Coal preparation plant equipment health state monitoring system based on machine learning

Abstract

The invention relates to the technical field of equipment monitoring, in particular to a coal preparation plant equipment health state monitoring system based on machine learning, and the system comprises a data collection module which is used for collecting equipment operation data and raw material parameters; the coal slime characteristic analysis module is used for constructing a load coefficient and judging a coal slime load state according to the load coefficient, and the structure monitoring module is used for constructing a structure state index and judging an equipment structure state according to the structure state index; the pressure monitoring module is used for carrying out conjoint analysis on the cylinder pressure, the cylinder pressure standard deviation and the cylinder pressure average value so as to construct a pressure state index and update judgment of an equipment structure state; the electrical safety monitoring module is used for judging the electrical safety state and updating the construction process of the pressure state index according to the electrical safety state, and the equipment state monitoring module is used for judging the health state of equipment and outputting an equipment maintenance prompt to a user. The equipment monitoring efficiency of the coal preparation plant is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment monitoring, and in particular to a coal preparation plant equipment health status monitoring system based on machine learning. Background Art

[0002] The core equipment of coal preparation plants is in a high-wear and high-load working condition for a long time. Traditional monitoring technologies rely on manual inspections and single-sensor threshold alarms, and have significant defects: Limitation of static threshold: Fixed thresholds cannot dynamically respond to the impact of changes in the characteristics of coal slime raw materials (such as moisture content, particle size) on equipment, resulting in false alarms. For example, abnormal screen differential pressure is judged only by a fixed pressure threshold, ignoring the non-linear impact of sudden changes in coal slime particle size on the differential pressure; Insufficient fault coupling: The correlation between mechanical wear, electrical aging and raw material load has not been quantified. Drum wear and screen blockage are often analyzed in isolation, making it difficult to locate the root cause of complex faults (such as wear causing material retention and pressure fluctuations); Lack of early warning: Traditional methods are less sensitive to progressive faults (such as motor insulation degradation), and usually alarm after the fault is serious. Dynamic parameters such as temperature gradient and wear rate are not included in the model, missing the maintenance window period.

[0003] Chinese Patent Publication No. CN114911206A discloses an intelligent predictive maintenance system for coal preparation plants based on the Internet of Things, including: IoT sensors for collecting real-time operation data of equipment; IoT repeaters for sending real-time operation data of equipment to a local server for storage and uploading; an intelligent cloud processing platform for analyzing various operation data of equipment in the real-time operation data of equipment through AI machine learning algorithms and models embedded in the cloud, comprehensively evaluating the operation status of the equipment and pushing; a digital twin control platform for building a virtual coal preparation plant using 3D laser scanner field scanning technology to achieve intelligent interaction functions, monitoring functions and warning functions, and pushing warning and alarm data; an intelligent control client for receiving data, viewing the operation status of the plant in real time, analyzing and implementing predictive maintenance and intelligent control; It can be seen that this solution only realizes "3D display + alarm push" for the digital twin platform, but lacks the ability of multi-dimensional root cause analysis, resulting in low efficiency of coal preparation plant equipment monitoring. Summary of the Invention

[0004] The purpose of the present invention is to provide a coal preparation plant equipment health status monitoring system based on machine learning to solve at least one of the problems existing in the prior art.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A coal preparation plant equipment health status monitoring system based on machine learning, including:

[0007] A structure monitoring module is used to jointly analyze the thickness of the inner lining of the equipment drum and the pressure difference across the screen collected within the monitoring period, so as to construct a structure status index and determine the equipment structure status according to the structure status index;

[0008] A pressure monitoring module is used to jointly analyze the cylinder pressure, the standard deviation of the cylinder pressure, and the average value of the cylinder pressure, so as to construct a pressure status index and update the determination of the equipment structure status;

[0009] An equipment status monitoring module is used to determine the equipment health status according to the coal slime load status, the equipment structure status, and the electrical safety status within the management period, and output equipment maintenance prompts to the user according to the determination result of the equipment health status.

[0010] Optionally, the structure monitoring module includes a wear analysis unit, which is used to calculate the wear rate Ms of the thickness m0 of the inner lining of the equipment drum and the reference thickness m1 collected within the monitoring period, and compare and analyze the wear rate Ms with the preset wear rate ms0 to determine the equipment wear status and construct a wear index according to the equipment wear status.

[0011] Optionally, the structure monitoring module further includes a pressure difference analysis unit, which is used to compare and analyze the screen pressure difference △P and the preset pressure difference p1 collected within the monitoring period to determine the equipment pressure difference status and construct a balance index according to the equipment pressure difference status, where:

[0012] If △P is less than or equal to p1, the pressure difference analysis unit determines that the equipment pressure difference status is normal in the current monitoring period, sets the equipment pressure difference index to YC1, and sets YC1 = 0. Otherwise, the pressure difference analysis unit sets the equipment pressure difference index to YC2 and sets YC2 = 3×[(△P - p1) / (△P + p1)]^2 - 2×[(△P - p1) / (△P + p1)]^3.

