A monitoring and optimization method for crusher working efficiency

Abstract

The present invention relates to the technical field of operating machinery, and in particular to a method for monitoring and optimizing the working efficiency of a crusher. The method comprises: obtaining monitoring data sequences corresponding to different monitoring data types, obtaining a time characteristic value of the monitoring data type based on the autocorrelation coefficient of the monitoring data sequence, obtaining a nonlinear characteristic value of the monitoring data type based on a fitting straight line and a Hurst exponent of the monitoring data sequence, obtaining a dimensional structure characteristic value of the monitoring data type based on the result of principal component analysis of the monitoring data sequence, obtaining a noise distribution characteristic value of the monitoring data type based on the result of smoothing the monitoring data sequence, obtaining a matching model of the monitoring data type at the current working efficiency monitoring moment based on the time characteristic value, the nonlinear characteristic value, the dimensional structure characteristic value, and the noise distribution characteristic value; monitoring the working efficiency of the crusher based on the matching model and the monitoring data sequence; and the present invention can improve the accuracy of monitoring the working efficiency of the crusher.

Description

Technical Field

[0001] The present invention relates to the technical field of operating machinery, and in particular to a method for monitoring and optimizing the working efficiency of a crusher. Background Art

[0002] Crusher is the core equipment in mining, building materials, metallurgy and other industries. Since its working efficiency directly affects the production capacity, energy consumption and operating costs of the production line, the working efficiency of the crusher is currently monitored during its operation.

[0003] In the existing process of monitoring the working efficiency of the crusher, the multidimensional data related to the working efficiency of the crusher is usually monitored and collected first, and then the prediction data is obtained based on the model and the multidimensional data segments related to the working efficiency of the crusher collected by monitoring, and then the predicted data is fused and analyzed to evaluate the working efficiency of the crusher; however, when performing data prediction, the model is usually randomly selected or a fixed model is used for prediction. Since data of different dimensions have different characteristics in different working scenarios, the existing model selection mode will result in the model used being incompatible with the data of the corresponding dimension, resulting in the selected model having poor ability to capture key features, which will lead to low accuracy of the prediction results or prediction errors. The effect is poor, and when the accuracy of the prediction results is low or the prediction effect is poor, the subsequent evaluation results of the crusher work efficiency will be less accurate. For example, if the data segment of a certain dimension has a strong nonlinear feature, if the model used to obtain the prediction data of this dimension is a linear regression model, the nonlinear trend in the data segment of this dimension cannot be effectively represented by the linear regression model, and the linear regression model's fitting of the actual law is seriously distorted. At this time, the reliability and accuracy of the prediction data of this dimension obtained based on the linear regression model are both low, which affects the subsequent results of the crusher work efficiency evaluation. Therefore, how to select models for data of different dimensions to improve the accuracy of the subsequent crusher work efficiency evaluation has become an urgent problem to be solved. Summary of the Invention

[0004] In order to solve the above problems, the present invention provides a method for monitoring and optimizing the working efficiency of a crusher. The technical solutions adopted are as follows:

[0005] One embodiment of the present invention provides a method for monitoring and optimizing the working efficiency of a crusher, comprising the following steps:

[0006] Acquire monitoring data sequences corresponding to different monitoring data types at the current working efficiency monitoring moment of the crusher;

[0007] For any monitoring data type, according to the autocorrelation coefficient of the monitoring data sequence corresponding to the monitoring data type, the time characteristic value of the monitoring data type at the current work efficiency monitoring moment is obtained; according to the fitting straight line and Hurst exponent of the monitoring data sequence corresponding to the monitoring data type, the nonlinear characteristic value of the monitoring data type at the current work efficiency monitoring moment is obtained; according to the result of principal component analysis of the monitoring data sequence corresponding to the monitoring data type, the dimensional structure characteristic value of the monitoring data type at the current work efficiency monitoring moment is obtained; according to the result of smoothing the monitoring data sequence corresponding to the monitoring data type, the noise distribution characteristic value of the monitoring data type at the current work efficiency monitoring moment is obtained; according to the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value, a matching model of the monitoring data type at the current work efficiency monitoring moment is obtained;

[0008] The working efficiency of the crusher is monitored according to the matching model and the monitoring data sequence.

[0009] Beneficial effects: The present invention first obtains monitoring data sequences corresponding to different monitoring data types at the current working efficiency monitoring moment of the crusher; then, according to the autocorrelation coefficient of the monitoring data sequence corresponding to the monitoring data type, obtains the time characteristic value of the monitoring data type at the current working efficiency monitoring moment; according to the fitting straight line and Hurst index of the monitoring data sequence corresponding to the monitoring data type, obtains the nonlinear characteristic value of the monitoring data type at the current working efficiency monitoring moment; according to the result of principal component analysis of the monitoring data sequence corresponding to the monitoring data type, obtains the dimensional structure characteristic value of the monitoring data type at the current working efficiency monitoring moment; according to the result of smoothing the monitoring data sequence corresponding to the monitoring data type, obtains the noise distribution characteristic value of the monitoring data type at the current working efficiency monitoring moment; according to the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value, obtains the matching model of the monitoring data type at the current working efficiency monitoring moment; finally, the working efficiency of the crusher is monitored according to the matching model and the monitoring data sequence; and the present invention improves the matching degree between the model and the monitoring data type through the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value, thereby improving the accuracy of monitoring the working efficiency of the crusher. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. 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 any creative work.

