A coal mill fault early warning method and system based on a data model and a storage medium

Abstract

The application provides a kind of based on data model evaluation coal mill fault early warning method, system and storage medium, analyze and determine the characteristic parameter reflecting the operation performance of coal mill, according to the characteristic parameter reflecting the operation performance of coal mill determined in advance, the history data of collection is carried out data preprocessing to the history data set collected, eliminate the abnormal data therein, as the input data of state matrix extraction, state matrix is extracted;Collect the real-time operation data of each characteristic parameter of coal mill, according to the real-time operation data of each characteristic parameter of coal mill and the vector in each typical state in state matrix, the real-time estimated value of each characteristic parameter is calculated, the analysis of the degradation degree of each characteristic parameter is carried out, the state of the characteristic parameter of coal mill is converted into the reliability evaluation index of coal mill, according to the change of reliability evaluation index trend, determine the area of sending early warning signal, this method is simple, engineering applicability is strong, intelligent realizes the reliability monitoring and fault early warning prompt of coal mill.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power plant intelligence, and particularly relates to a coal mill fault early warning method and system based on a data model and a storage medium. BACKGROUND

[0002] The pulverizing system is an important component of a thermal power unit, and its operation directly affects the stability of boiler combustion and the safety of unit load. Since the working environment of the coal mill is relatively harsh, impurity wear, damp coal blocking and the like can cause faults of the coal mill. The fault features are mainly embodied in monitoring parameters such as temperature, air pressure and ventilation volume. For example, coal blocking of the coal mill can cause a sharp decrease in ventilation volume, and the moisture of raw coal has a direct relationship with the outlet temperature. Therefore, the development of a coal mill monitoring and evaluation technology based on a data model has great practical significance for the safety production of the unit.

[0003] At present, various data modeling methods have been applied to the related reports of coal mill fault early warning and diagnosis, but in the actual application process, there are technical problems such as overly complex models and poor engineering applicability. SUMMARY

[0004] In order to solve the technical problems such as overly complex models and poor engineering applicability of the data modeling method in the actual application of the coal mill fault early warning and diagnosis process, the present application provides a coal mill fault early warning method and system based on a data model and a storage medium.

[0005] In order to achieve the above purpose, the technical scheme of the present application is as follows:

[0006] The present application provides a coal mill fault early warning method based on a data model, comprising the following steps:

[0007] S1. Analyzing and determining characteristic parameters reflecting the operation performance of the coal mill, collecting historical data according to the pre-determined characteristic parameters reflecting the operation performance of the coal mill, performing data preprocessing on the collected historical data set, eliminating abnormal data therein as input data for state matrix extraction, and extracting a state matrix;

[0008] S2. Collecting real-time operation data of each characteristic parameter of the coal mill, calculating real-time estimated values of each characteristic parameter according to the real-time operation data of each characteristic parameter of the coal mill and the vectors in each typical state in the state matrix, analyzing the degradation degree of each characteristic parameter, converting the state of the coal mill characteristic parameter into a reliability evaluation index of the coal mill, determining the area of the early warning signal according to the change trend of the reliability evaluation index, and realizing the reliability monitoring and fault early warning of the coal mill.

[0009] The application provides a coal mill fault early warning method based on a data model.

[0010] As a preferred technical solution, the step S1 of analyzing and determining the characteristic parameters reflecting the operation performance of the coal mill comprises the following steps.

[0011] The characteristic parameters reflecting the operation performance of the coal mill are analyzed by using expert experience and / or a correlation analysis method, and the characteristic parameters reflecting the operation performance of the coal mill include any one or several of a coal supply amount, a coal mill current, an inlet air pressure, an inlet air temperature, an inlet air flow, an outlet mixture pressure, an outlet mixture temperature, a cold air adjusting door opening degree and a hot air adjusting door opening degree.

[0012] As a preferred technical solution, the step S1 of collecting historical data according to the characteristic parameters reflecting the operation performance of the coal mill comprises the following steps.

