A method, equipment and system for recycling waste refractory materials

Abstract

This application relates to the field of refractory material processing technology, specifically to a method, equipment, and system for the reuse of waste refractory materials. The method includes: collecting compositional parameter data and flotation process parameter data of the waste refractory materials; extracting two potential factors from the material composition and flotation process through factor analysis; and constructing a separation efficiency and energy efficiency structure by combining grey relational analysis with the correlation between the actual value and target value of each parameter at any given time of data collection, thereby assessing the process stability at the current time of data collection and identifying key parameters for control. This application aims to improve separation efficiency and process stability, significantly optimize resource utilization, and support efficient and stable separation under complex operating conditions.

Description

Technical Field

[0001] This application relates to the field of refractory material processing technology, specifically to a method, equipment, and system for the reuse of waste refractory materials. Background Technology

[0002] Traditional high-temperature industries (such as steel, cement, and glass) generate millions of tons of waste refractory materials annually. In the past, these materials were mostly disposed of through landfill, which not only occupied land but also caused soil heavy metal pollution and resource waste due to the presence of non-degradable components such as alumina and magnesium oxide. Against this backdrop, waste refractory material recycling technology has developed rapidly. In the early days, it was mainly used as a low-end auxiliary material after simple crushing. Now, through breakthroughs in key technologies such as component analysis, intelligent sorting, and gradient purification, high-value-added recycling has been achieved. From an environmental perspective, each ton of recycled refractory material can reduce carbon emissions and significantly reduce the ecological burden. Currently, domestic and foreign companies are building a "recycling-sorting-regeneration-application" industrial chain. In the future, with breakthroughs in technologies such as intelligent sorting and in-situ remediation, the recycling of waste refractory materials will become a key fulcrum for achieving the green and low-carbon transformation of industry.

[0003] In the field of waste refractory material sorting, the core technical problem of existing flotation processes lies in the lag in dynamic parameter control caused by the nonlinear coupling of multidimensional process data, which leads to unstable sorting efficiency. Taking magnesia-carbon bricks as an example, due to the small density difference between magnesium oxide (MgO) and graphite (C) in the waste material and the difficulty in peeling off the surface oxide passivation layer, traditional flotation processes rely on static parameter settings and empirical control, making it difficult to analyze the strong nonlinear correlation between multidimensional process parameters such as pH value, bubble size, and reagent flow rate and material composition (such as MgO / C concentration), resulting in deviations in sorting efficiency prediction. At the same time, dynamic fluctuations in the flotation liquid environment (such as uneven bubble distribution and random peeling of surface oxides) and external interference (fluctuations in raw material composition) further exacerbate process instability. Existing technologies lack the ability to couple historical fluctuation risks with future trends, and it is difficult to quantify process anomalies by relying solely on fixed threshold alarms, leading to lag in parameter adjustment. Ultimately, due to the redundancy of multidimensional data and the disconnect between control, the execution units cannot form a synergy, resulting in low separation efficiency of the magnesium oxide and graphite interface and deterioration of the uniformity of recycled materials. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method, equipment, and system for the reuse of waste refractory materials. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of this application provide a method for reusing waste refractory materials, the method comprising the following steps:

[0006] Collect compositional parameter data and flotation process parameter data of waste refractory materials, and perform noise reduction and normalization preprocessing;

[0007] Two potential factors in the material composition and flotation process are extracted by factor analysis. The correlation between the actual value and the target value of each parameter at any collection time is analyzed by grey relational analysis. The two potential factors and the correlation of all parameters at any collection time are combined to construct the separation efficiency and energy efficiency.

[0008] By combining all the data collected from all parameters, the XGBoost regression model is used to predict the sorting efficiency and energy efficiency, and the coefficient of variation of sorting efficiency and energy efficiency at all collection times is analyzed by sliding window analysis to evaluate the process stability at the current collection time.