[0011] Figure 1 The present invention is a flow chart of a method for monitoring and optimizing the working efficiency of a crusher. DETAILED DESCRIPTION

[0012] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field fall within the scope of protection of the embodiments of the present invention.

[0013] 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 pertains.

[0014] This embodiment provides a method for monitoring and optimizing the working efficiency of a crusher, which is described in detail as follows:

[0015] like Figure 1 As shown, the method for monitoring and optimizing the working efficiency of the crusher includes the following steps:

[0016] Step S001: acquiring monitoring data sequences corresponding to different monitoring data types at the current working efficiency monitoring moment of the crusher.

[0017] The purpose of this embodiment is to select a model for the corresponding dimension based on the characteristics of the collected data segments of different dimensions related to the working efficiency of the crusher, so as to ensure the accuracy and reliability of subsequent data prediction, and then ensure the accuracy of the subsequent evaluation of the crusher's working efficiency based on the predicted data, that is, the model in this embodiment is mainly used for data prediction; and for the convenience of analysis, this embodiment will subsequently monitor the working efficiency of the crusher in any working process, that is, the crushers that appear subsequently are all the same crusher, and the monitoring data that appear subsequently are all data related to the working efficiency of the crusher.

[0018] This embodiment first obtains the monitoring data type related to the working efficiency of the crusher, and records the set constructed by the monitoring data type related to the working efficiency of the crusher as the monitoring data type set, and the monitoring data type related to the working efficiency of the crusher is generally divided into four categories, including operation performance monitoring data, energy efficiency monitoring data, product quality monitoring data and auxiliary monitoring data. Operation performance monitoring data generally refers to the bearing vibration data and gearbox vibration data of the crusher monitored by the vibration sensor, the motor temperature data and bearing temperature data of the crusher monitored by the temperature sensor, the pressure of the hydraulic system of the crusher monitored by the pressure sensor, the speed of the main shaft of the crusher monitored by the speed sensor, the load rate of the crusher monitored by the current sensor or power analyzer, the weight of the material crushed by the crusher monitored by the belt scale, the level meter or the flow sensor, etc. The energy efficiency data generally refers to the crushing The motor power of the crusher, the material filling rate in the crushing chamber of the crusher monitored by the laser radar or pressure sensor, etc. The product quality data generally refers to the particle size distribution data monitored by the linear particle size analyzer or laser diffraction method installed at the discharge port of the crusher, the weight of the crushed product monitored by the weighing sensor, etc. The particle size distribution is also called the proportion of particles of different particle sizes in the crushed product in the total material. The auxiliary monitoring data generally refers to the ambient dust concentration of the crusher when it is working monitored by the dust concentration sensor, the ambient noise data of the crusher when it is working monitored by the sound level meter, etc.; therefore, through the above analysis, it can be seen that the operating performance monitoring data, energy efficiency monitoring data, product quality monitoring data and auxiliary monitoring data are all composed of multiple monitoring data types, and the above-mentioned monitored data types can indirectly or directly reflect the working efficiency of the crusher. For example, the weight of the material crushed by the crusher monitored by the level meter or flow sensor can reflect the production efficiency of the crusher.

[0019] To facilitate subsequent analysis and understanding, this embodiment will be described below by taking the model selection process or model matching process of any monitoring data type A in the monitoring data type set at the current work efficiency monitoring moment as an example. For example, the model matching process of the ambient dust concentration at the current work efficiency monitoring moment can be described as an example. Before obtaining the matching model of the ambient dust concentration at the current work efficiency monitoring moment, it is necessary to first obtain the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment of the crusher. The specific acquisition process is as follows:

[0020] First, the time period from the last crusher work efficiency monitoring moment to the current work efficiency monitoring moment is obtained and recorded as the current monitoring time period. Then, all data belonging to monitoring data type A collected during the current monitoring time period is obtained, and the time series constructed by all data belonging to monitoring data type A collected during the current monitoring time period is recorded as the sequence to be processed. Then, the sequence to be processed is preprocessed, and the preprocessed sequence is recorded as the monitoring data sequence corresponding to monitoring data type A at the current work efficiency monitoring moment. In addition, the time interval between adjacent work efficiency monitoring moments is generally set by the equipment operation characteristics, generally set to 30 minutes to 2 hours, and the adjacent time interval for collecting monitoring data is generally set by the performance and type of the collection equipment, but it is required that the adjacent time interval for collecting monitoring data is much smaller than the time interval between adjacent work efficiency monitoring moments. For example, for the monitoring data type collected by sensors, the adjacent time interval for collecting monitoring data by sensors can be set to 1 second. The method for obtaining the monitoring data sequence corresponding to other monitoring data types at the current work efficiency monitoring moment is the same as the method for obtaining the monitoring data sequence corresponding to monitoring data type A at the current work efficiency monitoring moment.