[0013] The historical data are collected in the entire working condition interval of the coal mill, and the working conditions are uniformly distributed in the entire working condition interval.

[0014] As a preferred technical solution, the step S1 of performing data preprocessing on the collected historical data set to remove abnormal data in the historical data set as input data of state matrix extraction, and extracting a state matrix, specifically comprises the following steps.

[0015] The collected historical data set is subjected to data preprocessing to remove abnormal data in the historical data set.

[0016] The processed data set is subjected to clustering by using a data clustering method to obtain a typical working condition library representing each typical working condition of the coal mill, and a state matrix is extracted by using a clustering algorithm, and the state matrix is as follows.

[0017]

[0018] Wherein, the element x ij is historical operation data of a certain parameter, i represents a time, and j represents a parameter.

[0019] As a preferred technical solution, the step S2 comprises the following steps.

[0020] The step S201 of collecting real-time operation data of each characteristic parameter, and calculating real-time estimated values of each characteristic parameter according to the collected real-time operation data of each characteristic parameter and vectors in each typical state in the state matrix, specifically comprises the following steps.

[0021] The real-time running data of each characteristic parameter is collected, and the weight of each typical state vector in the state matrix and the real-time observation vector is calculated by using kernel function mapping, and the calculation formula is:

[0022]

[0023] Wherein: x obs represents the real-time observation vector, i.e. the measured value, x i represents the historical state vector of the i-th row in the state matrix, h is the selected model parameter, which is determined according to the actual situation;‖x obs -x i ‖ 2 is the kernel function distance operator of two vectors, and the calculation formula of the kernel function distance operator of two vectors is:

[0024] ‖x obs -x i ‖ 2 =(x obs -x i ) T S -1 (x obs -x i )

[0025] In the formula, is the covariance matrix, wherein σ j is the standard deviation of the j-th parameter, which is obtained by statistical measurement or instrument design value, the real-time estimated value of each characteristic parameter is calculated, and the calculation formula is:

[0026]

[0027] As a preferred technical solution, step S2 comprises the following steps:

[0028] S202, according to the deviation between the real-time estimated value of each characteristic parameter and the measured value of each characteristic parameter, analyzing the degradation degree of each characteristic parameter;

[0029] The deviation between the real-time estimated value of each parameter and the measured value of each parameter is calculated, and the calculation formula is:

[0030]

[0031] According to the calculated deviation of each parameter, the degradation degree of each parameter is calculated, and the calculation formula is:

[0032]

[0033] Wherein: l is the number of coal mill failures, m is the number of parameters; q j is the influence weight of the j-th parameter on the output of the coal mill when the fault occurs, a normal state threshold value of the jth parameter when the coal mill is malfunctioning, an emergency shutdown threshold value of the jth parameter when the coal mill is malfunctioning.

[0034] As a preferred technical solution, the step S2 comprises the following steps:

[0035] S203, after analyzing the degradation degrees of the characteristic parameters, converting the state of the characteristic parameters of the coal mill into a reliability evaluation index of the coal mill, determining a region for sending a warning signal according to the change in the trend of the reliability evaluation index, so as to realize reliability monitoring and fault warning of the coal mill, and specifically comprising the following steps:

[0036] After obtaining the degradation degrees of the characteristic parameters, the state of the characteristic parameters of the coal mill is converted into a reliability evaluation index of the coal mill, and the reliability evaluation index of the coal mill, i.e., the reliability, is defined as:

[0037]

[0038] The reliability of the coal mill is monitored through the trend of the reliability index of the coal mill, and when the reliability is closer to 1, the reliability of the coal mill is higher, and when there is a significant downward trend, a fault warning should be sent.

[0039] The application provides a coal mill fault warning system based on a data model, comprising:

[0040] An offline modeling module is configured to analyze and determine characteristic parameters reflecting the operating performance of the coal mill, collect historical data of the characteristic parameters reflecting the operating performance of the coal mill according to the pre-determined characteristic parameters, perform data preprocessing on the collected historical data set, eliminate abnormal data therefrom, use the abnormal data as input data for state matrix extraction, and extract a state matrix.

