Method and system for evaluating residual life of roller surface of roller press

Abstract

The invention discloses a method and a system for evaluating the residual life of a roller surface of a roller press, relates to the technical field of industrial predictive maintenance, and solves the technical problem of low accuracy of life evaluation caused by neglecting the limitation of a single mode in a variable working condition state because the single mode is often adopted for life evaluation in the prior art. A working condition portrait is constructed based on historical operation data; constructing a life prediction dual model based on the working condition portrait and the historical operation data; model labels and model influence parameters are determined based on real-time operation data and working condition portraits, roll surface remaining life and alarm signals are generated in combination with life prediction double models, clustering analysis is performed on historical operation data, and life prediction double models suitable for different working condition portraits are constructed, so that when roll surface life prediction is performed, roll surface life prediction accuracy is improved. The appropriate life prediction double models can be selected according to the current roller surface operation data so as to carry out accurate life evaluation, and the accuracy of life evaluation is improved.

Description

Technical Field

[0001] This application belongs to the field of industrial predictive maintenance technology, specifically a method and system for assessing the remaining life of a roller press roller surface. Background Technology

[0002] In modern industries such as cement, metallurgy, and mining, roller presses are critical grinding equipment, and their operating efficiency and reliability directly affect the capacity and energy consumption of the entire production line. The core component of the roller press—the roller surface—inevitably suffers wear and fatigue spalling damage under long-term high-pressure and high-wear conditions, severely impacting equipment performance and product quality. Roller surface remaining life assessment refers to predicting the usable operating time of the roller surface from its current state to the failure threshold based on the wear mechanism of the roller surface material, operating condition data, historical maintenance records, and online or offline testing results, using mathematical models, data analysis, or intelligent algorithms. This assessment not only helps to accurately grasp the health status of the equipment but also signifies a transformation and upgrade from traditional "periodic maintenance" to "condition-based maintenance" and even "predictive maintenance." Therefore, conducting "roller press roller surface remaining life assessment" has significant engineering and economic value.

[0003] Existing technologies often employ a single method for life assessment, neglecting the limitations of this single method under varying operating conditions, resulting in low accuracy and efficiency in life assessment. Therefore, further improvements are needed for methods to assess the remaining life of roller press roller surfaces. Summary of the Invention

[0004] This application aims to solve at least one of the technical problems existing in the prior art; to this end, this application proposes a method and system for assessing the remaining life of a roller press roller surface, which solves the technical problem that the prior art often uses a single method for life assessment, ignoring the limitations of the single method when the operating conditions are variable, resulting in low accuracy of life assessment.

[0005] To achieve the above objectives, the first aspect of this application provides a method for assessing the remaining life of a roller press roller surface, comprising: Acquire historical and real-time operating data of the roller press; A working condition profile is constructed based on historical operating data; the working condition profile refers to the characteristics of the roller surface of the roller press under different production conditions obtained after cluster analysis of historical operating data. A dual lifespan prediction model is constructed based on operating condition profiles and historical operating data; Model labels and model impact parameters are determined based on real-time operational data and operating condition profiles; the model labels and model impact parameters refer to the results obtained by matching several data points in the real-time operational data with the life prediction dual model; The remaining life of the roller surface and alarm signals are generated based on a dual model that combines model labels, model influence parameters, and life prediction.

[0006] The steps described in this application, based on historical operating data, conduct cluster analysis to construct a dual-model life prediction system for different operating conditions. When predicting the life of the roller surface, the system can match the corresponding operating condition profile according to the current equipment operating status and adaptively select the optimal prediction model combination to achieve a more accurate life assessment. This fully considers the diversity and complexity of operating conditions in actual operation and significantly improves the accuracy and applicability of the remaining life prediction of the roller surface.

[0007] Furthermore, the construction of the operating condition profile based on historical operating data includes: Extract the roller press working pressure, roller gap deviation, and material type identifier from historical operating data; N cluster centers were determined by cluster analysis of the working pressure, roll gap deviation, and material type identifier of the roller press using an unsupervised clustering algorithm. N operating condition profiles are determined based on N cluster centers, and these operating condition profiles are labeled GH. n Where n represents the number corresponding to the working condition profile, n is an integer, n∈[1,N].