[0021] In this embodiment, the process of preprocessing the sequence to be processed includes but is not limited to steps such as filtering, normalization, data cleaning, correction, and missing value processing of the sequence to be processed, such as using wavelet transform, Kalman filtering and other technologies to eliminate high-frequency interference and random noise in the sequence, filling missing values caused by interruption of sensor data transmission through interpolation, or using historical mean values to fill long-term missing data, etc., and the process of preprocessing data is a well-known technology, so this embodiment will not be described in detail.

[0022] Therefore, this embodiment can obtain the monitoring data sequence corresponding to the monitoring data type A at the current working efficiency monitoring moment of the crusher through the above process.

[0023] Step S002: For any monitoring data type, according to the autocorrelation coefficient of the monitoring data sequence corresponding to the monitoring data type, obtain the time characteristic value of the monitoring data type at the current work efficiency monitoring moment; according to the fitting straight line and Hurst exponent of the monitoring data sequence corresponding to the monitoring data type, obtain the nonlinear characteristic value of the monitoring data type at the current work efficiency monitoring moment; according to the result of principal component analysis of the monitoring data sequence corresponding to the monitoring data type, obtain the dimensional structure characteristic value of the monitoring data type at the current work efficiency monitoring moment; according to the result of smoothing the monitoring data sequence corresponding to the monitoring data type, obtain the noise distribution characteristic value of the monitoring data type at the current work efficiency monitoring moment; according to the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value, obtain a matching model of the monitoring data type at the current work efficiency monitoring moment.

[0024] Since the current model matching mode may not match the data of the corresponding dimension or data type, the matched model may have poor ability to capture key features, that is, the mismatch or low matching phenomenon will directly weaken the model's ability to capture key features, which will lead to low accuracy of the prediction results. When the accuracy of the prediction results is low, it will have a negative impact on the subsequent evaluation of the crusher's working efficiency. For example, the vibration signal of the crusher changes significantly over time, has time series characteristics or has strong time dependence. If the subsequent matched model is more suitable for low time dependence, then the subsequent prediction of the crusher's vibration signal will be ineffective. The result is poor. If the monitoring data sequence corresponding to a certain monitoring data type at the current working efficiency monitoring moment of the crusher has a strong nonlinear feature, if the model used when obtaining the prediction data of the monitoring data type is a linear regression model, the nonlinear trend of the monitoring data sequence corresponding to the monitoring data type cannot be effectively represented by the linear regression model, and the linear regression model's fitting of the actual law is seriously distorted. At this time, the reliability and accuracy of the prediction data of the monitoring data type obtained based on the linear regression model are both low. In order to ensure the accuracy of the prediction, this embodiment will analyze the characteristics of the monitoring data type at the working efficiency monitoring moment, and perform model matching based on the analysis results, so as to To ensure the accuracy of subsequent evaluation or monitoring of the crusher's working efficiency, that is, this embodiment will first analyze and obtain the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value of the monitoring data type A at the current working efficiency monitoring moment, and then obtain the matching model of the monitoring data type A at the current working efficiency monitoring moment based on the obtained time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value of the monitoring data type A at the current working efficiency monitoring moment. The reason for analyzing the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value is because different monitoring data types related to the crusher's working efficiency exhibit different characteristics. The same, but different characteristic performances are adapted to different models. For example, the wear, lubrication, feeding speed and other data of the crusher change with time, so the wear, lubrication, feeding speed and other data of the crusher have obvious time dependence, then the wear, lubrication, feeding speed and other data of the crusher are more adapted to the time series model, the electrode speed, feeding particle size, material hardness and other data of the crusher have strong nonlinear characteristics, then the electrode speed, feeding particle size, material hardness and other data of the crusher are more adapted to the deep learning model; then the specific process of obtaining the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring time in this embodiment is:

[0025] In this embodiment, it is necessary to first obtain the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment based on the autocorrelation coefficient of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment. The time characteristic value of the monitoring data type A at the current work efficiency monitoring moment is an important parameter for subsequently determining the matching model of the monitoring data type A at the current work efficiency monitoring moment. The specific process of obtaining the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment is as follows:

[0026] First, a preset lag period value interval of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is obtained, and recorded as the first interval. In this embodiment, the minimum value of the first interval is 1, and the maximum value of the first interval is T. The value of T generally does not exceed 25% of the total number of monitoring data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment. For example, in this embodiment, T can be set to an upward rounded value of 10% of the total number of monitoring data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment; then, the absolute values of the autocorrelation coefficients of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment at each lag period in the first interval are obtained, and are all recorded as the autocorrelation coefficients to be selected corresponding to the monitoring data type A at the current work efficiency monitoring moment; then, among all the autocorrelation coefficients to be selected corresponding to the monitoring data type A at the current work efficiency monitoring moment, the largest autocorrelation coefficient to be selected is selected as the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment. The process of obtaining the autocorrelation coefficient of any sequence at any lag period is a well-known technology, so this embodiment will not be described in detail. In addition, the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment can reflect the strength of the time dependence of the data changes in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment, and when the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment is larger, it indicates that the data changes in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment have stronger time dependence, or the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment have stronger time dependence characteristics. When the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment is smaller, it indicates that the data changes in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment have weaker time dependence, or the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment have weaker time dependence characteristics. Strong time dependence characteristics refer to data changes over time, such as the vibration signal of the crusher changes dramatically over time. Strong time dependence characteristics refer to data changes less over time, such as static parameters.

