A fault prediction alarm system and method for coke oven tamping hammer

Abstract

The present application relates to the technical field of tamping hammer monitoring, and particularly relates to a coke oven tamping hammer fault prediction alarm system and method, the method comprising: setting a monitoring point; collecting parameters in real time; screening a first temporary hammer head; screening a second temporary hammer head; determining an abnormal hammer head; dynamically adjusting a fluctuation threshold; adjusting parameters and giving an early warning. The present application establishes a fault prediction logic chain through multi-dimensional parameter cross-fusion, identifies mechanical motion abnormalities by reflecting the actual movement trajectory of the tamping hammer through image change values; further verifies the stability of the striking process through vibration acceleration; image definition is used to judge the strength and accuracy of the actual contact of the hammer head with the coal seam, and to corroborate the reliability of vibration and change data; the quality of the finished coal cake feeds back the comprehensive effect of the entire tamping process, effectively solving the problems of low monitoring accuracy and slow response speed caused by excessive reliance on simulation devices for monitoring, which in turn causes deviations from the real tamping environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tamping hammer monitoring, and in particular to a coke oven tamping hammer fault prediction and alarm system and method. BACKGROUND

[0002] In the production process of a coke oven, the tamping hammer, as a key device, directly affects the quality of coal briquette formation and the stability of subsequent carbonization process. However, due to the complex working environment of the coke oven, frequent impact load and long-term high-frequency operation of the equipment, the tamping hammer is prone to abnormal phenomena such as loosening, wear, and blockage. If these phenomena are not identified and handled in a timely manner, it will not only lead to a decrease in coke strength and an increase in energy consumption, but also may cause equipment failure and safety hazards, thereby restricting the continuous improvement of the production efficiency and product quality of the coke oven. Therefore, it is urgent to establish an efficient and accurate fault identification and early warning mechanism that can timely detect the deviation of the running state at an early stage of abnormality and dynamically evaluate the evolution trend, so as to realize the early prediction and intervention of potential faults and ensure the continuity and stability of the coke oven production system.

[0003] The patent document with the publication number CN110389025A discloses a tamping hammer simulation stop monitoring system, which comprises a tamping machine simulation device, a data acquisition device, a data analysis and transmission device, a cloud data monitoring platform, and at least one client. The tamping machine simulation device communicates wirelessly with the cloud data monitoring platform. The data acquisition device is arranged on the tamping machine simulation device to monitor the running state of the tamping machine simulation device in real time. The data analysis and transmission device is used to analyze the parameters collected by the data acquisition device and send the monitored information to the cloud data monitoring platform. The client exchanges data with the cloud data monitoring platform through wireless communication. When the data acquisition device detects an abnormal running state of the tamping machine simulation device, the cloud data monitoring platform sends a control instruction to the client for fault warning.

[0004] It can be seen that the tamping hammer simulation stop monitoring system has the following problems: the system relies on the simulation device for monitoring, which leads to a deviation from the real tamping environment and makes it difficult to accurately reflect the actual running state; the data acquisition is based only on the running state of the tamping machine simulation device, which limits the accuracy of abnormality identification and is prone to false positives or false negatives; the monitoring system mainly relies on a single data type for judgment, which makes it impossible to identify the root cause of the fault from multiple dimensions and weakens the prediction ability; the control instruction is passively issued only after detecting an abnormality, which lacks targeted adjustment means and makes it impossible to realize dynamic optimization control of parameters; the cloud platform mainly serves as an information transfer platform and lacks real-time analysis capabilities, which leads to a large delay in system response and low processing efficiency. SUMMARY

[0005] To this end, the present application provides a coke oven rammer fault prediction alarm system and method to overcome the low monitoring accuracy and slow response speed caused by excessive reliance on analog devices for monitoring in the prior art, by real-time collection and dynamic threshold adjustment of multi-source parameters.

[0006] To achieve the above-mentioned purpose, in one aspect, the present application provides a coke oven rammer fault prediction alarm method, comprising:

[0007] Setting a plurality of monitoring points on a coke oven ramming flow line comprising a plurality of rammers arranged side by side;

[0008] Real-time collection of image change value, vibration acceleration, image definition, texture roughness of output coal cake and compaction density of each rammer during operation at a preset striking frequency and a preset striking stroke;

[0009] Screening a plurality of first temporary hammer heads from all the rammers according to the image change value and a preset change fluctuation threshold;

[0010] Screening a plurality of second temporary hammer heads from all the first temporary hammer heads according to the vibration acceleration and the change of the image change value of each first temporary hammer head within a preset judgment duration;