[0013] Optionally, the structure monitoring module further includes a structure status judgment unit, which is used to construct a structure status index ZY according to the wear index and the balance index, and set ZY = lg(3×wear index + 1) / lg4×balance index;

[0014] The structure status judgment unit compares and analyzes the structure status index ZY with the preset status index zy to determine the equipment structure status. If ZY is less than or equal to zy, the structure status judgment unit determines that the equipment structure status is in a healthy state in the current monitoring period and does not give a structure health status warning to the user. Otherwise, the structure status judgment unit determines that the equipment structure status is in an abnormal state in the current monitoring period and gives a structure health status warning to the user.

[0015] Optionally, the pressure monitoring module compares and analyzes the i-th cylinder pressure Pi collected during the monitoring period with a preset cylinder pressure threshold p0 to determine whether the cylinder pressure exceeds the limit. If Pi is less than or equal to p0, the pressure monitoring module determines that the i-th cylinder pressure does not exceed the limit. Otherwise, the pressure monitoring module determines that the i-th cylinder pressure exceeds the limit, and counts the number Pn of data with the cylinder pressure exceeding the limit. The proportion of the cylinder pressure exceeding the limit is set as β, and β is the ratio of Pn to Pu, where Pu is the number of cylinder pressures collected during the monitoring period.

[0016] The pressure monitoring module jointly analyzes the proportion β of the cylinder pressure exceeding the limit, the standard deviation σP of the cylinder pressure, and the average value Pavg of the cylinder pressure to construct a pressure state index PC, and sets: PC = σP / Pavg × β;

[0017] When the equipment structure state is in a healthy state during the current monitoring period, if the pressure state index PC is greater than or equal to a preset pressure state index pc0, it is determined that the equipment structure state in the current monitoring period is in an abnormal state, and a structural health status warning is given to the user.

[0018] Optionally, the system further includes: an electrical safety monitoring module, which is used to perform data fusion on the motor insulation resistance and winding temperature collected during the monitoring period to construct an insulation degradation index, and judge the electrical safety state according to the insulation degradation index, and is also used to update the construction process of the pressure state index according to the electrical safety state.

[0019] Optionally, the electrical safety monitoring module includes a construction unit, which is used to calculate the winding temperature gradient △T according to the winding temperature collected during the monitoring period, and the winding temperature gradient △T is the difference between the maximum value and the minimum value of the winding temperature collected during the monitoring period;

[0020] The construction unit performs data fusion on the winding temperature gradient △T and the motor insulation resistance RI to construct an insulation degradation index HI.

[0021] Optionally, the electrical safety monitoring module includes an electrical safety monitoring unit, which is used to compare and analyze the insulation degradation index HI with an electrical safety threshold IDI to judge the electrical safety state. If HI is less than or equal to IDI, the electrical safety monitoring unit determines that the electrical safety state is in a normal state. Otherwise, the electrical safety monitoring unit determines that the electrical safety state is in an insulation aging risk state, and updates the construction process of the pressure state index by updating the preset pressure state index. The updated preset pressure state index is set as pc1, and the calculation formula of pc1 is set as pc1 = pc0 × exp[-(HI - IDI)].

[0022] Optionally, the equipment status monitoring module includes a status monitoring unit, and when the coal slime load state is normal, if L1×j1 / J+L2×j2 / J is less than or equal to the first equipment status threshold u1, the status monitoring unit determines that the equipment health state of the current management cycle is a normal state, otherwise, the status monitoring unit determines that the equipment health state of the current management cycle is an abnormal state, L1 is a structural weight, L2 is an electrical safety weight, L1+L2=1, j1 is the number of monitoring cycles in which the equipment structure state is an abnormal state within the management cycle, j2 is the number of monitoring cycles in which the electrical safety state is an insulation aging risk state, and J is the number of monitoring cycles within the management cycle;

[0023] When the coal slime load state is normal, if L1×j1 / J+L2×j2 / J is less than or equal to the second equipment state threshold u2, the state monitoring unit determines that the equipment health state of the current management cycle is normal; otherwise, the state monitoring unit determines that the equipment health state of the current management cycle is abnormal.