[0027] After obtaining the time characteristic value, the nonlinear characteristic value of the monitoring data type A at the current work efficiency monitoring moment is obtained based on the fitting straight line and Hurst exponent of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment. The nonlinear characteristic value of the monitoring data type A at the current work efficiency monitoring moment is an important parameter for subsequently determining the matching model of the monitoring data type A at the current work efficiency monitoring moment. The specific process of obtaining the nonlinear characteristic value of the monitoring data type A at the current work efficiency monitoring moment is as follows:

[0028] First, the Hurst index of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is obtained by using the rescaled range analysis method, and the result of subtracting the Hurst index of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment from the preset constant is used as the first indicator value, and the process of obtaining the Hurst index is a well-known calculation factor, which will not be described in detail in this embodiment; then, a mapping space of the monitoring data type A is constructed, and the horizontal axis of the mapping space of the monitoring data type A is time, and the vertical axis is the data value belonging to the monitoring data type A, and then each monitoring data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment and the collection time of each monitoring data are mapped to the mapping space of the monitoring data type A, and the mapping data points of each monitoring data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are obtained, that is, the horizontal axis value of the mapping data point of the monitoring data is the collection time of the corresponding monitoring data, and the vertical axis value is the value of the corresponding monitoring data; then, a straight line fitting is performed on all the mapping data points of the monitoring data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment, and the straight line obtained by fitting is recorded as the first straight line, and the least squares method can be selected for Line fitting is performed; then, according to the mapping data points of each monitoring data in the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment and the first straight line, a second index value is obtained, and the process of obtaining the second index value is as follows: on the first straight line, a data point with the same horizontal coordinate value as the mapping data point of the monitoring data in the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is obtained, and recorded as the fitting data point of the corresponding monitoring data, that is, the horizontal coordinate value of the fitting data point of any monitoring data is the same as the horizontal coordinate value of the mapping data point of the monitoring data, and the fitting data point of the monitoring data is located on the first straight line, and then the Euclidean distance between the fitting data point of each monitoring data in the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment and the mapping data point of the corresponding monitoring data is obtained, and recorded as the fitting distance of the corresponding monitoring data, and then the normalized value of the mean of the fitting distances of all monitoring data in the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is obtained, and recorded as the second index value, the normalization process here uses the normalization function Norm(); finally, the first index value and the second index value are weighted summed, and the result of the weighted summation is used as the nonlinear characteristic value of the monitoring data type A at the current work efficiency monitoring moment. The specific expression for obtaining the nonlinear characteristic value of the monitoring data type A at the current work efficiency monitoring moment is:

[0029] Among them, F is the nonlinear characteristic value of the monitoring data type A at the current work efficiency monitoring moment, H is the Hurst index of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment, is the first weight value, is the second weight value, M is the total number of monitoring data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment, is the fitting distance of the bottom m monitoring data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment, Norm() is the normalization function, C is the preset constant, and the value of the preset constant in this embodiment needs to be determined based on the maximum value of the Hurst exponent. Under normal circumstances, the maximum value of the Hurst exponent is 1, so the value of the preset constant in this embodiment is 1.

[0030] In addition, when the value of H is closer to 0, it indicates that the anti-persistence is stronger or the nonlinear correlation characteristics of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are more obvious, so The larger the value is, the more obvious the nonlinear correlation characteristics of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are. When the value of F is larger, it also indicates that the nonlinear characteristics of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are more obvious; then based on the above description, it can be seen that when the value of F is larger, it indicates that the nonlinear characteristics of the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are more obvious, or the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has a stronger nonlinear characteristic; when the value of F is smaller, it indicates that the nonlinear characteristics of the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are less obvious, or the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has a weaker nonlinear characteristic; and since the Hurst exponent has better robustness in different scenarios, this embodiment requires Greater than , such as you can set 0.7, is 0.3.

[0031] After obtaining the nonlinear characteristic value, the result of principal component analysis is performed on the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment to obtain the dimensional structure characteristic value of the monitoring data type A at the current work efficiency monitoring moment. The dimensional structure characteristic value of the monitoring data type A at the current work efficiency monitoring moment is an important parameter for subsequently determining the matching model of the monitoring data type A at the current work efficiency monitoring moment, and the specific process of obtaining the dimensional structure characteristic value of the monitoring data type A at the current work efficiency monitoring moment is as follows:

[0032] First, principal component analysis is used to perform principal component analysis on the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment, and the sequence constructed by all principal components of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment obtained by principal component analysis is recorded as the sequence to be analyzed, and the principal components in the sequence to be analyzed are arranged according to the variance of the principal components. The variance of the principal components is the eigenvalue corresponding to the corresponding principal component, that is, the variance of the i-th principal component in the sequence to be analyzed is not greater than the variance of the i-1-th principal component in the sequence to be analyzed, and the variance of the i-th principal component in the sequence to be analyzed is not less than the variance of the i+1-th principal component in the sequence to be analyzed, i Greater than 1 and less than N, where N is the total number of principal components in the sequence to be analyzed; then the variance contribution rate of each principal component in the sequence to be analyzed is obtained, and the variance contribution rate of the i-th principal component in the sequence to be analyzed is the ratio of the variance of the i-th principal component to the cumulative sum of the variances of all principal components in the sequence to be analyzed; and if the cumulative sum of the variance contribution rates of the first K principal components in the sequence to be analyzed is greater than the preset cumulative contribution rate threshold, and the cumulative sum of the variance contribution rates of the first K-1 principal components in the sequence to be analyzed is not greater than the preset cumulative contribution rate threshold, then the normalized value of K is used as the dimensional structure characteristic value of the monitoring data type A at the current work efficiency monitoring moment, and the normalized value of K refers to the ratio of K to N.