[0009] Based on the regression equation output by the XGBoost regression model, the coefficient of each parameter in the regression equation is determined. The parameter corresponding to the maximum absolute value of the coefficient is taken as the key parameter, and the process stability at the current acquisition time is used to adjust the key parameter.

[0010] Preferably, during the data acquisition process, X-ray fluorescence spectrometry (XRF) is used to collect the mass fraction data of magnesium oxide (MgO) and graphite (C), and pH sensor, nanobubble size detector and flow meter are used to collect the pH value of flotation solution, bubble size and reagent dosage data.

[0011] Preferably, the two potential factors extracted through factor analysis include a material composition purity factor extracted from composition parameter data and a flotation process stability factor extracted from flotation process parameter data.

[0012] Preferably, the method for constructing the sorting efficiency and energy efficiency is as follows:

[0013] Calculate the product of the two potential factors; calculate the mean of the correlation of all parameters at any given time; and use the ratio of the product to the mean as the sorting efficiency.

[0014] Preferably, the method for predicting the sorting efficiency and energy efficiency prediction values ​​using the XGBoost regression model, combining all data collected from all types of parameters, is as follows:

[0015] The sorting efficiency and energy efficiency sequences calculated at all collection times and the data sequences of all parameters are used as independent variables. At each data collection time, the absolute value of the difference between each collected parameter and the target value of the corresponding parameter is calculated and recorded as the first difference. The mean of the first differences of all parameters is calculated and used as the dependent variable at the corresponding data collection time. The XGBoost regression model is used to predict and output the predicted value Y of sorting efficiency.

[0016] Preferably, the method for evaluating the process stability at the current acquisition time is as follows:

[0017]

[0018] Where B represents the process stability at the current acquisition time; Y represents the predicted sorting efficiency; J represents the number of windows in the sequence composed of the sorting efficiency energy efficiency A calculated over all acquisition times; t j The time interval between the rightmost acquisition time of the j-th window and the current acquisition time; L is the sequence length composed of the sorting efficiency energy efficiency A calculated from all acquisition times; e is the natural constant; Cv j Let be the coefficient of variation for the j-th window.

[0019] Preferably, the method for adjusting key parameters is as follows:

[0020] When the key parameter is the pH value of the flotation solution, acidic / alkaline reagents are injected by starting the automatic dosing pump to adjust the pH to the target value of the key parameter in real time.

[0021] When the key parameter is bubble size, the voltage and frequency of the nanobubble generator are adjusted to make the bubble size meet the target value of the key parameter.

[0022] When the key parameter is the dosage of the agent, the agent flow rate is dynamically calibrated to the target value of the key parameter through the linkage proportional valve.

[0023] Preferably, when adjusting key parameters based on the process stability at the current acquisition time, the deviation between the measured value and the target value of the key parameter is used as the input to the PID control algorithm. The output control signal is applied to the corresponding device according to the key parameter. Before adjusting using the PID control algorithm, the proportional gain Kp in the PID parameters is dynamically adjusted based on the process stability at the current acquisition time. The specific adjustment formula is as follows: Where Kp′ and Kp are the improved and preset initial proportional gains, respectively, and B is the process stability at the current acquisition time.

[0024] Secondly, embodiments of this application provide a waste refractory material recycling device to implement the waste refractory material recycling method described in any one of the above claims, the device comprising:

[0025] The data acquisition module is used to collect compositional parameter data and flotation process parameter data of waste refractory materials, and to perform noise reduction and normalization preprocessing.

[0026] The data processing module is used to extract two potential factors in the material composition and flotation process through factor analysis, and to analyze the correlation between the actual value and the target value of each parameter at any collection time through grey relational analysis. The correlation between the two potential factors and all parameters at any collection time is combined to construct the separation efficiency.

[0027] The prediction and evaluation module combines all data collected from all parameters, uses the XGBoost regression model to predict the sorting efficiency and energy efficiency, and analyzes the coefficient of variation of sorting efficiency and energy efficiency at all collection times through a sliding window to evaluate the process stability at the current collection time.