[0024] Optionally, the system further comprises: a data acquisition module for collecting equipment operation data and raw material parameters;

[0025] The coal slime characteristic analysis module is used to perform data fusion analysis on the coal slime moisture content and coal slime particle size to construct a load coefficient, and judge the coal slime load state according to the load coefficient; the coal slime characteristic analysis module constructs the load coefficient LC according to the collected coal slime moisture content s0 and coal slime particle size d0, and compares and analyzes the load coefficient LC with the preset load threshold LC0 to judge the coal slime load state, wherein:

[0026] If LC is less than or equal to LC0, the coal slime characteristic analysis module determines that the coal slime load state is normal, otherwise, the coal slime characteristic analysis module determines that the coal slime load state is abnormal.

[0027] The beneficial effects of the present invention are as follows: This system significantly improves the accuracy and reliability of equipment health monitoring in coal preparation plants through multi-source data fusion and dynamic threshold adjustment mechanism. First, based on the load coefficient model of coal slime moisture content and particle size, it breaks through the limitations of the traditional single-parameter threshold method, can identify the risk of raw material overload in advance, and avoid equipment wear or efficiency loss caused by abnormal load. Secondly, the structural monitoring module adopts a joint analysis of wear rate and pressure difference, combined with logarithmic function and exponential formula to enhance abnormal sensitivity and achieve early warning. The pressure monitoring module introduces dynamic calculation of over-limit proportion and fluctuation value, effectively identifies cylinder deformation or sealing failure, and fills the blind spot of traditional structural monitoring. The electrical safety module uses a fusion analysis of insulation resistance and temperature gradient. When the insulation resistance of the motor does not reach the critical value, it can warn of potential aging risks through temperature mutation, and adjust the pressure threshold in a linked manner to form cross-module collaborative protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0029] Figure 1 It is a schematic structural diagram of the coal preparation plant equipment health status monitoring system based on machine learning in this embodiment.

[0030] Figure 2 It is a schematic structural diagram of the structure monitoring module in this embodiment.

[0031] Figure 3 It is a schematic structural diagram of the electrical safety monitoring module in this embodiment.

[0032] Figure 4 It is a schematic structural diagram of the equipment status monitoring module in this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] To more clearly illustrate the present invention, the following further describes the present invention in conjunction with preferred embodiments and drawings. Similar components in the drawings are denoted by the same reference numerals. Those skilled in the art should understand that the content specifically described below is illustrative rather than restrictive, and should not limit the protection scope of the present invention.

[0034] It should be noted that although terms such as first, second, and third may be used in the embodiments of the present application for description, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, without departing from the scope of the embodiments of the present application, the first can also be referred to as the second, and similarly, the second can also be referred to as the first.

[0035] Please refer to Figure 1 as shown, which is a schematic structural diagram of the coal preparation plant equipment health status monitoring system based on machine learning in this embodiment. The system includes,

[0036] A data acquisition module is used to collect equipment operation data and raw material parameters. The equipment operation data includes the thickness of the inner lining of the equipment drum, the pressure difference across the screen, the cylinder pressure, the insulation resistance of the motor, and the winding temperature. The raw material parameters include the moisture content of the slime and the particle size of the slime. It can be understood that in this embodiment, the acquisition methods of the equipment operation data and the raw material parameters are not specifically limited, and those skilled in the art can freely set them as long as the acquisition requirements of the equipment operation data and the raw material parameters are met. Among them, a laser distance sensor or an ultrasonic sensor can be installed on the inner wall of the drum to measure the wear thickness of the inner lining in real time. A differential pressure sensor can be installed at both ends of the screen to monitor the pressure difference before and after the screen. Pressure sensors can be deployed at key positions of the cylinder to collect pressure data in real time. An insulation resistance tester can be used to monitor the insulation performance of the motor winding regularly or online. A thermocouple or an infrared temperature sensor can be embedded in the motor winding to collect temperature data in real time. A microwave moisture analyzer or a drying method can be used, combined with an online sensor to detect the moisture content of the slime in real time. A laser particle size analyzer or a sieving method can be used to analyze the particle size of the slime sample.

[0037] Please continue to refer to Figure 1 as shown, the system further includes:

[0038] A slime characteristic analysis module, which is connected to the data acquisition module, is used to perform data fusion analysis on the moisture content of the slime and the particle size of the slime to construct a load coefficient, and judge the slime load status according to the load coefficient.

[0039] Specifically, the slime characteristic analysis module constructs the load coefficient LC according to the collected moisture content s0 of the slime and the particle size d0 of the slime, and sets LC = w1 × ln(s0 / s1 + 1) + w2 × (d0 / d1) 1.5 ; w1 is the moisture content weight, w2 is the particle size weight, w1 + w2 = 1, s1 is the preset moisture content, and d1 is the preset particle size;

[0040] The slime characteristic analysis module compares and analyzes the load coefficient LC with the preset load threshold LC0 to judge the slime load status, where:

[0041] If LC is less than or equal to LC0, the slime characteristic analysis module determines that the slime load status is normal. Otherwise, the slime characteristic analysis module determines that the slime load status is abnormal; through the multi-source data fusion of the moisture content and the particle size, the construction of the load coefficient LC realizes the dynamic load assessment. The logarithmic function is used to process the non-linear influence of the moisture content, and the exponential function strengthens the abnormal response of the particle size. Compared with the traditional single-parameter threshold method, the accuracy of the slime load status analysis is improved.