[0011] Screening a plurality of abnormal hammer heads from all the second temporary hammer heads according to the vibration acceleration and the image definition of each second temporary hammer head and the next second temporary hammer head adjacent thereto;

[0012] Adjusting the preset change fluctuation threshold according to the compaction density, arrangement position and number of the abnormal hammer heads to obtain an adjusted change fluctuation threshold;

[0013] Adjusting the preset striking frequency or the preset striking stroke according to the texture roughness collected within a preset correction duration based on the abnormal hammer heads re-screened based on the adjusted change fluctuation threshold;

[0014] Issuing an alarm for the abnormal hammer heads re-screened based on the adjusted preset striking frequency or the adjusted preset striking stroke.

[0015] Further, the process of screening a plurality of second temporary hammer heads from all the first temporary hammer heads according to the vibration acceleration and the change of the image change value of each first temporary hammer head within a preset judgment duration comprises:

[0016] Drawing a curve of the image change value changing with time within the preset judgment duration to obtain a change value change curve;

[0017] calculate slopes of the change value change curve at each adjacent time, to obtain a plurality of slopes;

[0018] When the absolute value of the slope is greater than a preset absolute value of slope threshold, or when positive and negative changes exceeding a preset number of flips occur continuously within a preset observation period, calculate a difference between the image change value at the final time and the image change value at the initial time within the preset determination duration, to obtain a change difference;

[0019] When the change difference is greater than a preset change difference threshold, determine the first temporary hammer head to be the second temporary hammer head according to the change of the vibration acceleration, to screen a plurality of second temporary hammer heads from all first temporary hammer heads.

[0020] Further, the process of determining the first temporary hammer head to be the second temporary hammer head according to the change of the vibration acceleration comprises:

[0021] Calculate the average value of all the vibration accelerations within the preset determination duration, to obtain an average acceleration, and mark the vibration acceleration at the last time of the preset determination duration, to obtain a marked acceleration;

[0022] Calculate the ratio of the difference between the average acceleration and the marked acceleration to the average acceleration, to obtain a vibration acceleration attenuation rate;

[0023] When the vibration acceleration attenuation rate is greater than a preset attenuation rate threshold, determine the first temporary hammer head to be the second temporary hammer head.

[0024] Further, the process of screening a plurality of abnormal hammer heads from all second temporary hammer heads according to the vibration acceleration and the image definition of each second temporary hammer head and the next second temporary hammer head adjacent thereto comprises:

[0025] Calculate the absolute value of the ratio of the difference between the image definition of each second temporary hammer head and the next second temporary hammer head adjacent thereto and the image definition of the second temporary hammer head, to obtain a definition change rate;

[0026] Calculate the standard deviation of all the definition change rates within a preset observation duration, to obtain a definition change fluctuation value;

[0027] When the definition change fluctuation value is greater than a preset definition change fluctuation threshold, screen a plurality of abnormal hammer heads from all second temporary hammer heads according to the vibration acceleration of each second temporary hammer head and the next second temporary hammer head adjacent thereto.

[0028] Further, the process of screening a plurality of abnormal hammer heads from all second temporary hammer heads according to the vibration acceleration of each second temporary hammer head and the next second temporary hammer head adjacent thereto comprises:

[0029] calculate the average value of the vibration acceleration of each of the second temporary hammer head in the preset observation duration, to obtain an observation average acceleration;

[0030] calculate the average value of the vibration acceleration of each of the next second temporary hammer head in the preset observation duration, to obtain a comparative average acceleration;

[0031] calculate the ratio of the difference between the observation average acceleration and the comparative average acceleration to the observation average acceleration, to obtain a vibration acceleration attenuation rate;

[0032] when the vibration acceleration attenuation rate is greater than a preset attenuation rate threshold, determine that the corresponding two second temporary hammer heads are both the abnormal hammer heads, to screen out a plurality of abnormal hammer heads from all second temporary hammer heads.

[0033] Further, the process of adjusting the preset variation fluctuation threshold according to the compaction density, the arrangement position and the number of the abnormal hammer heads includes:

[0034] when the number of the abnormal hammer heads is greater than a preset number threshold, obtain the distance between every two abnormal hammer heads according to the arrangement position, to obtain a plurality of hammer head spacings;

[0035] calculate the standard deviation of all the hammer head spacings, to obtain a hammer head distribution degree;

[0036] when the hammer head distribution degree is less than a preset distribution degree threshold, adjust the preset variation fluctuation threshold according to the number of the abnormal hammer heads and the compaction density, to obtain an adjusted variation fluctuation threshold.