[0008] Furthermore, the dual-model for life prediction based on operating condition profiles and historical operating data includes: For each operating condition profile, the historical operating data corresponding to the operating condition profile is selected from the historical operating data as the historical operating data of the operating condition. Extract historical data from the operating conditions, including the original thickness of the roller surface, the measured thickness, the cumulative amount of material processed, and the operating pressure P of the roller press. i Historical roll gap deviation ΔG i ; The historical equivalent wear thickness is calculated by comparing the original historical roller surface thickness with the historical measured thickness. ; A trend life prediction model is constructed by using historical cumulative material processing volume and historical equivalent wear thickness from historical operating data of several working conditions. The trend life prediction model is based on a univariate linear regression model, with historical cumulative material processing volume as the independent variable and historical equivalent wear thickness as the dependent variable, and is fitted and trained to finally obtain the trend life prediction model. The historical wear uniformity index is calculated using a formula. The formula satisfies: Where i represents the number corresponding to the historical time point in the historical operating data; It is a very small constant; Through formula Calculate historical dynamic wear rate ; A state correction model is obtained by constructing and training a random forest model using historical wear uniformity index and historical dynamic wear rate corresponding to several historical operating data under various working conditions; the training objective of the state correction model is the prediction residual of the trend life prediction model. A dual-model for lifetime prediction is determined based on a trend lifetime prediction model and a state correction model.

[0009] This application obtains several operating condition profiles by performing cluster analysis on specific parameters in historical operating data, and uses the historical operating data corresponding to these operating condition profiles to determine several parameters for constructing a dual life prediction model. Different training methods are used for different models to obtain the final dual life prediction model, so that the dual life prediction model can be applied to the corresponding operating condition profiles and improve the accuracy of life assessment.

[0010] Furthermore, the determination of model labels and model impact parameters based on real-time operational data and operational condition profiles includes: Extract the original thickness of the roller surface, the actual measured thickness, the cumulative amount of material processed, the working pressure of the roller press, and the roller gap deviation from the real-time operating data; Assign a trend label to the cumulative amount of material processed, and determine the model influence parameters corresponding to the trend label based on the cumulative amount of material processed. The equivalent wear thickness is determined based on the original thickness of the roller surface and the actual measured thickness. The wear uniformity index is determined based on the working pressure of the roller press and the roller gap deviation. The dynamic wear rate is determined based on the equivalent wear thickness and the wear uniformity index. The wear uniformity index and dynamic wear rate are assigned to state labels; the model influence parameters corresponding to the state labels are determined based on the wear uniformity index and dynamic wear rate.

[0011] This application does not use all the data from historical operating data when constructing the dual life prediction model. Instead, it adjusts some of the data to obtain the parameters of the life prediction model. Therefore, when constructing the dual life prediction model, it is necessary to plan and adjust the actual operating data to determine several parameters that are suitable for the dual life prediction model, thereby improving the accuracy and efficiency of life assessment.

[0012] Furthermore, the generation of remaining roll life and alarm signals based on the dual model of model labels, model influence parameters, and life prediction includes: Extract the roller press working pressure, roller gap deviation, and material type identifier from the real-time operating data; Calculate the similarity between the working pressure, roll gap deviation, and material type identification of the roller press and several working condition profiles; Extract the life prediction dual model corresponding to the working condition profile with the maximum similarity; the life prediction dual model includes a trend life prediction model and a state correction model. Input the model influence parameters corresponding to the model label into the corresponding lifetime prediction dual model to obtain the prediction results; The remaining life of the roller surface and alarm signals are generated based on the prediction results.

[0013] When conducting life assessment, this application first matches and filters the most relevant operating condition profile based on the current operating condition data, and then calls a dedicated life prediction dual model that is compatible with the profile. Subsequently, the current key influencing parameters are input into the dual model for collaborative reasoning to obtain the corresponding prediction results. The prediction results are then comprehensively analyzed to finally achieve a quantitative assessment of the remaining life of the roller surface. This fully considers the dynamic changes in actual operating conditions and significantly improves the accuracy, robustness, and engineering applicability of life prediction.