[0028] The parameter control module is used to determine the coefficient of each parameter in the regression equation based on the regression equation output by the XGBoost regression model, take the parameter corresponding to the maximum absolute value of the coefficient as the key parameter, and control the key parameter using the process stability at the current acquisition time.

[0029] Thirdly, embodiments of this application also provide a waste refractory material recycling system, including the waste refractory material recycling equipment described above, and an industrial Internet of Things (IIoT) platform, the platform being used to realize data interaction and collaborative control between devices.

[0030] As can be seen from the above embodiments, the method, equipment, and system for recycling waste refractory materials provided in this application have at least the following beneficial effects:

[0031] (1) To address the issues of multidimensional data redundancy and nonlinearity in the sorting process of waste refractory materials, this paper combines factor analysis and grey relational analysis, considers the nonlinear synergistic effect of composition and process, strengthens its contribution weight to the ideal sorting efficiency, breaks through the bottleneck of inefficient processing of redundant data in traditional processes, and accurately identifies the core variables affecting sorting efficiency.

[0032] (2) To address the problem of poor stability caused by dynamic fluctuations in process parameters and environmental disturbances, the XGBoost regression model and sliding window coefficient of variation analysis are used to capture the nonlinear relationship between parameters and efficiency, quantify the risk of historical fluctuations, eliminate the interference of short-term random disturbances on the assessment, thereby suppressing the misleading effect of environmental variables on steady-state assessment and improving the accuracy of process anomaly identification.

[0033] (3) Based on the feature importance dynamic identification of the XGBoost regression model, the actuator is linked through the IIoT platform to realize hierarchical control, forming a “monitoring-analysis-execution-verification”, thereby improving sorting efficiency and process stability, significantly optimizing resource utilization, and supporting efficient and stable sorting under complex working conditions. Attached Figure Description

[0034] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 A flowchart illustrating the steps of a method for recycling waste refractory materials, as provided in one embodiment of this application;

[0036] Figure 2 This is a schematic diagram of a waste refractory material recycling device provided in one embodiment of this application. Detailed Implementation

[0037] To further illustrate the technical means and effects adopted by this application to achieve the intended inventive purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method, apparatus, and system for recycling waste refractory materials proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0038] Unless otherwise specified and limited, terms such as “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a circuit structure, article, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the article or device that includes said element. Furthermore, the term “and / or” as used herein includes any and all combinations of one or more of the associated listed items. 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 application pertains.

[0039] The following description, in conjunction with the accompanying drawings, details a specific scheme for a method, equipment, and system for the reuse of waste refractory materials provided in this application.

[0040] Please see Figure 1 The diagram illustrates a flowchart of a method for recycling waste refractory materials according to an embodiment of this application. The method specifically includes the following steps:

[0041] Collect compositional parameter data and flotation process parameter data of waste refractory materials, and perform noise reduction and normalization preprocessing;

[0042] Two potential factors in the material composition and flotation process are extracted by factor analysis. The correlation between the actual value and the target value of each parameter at any collection time is analyzed by grey relational analysis. The two potential factors and the correlation of all parameters at any collection time are combined to construct the separation efficiency and energy efficiency.

[0043] By combining all the data collected from all parameters, the XGBoost regression model is used to predict the sorting efficiency and energy efficiency, and the coefficient of variation of sorting efficiency and energy efficiency at all collection times is analyzed by sliding window analysis to evaluate the process stability at the current collection time.

[0044] Based on the regression equation output by the XGBoost regression model, the coefficient of each parameter in the regression equation is determined, and the parameter corresponding to the maximum absolute value of the coefficient is taken as the key parameter, and the key parameter is adjusted.

[0045] Specifically, the details of each step are described below:

[0046] Step 1: Data acquisition and preprocessing.