[0037] Further, the process of adjusting the preset variation fluctuation threshold according to the number of the abnormal hammer heads and the compaction density includes:

[0038] when the compaction density is less than a preset density threshold, calculate the relative deviation between the compaction density and the preset density threshold, to obtain a first adjustment factor;

[0039] calculate the ratio of the number of the abnormal hammer heads to the number of all the ramming hammers, to obtain a second adjustment factor;

[0040] adjust the preset variation fluctuation threshold according to the first adjustment factor, a preset first adjustment weight, the second adjustment factor and a preset second adjustment weight, to obtain an adjusted variation fluctuation threshold.

[0041] Further, the process of adjusting the preset striking frequency or the preset striking stroke according to the texture roughness collected from the abnormal hammer heads re-screened according to the adjusted variation fluctuation threshold within a preset correction duration includes:

[0042] calculating an average value of all the texture roughnesses in the preset correction duration to obtain an average roughness;

[0043] when the average roughness is greater than a maximum value of the preset roughness range, decreasing the preset striking frequency according to a relative deviation between the average roughness and the maximum value of the preset roughness range and a preset correction coefficient;

[0044] when the average roughness is less than a minimum value of the preset roughness range, increasing the preset striking stroke according to a relative deviation between the minimum value of the preset roughness range and the average roughness and the preset correction coefficient.

[0045] Further, the process of screening a plurality of first temporary hammer heads from all the tamping hammers according to the image change value and a preset change fluctuation threshold value comprises:

[0046] calculating a standard deviation of the image change value in a preset first temporary duration to obtain an image change fluctuation value;

[0047] when the image change fluctuation value is greater than the preset change fluctuation threshold value, determining the tamping hammer as the first temporary hammer head to screen a plurality of first temporary hammer heads.

[0048] On the other hand, the present application also provides a fault prediction and alarm system for coke oven tamping hammers, comprising:

[0049] a collection module, configured to collect, at a plurality of monitoring points arranged on a coke oven tamping flow line comprising a plurality of tamping hammers arranged side by side, image change values, vibration accelerations, image definition, texture roughnesses and compaction densities of output coal cakes of each tamping hammer in a running process at a preset striking frequency and a preset striking stroke;

[0050] a first temporary screening module, connected with the collection module, configured to screen a plurality of first temporary hammer heads from all the tamping hammers according to the image change value and a preset change fluctuation threshold value;

[0051] a second temporary screening module, connected with the collection module and the first temporary screening module respectively, configured to screen a plurality of second temporary hammer heads from all the first temporary hammer heads according to changes of the vibration acceleration and the image change value of each first temporary hammer head in a preset determination duration;

[0052] an abnormality screening module, connected with the collection module and the second temporary screening module respectively, configured to screen a plurality of abnormal hammer heads from all the second temporary hammer heads according to the vibration acceleration and the image definition of each second temporary hammer head and a next second temporary hammer head adjacent thereto;

[0053] a first adjusting module, connected with the collecting module and the anomaly screening module respectively, configured to adjust the preset fluctuation threshold according to the compaction density, the arrangement position and the number of the anomaly hammer head, so as to obtain an adjusted fluctuation threshold;

[0054] a second adjusting module, connected with the collecting module, configured to adjust the preset striking frequency or the preset striking stroke according to the texture roughness collected when the anomaly hammer head is re-screened based on the adjusted fluctuation threshold within a preset correction time length;

[0055] an alarm module, connected with the anomaly screening module, configured to issue an alarm for the anomaly hammer head re-screened based on the adjusted preset striking frequency or the adjusted preset striking stroke.

[0056] Compared with the prior art, the beneficial effects of the present application are that the multi-dimensional parameter cross-fusion establishes a fault prediction logic chain, and the precision and response speed of the rammer fault identification are improved. The image change value reflects the actual movement trajectory of the rammer, and the deviation from the preset fluctuation threshold can identify mechanical movement anomalies; the vibration acceleration further verifies the stability of the striking process and captures potential structural faults; the image definition is used to judge the force and precision of the actual contact of the hammer head with the coal seam, and to prove the credibility of the vibration and change data; the compaction density and the texture roughness reflect the quality of the finished coal cake, and feedback the comprehensive effect of the entire ramming process. Based on the multi-level screening and closed-loop control logic of "primary screening - secondary screening - comprehensive determination - self-adaptive adjustment", three layers of data linkage from mechanical characteristics - finished product quality - operating parameters are realized, false positives and false negatives are effectively avoided, the continuous and stable operation of the coke oven is ensured, and the service life of the equipment is prolonged, effectively solving the problems of low monitoring accuracy and slow response speed caused by excessive reliance on simulation devices for monitoring and deviation from the real ramming environment.