[0033] In addition, when the dimensional structure characteristic value of the monitoring data type A at the current work efficiency monitoring moment is larger, it indicates that the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has more high-dimensional complex structural characteristics. If the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is multi-sensor fusion data, then the high-dimensional complex structural characteristics of the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are more obvious. When the dimensional structure characteristic value of the monitoring data type A at the current work efficiency monitoring moment is smaller, it indicates that the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has more high-dimensional complex structural characteristics. The more the data has low-dimensional complex structural characteristics, if the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is single-sensor monitoring data, then the low-dimensional complex structural characteristics of the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are more obvious; in this embodiment, the implementer needs to set the preset cumulative contribution rate threshold according to the actual situation such as the accuracy of the analysis dimension structure or conventional experience. For example, the preset cumulative contribution rate threshold can be set to the empirical value of 0.8, that is, it is generally believed that when the cumulative result of the variance contribution rate of the first K principal components is greater than or equal to 0.8, it indicates that the first K principal components retain the main information of the original data.

[0034] After obtaining the dimensional structure characteristic value, the noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment is obtained by smoothing the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment. The noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment is an important parameter for subsequently determining the matching model of the monitoring data type A at the current work efficiency monitoring moment. The specific process of obtaining the noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment is as follows:

[0035] First, the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is smoothed using the moving average method, and the sequence obtained by smoothing is recorded as the smoothed sequence to be analyzed, and the data in the smoothed sequence to be analyzed are all recorded as smoothed data, and the process of smoothing using the moving average method is well known; then, the residual sequence is obtained based on the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment and the smoothed sequence to be analyzed, and the a-th residual in the residual sequence is the result of subtracting the a-th monitoring data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment from the a-th smoothed data in the smoothed sequence to be analyzed, and the normalized value of the mean of the residual sequence is obtained, and recorded as the noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment. The normalization process here also uses the normalization function Norm(). And when the noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment is larger, it indicates that the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has a high noise characteristic, and the data in the monitoring data sequence has a high noise characteristic means that the interference encountered when the data in the monitoring data sequence is collected is greater. When the noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment is smaller, it indicates that the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has a low noise characteristic, and the data in the monitoring data sequence has a low noise characteristic means that the interference encountered when the data in the monitoring data sequence is collected is smaller.

[0036] After obtaining the noise distribution characteristic value, the matching model of the monitoring data type A at the current work efficiency monitoring moment is obtained based on the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment. The specific acquisition process of the matching model of the monitoring data type A at the current work efficiency monitoring moment is:

[0037] First, the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked according to the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment. The specific marking process is as follows:

[0038] If it is judged that the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment is greater than the preset strong time characteristic threshold, then it is judged that the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has a strong time dependence characteristic, and therefore the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked as a strong time dependence characteristic; if it is judged that the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment is less than the preset weak time characteristic threshold, then it is judged that the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has a weak time dependence characteristic, and therefore the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked as a weak time dependence characteristic; if it is judged that the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment belongs to the interval formed by the preset strong time characteristic threshold and the preset weak time characteristic threshold, then it indicates that the time dependence strength and weakness characteristics of the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are relatively not obvious, so at this time, the strength and weakness of the time dependence characteristics of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are not marked.

[0039] If it is determined that the nonlinear characteristic value of monitoring data type A at the current work efficiency monitoring moment is greater than a preset strong nonlinear characteristic threshold, then the data in the monitoring data sequence corresponding to monitoring data type A at the current work efficiency monitoring moment is determined to have a strong nonlinear characteristic, and the monitoring data sequence corresponding to monitoring data type A at the current work efficiency monitoring moment is marked as having a strong nonlinear characteristic. If it is determined that the nonlinear characteristic value of monitoring data type A at the current work efficiency monitoring moment is less than a preset weak nonlinear characteristic threshold, then the data in the monitoring data sequence corresponding to monitoring data type A at the current work efficiency monitoring moment is determined to have a weak nonlinear characteristic, and the monitoring data sequence corresponding to monitoring data type A at the current work efficiency monitoring moment is marked as having a weak nonlinear characteristic. If it is determined that the time characteristic value of monitoring data type A at the current work efficiency monitoring moment falls within the interval formed by the preset strong nonlinear characteristic threshold and the preset weak nonlinear characteristic threshold, then it indicates that the nonlinear strength and weakness characteristics of the data in the monitoring data sequence corresponding to monitoring data type A at the current work efficiency monitoring moment are relatively insignificant, and in this case, the strength and weakness of the nonlinear characteristic of the monitoring data sequence corresponding to monitoring data type A at the current work efficiency monitoring moment are not marked.