[0041] An online monitoring module is configured to collect real-time operating data of the characteristic parameters, calculate real-time estimated values of the characteristic parameters according to the real-time operating data of the characteristic parameters and vectors in each typical state in the state matrix, analyze the degradation degrees of the characteristic parameters, convert the state of the characteristic parameters of the coal mill into a reliability evaluation index of the coal mill, determine a region for sending a warning signal according to the change in the trend of the reliability evaluation index, and realize reliability monitoring and fault warning of the coal mill.

[0042] As a preferred technical solution, the offline modeling module comprises:

[0043] A characteristic parameter determination module is configured to analyze characteristic parameters reflecting the operating performance of the coal mill, wherein the characteristic parameters reflecting the operating performance of the coal mill include any one or several of the following: coal supply amount, coal mill current, inlet air pressure, inlet air temperature, inlet air flow, outlet mixture pressure, outlet mixture temperature, cold air adjusting door opening degree, and hot air adjusting door opening degree.

[0044] a historical data collection module configured to collect historical data according to predetermined characteristic parameters reflecting the operation performance of the coal mill, the historical data being collected in the entire working condition range of the coal mill, and being uniformly distributed in each working condition in the entire working condition range;

[0045] a data preprocessing module configured to perform data preprocessing on the collected historical data set to eliminate abnormal data therefrom;

[0046] a state matrix extraction module configured to take the historical data from which the abnormal data are eliminated as input data for state matrix extraction, and to extract a state matrix;

[0047] the online monitoring module comprises:

[0048] a parameter real-time estimation module configured to collect real-time operation data of each characteristic parameter, and to calculate real-time estimation values of each characteristic parameter according to the real-time operation data of each characteristic parameter and the vectors in each typical state in the state matrix;

[0049] a parameter degradation analysis module configured to analyze the degradation degree of each characteristic parameter according to the deviation between the real-time estimation values of each characteristic parameter and the measured values of each characteristic parameter;

[0050] a reliability monitoring module configured to convert the state of the characteristic parameters of the coal mill into a reliability evaluation index of the coal mill, to determine a region for sending a warning signal according to the change in the reliability evaluation index trend, and to realize reliability monitoring and fault warning of the coal mill.

[0051] The application also provides a computer readable storage medium having a computer program stored thereon, the program being executed by a processor to realize the data model-based coal mill fault warning method according to any one of the preceding embodiments.

[0052] The data model-based coal mill fault warning method, system and storage medium provided by the application have the following beneficial effects:

[0053] 1) The data modeling method is simple and has strong engineering applicability in the actual application of the coal mill fault warning and diagnosis process, and realizes intelligent reliability monitoring and fault warning of the coal mill.

[0054] 2) First, the estimated values of each characteristic parameter are calculated by using the multi-parameter state estimation, then a method for calculating the parameter degradation degree based on the deviation between the estimated values and the measured values is proposed, and then the overall reliability of the coal mill is determined according to the degradation degree, and finally it is determined whether to give a warning according to the reliability trend of the coal mill, which provides a basis for intelligent monitoring of the coal mill, and realizes intelligent reliability monitoring and fault warning of the coal mill. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 A flowchart of a data model-based method for evaluating coal mill fault early warning is provided by the present invention;

[0056] Figure 2 A flowchart of a data model-based evaluation system for early warning of coal mill faults is provided for this invention.

[0057] Figure 3 The reliability trend chart of the coal mill provided by this invention. Detailed Implementation

[0058] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0059] like Figure 1 As shown, the present invention provides a data model-based coal mill fault early warning system, comprising:

[0060] The offline modeling module is used to analyze and determine the characteristic parameters that reflect the operating performance of the coal mill. Based on the pre-determined characteristic parameters that reflect the operating performance of the coal mill, historical data is collected. The collected historical dataset is preprocessed to remove abnormal data, which is then used as input data for state matrix extraction to obtain the state matrix.

