optimal mix ratio vector.

2. The method for predicting the optimal mix ratio of construction waste recycling according to claim 1 is characterized in that: The method for obtaining the distribution weight includes: In each of the clusters, the absolute value of the first ratio difference between each of the matching ratio vectors and the same category of garbage in the cluster center vector is obtained; in each of the matching ratio vectors, the difference garbage whose absolute value of the first ratio difference is greater than a preset threshold is screened out, and the ratio of the number of the difference garbage in the matching ratio vector to the number of all garbage is obtained; Obtain the absolute value mean of the ratio difference between each of the matching ratio vectors and all the same category garbage ratios in the corresponding cluster center vector; multiply the quantity ratio by the corresponding absolute value mean of the ratio difference to obtain the distribution weight corresponding to the matching ratio vector.

3. The method for predicting the optimal mix ratio of construction waste recycling according to claim 1 is characterized in that: The method for obtaining the target degree includes: In each of the clusters, the distribution weight of each of the mix ratio vectors is used as the weight of the corresponding performance indicator to obtain the weighted performance indicator of all the mix ratio vectors in the cluster; the coefficient of variation of the weighted performance indicator is obtained according to the coefficient of variation formula, and the coefficient of variation is the target degree of the cluster.

4. The method for predicting the optimal mix ratio of construction waste recycling according to claim 1 is characterized in that: The method of adding a new matching ratio vector to each of the clusters includes: The target degree of each cluster is multiplied by a preset second number and then rounded up to obtain the additional number of the mix ratio vectors added to the corresponding cluster in each increase process; in each cluster, the standard deviation of the ratio of each type of garbage in all the mix ratio vectors is obtained; among the preset second number of new mix ratio vectors, the optional mix ratio vectors whose ratios of all types of garbage are less than the corresponding mix ratio standard deviation are screened, and the optional mix ratio vectors corresponding to the increased number are added to each cluster.

5. The method for predicting the optimal mix ratio of construction waste recycling according to claim 4 is characterized in that: The method for obtaining the performance curve includes: All the mix ratio vectors are sorted in ascending order according to the ratio means of all garbage ratios in the mix ratio vectors, and the mix ratio vectors and the corresponding performance indicators are fitted using a multivariate nonlinear regression analysis method to obtain an initial performance curve; The optional mix ratio vector is added to each of the clusters according to the adding method, the initial performance curve is updated, and the similarity of the initial performance curve before and after the update is calculated; new optional mix ratio vectors are continuously added to each cluster until a preset cutoff condition is met, and the initial performance curve after the last update is used as the performance curve.

6. The method for predicting the optimal mix ratio of construction waste recycling according to claim 1 is characterized in that: The method for obtaining the segmentation points includes: In the performance curve, the tangent slopes of all data points corresponding to the mix ratio vectors are obtained; the tangent slope ratio of the data point corresponding to the mix ratio vector and the data point corresponding to the next adjacent mix ratio vector is calculated, and the absolute value of the difference between the tangent slope ratio and a preset first normal constant is obtained and normalized to obtain a segmentation point reference value; The data point corresponding to the mix ratio vector whose segmentation point reference value is greater than the preset segmentation reference value threshold is used as the segmentation point.

7. The method for predicting the optimal mix ratio of construction waste recycling according to claim 1 is characterized in that: The method for obtaining the influence contribution of each type of garbage in the performance segment on the performance index includes: In each of the performance segments, the cosine similarities between all the mix ratio vectors and the cluster center vector in each cluster are obtained; in each cluster corresponding to the performance segment, the ratio similarity ratio of the cosine similarity of the mix ratio vector and the sum of the cosine similarities of all the mix ratio vectors is obtained, the ratio similarity ratios of all clusters are averaged and then added with a preset second positive constant, to obtain the segment ratio weights of all clusters in the corresponding performance segment; The segment ratio weight is multiplied by the change correlation to obtain the contribution of each type of garbage in the performance segment to the performance index.

8. The method for predicting the optimal mix ratio of construction waste recycling according to claim 1 is characterized in that: The method for predicting the performance index of the mix ratio vector to be predicted includes: The influence contribution of each garbage ratio in the performance segment is used as a weight, the weighted average of the differences to be predicted is calculated, and the weighted average is used as the feature similarity between the mix vector to be predicted and the corresponding mix vector; the performance index corresponding to the mix vector with the largest feature similarity is used as the predicted performance index of the mix vector to be predicted.

9. The method for predicting the optimal mix ratio of construction waste recycling according to claim 1 is characterized in that: The method for obtaining the optimal mix ratio vector includes: All mix ratio vectors corresponding to the maximum value of the performance index in the performance curve and the mix ratio vector to be predicted are obtained as the preferred mix ratio vector, and the preferred mix ratio vector corresponding to the lowest sum of all garbage ratios in the preferred mix ratio vector is taken as the optimal mix ratio vector.

10. A construction waste recycling optimal mix ratio prediction system, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for predicting the optimal mix ratio of construction waste recycling as described in any one of claims 1 to 9 are implemented.

Images