[0040] If it is determined that the dimensional structure characteristic value of the monitoring data type A at the current work efficiency monitoring moment is greater than the preset high-dimensional structure characteristic threshold, then it is determined that the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has high-dimensional complex structure characteristics, and the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked as high-dimensional complex structure characteristics. If it is determined that the dimensional structure characteristic value of the monitoring data type A at the current work efficiency monitoring moment is less than the preset low-dimensional structure characteristic threshold, then it is determined that the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has low-dimensional complex structure characteristics, and the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked as low-dimensional complex structure characteristics. If it is determined that the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment falls within the interval formed by the preset high-dimensional structure characteristic threshold and the preset low-dimensional structure characteristic threshold, then it indicates that the dimensional structure characteristics of the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are relatively insignificant, and at this time, the high or low dimensional structure characteristics of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment are not marked.

[0041] If it is determined that the noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment is greater than the preset high noise distribution characteristic threshold, then it is determined that the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has a high noise characteristic, and the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked as a high noise characteristic. If it is determined that the noise distribution characteristic value of the monitoring data type A at the current work efficiency monitoring moment is less than the preset low noise distribution characteristic threshold, then it is determined that the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment has a low noise characteristic, and the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked as a low noise characteristic. If it is determined that the time characteristic value of the monitoring data type A at the current work efficiency monitoring moment falls within the interval formed by the preset high noise distribution characteristic threshold and the preset low noise distribution characteristic threshold, then it indicates that the noise characteristic of the data in the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is relatively insignificant, and at this time, the high or low noise characteristic of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is not marked.

[0042] In this embodiment, the implementer needs to set the preset strong time characteristic threshold, the preset weak time characteristic threshold, the preset strong nonlinear characteristic threshold, the preset weak nonlinear characteristic threshold, the preset high-dimensional structure characteristic threshold, the preset low-dimensional structure characteristic threshold, the preset high noise distribution characteristic threshold, and the preset low noise distribution characteristic threshold according to actual conditions. For example, in this embodiment, the preset strong time characteristic threshold and the preset weak time characteristic threshold can be set to 0.7 and 0.3 respectively, the preset strong nonlinear characteristic threshold and the preset weak nonlinear characteristic threshold can be set to 0.6 and 0.4 respectively, the preset high-dimensional structure characteristic threshold and the preset low-dimensional structure characteristic threshold can be set to 0.75 and 0.25 respectively, and the preset high noise distribution characteristic threshold and the preset low noise distribution characteristic threshold can be set to 0.8 and 0.2 respectively.

[0043] Since the models adapted to strong time-dependent characteristics are time series models, such as LSTM, GRU and other models, the models adapted to weak time-dependent characteristics are traditional machine learning models, such as XGBoost, random forest and other models, the models adapted to strong nonlinear characteristics are deep learning models, such as MLP, CNN and other models, the models adapted to weak nonlinear characteristics are traditional machine learning models, such as SVM, linear kernel and other models, the models adapted to high-dimensional complex structure characteristics are dimensionality reduction + deep learning models, such as PCA + CNN models, etc., the models adapted to low-dimensional complex structure characteristics are traditional machine learning models, such as SVM, decision tree models, etc., the models adapted to high-noise characteristics are robust models, such as ensemble learning, robust regression models, etc., and the models adapted to low-noise characteristics are lightweight models, such as linear regression models. Therefore, this embodiment will then obtain the labeling results of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment, and then it will be able to obtain the matching model of the monitoring data type A at the current work efficiency monitoring moment based on the labeling results of the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment and the models adapted to different characteristics. If the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked as a strong time-dependent characteristic, a strong nonlinear characteristic, a high-dimensional complex structure characteristic, or a high-noise characteristic, then the time series model, the deep learning model, the dimensionality reduction + deep learning model, and the robust model will be used as the matching model of the monitoring data type A at the current work efficiency monitoring moment. However, when subsequently obtaining the predicted data of the monitoring data type A, the current work efficiency monitoring moment will be selected. One of the matching models of the monitoring data type A at the current work efficiency monitoring moment can be used for data prediction. If the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is marked as weak time dependence characteristics, strong nonlinear characteristics, and high-dimensional complex structure characteristics, then the traditional machine learning model, deep learning model, and dimensionality reduction + deep learning model are used as the matching models of the monitoring data type A at the current work efficiency monitoring moment; in addition, if the monitoring data sequence corresponding to the monitoring data type A at the current work efficiency monitoring moment is not marked, then any one of the above models can be selected as the matching model of the monitoring data type A at the current work efficiency monitoring moment; and the matching model of the monitoring data type A at the current work efficiency monitoring moment in this embodiment has better prediction effect and prediction accuracy for the monitoring data type A.

[0044] In addition, LSTM is a special recurrent neural network, the GRU model (Gated Recurrent Unit) is an improved recurrent neural network (RNN), the XGBoost model (eXtreme Gradient Boosting) is a supervised learning integration model, the PCA+CNN model refers to a model that combines principal component analysis (PCA) and convolutional neural network (CNN), the MLP model (Multilayer Perceptron) is a feedforward neural network model, and CNN is a convolutional neural network (CNN).