[0061] The offline modeling module includes:

[0062] The characteristic parameter determination module is used to analyze the characteristic parameters that reflect the operating performance of the coal mill. The characteristic parameters that reflect the operating performance of the coal mill include any one or more of the following: coal feed rate, coal mill current, inlet air pressure, inlet air temperature, inlet air flow rate, outlet mixture pressure, outlet mixture temperature, cold air regulating damper opening, and hot air regulating damper opening.

[0063] The historical data acquisition module is used to collect historical data based on predetermined characteristic parameters that reflect the operating performance of the coal mill. The historical data is collected throughout the entire operating range of the coal mill, and the data is evenly distributed across all operating conditions within the entire operating range.

[0064] The data preprocessing module is used to preprocess the collected historical datasets and remove outlier data.

[0065] The state matrix extraction module uses historical data, after removing abnormal data, as input data for state matrix extraction to obtain the state matrix.

[0066] An online monitoring module is configured to collect real-time operation data of each characteristic parameter, calculate real-time estimation values of each characteristic parameter according to the real-time operation data of each characteristic parameter and vectors in each typical state in the state matrix, analyze the degradation degree of each characteristic parameter, convert the state of the characteristic parameter of the coal mill into a reliability evaluation index of the coal mill, determine a region for sending a warning signal according to the change in the trend of the reliability evaluation index, and realize reliability monitoring and fault warning of the coal mill.

[0067] The online monitoring module comprises:

[0068] A parameter real-time estimation module is configured to collect real-time operation data of each characteristic parameter, and calculate real-time estimation values of each characteristic parameter according to the real-time operation data of each characteristic parameter and vectors in each typical state in the state matrix.

[0069] A parameter degradation analysis module is configured to analyze the degradation degree of each characteristic parameter according to the deviation between the real-time estimation values of each characteristic parameter and the measured values of each characteristic parameter.

[0070] A reliability monitoring module is configured to convert the state of the characteristic parameter of the coal mill into a reliability evaluation index of the coal mill, determine a region for sending a warning signal according to the change in the trend of the reliability evaluation index, and realize reliability monitoring and fault warning of the coal mill.

[0071] The application provides a data model-based evaluation method for coal mill fault warning, which comprises the following steps:

[0072] S1: expert experience and / or correlation analysis method are used to analyze characteristic parameters reflecting the operation performance of the coal mill, wherein the characteristic parameters reflecting the operation performance of the coal mill include any one or several of the following: coal supply amount, coal mill current, inlet air pressure, inlet air temperature, inlet air flow, outlet mixture pressure, outlet mixture temperature, cold air adjusting door opening degree and hot air adjusting door opening degree.

[0073] S2: historical data of the characteristic parameters reflecting the operation performance of the coal mill are collected according to the pre-determined characteristic parameters, the historical data are collected in the entire working condition interval of the coal mill, and the working conditions in the entire working condition interval are uniformly distributed,

[0074] S3: the collected historical data set is pre-processed to remove abnormal data, the data clustering method is used to cluster the processed data set, a typical working condition library representing each typical operation working condition of the coal mill is obtained, and a state matrix is obtained through a clustering algorithm, wherein the state matrix is as follows:

[0075]

[0076] wherein element x ij is the historical running data of a parameter, i represents the time, and j represents the parameter;

[0077] S4 collects real-time running data of each characteristic parameter, and calculates the weight of each typical state vector in the state matrix and the real-time observation vector by using kernel function mapping, and the calculation formula is:

[0078]

[0079] wherein: x obs represents the real-time observation vector, i.e. the measured value, x i represents the historical state vector of the i-th row in the state matrix, h is the selected model parameter, and is determined according to the actual situation;‖x obs -x i ‖ 2 is the kernel function distance operator of two vectors, and the calculation formula of the kernel function distance operator of two vectors is:

[0080] ‖x obs -x i ‖ 2 =(x obs -x i ) T S -1 (x obs -x i )