[0042] Specifically, in this embodiment, no specific limitations are imposed on the settings of each weight, the preset moisture content, and the preset particle size. Those skilled in the art can freely set them as long as the setting requirements for each weight, the preset moisture content, the preset load threshold, and the preset particle size are met. Among them, the optimal value of s1 is 0.2, the optimal value of d1 is 0.5 mm, the optimal value of w1 is 0.6, the optimal value of w2 is 0.4, and the optimal value of LC0 is 0.8.

[0043] Please continue to refer to Figure 1 As shown, the system further includes:

[0044] A structure monitoring module, which is connected to the slime property analysis module, is used to jointly analyze the thickness of the inner lining of the equipment drum and the screen differential pressure collected within the monitoring period to construct a structure state index, and judge the equipment structure state according to the structure state index.

[0045] Specifically, in this embodiment, no specific limitations are imposed on the setting of the monitoring period. Those skilled in the art can freely set it as long as the setting requirements of the monitoring period are met. Among them, the monitoring period can be set to 5 minutes.

[0046] Please refer to Figure 2 As shown, it is a schematic structural diagram of the structure monitoring module of this embodiment, including

[0047] A wear analysis unit, which is used to calculate the wear rate Ms of the thickness m0 of the inner lining of the equipment drum and the reference thickness m1 collected within the monitoring period, set Ms = (m1 - m0) / m1, and compare and analyze the wear rate Ms with the preset wear rate ms0 to judge the equipment wear state, and construct a wear index according to the equipment wear state, where:

[0048] If Ms is less than or equal to ms0, the wear analysis unit determines that the equipment wear state is normal in the current monitoring period, and sets the wear index to MZ1, setting MZ1 = 0. On the contrary, the wear analysis unit determines that the equipment wear state is abnormal in the current monitoring period, and sets the wear index to MZ2, setting MZ2 = lg[(Ms - ms0) + 1] / lg2; based on the calculation of the wear rate based on the change in the inner lining thickness, an exponential formula is used to enhance the sensitivity of abnormal wear. When the wear rate exceeds the threshold, the degree of abnormal wear is quantified, thereby improving the accuracy of structural health state monitoring.

[0049] Specifically, the reference thickness in this embodiment is the initial thickness of the inner lining of the equipment drum. In this embodiment, no specific limitations are imposed on the setting of the preset wear rate. Those skilled in the art can freely set it as long as the setting requirements of the preset wear rate are met. Among them, the optimal value of ms0 is 0.05.

[0050] Please continue to refer to Figure 2As shown, the structure monitoring module further includes:

[0051] A differential pressure analysis unit, which is connected to the wear analysis unit, is used to compare and analyze the screen differential pressure △P and the preset differential pressure p1 collected during the monitoring period to judge the equipment differential pressure state, and construct a balance index according to the equipment differential pressure state, where:

[0052] If △P is less than or equal to p1, the differential pressure analysis unit determines that the equipment differential pressure state is normal during the current monitoring period, and sets the equipment differential pressure index to YC1, setting YC1 = 0. Conversely, the differential pressure analysis unit sets the equipment differential pressure index to YC2, setting YC2 = 3×[(△P - p1) / (△P + p1)] 2 - 2×[(△P - p1) / (△P + p1)] 3 ; Combining the preset threshold and the non-linear differential pressure response function can accurately identify the trend of screen blockage. For example, when the screen is slightly blocked, the system can trigger an early warning to avoid sudden shutdown caused by complete blockage.

[0053] Specifically, in this embodiment, the setting of the preset differential pressure is not specifically limited, and those skilled in the art can freely set it as long as the setting requirements of the preset differential pressure are met. Among them, the best value of p1 is 0.2MPa.

[0054] Please continue to refer to Figure 2 As shown, the structure monitoring module further includes:

[0055] A structure state judgment unit, which is connected to the differential pressure analysis unit, is used to construct a structure state index according to the wear index and the balance index, and judge the equipment structure state according to the structure state index, where:

[0056] The structure state judgment unit sets the structure state index to ZY, setting ZY = lg(3×wear index + 1) / lg4×balance index;

[0057] The structure state judgment unit compares and analyzes the structure state index ZY and the preset state index zy to judge the equipment structure state. If ZY is less than or equal to zy, the structure state judgment unit determines that the equipment structure state is in a healthy state during the current monitoring period and does not give a structure health state warning to the user. Conversely, the structure state judgment unit determines that the equipment structure state is in an abnormal state during the current monitoring period and gives a structure health state warning to the user; Through the combined analysis of the wear index and the differential pressure index, the mechanical structure health degree is comprehensively evaluated, and the detection efficiency for compound faults such as bearing damage and screen rupture is higher.