[0045] Therefore, this embodiment obtains the matching model of the monitoring data type A at the current work efficiency monitoring moment through the above process.

[0046] Step S003: monitoring the working efficiency of the crusher according to the matching model and the monitoring data sequence.

[0047] After obtaining the matching model of each monitoring data type in the monitoring data type set at the current working efficiency monitoring moment, the working efficiency of the crusher is monitored according to the matching model of each monitoring data type at the current working efficiency monitoring moment and the monitoring data sequence corresponding to each monitoring data type at the current working efficiency monitoring moment, specifically:

[0048] The monitoring data sequence corresponding to each monitoring data type at the current work efficiency monitoring moment is input into the matching model of the corresponding monitoring data type at the current work efficiency monitoring moment, and the prediction data corresponding to the corresponding monitoring data type is output, and the output prediction data is the data between the current work efficiency monitoring moment and the next work efficiency monitoring moment, and then the work efficiency of the crusher is evaluated according to the obtained prediction data corresponding to the monitoring data type. For example, the vibration intensity, energy consumption ratio, particle size qualification rate and other index values can be calculated based on the prediction data, and then all the calculated index values are weighted and fused to obtain the work efficiency evaluation result of the crusher; and after obtaining the prediction data corresponding to the monitoring data type, the process of fusing and analyzing the prediction data to obtain the work efficiency evaluation result of the crusher is a well-known technology, so this embodiment will not be described in detail; in addition, the models mentioned above in this embodiment are all existing prediction models, and the process of performing data prediction and obtaining prediction data based on the existing prediction model is a well-known technology, so this embodiment will not be described in detail.

[0049] At this point, this embodiment has completed the monitoring of the crusher's working efficiency, and the model matching process in this embodiment can make the obtained prediction data more accurate and reliable, thereby improving the monitoring accuracy of the crusher's working efficiency; in addition, monitoring the crusher's working efficiency can not only timely adjust equipment parameters (such as feed speed, discharge port size), avoid overload or inefficient operation, and ensure that the equipment is always in the best working condition, but also optimize operating strategies and reduce electricity expenses by identifying high-energy consumption links (such as no-load operation and equipment aging).

[0050] To summarize, this embodiment first obtains monitoring data sequences corresponding to different monitoring data types at the current working efficiency monitoring moment of the crusher; then, based on the autocorrelation coefficient of the monitoring data sequence corresponding to the monitoring data type, obtains the time characteristic value of the monitoring data type at the current working efficiency monitoring moment; based on the fitting straight line and Hurst index of the monitoring data sequence corresponding to the monitoring data type, obtains the nonlinear characteristic value of the monitoring data type at the current working efficiency monitoring moment; based on the result of principal component analysis of the monitoring data sequence corresponding to the monitoring data type, obtains the dimensional structure characteristic value of the monitoring data type at the current working efficiency monitoring moment; based on the result of smoothing the monitoring data sequence corresponding to the monitoring data type, obtains the noise distribution characteristic value of the monitoring data type at the current working efficiency monitoring moment; based on the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value, obtains a matching model of the monitoring data type at the current working efficiency monitoring moment; finally, the working efficiency of the crusher is monitored according to the matching model and the monitoring data sequence; and this embodiment improves the matching degree between the model and the monitoring data type through the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value, thereby improving the accuracy of monitoring the working efficiency of the crusher.

[0051] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for monitoring and optimizing the working efficiency of a crusher, characterized in that: The method comprises the following steps: Acquire monitoring data sequences corresponding to different monitoring data types at the current working efficiency monitoring moment of the crusher; For any monitoring data type, according to the autocorrelation coefficient of the monitoring data sequence corresponding to the monitoring data type, the time characteristic value of the monitoring data type at the current work efficiency monitoring moment is obtained; according to the fitting straight line and Hurst exponent of the monitoring data sequence corresponding to the monitoring data type, the nonlinear characteristic value of the monitoring data type at the current work efficiency monitoring moment is obtained; according to the result of principal component analysis of the monitoring data sequence corresponding to the monitoring data type, the dimensional structure characteristic value of the monitoring data type at the current work efficiency monitoring moment is obtained; according to the result of smoothing the monitoring data sequence corresponding to the monitoring data type, the noise distribution characteristic value of the monitoring data type at the current work efficiency monitoring moment is obtained; according to the time characteristic value, nonlinear characteristic value, dimensional structure characteristic value, and noise distribution characteristic value, a matching model of the monitoring data type at the current work efficiency monitoring moment is obtained; monitoring the working efficiency of the crusher according to the matching model and the monitoring data sequence; A method for obtaining the nonlinear characteristic value of the monitoring data type at the current work efficiency monitoring moment includes: taking the result of subtracting the Hurst exponent of the monitoring data sequence corresponding to the monitoring data type from a preset constant as a first indicator value; constructing a mapping space for the monitoring data type, wherein the horizontal axis of the mapping space is time and the vertical axis is a data value consistent with the monitoring data type, mapping each monitoring data in the monitoring data sequence corresponding to the monitoring data type and the acquisition time of each monitoring data to the mapping space of the monitoring data type, obtaining the mapping data point of each monitoring data in the monitoring data sequence corresponding to the monitoring data type, performing straight line fitting on all the mapping data points of the monitoring data in the monitoring data sequence corresponding to the monitoring data type, and recording the straight line obtained by fitting as a first straight line; obtaining a second indicator value based on the mapping data points of each monitoring data in the monitoring data sequence corresponding to the monitoring data type and the first straight line, and taking the weighted sum of the first indicator value and the second indicator value as the time characteristic value of the monitoring data type at the current work efficiency monitoring moment; The method for obtaining the second indicator value includes: recording the data point on the first straight line with the same horizontal coordinate value as the mapping data point of the monitoring data as the fitting data point of the corresponding monitoring data, recording the distance between the fitting data point of the monitoring data and the mapping data point of the corresponding monitoring data as the fitting distance of the corresponding monitoring data, and recording the normalized value of the mean of the fitting distances of all monitoring data in the monitoring data sequence corresponding to the monitoring data type as the second indicator value.