[0081] wherein is a covariance matrix, wherein σ j is the standard deviation of the j-th parameter, which is obtained by statistical measurement or instrument design value, the real-time estimated value of each characteristic parameter is calculated, and the calculation formula is:

[0082]

[0083] S5 analyzes the degradation degree of each characteristic parameter according to the deviation between the real-time estimated value of each characteristic parameter and the measured value of each characteristic parameter;

[0084] The deviation between the real-time estimated value of each characteristic parameter and the measured value of each characteristic parameter is calculated, and the calculation formula is:

[0085]

[0086] According to the calculated deviation of each parameter, the degradation degree of each parameter is calculated, and the calculation formula is:

[0087]

[0088] wherein: l is the number of coal mill failures, and m is the number of parameters; qj a weight of the jth parameter on the mill output when a fault occurs, a normal state threshold of the jth parameter when the mill is in fault, an emergency shutdown threshold of the jth parameter when the mill is in fault;

[0089] S6 converts the state of the mill characteristic parameter into a reliability evaluation index of the mill after obtaining the degradation degree of the characteristic parameter, and defines the reliability evaluation index of the mill, i.e., the reliability, as follows:

[0090]

[0091] wherein l is the number of mill faults, m is the number of parameters; q j a weight of the jth parameter on the mill output when a fault occurs, a normal state threshold of the jth parameter when the mill is in fault, an emergency shutdown threshold of the jth parameter when the mill is in fault;

[0092] The mill reliability index is monitored, and when the reliability is closer to 1, the reliability of the mill operation is higher, and the area (e.g., as shown in FIG. 6) where the early warning signal is sent is determined. When there is a significant downward trend, a fault early warning prompt should be sent. Figure 2

[0093] The application also provides a computer readable storage medium having a computer program stored thereon, characterized in that the program, when executed by a processor, implements the mill fault early warning method based on a data model according to any one of the above.

[0094] The mill fault early warning method based on a data model, the system and the storage medium provided by the application have the following beneficial effects:

[0095] 1) The data modeling method is simple and has strong engineering applicability in the actual application of the mill fault early warning and diagnosis process, and the reliability monitoring and fault early warning prompt of the mill are intelligently realized.

[0096] 2) First, the estimated value of each characteristic parameter is calculated by using the multi-parameter state estimation, and then the method of calculating the parameter degradation degree based on the deviation between the estimated value and the measured value is proposed, and then the overall reliability of the mill is determined according to the degradation degree, and finally whether to give a warning is determined according to the trend of the mill reliability, which provides a basis for intelligent monitoring of the mill, and the reliability monitoring and fault early warning prompt of the mill are intelligently realized.

[0097] ​It is to be understood that the present application is described by way of example only, and that modifications or alterations can be made to the features and embodiments described without departing from the spirit and scope of the application. In addition, modifications can be made to the features and embodiments described to adapt them to particular situations and materials without departing from the spirit and scope of the application. Accordingly, the present application is not limited to the specific embodiments described herein, but rather only by the claims which follow, all variations and equivalents which fall within the ranges of the claims being intended to be embraced herein.