[0014] Furthermore, the generation of remaining roll life and alarm signals based on the prediction results includes: Extract the predicted wear thickness YMH and the condition correction ΔH from the prediction results; Through formula Calculate the comprehensive equivalent wear thickness ZDMH; where α is a correction factor; The roller surface health degree (GJD) is calculated using a formula; the formula satisfies: Where GCH represents the initial thickness of the roll surface, This is expressed as the roller surface scrap threshold; Extracting model parameters from the trend life prediction model in the dual life prediction model. The model parameters This represents the slope parameter corresponding to the trend life prediction model under the current working condition profile. Different working condition profiles correspond to different trend life prediction models, so the slope parameter will also be different as the working condition profile changes. The remaining roller surface treatment capacity GSCL is calculated using a formula; the formula satisfies: ; The remaining life of the roller surface is determined based on the remaining amount of material that can be processed on the roller surface. An alarm signal is generated based on the health status of the roller surface.

[0015] Furthermore, the correction factor is obtained through the following methods: Extract the standard deviation BC corresponding to the state correction in the prediction results; the standard deviation is the core indicator for measuring the confidence of the current state correction. The smaller BC is, the more consistent the results of the state correction model are, and the higher the confidence of the results. The correction factor is calculated using a formula; the formula satisfies: ;in, and These represent the upper and lower limits of the correction factor, respectively; the specific values ​​are set based on experience; t represents the decay coefficient, t>0; the specific value is set based on experience; the decay coefficient is set to control the rate at which the correction factor decays as the standard deviation increases; Represented as the reference standard deviation, it is used to normalize the standard deviation, thereby making... It becomes a dimensionless ratio.

[0016] Furthermore, the generation of alarm signals based on roller surface health includes: When the health status of the roller surface is less than D times the health threshold, an alarm signal for short roller surface life is generated. When the health status of the roller surface is greater than or equal to D times the health threshold, and the health status of the roller surface is less than the health threshold, an early warning signal for accelerated roller surface wear is generated; where D is a proportional coefficient, D∈(0,1); the specific value is set based on experience.

[0017] A second aspect of the present invention provides a system for assessing the remaining life of a roller press roller surface, comprising: a data acquisition module, a data analysis module, and an early warning module; the data acquisition module and the data analysis module are connected together; the data analysis module and the early warning module are connected together. The data acquisition module acquires historical and real-time operating data of the roller press through data acquisition equipment. The data analysis module: constructs a working condition profile based on historical operating data; constructs a dual life prediction model based on the working condition profile and historical operating data; determines model labels and model influence parameters based on real-time operating data and the working condition profile; and generates the remaining life of the roller surface and alarm signals based on the model labels, model influence parameters, and the dual life prediction model. The early warning module provides corresponding prompts based on alarm signals.

[0018] Another aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements a roller surface remaining life assessment system according to a second aspect of the present invention.

[0019] Compared with the prior art, the beneficial effects of this application are: 1. This application constructs a working condition profile based on historical operating data; constructs a dual life prediction model based on the working condition profile and historical operating data; determines model labels and model influence parameters based on real-time operating data and working condition profile; and generates the remaining life of the roll surface and alarm signals based on the model labels, model influence parameters, and the dual life prediction model. By performing cluster analysis on historical operating data, and constructing a dual life prediction model suitable for different working condition profiles, this application enables the selection of an appropriate dual life prediction model based on the current roll surface operating data to conduct accurate life assessment, thereby improving the accuracy of life assessment.

[0020] 2. When conducting life assessment, this application first filters the working condition profile corresponding to the current working condition data, extracts the life prediction dual model corresponding to the working condition profile, and then inputs the current model influence parameters into the corresponding life prediction dual model to obtain the prediction result. The remaining life of the roller surface is quantified by the prediction result, which improves the accuracy of life assessment.

[0021] 3. This application uses the prediction results of a dual-model life prediction system and combines them with a dynamic correction factor to adaptively quantify the remaining life of the roll surface, making the results of the remaining life of the roll surface more consistent with the actual situation and improving the accuracy of life assessment. Attached Figure Description

[0022] To more clearly illustrate the technical solutions 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.