[0047] An X-ray fluorescence spectrometer (XRF) is installed at the end of the data acquisition module of the waste refractory material recycling equipment to collect mass fraction data of magnesium oxide (MgO) and graphite (C) for accurate analysis of the composition of the waste material. A pH sensor, a nanobubble size detector and a flow meter are integrated in the flotation cell to collect data on the pH value of the flotation solution, bubble size and reagent dosage, so as to dynamically control the flotation environment to enhance the graphite adsorption selectivity.

[0048] In this embodiment, the data acquisition frequency of all sensors is set to 100Hz. Process data from each module is synchronously integrated through an Industrial Internet of Things (IIoT) platform. Preprocessing is performed using wavelet transform denoising and Min-Max normalization to eliminate measurement noise, unify data scale, and construct a highly reliable dataset. The preprocessed multidimensional data from each acquisition moment is then combined into a multidimensional data sequence corresponding to that acquisition moment, providing a foundation for subsequent sorting efficiency index modeling and process optimization. The wavelet transform denoising and Min-Max normalization methods are well-known techniques and will not be described in detail here.

[0049] Step 2: Multidimensional data analysis and model building.

[0050] Because the multidimensional data in the waste refractory material sorting process is highly redundant and has nonlinear correlation, traditional processes struggle to accurately identify key influencing factors, resulting in significant fluctuations in sorting efficiency and component recovery rate.

[0051] Therefore, using the data sequence of each parameter composed of all preprocessed acquisition times as input, factor analysis (FA) was used. Based on the Kaiser criterion of eigenvalue > 1, the number of factors was set to 2 to extract two core latent factors in the material composition and flotation process. These two latent factors extracted by factor analysis include the material purity factor extracted from the composition parameter data and the flotation process stability factor extracted from the flotation process parameter data. The maximum number of iterations was set to 100 to ensure model convergence. Factor scores F1 and F2 were output. Factor analysis compressed highly correlated variables in the original data (such as MgO, C concentration and pH, bubble size, and reagent dosage) into two independent latent factors through a loading matrix. Their physical meaning is defined by the variable with the highest loading weight. F1 reflects the purity of the material composition, and F2 characterizes the stability of the flotation process. Dimensionality reduction eliminated noise and quantified the synergistic effect between composition and process. Factor analysis is a well-known technique and will not be elaborated further.

[0052] Subsequently, an ideal separation efficiency sequence is set. Specifically, the ideal separation efficiency is based on industry standards or historical best separation data, and target values ​​for key parameters are set. In this embodiment, the target values ​​are set as follows: MgO recovery rate = 92%, graphite recovery rate = 88%, flotation solution pH = 6.0, bubble size = 25μm, and reagent dosage = 55g / ton of raw material. The sequence composed of all target values ​​is taken as the ideal separation efficiency sequence, which is aligned with the element types in the corresponding positions of the multidimensional data sequence.

[0053] Using the multidimensional data sequence at any actual acquisition time as the comparison sequence and the ideal sorting efficiency sequence as the reference sequence, grey relational analysis (GRA) is used. To balance the sensitivity and anti-interference of the correlation, the resolution coefficient is set to 0.5. The grey relational degree γ between the multidimensional data sequence at any acquisition time and the ideal sorting efficiency sequence at the i-th parameter is output. i That is, to analyze the correlation between the actual value and the target value of each parameter through grey relational analysis, where γ i A larger value indicates a stronger correlation between the parameter and the ideal sorting efficiency. Grey relational analysis is a well-known technique and will not be elaborated further.

[0054] Based on the above analysis, the sorting efficiency and energy efficiency A are constructed, and the specific calculation formula is as follows:

[0055]

[0056] F1 and F2 represent two potential factors in the material composition and flotation process, respectively. F1, extracted from multidimensional data through factor analysis (FA), reflects the purity characteristics of the target component in waste refractory materials. F2 characterizes the stability of the flotation process, reflecting the dynamic balance capability of process parameters. The product of the two reflects the nonlinear coupling effect between material purity and process stability. The larger the value, the more significant the synergistic improvement effect of separation efficiency and process stability, indicating that material purity and process environment tend to approach the optimal state, and the higher the overall energy efficiency of the separation process. N represents the parameter type, r i The grey relational degree between the multidimensional data sequence and the ideal sorting efficiency sequence for the i-th parameter quantifies the similarity between the actual value and the target value. The larger the value, the more significant the contribution of the i-th parameter to the ideal sorting efficiency, and the more likely it is to be the key control point for process optimization. i is the summation index, and the value can be from 1 to N for the parameter type.