Claims

1. A method for early warning of coal mill faults based on a data model, characterized in that, Includes the following steps: S1 analyzes and determines the characteristic parameters that reflect the operating performance of the coal mill. Based on the predetermined characteristic parameters that reflect the operating performance of the coal mill, historical data is collected. The collected historical dataset is preprocessed to remove abnormal data and used as input data for state matrix extraction. The state matrix is ​​then extracted. S2 collects real-time operating data of various characteristic parameters of the coal mill. Based on the collected real-time operating data of various characteristic parameters of the coal mill and the vectors in each typical state in the state matrix, it calculates the real-time estimated value of each characteristic parameter, analyzes the deterioration degree of each characteristic parameter, converts the state of the coal mill characteristic parameters into the reliability evaluation index of the coal mill, and determines the area to issue early warning signals based on the change of the reliability evaluation index, so as to realize the reliability monitoring and fault early warning of the coal mill. Step S2 includes the following steps: S201 collects real-time operational data for each feature parameter, and calculates real-time estimated values ​​for each feature parameter based on the collected real-time operational data and the vectors in each typical state of the state matrix. This specifically includes the following steps: Real-time operational data for each feature parameter is collected, and the weights of each typical state vector and the real-time observation vector in the state matrix are calculated using kernel function mapping. The calculation formula is as follows: ； in: The real-time observation vector represents the measured value. Let h represent the historical state vector in the i-th row of the state matrix, where h is the selected model parameter. Let be the kernel function distance operator for two vectors. The formula for calculating the kernel function distance operator for two vectors is: ； In the formula Let be the covariance matrix, where... Let be the standard deviation of the j-th parameter, obtained through statistical analysis of measured data or instrument design. Calculate the real-time estimated values ​​of each characteristic parameter using the following formula: ； Among them, elements This represents the historical data of a certain parameter, where i represents the time point and j represents the parameter. S202 analyzes the degree of degradation of each characteristic parameter based on the deviation between the real-time estimated value and the measured value of each characteristic parameter; The deviation between the real-time estimated values ​​of the parameters and the measured values ​​of each characteristic parameter is calculated using the following formula: ； Based on the calculated deviations of each parameter, the degree of degradation of each parameter is calculated using the following formula: ； Where: m is the number of parameters; Let j be the weight of the influence of the j-th parameter on the output of the coal mill when the fault occurs. Let be the normal state threshold value of the j-th parameter when the coal mill malfunctions. Let be the emergency shutdown threshold for the j-th parameter when the coal mill malfunctions; After analyzing the degradation degree of each characteristic parameter, S203 converts the state of the coal mill characteristic parameters into a reliability evaluation index for the coal mill. Based on the changes in the trend of the reliability evaluation index, it determines the area where a warning signal is issued, so as to realize the reliability monitoring and fault early warning of the coal mill. Specifically, it includes the following steps: After obtaining the degradation degree of the characteristic parameters, the state of the coal mill's characteristic parameters is converted into a reliability evaluation index for the coal mill. The reliability evaluation index of the coal mill, namely the reliability, is defined as: ； By monitoring the trend of the coal mill reliability index, the higher the reliability of the coal mill operation, the closer the reliability is to 1. When there is a significant downward trend, a fault warning should be issued.

2. The method for early warning of coal mill faults based on data model evaluation according to claim 1, characterized in that, Step S1 involves analyzing and determining the characteristic parameters that reflect the operating performance of the coal mill, including the following steps: The characteristic parameters reflecting the operating performance of the coal mill were analyzed using expert experience and correlation analysis methods. These characteristic parameters include any one or more of the following: coal feed rate, coal mill current, inlet air pressure, inlet air temperature, inlet air flow rate, outlet mixture pressure, outlet mixture temperature, cold air regulating damper opening, and hot air regulating damper opening.

3. The method for early warning of coal mill faults based on data model evaluation according to claim 2, characterized in that, Step S1 involves collecting historical data based on predetermined characteristic parameters reflecting the operating performance of the coal mill, including the following steps: Historical data is collected for the entire operating range of the coal mill, and the operating conditions are evenly distributed throughout the entire operating range.

4. The method for early warning of coal mill faults based on data model evaluation according to claim 3, characterized in that, Step S1 involves preprocessing the collected historical dataset to remove outliers, using this data as input for state matrix extraction. The process includes the following steps: The collected historical dataset is preprocessed to remove outlier data. The processed dataset is clustered using a data clustering method to obtain a typical operating condition database representing various typical operating conditions of the coal mill. A state matrix is ​​then extracted using a clustering algorithm. The state matrix is ​​as follows: ； Among them, elements This represents the historical data of a certain parameter, where i represents the time and j represents the parameter.