[0023] Figure 1 This is a flowchart of a method for assessing the remaining life of a roller press roller surface according to this application; Figure 2 This is a schematic diagram of the principle of a roller press roll surface remaining life assessment system according to this application. Detailed Implementation

[0024] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0025] Please see Figure 1 The first aspect of this application provides a method for assessing the remaining life of a roller press roller surface, comprising: Acquire historical and real-time operating data of the roller press; A working condition profile is constructed based on historical operating data. The working condition profile refers to the characteristics of the roller surface of the roller press under different production conditions obtained after cluster analysis of historical operating data. A dual lifespan prediction model is constructed based on operating condition profiles and historical operating data; Model labels and model impact parameters are determined based on real-time operational data and operating condition profiles; model labels and model impact parameters refer to the results obtained by matching several data points from real-time operational data with the life prediction dual model; The remaining life of the roller surface and alarm signals are generated based on a dual model that combines model labels, model influence parameters, and life prediction.

[0026] The construction of a working condition profile based on historical operating data in this embodiment includes: Extract the roller press working pressure, roller gap deviation, and material type identifier from historical operating data; The working pressure, roll gap deviation, and material type identifier of the roller press are clustered using an unsupervised clustering algorithm to determine N cluster centers; in this embodiment, the unsupervised clustering algorithm selected is K-Means. N operating condition profiles are determined based on N cluster centers, and these operating condition profiles are labeled GH. n Where n represents the number corresponding to the working condition profile, n is an integer and n∈[1,N]; the value of N is set according to experience. In this embodiment, N is set to 3, that is, there are 3 types of working condition profiles. In this embodiment, the working condition profiles are set as high pressure slag working condition, normal limestone working condition and eccentric roller wear working condition.

[0027] The dual-model for lifespan prediction based on operating condition profiles and historical operating data in this embodiment includes: For each operating condition profile, the historical operating data corresponding to the operating condition profile is selected from the historical operating data as the historical operating data of the operating condition. Extract historical data from the operating conditions, including the original thickness of the roller surface, the measured thickness, the cumulative amount of material processed, and the operating pressure P of the roller press. i Historical roll gap deviation ΔG i ; The historical equivalent wear thickness is calculated by comparing the original historical roller surface thickness with the historical measured thickness. ; A trend life prediction model is constructed by using historical cumulative material processing volume and historical equivalent wear thickness from historical operating data of several working conditions. The trend life prediction model is based on a univariate linear regression model, with historical cumulative material processing volume as the independent variable and historical equivalent wear thickness as the dependent variable, and is fitted and trained to finally obtain the trend life prediction model. The model parameters in the trend life prediction model in this embodiment are determined by fitting using the least squares method, and the model parameters in the trend life prediction model under different working conditions will be different. The historical wear uniformity index is calculated using a formula. The formula satisfies: Where i represents the number corresponding to the historical time point in the historical operating data; It is a very small constant, and the specific value is set based on experience. In this embodiment, it will be... Set to 0.001; Set This is to avoid the phenomenon of the denominator being 0 in the formula and to maintain the rationality of the formula; Through formula Calculate historical dynamic wear rate ; The state correction model is obtained by constructing and training a random forest model based on the historical wear uniformity index and historical dynamic wear rate corresponding to several historical operating data under various working conditions; the training objective of the state correction model is the prediction residual of the trend life prediction model. A dual-model for lifetime prediction is determined based on a trend lifetime prediction model and a state correction model.

[0028] This embodiment extracts several typical operating condition profiles by clustering key parameters in historical operating data, and determines the key construction parameters required for the dual life prediction model based on the historical datasets corresponding to each operating condition profile. Differentiated modeling strategies and training methods are adopted for different operating condition characteristics to optimize the structure and parameters of the dual model respectively. Finally, a dedicated dual life prediction model matching various operating condition profiles is constructed, realizing accurate adaptation between the model and actual operating conditions, and significantly improving the pertinence and prediction accuracy of the roll surface remaining life assessment.

[0029] In this embodiment, the determination of model labels and model impact parameters based on real-time operational data and operational condition profiles includes: Extract the original thickness of the roller surface, the actual measured thickness, the cumulative amount of material processed, the working pressure of the roller press, and the roller gap deviation from the real-time operating data; Assign a trend label to the cumulative amount of material processed, and determine the model influence parameters corresponding to the trend label based on the cumulative amount of material processed. The equivalent wear thickness is determined based on the original thickness of the roller surface and the actual measured thickness. The wear uniformity index is determined based on the working pressure of the roller press and the roller gap deviation. The dynamic wear rate is determined based on the equivalent wear thickness and the wear uniformity index. The wear uniformity index and dynamic wear rate are assigned to state labels; the model influence parameters corresponding to the state labels are determined based on the wear uniformity index and dynamic wear rate.