[0057] It should be understood that the sorting efficiency energy efficiency A comprehensively considers the purity of material composition, process stability, and the correlation between the actual and target values ​​of each parameter to construct the energy efficiency situation in the sorting process. The larger the value, the more it represents the synergistic optimization of material purity and process stability, and the key parameters are close to the ideal state, resulting in high sorting efficiency.

[0058] Step 3: Process prediction and steady-state assessment.

[0059] Because the process parameters and sorting efficiency exhibit a strong nonlinear correlation during the sorting of waste refractory materials, and random disturbances in environmental variables cause dynamic fluctuations in sorting efficiency, traditional linear models struggle to accurately predict process trends, resulting in poor sorting process stability and lower-than-expected resource recovery rates.

[0060] Therefore, the sequence of sorting efficiency A calculated at all collection times and the data sequence of all parameters are used as independent variables. At each data collection time, the absolute value of the difference between each collected parameter and the target value of the corresponding parameter is calculated and recorded as the first difference. The mean of the first differences of all parameters is calculated and used as the dependent variable at the corresponding data collection time. The XGBoost regression model is used for prediction. In this embodiment, the learning rate is set to 0.05 to avoid overfitting, the maximum tree depth is 6 to balance model complexity and generalization ability, and the number of iterations is 500 to ensure convergence. The predicted value of sorting efficiency Y is output to quantify the quantitative trend of sorting efficiency in the future period. The marginal regulation effect of key parameters is revealed by ranking the importance of features.

[0061] Subsequently, the sequence of sorting efficiency and energy efficiency A calculated at all acquisition times is used as input. Sliding window coefficient of variation analysis is used. In this embodiment, the window length is set to 10s. Assume that there are J windows in the dataset. The coefficient of variation Cv of each window is output, and the difference t between the rightmost acquisition time of each window and the current acquisition time is calculated, thereby quantifying the fluctuation amplitude and identifying process anomalies.

[0062] Based on the above analysis, the process stability B at the current acquisition moment is constructed, and the specific calculation formula is as follows:

[0063]

[0064] Where Y is the predicted sorting efficiency. Since there will inevitably be deviations between parameters and target values ​​during the actual sorting process, the predicted value is generated based on the XGBoost regression model analyzing continuously dynamically fluctuating process data. Therefore, a completely unbiased (Y=0) ideal steady state cannot be achieved in actual industrial scenarios. By analyzing the deviation of each parameter from the target value in historical data, the nonlinear correlation between process parameters and sorting efficiency is captured, predicting the dynamic trend of future sorting efficiency and quantifying the expected level of sorting efficiency in the future period. A smaller Y value indicates that the model predicts a more efficient sorting process; J represents the number of windows in the sequence composed of sorting efficiency energy efficiency A calculated at all collection times; t j is the time interval between the rightmost acquisition time of the j-th window and the current acquisition time, used to measure the temporal proximity of the data in that window; L is the length of the sequence composed of the sorting efficiency and energy efficiency A calculated from all acquisition times. The time difference is standardized as a proportion of the total sequence length L to eliminate the influence of absolute time units. The further the window is from the current time, the smaller its value, thus giving less weight to subsequent fluctuations; e is the natural constant; Cv j is the coefficient of variation for the j-th window, used to quantify the fluctuation range of process parameters within the window, directly reflecting the risk of process abnormality or runaway. The larger its value, the greater the fluctuation of sorting efficiency within the window.