[0058] Specifically, in this embodiment, the setting of the preset state index is not specifically limited, and those skilled in the art can set it freely as long as the setting requirements of the preset state index are met. Among them, the optimal value of zy is 0.342.

[0059] Please continue to refer to Figure 1 as shown, the system further includes:

[0060] A pressure monitoring module, which is connected to the structure monitoring module, is used to compare and analyze the cylinder pressure collected during the monitoring period to determine the proportion of cylinder pressure exceeding the limit, and jointly analyze the proportion of cylinder pressure exceeding the limit, the standard deviation of cylinder pressure, and the average value of cylinder pressure to construct a pressure state index and update the judgment of the equipment structure state.

[0061] Specifically, the pressure monitoring module compares and analyzes the i-th cylinder pressure Pi collected during the monitoring period with the preset cylinder pressure threshold p0 to determine whether the cylinder pressure exceeds the limit. If Pi is less than or equal to p0, the pressure monitoring module determines that the i-th cylinder pressure does not exceed the limit. Otherwise, the pressure monitoring module determines that the i-th cylinder pressure exceeds the limit, and counts the number of data Pn of the cylinder pressure exceeding the limit, and sets the proportion of cylinder pressure exceeding the limit as β, where β is the ratio of Pn to Pu, and Pu is the number of cylinder pressures collected during the monitoring period;

[0062] The pressure monitoring module jointly analyzes the proportion of cylinder pressure exceeding the limit β, the standard deviation of cylinder pressure σP, and the average value of cylinder pressure Pavg to construct a pressure state index PC, and sets: PC = σP / Pavg × β;

[0063] When the equipment structure state is in a healthy state during the current monitoring period, if the pressure state index PC is greater than or equal to the preset pressure state index pc0, it is determined that the equipment structure state is in an abnormal state during the current monitoring period, and a structure health status warning is sent to the user; introducing the joint calculation of the proportion of pressure exceeding the limit, the average value, and the fluctuation value to quantify the abnormal risk of the cylinder pressure can effectively identify the cylinder deformation or seal failure and supplement the monitoring blind area of the structure state.

[0064] Specifically, in this embodiment, the setting of the preset cylinder pressure threshold and the preset pressure state index is not specifically limited, and those skilled in the art can set it freely as long as the setting requirements of the preset cylinder pressure threshold and the preset pressure state index are met. Among them, the optimal value of p0 is 1.1 times the rated cylinder pressure, and the optimal value of pc0 is 0.2.

[0065] Please continue to refer to Figure 1 as shown, the system further includes:

[0066] An electrical safety monitoring module, which is connected to the pressure monitoring module, is used to perform data fusion on the motor insulation resistance and winding temperature collected within a monitoring period to construct an insulation degradation index, and to judge the electrical safety status according to the insulation degradation index. It is also used to update the construction process of the pressure status index according to the electrical safety status.

[0067] Please refer to Figure 3 As shown, the electrical safety monitoring module includes:

[0068] A construction unit, which is used to calculate the winding temperature gradient △T according to the winding temperature collected within a monitoring period. The winding temperature gradient △T is the difference between the maximum value and the minimum value of the winding temperature collected within the monitoring period;

[0069] The construction unit performs data fusion on the winding temperature gradient △T and the motor insulation resistance RI to construct an insulation degradation index HI. The calculation formula of HI is HI = (RI / RM) × (1 + 0.05 × △T / TY), where RM is the motor insulation resistance threshold and TY is the temperature difference threshold; by calculating the fusion analysis of the winding temperature gradient (△T) and the insulation resistance (RI) in real time, an insulation degradation index (HI) is constructed. This model breaks through the limitation of traditional single-threshold judgment. For example, when the motor is intermittently overloaded and the temperature rises suddenly, even if the insulation resistance does not reach the critical value, the HI index can still give an early warning through the change of the temperature gradient, avoiding the accelerated insulation aging caused by temperature fluctuations.

[0070] Specifically, in this embodiment, the settings of the motor insulation resistance threshold and the temperature difference threshold are not specifically limited, and those skilled in the art can set them freely as long as they meet the setting requirements of the motor insulation resistance threshold and the temperature difference threshold. Among them, the optimal value of RM is 1 MΩ / kV, and the optimal value of TY is 50 °C.