[0030] In constructing the dual-model for lifetime prediction in this embodiment, the original historical operating data is not used directly. Instead, some key data are adjusted accordingly to improve the model's generalization ability and prediction performance. Therefore, when applying this dual-model for lifetime prediction, the actual operating data needs to be processed to generate adaptive parameters that meet the model's input requirements. This not only enhances the model's adaptability to complex operating conditions but also effectively improves the accuracy and computational efficiency of lifetime assessment.

[0031] In this embodiment, the generation of remaining roll life and alarm signals based on a dual model of model labels, model influence parameters, and life prediction includes: Extract the roller press working pressure, roller gap deviation, and material type identifier from the real-time operating data; The similarity between the working pressure, roll gap deviation, and material type identifier of the roller press and several working condition profiles is calculated. In this embodiment, the Euclidean distance between the current real-time running data and the N cluster centers in several working condition profiles is selected as the calculation index for the similarity calculation. Extract the life prediction dual model corresponding to the working condition profile with the maximum similarity; the life prediction dual model includes a trend life prediction model and a state correction model. Input the model influence parameters corresponding to the model label into the corresponding lifetime prediction dual model to obtain the prediction results; The remaining life of the roller surface and alarm signals are generated based on the prediction results.

[0032] In this embodiment, the generation of remaining roller life and alarm signals based on the prediction results includes: Extract the predicted wear thickness YMH and the condition correction ΔH from the prediction results; Through formula Calculate the comprehensive equivalent wear thickness ZDMH; where α is a correction factor; The roller surface health degree (GJD) is calculated using a formula; the formula satisfies: Where GCH represents the initial thickness of the roll surface, This represents the roll surface scrap threshold; in this embodiment, the initial roll surface thickness is 100mm, and the corresponding roll surface scrap threshold is 60mm. Extracting model parameters from the trend life prediction model in the dual life prediction model. Model parameters This represents the slope parameter corresponding to the trend life prediction model under the current working condition profile. In this embodiment, different working condition profiles correspond to different trend life prediction models, so the slope parameter will also be different as the working condition profile changes. The remaining roller surface throughput (GSCL) is calculated using the formula; the formula satisfies: ; The remaining life of the roller surface is determined based on the remaining amount of material that can be processed on the roller surface. An alarm signal is generated based on the health status of the roller surface.

[0033] The correction factor in this embodiment is obtained through the following methods: Extract the standard deviation BC corresponding to the state correction in the prediction results; the standard deviation is the core indicator for measuring the confidence of the current state correction. The smaller BC is, the more consistent the results of the state correction model are, and the higher the confidence of the results. The correction factor is calculated using the formula; the formula satisfies: ;in, and These represent the upper and lower limits of the correction factor, respectively; the specific values ​​are set based on experience, and in this embodiment, they will be... and Set to 1 and 0.1 respectively; t represents the attenuation coefficient, t>0; the specific value is set according to experience, and in this embodiment, t is set to 2; the attenuation coefficient is set to control the rate at which the correction factor decays as the standard deviation increases; Represented as the reference standard deviation, it is used to normalize the standard deviation, thereby making... It becomes a dimensionless ratio; the specific value is set according to experience. In this embodiment, the median of the standard deviations of all prediction results when using the random forest model to predict residuals on the validation set is used as the reference standard deviation.

[0034] Based on the prediction results of the dual life prediction model, this embodiment introduces a dynamic correction factor to adaptively adjust the prediction results. This effectively compensates for the prediction deviation of the model under complex working conditions, enhances the ability to track the actual degradation trend, and makes the quantitative result of the remaining life of the roller surface closer to the real degradation process. This significantly improves the accuracy, timeliness and engineering applicability of life assessment.

[0035] In this embodiment, the alarm signal generation based on the roller surface health includes: When the health status of the roller surface is less than D times the health threshold, an alarm signal for short roller surface life is generated. When the health of the roller surface is greater than or equal to D times the health threshold, and the health of the roller surface is less than the health threshold, an early warning signal for increased roller surface wear is generated; where D is a proportional coefficient, D∈(0,1); the specific value is set according to experience, and in this embodiment, D is set to 0.6; the health threshold is set according to experience, and in this embodiment, the health threshold is set to 70%.