[0065] It should be understood that process stability B is a quantitative representation of the dynamic stability of the sorting system. It comprehensively reflects the coupling effect of the expected efficiency of the process and the historical fluctuation risk. The larger the B value, the more stable the process parameters are in operation with low fluctuation over time. By strengthening the contribution of recent data through time decay weight, the B value can capture the real-time process stability decay signal and suppress the interference of early random fluctuations.

[0066] Step 4: Parameter optimization and control.

[0067] Because the density difference between magnesium oxide and graphite in waste refractory materials is small and the surface oxide coverage reduces hydrophobicity, coupled with dynamic fluctuations in flotation process parameters and environmental interference, traditional separation technology is difficult to achieve efficient separation of complex multiphase systems, resulting in problems such as low separation efficiency, poor process stability, and substandard performance of recycled materials.

[0068] Therefore, dynamic parameter optimization and control are required based on process stability B. The specific hierarchical control strategy is designed as follows:

[0069] By monitoring the B value in real time and analyzing the feature importance of the XGBoost regression model in step 3 above, the coefficient of each parameter in the regression equation is determined based on the regression equation output by the XGBoost regression model. This coefficient represents the direct influence weight of the corresponding parameter on the target variable. Then, the absolute values ​​of the coefficients of each parameter are sorted, and the parameter corresponding to the maximum absolute value of the coefficient is taken as the key parameter to identify the key parameters that have a significant impact on sorting efficiency.

[0070] In the sorting of waste refractory materials, the concentrations of magnesium oxide (MgO) and graphite (C) are inherent properties of the materials themselves, belonging to the uncontrollable characteristics of the input materials, and need to be indirectly optimized through the sorting process. On the other hand, the pH value of the flotation solution, the bubble size, and the reagent dosage are dynamic parameters that can be controlled by the process, directly affecting the selectivity and adsorption efficiency of the flotation environment. The pH value determines the surface charge and reagent activity, the bubble size is related to the graphite capture probability, and the reagent dosage regulates the interfacial chemical reaction. The sorting deviation can be quickly corrected through real-time feedback control of the three. However, the material concentration can only be adjusted through the initial batching or subsequent process, and cannot be directly intervened in the current sorting stage. Therefore, the process parameters should be adjusted first to achieve indirect optimization of uncontrollable components.

[0071] Secondly, tiered adjustments are performed to address parameter deviations. Specifically, the key parameter identification and adjustment methods are as follows:

[0072] 1. If the pH value of the flotation solution is a key parameter, its deviation from the target range will weaken the selective adsorption capacity of the flotation agent; in this case, acidic / alkaline reagents are injected by starting the automatic dosing pump to adjust the pH to the target value of the key parameter in real time.

[0073] 2. If bubble size is a key parameter, it directly affects the graphite adsorption efficiency. In this case, by adjusting the voltage and frequency of the nanobubble generator, the bubble size can be made to meet the target value of the key parameter, thereby optimizing the bubble distribution density and size stability.

[0074] 3. If the reagent dosage is a key parameter, excessive or insufficient dosage will lead to a decrease in the contact efficiency between the flotation agent and the material; in this case, the reagent flow rate is dynamically calibrated to the target value of the key parameter through the linkage proportional valve to match the real-time separation requirements.

[0075] The above three parameters are all controlled by a PID control algorithm. In this embodiment, the initial parameters of the PID control algorithm are set as Kp = 1.2, Ki = 0.05, and Kd = 0.3. The deviation between the measured value and the target value of the key parameter is used as the input of the PID control algorithm. The output control signal is applied to the corresponding device according to the key parameter, thereby realizing the adjustment of the key parameter.

[0076] However, due to the strong nonlinearity, time-varying disturbances, and equipment response delays in the flotation process, fixing PID parameters may lead to adjustment lag or overshoot, reducing process stability.