[0057] Further, through joint determination of multi-dimensional parameters, the trend of change (slope and number of reversals) and the change of vibration acceleration are combined to effectively capture subtle and abnormal dynamic fluctuations of the hammer head in operation, avoiding misjudgment caused by a single parameter. The preset slope threshold reflects the degree of change in change, the number of reversals reflects the instability of change, the change difference quantifies the overall deviation amplitude, and the vibration acceleration as a dynamic response parameter assists in confirming the anomaly. The logical correlation between these parameters through time series changes constitutes a multi-level dynamic monitoring mechanism for the equipment state, thereby improving the sensitivity and accuracy of fault identification and ensuring the stable operation and safe maintenance of the coke oven rammer.

[0058] Further, by comparing the average acceleration with the decay rate of the acceleration at the end time, not only the overall trend of the vibration response can be quantified, but also the sudden drop of the end acceleration can be sensitively captured. The average value reflects the normal inertia level, the end value reveals the instantaneous state of the current output, and the difference and ratio of the two values are related to the hammer dynamics and the real-time mechanical load change, forming a three-layer monitoring of "baseline-instantaneous response-energy decay". This effectively avoids single-point noise interference and improves the robustness and accuracy of fault identification.

[0059] Further, by jointly determining the clarity change rate and its standard deviation, the stability of image blurring and imaging jitter can be quantified, reflecting the microscopic inconsistency of the hammer and the coal cake contact. When the clarity fluctuation is abnormally triggered, the vibration acceleration is further investigated, which can closely link the "visual abnormalities" at the image level with the "dynamic abnormalities" at the mechanical level, forming a three-level screening of "visual clues-fluctuation threshold-vibration verification". The multi-parameter and multi-level fusion determination can not only eliminate false alarms caused by single sensor noise, but also accurately capture underlying faults such as poor contact or mechanical looseness, significantly improving the reliability and timeliness of the alarm.

[0060] Further, through "front and back hammer comparison", the influence of global vibration baseline drift is effectively eliminated: the observation average acceleration reflects the normal dynamic level, the comparison average acceleration reveals the instantaneous response difference of adjacent hammers, and the decay rate of the two directly quantifies the decline of the striking energy; When the decay rate exceeds the threshold, it indicates that the mechanical transmission or striking contact is abnormal. Through dynamic monitoring of adjacent comparison, it can accurately distinguish between general fluctuations caused by overall equipment vibration or environmental interference and local attenuation caused by local hammer failure, significantly improving the reliability and timeliness of abnormal identification.

[0061] Further, through the closed-loop self-adaptation of the three parameters of "abnormal number-spatial distribution-compaction density", it can distinguish between single-point occasional faults and regional concentrated failures: the number of abnormal hammers reflects the overall fault degree, and the distribution quantifies whether the fault is clustered; When the fault is concentrated, local vibration or fluctuation may overlap, and the threshold should be relaxed to reduce false alarms; At the same time, the compaction density is closely related to the striking quality, and its deviation participates in threshold adjustment to maintain the sensitivity to the output quality. Through the bottom-up parameter linkage and closed-loop correction, the intelligent optimization of the fault determination threshold is realized, which significantly improves the discrimination accuracy and robustness of abnormal hammers under complex working conditions.

[0062] Further, the finished product quality (compaction density) is combined with the fault severity (abnormal quantity), the potential impact of insufficient compaction on system stability is quantified by a first adjustment factor, the fault diffusion degree is reflected by a second adjustment factor, and then the three of "quality-fault-threshold value" are flexibly integrated according to the weight to build a closed-loop adaptive mechanism. In this way, when the coal cake is too loose or the faults are concentrated, the change fluctuation threshold can be dynamically relaxed or tightened to accurately match the current working condition, realizing high sensitivity to real faults and high robustness to environmental interference.

[0063] Further, the standard deviation of the image change value is used as a quantitative indicator of running stability, which can objectively reflect the position fluctuation during the hammer hitting process. The larger the standard deviation, the more unstable the hammer hitting behavior, and there may be problems such as structural loosening, obstruction or abnormal energy transmission. The preset change fluctuation threshold is used as a screening benchmark, combined with the preset first temporary time length to ensure that the data sampling has time continuity and statistical representativeness, thereby avoiding false judgments caused by short-term disturbances, and building an initial screening mechanism based on image recognition and statistical analysis, providing a data basis and target hammer range for subsequent higher-level fault identification, with the advantages of fast response, efficient screening and low cost.