[0036] Please see Figure 2 A second aspect of this application provides a system for assessing the remaining life of a roller press roller surface, comprising: a data acquisition module, a data analysis module, and an early warning module; the data acquisition module and the data analysis module are electrically and / or communicatively connected; the data analysis module and the early warning module are electrically and / or communicatively connected; the data analysis module and the early warning module are connected together. Data acquisition module: Acquires historical and real-time operating data of the roller press through data acquisition equipment; the data acquisition equipment includes several sensors, etc. Data analysis module: Constructs operating condition profiles based on historical operating data; constructs dual life prediction models based on operating condition profiles and historical operating data; determines model labels and model influence parameters based on real-time operating data and operating condition profiles; generates remaining roll life and alarm signals based on model labels, model influence parameters, and the dual life prediction models. Early warning module: Provides corresponding prompts based on alarm signals.

[0037] Another embodiment of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements a roller surface remaining life assessment system according to a second aspect embodiment of this application.

[0038] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0039] The working principle of this application is as follows: Historical and real-time operating data of the roller press are acquired; a working condition profile is constructed based on the historical operating data; a dual-model for life prediction is constructed based on the working condition profile and historical operating data; model labels and model influence parameters are determined based on real-time operating data and the working condition profile; the remaining life of the roller surface and alarm signals are generated based on the model labels, model influence parameters, and the dual-model for life prediction. By performing cluster analysis on historical operating data, dual-models for life prediction suitable for different working condition profiles are constructed. This allows for the selection of an appropriate dual-model for life prediction based on the current roller surface operating data when predicting the life of the roller surface, thus improving the accuracy of life assessment. This avoids the problem that existing technologies often use a single method for life assessment, ignoring the limitations of a single method in situations with changing working conditions, resulting in low accuracy in life assessment.

[0040] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.

Claims

1. A method for assessing the remaining life of a roller press roller surface, characterized in that, include: Acquire historical and real-time operating data of the roller press; Building operational profiles based on historical operational data; The operating condition profile refers to the characteristics of the roller surface of the roller press under different production conditions obtained by clustering analysis of historical operating data. A dual lifespan prediction model is constructed based on operating condition profiles and historical operating data; Model labels and model impact parameters are determined based on real-time operational data and operating condition profiles; the model labels and model impact parameters refer to the results obtained by matching several data points in the real-time operational data with the life prediction dual model; The remaining life of the roller surface and alarm signals are generated based on a dual model that combines model labels, model influence parameters, and life prediction.

2. The method for assessing the remaining life of a roller press surface according to claim 1, characterized in that, The construction of the operating condition profile based on historical operating data includes: Extract the roller press working pressure, roller gap deviation, and material type identifier from historical operating data; N cluster centers were determined by cluster analysis of the working pressure, roll gap deviation, and material type identifier of the roller press using an unsupervised clustering algorithm. N operating condition profiles are determined based on N cluster centers, and these operating condition profiles are labeled GH. n Where n represents the number corresponding to the working condition profile, n is an integer, n∈[1,N].

3. The method for assessing the remaining life of a roller press surface according to claim 1, characterized in that, The dual-model for lifespan prediction based on operating condition profiles and historical operating data includes: For each operating condition profile, the historical operating data corresponding to the operating condition profile is selected from the historical operating data as the historical operating data of the operating condition. Extract historical data from the operating conditions, including the original thickness of the roller surface, the measured thickness, the cumulative amount of material processed, and the operating pressure P of the roller press. i Historical roll gap deviation ΔG i ; The historical equivalent wear thickness is calculated by comparing the original historical roller surface thickness with the historical measured thickness. ; A trend life prediction model is constructed by using historical cumulative material processing volume and historical equivalent wear thickness from historical operating data of several working conditions. The trend life prediction model is based on a univariate linear regression model, with historical cumulative material processing volume as the independent variable and historical equivalent wear thickness as the dependent variable, and is fitted and trained to finally obtain the trend life prediction model. The historical wear uniformity index is calculated using a formula. The formula satisfies: Where i represents the number corresponding to the historical time point in the historical operating data; It is a very small constant; Through formula Calculate historical dynamic wear rate ; A state correction model is obtained by constructing and training a random forest model using historical wear uniformity index and historical dynamic wear rate corresponding to several historical operating data under various working conditions; the training objective of the state correction model is the prediction residual of the trend life prediction model. A dual-model for lifetime prediction is determined based on a trend lifetime prediction model and a state correction model.