[0077] Therefore, before using the PID control algorithm for regulation, Kp needs to be dynamically adjusted according to the process stability B to enhance the adaptive capability of the control system. When the system is stable (high B value), Kp should be reduced to avoid oscillations; when the system is fluctuating (low B value), Kp should be increased to accelerate the response. The specific adjustment formula is as follows:

[0078]

[0079] Where Kp′ and Kp are the improved and preset initial proportional gains, respectively, and B is the process stability at the current acquisition time. When its value is larger, it means that the process is highly stable. At this time, the improved proportional gain is closer to the initial proportional gain to maintain mild adjustment. Conversely, when B is smaller, it means that the process fluctuation is aggravated. At this time, the proportional gain increases inversely to improve the adjustment speed and suppress deviation.

[0080] By integrating an Industrial Internet of Things (IIoT) platform, a unified control system is built to achieve coordinated control of three key parameters: the system collects flotation solution pH, bubble size, and reagent dosage in real time, analyzes the coefficients of parameters in the regression equation based on the XGBoost regression model, and dynamically generates priority adjustment instructions—automatic dosing pump adjustment, nanobubble generator optimization of bubble size, and proportional valve calibration of reagent flow matching requirements, ultimately achieving dynamic balance of multiple parameters and stable improvement of sorting efficiency.

[0081] Based on the same inventive concept as the above method, a schematic diagram of a waste refractory material recycling device is provided in one embodiment of this application, as shown in the attached figure. Figure 2 As shown, the device includes:

[0082] The data acquisition module is used to collect compositional parameter data and flotation process parameter data of waste refractory materials, and to perform noise reduction and normalization preprocessing.

[0083] The data processing module is used to extract two potential factors in the material composition and flotation process through factor analysis, and to analyze the correlation between the actual value and the target value of each parameter at any collection time through grey relational analysis. The correlation between the two potential factors and all parameters at any collection time is combined to construct the separation efficiency.

[0084] The prediction and evaluation module combines all data collected from all parameters, uses the XGBoost regression model to predict the sorting efficiency and energy efficiency, and analyzes the coefficient of variation of sorting efficiency and energy efficiency at all collection times through a sliding window to evaluate the process stability at the current collection time.

[0085] The parameter control module is used to determine the coefficient of each parameter in the regression equation based on the regression equation output by the XGBoost regression model, take the parameter corresponding to the maximum absolute value of the coefficient as the key parameter, and control the key parameter using the process stability at the current acquisition time.

[0086] Based on the same inventive concept as the above method, another embodiment of this application also provides a waste refractory material recycling system, including the above-mentioned waste refractory material recycling equipment and an industrial Internet of Things (IIoT) platform, which is used to realize data interaction and collaborative control between devices.

[0087] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0088] It should be noted that, unless otherwise specified and limited, terms such as “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a circuit structure, article, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such article or device. Without further limitations, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the article or device that includes said element. Furthermore, the term “and / or” as used herein includes any and all combinations of one or more of the associated listed items.

[0089] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not invented in this application.

[0090] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A method for recycling waste refractory materials, characterized in that, The method includes the following steps: Collect compositional parameter data and flotation process parameter data of waste refractory materials, and perform noise reduction and normalization preprocessing; Two potential factors in the material composition and flotation process are extracted by factor analysis. The correlation between the actual value and the target value of each parameter at any collection time is analyzed by grey relational analysis. The two potential factors and the correlation of all parameters at any collection time are combined to construct the separation efficiency and energy efficiency. By combining all the data collected from all parameters, the XGBoost regression model is used to predict the sorting efficiency and energy efficiency, and the coefficient of variation of sorting efficiency and energy efficiency at all collection times is analyzed by sliding window analysis to evaluate the process stability at the current collection time. Based on the regression equation output by the XGBoost regression model, the coefficient of each parameter in the regression equation is determined. The parameter corresponding to the maximum absolute value of the coefficient is taken as the key parameter, and the key parameter is adjusted using the process stability at the current acquisition time. During the data acquisition process, X-ray fluorescence spectrometry (XRF) was used to collect the mass fraction data of magnesium oxide (MgO) and graphite (C), and pH sensor, nanobubble size detector and flow meter were used to collect the pH value, bubble size and reagent dosage data of flotation solution. The method for adjusting key parameters is as follows: When the key parameter is the pH value of the flotation solution, acidic / alkaline reagents are injected by starting the automatic dosing pump to adjust the pH to the target value of the key parameter in real time. When the key parameter is bubble size, the voltage and frequency of the nanobubble generator are adjusted to make the bubble size meet the target value of the key parameter. When the key parameter is the dosage of the agent, the agent flow rate is dynamically calibrated to the target value of the key parameter through the linkage proportional valve. When adjusting key parameters based on the process stability at the current acquisition time, the deviation between the measured and target values ​​of the key parameters is used as the input to the PID control algorithm. The output control signal is applied to the corresponding device according to the key parameters. Before adjusting using the PID control algorithm, the proportional gain Kp in the PID parameters is dynamically adjusted based on the process stability at the current acquisition time. The specific adjustment formula is as follows: ;in, and These represent the improved and preset initial proportional gains, respectively, and B represents the process stability at the current acquisition time.