5. A data model-based fault early warning system for coal mills, characterized in that, include: The offline modeling module is used to analyze and determine the characteristic parameters that reflect the operating performance of the coal mill. Based on the pre-determined characteristic parameters that reflect the operating performance of the coal mill, historical data is collected. The collected historical dataset is preprocessed to remove abnormal data, which is then used as input data for state matrix extraction to obtain the state matrix. The online monitoring module is used to collect real-time operating data of various characteristic parameters. Based on the collected real-time operating data of various characteristic parameters and the vectors of each typical state in the state matrix, it calculates the real-time estimated value of each characteristic parameter, analyzes the deterioration degree of each characteristic parameter, converts the state of the coal mill characteristic parameters into the reliability evaluation index of the coal mill, and determines the area to issue early warning signals based on the change of the reliability evaluation index, so as to realize the reliability monitoring and fault early warning of the coal mill. The online monitoring module includes: The real-time parameter estimation module is used to collect real-time operational data for each feature parameter. Based on the collected real-time operational data and the vectors in each typical state of the state matrix, it calculates the real-time estimated value of each feature parameter. Specifically, it includes the following steps: Real-time operational data for each feature parameter is collected, and the weights of each typical state vector and the real-time observation vector in the state matrix are calculated using kernel function mapping. The calculation formula is as follows: ； in: The real-time observation vector represents the measured value. Let h represent the historical state vector in the i-th row of the state matrix, where h is the selected model parameter. Let be the kernel function distance operator for two vectors. The formula for calculating the kernel function distance operator for two vectors is: ； In the formula Let be the covariance matrix, where... Let be the standard deviation of the j-th parameter, obtained through statistical analysis of measured data or instrument design. Calculate the real-time estimated values ​​of each characteristic parameter using the following formula: ； The parameter degradation analysis module is used to analyze the degradation degree of each characteristic parameter based on the deviation between the real-time estimated value and the measured value of each characteristic parameter. The deviation between the real-time estimated value and the measured value of each characteristic parameter is calculated using the following formula: ； Based on the calculated deviations of each parameter, the degree of degradation of each parameter is calculated using the following formula: ； Where: m is the number of parameters; Let j be the weight of the influence of the j-th parameter on the output of the coal mill when the fault occurs. Let be the normal state threshold value of the j-th parameter when the coal mill malfunctions. Let be the emergency shutdown threshold for the j-th parameter when the coal mill malfunctions; The reliability monitoring module converts the state of the coal mill's characteristic parameters into reliability evaluation indicators. Based on the changes in the trend of the reliability evaluation indicators, it determines the area where early warning signals are issued, thereby realizing the reliability monitoring and fault early warning of the coal mill. Specifically, it includes the following steps: After obtaining the degradation degree of the characteristic parameters, the state of the coal mill's characteristic parameters is converted into a reliability evaluation index for the coal mill. The reliability evaluation index of the coal mill, namely the reliability, is defined as: ； By monitoring the trend of the coal mill reliability index, the higher the reliability of the coal mill operation, the closer the reliability is to 1. When there is a significant downward trend, a fault warning should be issued.

6. The data model-based coal mill fault early warning system according to claim 5, characterized in that, The offline modeling module includes: The characteristic parameter determination module is used to analyze the characteristic parameters that reflect the operating performance of the coal mill. The characteristic parameters that reflect the operating performance of the coal mill include any one or more of the following: coal feed rate, coal mill current, inlet air pressure, inlet air temperature, inlet air flow rate, outlet mixture pressure, outlet mixture temperature, cold air regulating damper opening, and hot air regulating damper opening. The historical data acquisition module is used to collect historical data based on predetermined characteristic parameters that reflect the operating performance of the coal mill. The historical data is collected throughout the entire operating range of the coal mill, and the data is evenly distributed across all operating conditions within the entire operating range. The data preprocessing module is used to preprocess the collected historical datasets and remove outlier data. The state matrix extraction module uses historical data, after removing abnormal data, as input data to extract the state matrix.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the data model-based method for evaluating coal mill fault early warning as described in any one of claims 1-4.

Images