4. The method for assessing the remaining life of a roller press surface according to claim 1, characterized in that, The determination of model labels and model impact parameters based on real-time operational data and operational condition profiles includes: Extract the original thickness of the roller surface, the actual measured thickness, the cumulative amount of material processed, the working pressure of the roller press, and the roller gap deviation from the real-time operating data; Assign a trend label to the cumulative amount of material processed, and determine the model influence parameters corresponding to the trend label based on the cumulative amount of material processed. The equivalent wear thickness is determined based on the original thickness of the roller surface and the actual measured thickness. The wear uniformity index is determined based on the working pressure of the roller press and the roller gap deviation. The dynamic wear rate is determined based on the equivalent wear thickness and the wear uniformity index. The wear uniformity index and dynamic wear rate are assigned to state labels; the model influence parameters corresponding to the state labels are determined based on the wear uniformity index and dynamic wear rate.

5. The method for assessing the remaining life of a roller press surface according to claim 1, characterized in that, The method of generating remaining roll life and alarm signals based on a dual model of model labels, model influence parameters, and life prediction includes: Extract the roller press working pressure, roller gap deviation, and material type identifier from the real-time operating data; Calculate the similarity between the working pressure, roll gap deviation, and material type identification of the roller press and several working condition profiles; Extract the life prediction dual model corresponding to the working condition profile with the maximum similarity; the life prediction dual model includes a trend life prediction model and a state correction model. Input the model influence parameters corresponding to the model label into the corresponding lifetime prediction dual model to obtain the prediction results; The remaining life of the roller surface and alarm signals are generated based on the prediction results.

6. The method for assessing the remaining life of a roller press surface according to claim 5, characterized in that, The generation of remaining roll life and alarm signals based on the prediction results includes: Extract the predicted wear thickness YMH and the condition correction ΔH from the prediction results; Through formula Calculate the comprehensive equivalent wear thickness ZDMH; where α is a correction factor; The roller surface health degree (GJD) is calculated using a formula; the formula satisfies: Where GCH represents the initial thickness of the roll surface, This is expressed as the roller surface scrap threshold; Extracting model parameters from the trend life prediction model in the dual life prediction model. The model parameters This represents the slope parameter corresponding to the trend life prediction model under the current working condition profile; The remaining roller surface treatment capacity GSCL is calculated using a formula; the formula satisfies: ； The remaining life of the roller surface is determined based on the remaining amount of material that can be processed on the roller surface. An alarm signal is generated based on the health status of the roller surface.

7. The method for assessing the remaining life of a roller press surface according to claim 6, characterized in that, The correction factor is obtained through the following methods: Extract the standard deviation BC corresponding to the state correction in the prediction results; The correction factor is calculated using a formula; the formula satisfies: ;in, and These represent the upper and lower limits of the correction factor, respectively; the specific values ​​are set based on experience; t represents the attenuation coefficient, t>0; It is expressed as the reference standard deviation.

8. The method for assessing the remaining life of a roller press surface according to claim 6, characterized in that, The generation of alarm signals based on roller surface health includes: When the health status of the roller surface is less than D times the health threshold, an alarm signal for short roller surface life is generated. When the health status of the roller surface is greater than or equal to D times the health threshold, and the health status of the roller surface is less than the health threshold, an early warning signal for accelerated roller surface wear is generated; where D is a proportionality coefficient, D∈(0,1).

9. A system for assessing the remaining life of a roller press roller surface, characterized in that, include: Data acquisition module, data analysis module, and early warning module; The data acquisition module is connected to the data analysis module; the data analysis module is connected to the early warning module. The data acquisition module acquires historical and real-time operating data of the roller press through data acquisition equipment. The data analysis module: constructs a working condition profile based on historical operating data; A dual life prediction model is constructed based on operating condition profiles and historical operating data; model labels and model impact parameters are determined based on real-time operating data and operating condition profiles; and the remaining life of the roller surface and alarm signals are generated based on the model labels, model impact parameters, and the dual life prediction model. The early warning module provides corresponding prompts based on alarm signals.

10. A computer-readable storage medium applied to a roller press roll surface remaining life assessment system as described in claim 9, characterized in that, The computer program is stored on the computer-readable storage medium.

Images