2. The method for recycling waste refractory materials as described in claim 1, characterized in that, The two potential factors extracted through factor analysis include the material composition purity factor extracted from the composition parameter data and the flotation process stability factor extracted from the flotation process parameter data.

3. The method for recycling waste refractory materials as described in claim 1, characterized in that, The method for constructing the sorting efficiency and energy efficiency is as follows: Calculate the product of the two potential factors; calculate the mean of the correlation of all parameters at any given time; and use the ratio of the product to the mean as the sorting efficiency.

4. The method for recycling waste refractory materials as described in claim 1, characterized in that, The method for predicting sorting efficiency and energy efficiency using the XGBoost regression model, combining all data collected from all parameters, is as follows: The sorting efficiency and energy efficiency sequences calculated at all collection times and the data sequences of all parameters are used as independent variables. At each data collection time, the absolute value of the difference between each collected parameter and the target value of the corresponding parameter is calculated and recorded as the first difference. The mean of the first differences of all parameters is calculated and used as the dependent variable at the corresponding data collection time. The XGBoost regression model is used to predict and output the predicted value Y of sorting efficiency.

5. A method for recycling waste refractory materials as described in claim 1, characterized in that, The method for evaluating the process stability at the current acquisition time is as follows: Where B represents the process stability at the current acquisition time; Y represents the predicted sorting efficiency; and J represents the number of windows in the sequence composed of the sorting efficiency energy efficiency A calculated at all acquisition times. is the time interval between the rightmost acquisition time of the j-th window and the current acquisition time; L is the sequence length composed of the sorting efficiency and energy efficiency A calculated from all acquisition times; e is the natural constant; Let be the coefficient of variation for the j-th window.

6. A device for recycling waste refractory materials, characterized in that, The equipment for implementing the method for recycling waste refractory materials as described in any one of claims 1-5 includes: The data acquisition module is used to collect compositional parameter data and flotation process parameter data of waste refractory materials, and to perform noise reduction and normalization preprocessing. The data processing module is used to extract two potential factors in the material composition and flotation process through factor analysis, and to analyze the correlation between the actual value and the target value of each parameter at any collection time through grey relational analysis. The correlation between the two potential factors and all parameters at any collection time is combined to construct the separation efficiency. The prediction and evaluation module combines all data collected from all parameters, uses the XGBoost regression model to predict the sorting efficiency and energy efficiency, and analyzes the coefficient of variation of sorting efficiency and energy efficiency at all collection times through a sliding window to evaluate the process stability at the current collection time. The parameter control module is used to determine the coefficient of each parameter in the regression equation based on the regression equation output by the XGBoost regression model, take the parameter corresponding to the maximum absolute value of the coefficient as the key parameter, and control the key parameter using the process stability at the current acquisition time.

7. A waste refractory material recycling system, comprising the waste refractory material recycling equipment as described in claim 6, and an industrial Internet of Things (IIoT) platform, wherein the platform is used to realize data interaction and collaborative control between the equipment.

Images