[0071] Please continue to refer to Figure 3 As shown, the electrical safety monitoring module includes:

[0072] An electrical safety monitoring unit, which is connected to the construction unit, is used to compare and analyze the insulation degradation index HI and the electrical safety threshold IDI to judge the electrical safety status. If HI is less than or equal to IDI, the electrical safety monitoring unit determines that the electrical safety status is in a normal state. Otherwise, the electrical safety monitoring unit determines that the electrical safety status is in an insulation aging risk state, and updates the construction process of the pressure status index by updating the preset pressure status index. Set the updated preset pressure status index as pc1, and set the calculation formula of pc1 as pc1 = pc0 × exp[-(HI - IDI)]; by linking the HI index with the pressure monitoring module, when insulation degradation is detected, the pressure status threshold is dynamically adjusted to avoid the mechanical overload risk indirectly caused by electrical abnormalities.

[0073] Specifically, in this embodiment, the setting of the electrical safety threshold is not specifically limited, and those skilled in the art can set it freely as long as the setting requirements of the electrical safety threshold are met. Among them, the optimal value of IDI is 0.8.

[0074] Please continue to refer to Figure 1 As shown, the system further includes:

[0075] An equipment status monitoring module, which is connected to the electrical safety monitoring module, is used to judge the equipment health status according to the slime load status, equipment structure status and electrical safety status within the management cycle, and output equipment maintenance prompts to the user according to the judgment result of the equipment health status.

[0076] Specifically, in this embodiment, the setting of the management cycle is not specifically limited, and those skilled in the art can set it freely as long as the setting requirements of the management cycle are met. Among them, the management cycle can be set to 2h.

[0077] Please refer to Figure 4 As shown, the equipment status monitoring module includes:

[0078] A status monitoring unit, which is used to judge the equipment health status according to the slime load status, equipment structure status and electrical safety status within the management cycle.

[0079] Specifically, when the slime load status is normal, if L1×j1 / J + L2×j2 / J is less than or equal to the first equipment status threshold u1, the status monitoring unit determines that the equipment health status in the current management cycle is in a normal state; otherwise, the status monitoring unit determines that the equipment health status in the current management cycle is in an abnormal state. Here, L1 is the structure weight, L2 is the electrical safety weight, L1 + L2 = 1, j1 is the number of monitoring cycles in which the equipment structure status is abnormal within the management cycle, j2 is the number of monitoring cycles in which the electrical safety status is in an insulation aging risk state, and J is the number of monitoring cycles within the management cycle.

[0080] When the slime load status is normal, if L1×j1 / J + L2×j2 / J is less than or equal to the second equipment status threshold u2, the status monitoring unit determines that the equipment health status in the current management cycle is in a normal state; otherwise, the status monitoring unit determines that the equipment health status in the current management cycle is in an abnormal state. By integrating multi-dimensional real-time data such as mechanical load, structural wear, and electrical parameters (such as vibration, temperature, insulation index, etc.), a comprehensive health assessment model is established to improve the efficiency of equipment health status monitoring.

[0081] Specifically, in this embodiment, no specific limitations are imposed on the settings of each weight and each state threshold. Those skilled in the art can set them freely as long as the setting requirements for the settings of each weight and each state threshold are met. Among them, the optimal value of L1 is 0.7, the optimal value of L2 is 0.3, the optimal value of u1 is 0.3, and the optimal value of u2 is 0.26.

[0082] Please continue to refer to Figure 4 As shown, the device status monitoring module further includes

[0083] An output unit, which is connected to the status monitoring unit and is used to output a device maintenance prompt to the user when the device health status is an abnormal state.

[0084] Specifically, in an exemplary application scenario, during the monitoring of a coal preparation plant device, the data acquisition module collects device operation data: the current thickness measured by an ultrasonic sensor is 47 mm (the initial thickness is 50 mm); the differential pressure sensor shows the differential pressure on both sides of the screen as 0.3 MPa in real time; the data collected by five pressure sensors are 0.9 MPa, 1.2 MPa, 1.3 MPa, 0.8 MPa, 1.1 MPa (the rated pressure is 1.0 MPa, and the trigger threshold is 1.1 times the rated value); motor parameters: the insulation resistance tester shows the insulation resistance as 0.9 MΩ / kV (the rated threshold is 1 MΩ / kV), and the thermocouple measures that the winding temperature rises from 35 °C to 90 °C within 5 minutes; raw material parameters are collected: the detection value of the online microwave moisture analyzer is 30% (the preset threshold is 20%); the particle size detected by the laser particle size analyzer is 0.8 mm (the preset threshold is 0.5 mm);