2. The method for monitoring and optimizing the working efficiency of a crusher according to claim 1, wherein: The method for obtaining the time characteristic value of the monitoring data type at the current work efficiency monitoring time includes: Obtain a preset lag period value interval of the monitoring data sequence corresponding to the monitoring data type, and record it as the first interval; obtain the absolute value of the autocorrelation coefficient of the monitoring data sequence corresponding to the monitoring data type at each lag period in the first interval, and record them as the autocorrelation coefficients to be selected corresponding to the monitoring data type at the current work efficiency monitoring moment; among all the autocorrelation coefficients to be selected corresponding to the monitoring data type at the current work efficiency monitoring moment, select the largest autocorrelation coefficient to be selected as the time characteristic value of the monitoring data type at the current work efficiency monitoring moment.

3. The method for monitoring and optimizing the working efficiency of a crusher according to claim 1, wherein: The method for obtaining the dimensional structure characteristic value of the monitoring data type at the current work efficiency monitoring moment includes: The sequence constructed by all principal components obtained by performing principal component analysis on the monitoring data sequence corresponding to the monitoring data type using the principal component analysis method is recorded as the sequence to be analyzed, and the variance contribution rate of each principal component in the sequence to be analyzed is obtained. The principal components in the sequence to be analyzed are arranged in descending order according to the principal component variance; if the cumulative sum of the variance contribution rates of the first K principal components in the sequence to be analyzed is greater than the preset cumulative contribution rate threshold, and the cumulative sum of the variance contribution rates of the first K-1 principal components in the sequence to be analyzed is not greater than the preset cumulative contribution rate threshold, then the normalized value of K is used as the dimensional structure characteristic value of the monitoring data type at the current work efficiency monitoring moment.

4. The method for monitoring and optimizing the working efficiency of a crusher according to claim 1, wherein: The method for obtaining the noise distribution characteristic value of the monitoring data type at the current work efficiency monitoring moment includes: The sequence obtained by smoothing the monitoring data sequence corresponding to the monitoring data type is recorded as the smoothed sequence to be analyzed. The residual sequence is obtained based on the monitoring data sequence corresponding to the monitoring data type and the smoothed sequence to be analyzed. The normalized value of the mean of the residual sequence is recorded as the noise distribution characteristic value of the monitoring data type at the current work efficiency monitoring moment.

5. The method for monitoring and optimizing the working efficiency of a crusher according to claim 1, characterized in that: The method for matching the monitoring data type with the model at the current work efficiency monitoring moment includes: If the time characteristic value of the monitoring data type is greater than the preset strong time characteristic threshold, the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is marked as a strong time dependence characteristic; if the time characteristic value of the monitoring data type is less than the preset weak time characteristic threshold, the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is marked as a weak time dependence characteristic; if the nonlinear characteristic value of the monitoring data type is greater than the preset strong nonlinear characteristic threshold, the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is marked as a strong nonlinear characteristic; if the nonlinear characteristic value of the monitoring data type is less than the preset weak nonlinear characteristic threshold, the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is marked as a weak nonlinear characteristic; if the monitoring data If the dimensional structure characteristic value of the monitoring data type is greater than the preset high dimensional structure characteristic threshold, the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is marked as a high dimensional complex structure characteristic; if the dimensional structure characteristic value of the monitoring data type is less than the preset low dimensional structure characteristic threshold, the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is marked as a low dimensional complex structure characteristic; if the noise distribution characteristic value of the monitoring data type is greater than the preset high noise distribution characteristic threshold, the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is marked as a high noise characteristic; if the noise distribution characteristic value of the monitoring data type is less than the preset low noise distribution characteristic threshold, the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment is marked as a low noise characteristic; A matching model of the monitoring data type at the current work efficiency monitoring moment is obtained according to the labeling result of the monitoring data sequence corresponding to the monitoring data type at the current work efficiency monitoring moment.

6. The method for monitoring and optimizing the working efficiency of a crusher according to claim 1, wherein: The method for monitoring the working efficiency of a crusher according to the matching model and the monitoring data sequence comprises: The monitoring data sequence corresponding to each monitoring data type at the current work efficiency monitoring moment is input into the matching model of the corresponding monitoring data type at the current work efficiency monitoring moment, and the prediction data corresponding to the corresponding monitoring data type is output. The work efficiency of the crusher is evaluated and monitored based on the prediction data.

Images