[0085] The slime property analysis module inputs the slime moisture content (30%) and particle size (0.8 mm) into the load model, and through weighted fusion analysis (the moisture content weight is 60%, and the particle size weight is 40%), calculates that the load coefficient is 1.358, which is significantly higher than the preset threshold of 0.8, and determines that the slime load is abnormal;

[0086] The structure monitoring module evaluates the wear of the drum: the current lining thickness (47 mm) is worn by 6% compared with the initial value (50 mm), exceeding the preset wear rate threshold of 5%. The system calculates through the wear index and confirms that there is an abnormal wear trend;

[0087] Risk assessment of the screen differential pressure: the screen differential pressure (0.3 MPa) exceeds the safety threshold (0.2 MPa), triggers the analysis of the differential pressure index formula, and determines that there is a risk of screen blockage (the index tends to the warning range);

[0088] Judgment of the structural health: By integrating the wear index and the differential pressure index, a structural state index of 0.003 is generated, which is much lower than the health threshold (0.342), and the system maintains the judgment of "normal structural state";

[0089] The pressure monitoring module analyzed the pressure exceeding the limit: the pressure exceeded the limit 3 times (1.2MPa, 1.3MPa, 1.1MPa) within 5 minutes, the exceeding limit accounted for 60%, the pressure fluctuation standard deviation was 0.18MPa, and the pressure state index was generated as 0.101 (threshold 0.2), and no warning was triggered;

[0090] Correction to linkage electrical safety: Due to insulation degradation detected by the electrical module, the system dynamically adjusted the pressure warning threshold to 0.171. After the update, the pressure index is still below the threshold and remains normal;

[0091] The electrical safety module conducted a joint analysis of temperature and insulation, and found that the winding temperature soared by 55°C within 5 minutes, and the insulation resistance (0.9MΩ / kV) did not reach the safety threshold;

[0092] The system generates an insulation degradation index of 0.949 (threshold 0.8) to determine the insulation aging risk;

[0093] Global impact: Dynamically reduce the pressure warning threshold and enhance the sensitivity to the pressure state of the cylinder;

[0094] The equipment status monitoring module comprehensive health decision management cycle (2 hours) summary: coal slime load status: in 24 monitoring cycles, 20 load abnormalities (accounting for 83%); structural abnormalities: 8 wear or pressure difference exceeding the standard (accounting for 33%); electrical abnormalities: 15 insulation degradation warnings (accounting for 62%);

[0095] Health determination logic:

[0096] Weight distribution: structural health weight 70%, electrical safety weight 30%.

[0097] Comprehensive score: 0.421 (threshold 0.26), far beyond the safety range;

[0098] System output: Push maintenance instructions: "The equipment health status is abnormal, recommended priorities: ① coal slime dehydration; ② check screen blockage and drum wear; ③ motor insulation retest."

[0099] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the embodiments here. All obvious changes or modifications derived from the technical solution of the present invention are still within the protection scope of the present invention.

Claims

1. A coal preparation plant equipment health status monitoring system based on machine learning, characterized in that: include: The structural monitoring module is used to jointly analyze the equipment drum lining thickness and screen pressure difference collected during the monitoring period to construct a structural status index and judge the equipment structural status based on the structural status index; The pressure monitoring module is used to jointly analyze the cylinder pressure, cylinder pressure standard deviation and cylinder pressure average value to construct a pressure status index and update the judgment of the equipment structure status; The equipment status monitoring module is used to determine the equipment health status according to the coal slime load status, equipment structure status and electrical safety status within the management cycle, and output equipment maintenance prompts to the user based on the determination results of the equipment health status.

2. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 1 is characterized in that: The structure monitoring module includes a wear analysis unit, which is used to calculate the wear rate Ms for the equipment drum lining thickness m0 and the reference thickness m1 collected during the monitoring period, and compare and analyze the wear rate Ms with the preset wear rate ms0 to determine the equipment wear status, and construct a wear index based on the equipment wear status.

3. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 2 is characterized in that: The structure monitoring module also includes a pressure difference analysis unit for comparing and analyzing the screen pressure difference △P collected during the monitoring period with the preset pressure difference p1 to determine the pressure difference state of the equipment and construct a balance index according to the pressure difference state of the equipment, wherein: If △P is less than or equal to p1, the pressure difference analysis unit determines that the pressure difference state of the equipment in the current monitoring period is normal, and sets the equipment pressure difference index to YC1, setting YC1=0; otherwise, the pressure difference analysis unit sets the equipment pressure difference index to YC2, setting YC2=3×[(△P-p1) / (△P+p1)] 2 -2×[(△P-p1) / (△P+p1)] 3 .

4. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 3 is characterized in that: The structure monitoring module also includes a structure state judgment unit, which is used to construct a structure state index ZY according to the wear index and the balance index, and set ZY=lg(3×wear index+1) / lg4×balance index; The structural state judgment unit compares and analyzes the structural state index ZY and the preset state index zy to judge the structural state of the equipment. If ZY is less than or equal to zy, the structural state judgment unit determines that the structural state of the equipment in the current monitoring period is a healthy state, and does not issue a structural health state warning to the user. Otherwise, the structural state judgment unit determines that the structural state of the equipment in the current monitoring period is an abnormal state, and issues a structural health state warning to the user.

5. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 4 is characterized in that: The pressure monitoring module compares and analyzes the i-th cylinder pressure Pi collected during the monitoring period with the preset cylinder pressure threshold p0 to determine whether the cylinder pressure exceeds the limit. If Pi is less than or equal to p0, the pressure monitoring module determines that the i-th cylinder pressure does not exceed the limit. Otherwise, the pressure monitoring module determines that the i-th cylinder pressure exceeds the limit, and counts the number of data Pn of cylinder pressure exceeding the limit, and sets the cylinder pressure exceeding the limit ratio to β, β is the ratio of Pn to Pu, and Pu is the number of cylinder pressures collected during the monitoring period; The pressure monitoring module jointly analyzes the cylinder pressure over-limit ratio β, the cylinder pressure standard deviation σP and the cylinder pressure average value Pavg to construct a pressure state index PC, which is set as follows: PC = σP / Pavg × β; When the equipment structure state in the current monitoring period is in a healthy state, if the pressure state index PC is greater than or equal to the preset pressure state index pc0, the equipment structure state in the current monitoring period is determined to be in an abnormal state, and a structural health status warning is issued to the user.

6. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 5, characterized in that: The system also includes: an electrical safety monitoring module, which is used to fuse the motor insulation resistance and winding temperature collected during the monitoring period to construct an insulation degradation index, and to judge the electrical safety status based on the insulation degradation index, and to update the construction process of the pressure state index based on the electrical safety status.

7. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 6, characterized in that: The electrical safety monitoring module includes a construction unit for calculating a winding temperature gradient ΔT according to the winding temperature collected during a monitoring period, wherein the winding temperature gradient ΔT is a difference between a maximum value and a minimum value of the winding temperature collected during the monitoring period; The construction unit performs data fusion on the winding temperature gradient ΔT and the motor insulation resistance RI to construct the insulation degradation index HI.

8. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 2, characterized in that: The electrical safety monitoring module includes an electrical safety monitoring unit, which is used to compare and analyze the insulation degradation index HI and the electrical safety threshold IDI to determine the electrical safety state. If HI is less than or equal to IDI, the electrical safety monitoring unit determines that the electrical safety state is a normal state. Otherwise, the electrical safety monitoring unit determines that the electrical safety state is an insulation aging risk state, and updates the construction process of the pressure state index by updating the preset pressure state index, sets the updated preset pressure state index to pc1, and sets the calculation formula of pc1 to pc1=pc0×exp[-(HI-IDI)].

9. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 8, characterized in that: The equipment status monitoring module includes a status monitoring unit. When the coal slime load state is normal, if L1×j1 / J+L2×j2 / J is less than or equal to the first equipment status threshold u1, the status monitoring unit determines that the equipment health state of the current management cycle is a normal state. Otherwise, the status monitoring unit determines that the equipment health state of the current management cycle is an abnormal state. L1 is a structural weight, L2 is an electrical safety weight, L1+L2=1, j1 is the number of monitoring cycles in which the equipment structure state is an abnormal state within the management cycle, j2 is the number of monitoring cycles in which the electrical safety state is an insulation aging risk state, and J is the number of monitoring cycles within the management cycle; When the coal slime load state is normal, if L1×j1 / J+L2×j2 / J is less than or equal to the second equipment state threshold u2, the state monitoring unit determines that the equipment health state of the current management cycle is normal; otherwise, the state monitoring unit determines that the equipment health state of the current management cycle is abnormal.

10. The machine learning-based coal preparation plant equipment health status monitoring system according to claim 9, characterized in that: The system also includes: a data acquisition module for collecting equipment operation data and raw material parameters; The coal slime characteristic analysis module is used to perform data fusion analysis on the coal slime moisture content and coal slime particle size to construct a load coefficient, and judge the coal slime load state according to the load coefficient; the coal slime characteristic analysis module constructs the load coefficient LC according to the collected coal slime moisture content s0 and coal slime particle size d0, and compares and analyzes the load coefficient LC with the preset load threshold LC0 to judge the coal slime load state, wherein: If LC is less than or equal to LC0, the coal slime characteristic analysis module determines that the coal slime load state is normal, otherwise, the coal slime characteristic analysis module determines that the coal slime load state is abnormal.

Images