[0064] Further, the relative deviation between the average roughness and the upper and lower limits of the preset roughness is used to establish a direct correlation between roughness feedback and device parameter adjustment, and the correction coefficient is used as the core variable of adjustment sensitivity to realize quantitative coupling between roughness feedback and hitting frequency / travel. The underlying logic is to use surface texture as an indirect representation of compaction effect, combined with the trend of roughness exceeding or being insufficient, to dynamically optimize the hitting parameters and improve the system's adaptive adjustment capability and compaction quality consistency. BRIEF DESCRIPTION OF DRAWINGS

[0065] Figure 1 Flowchart of the fault prediction and alarm method of the coke oven rammer in the embodiment;

[0066] Figure 2 Determination logic diagram for determining the second temporary hammer in the embodiment;

[0067] Figure 3 Determination logic diagram for determining the abnormal hammer in the embodiment;

[0068] Figure 4 Determination logic diagram for adjusting the preset change fluctuation threshold in the embodiment. DETAILED DESCRIPTION

[0069] In order to make the purpose and advantages of the present application clearer and more apparent, the present application will be further described below in conjunction with the embodiments; it should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0070] The preferred embodiments of the present application will be described below with reference to the accompanying drawings. Those skilled in the art will appreciate that the embodiments are only used to explain the technical principles of the present application, and are not intended to limit the protection scope of the present application.

[0071] In one aspect, referring to Figure 1 shown, which is a flowchart of the failure prediction alarm method of the coke oven rammer in the embodiment;

[0072] The embodiment provides a failure prediction alarm system of a coke oven rammer, comprising:

[0073] A plurality of monitoring points are arranged on a coke oven ramming flow line comprising a plurality of rammers arranged side by side;

[0074] Real-time image change values, vibration accelerations, image definition, texture roughness and compaction density of each rammer during operation at a preset striking frequency and a preset striking stroke are collected;

[0075] A plurality of first temporary hammer heads are selected from all the rammers according to the image change values and a preset change fluctuation threshold;

[0076] A plurality of second temporary hammer heads are selected from all the first temporary hammer heads according to the vibration accelerations and the image change values of each first temporary hammer head within a preset judgment duration;

[0077] A plurality of abnormal hammer heads are selected from all the second temporary hammer heads according to the vibration accelerations and the image definition of each second temporary hammer head and the next second temporary hammer head adjacent thereto;

[0078] The preset change fluctuation threshold is adjusted according to the compaction density, arrangement position and number of the abnormal hammer heads, to obtain an adjusted change fluctuation threshold;

[0079] The preset striking frequency or the preset striking stroke is adjusted according to the texture roughness collected during the abnormal hammer heads reselected based on the adjusted change fluctuation threshold within a preset correction duration;

[0080] An alarm is sent to the abnormal hammer heads reselected based on the adjusted preset striking frequency or the adjusted preset striking stroke.

[0081] On the coke oven ramming flow line, multiple ramming hammers are arranged side by side at a predetermined interval, and a plurality of monitoring points are arranged at the corresponding positions of the ramming hammers. Each monitoring point is provided with a high-frame-rate industrial camera, a three-axis acceleration sensor and an image processing module for multi-dimensional perception of the running state of the ramming hammer. The industrial camera is installed at an optimal upward or lateral position in terms of visual angle, and is used to collect image sequences of the ramming hammer in the working process in real time. The image recognition algorithm is used to calculate the image change value of the hammer head and the image definition in the striking process, so as to determine the motion trajectory and contact stability of the hammer head. The acceleration sensor is installed on the hammer body structure, and is used to obtain vibration acceleration data in the striking process in real time, so as to reflect the mechanical impact stability and structural integrity. In addition, an intelligent image detection unit and a density sensor are arranged at the discharge end, and are used to perform texture image analysis on the coal cake formed by each strike, extract the surface texture roughness features of the coal cake, and simultaneously measure the compaction density, so as to evaluate the ramming quality. All the collected data are preprocessed by an on-site edge computing device and then uploaded to a fault analysis system, so as to provide full-quantity basic data support for subsequent fault screening, dynamic adjustment and early warning. The above collection techniques are all prior art, and will not be described here.

[0082] For the abnormal hammer head re-screened based on the adjusted preset striking frequency or preset striking stroke, the system sends a vibration abnormality alarm in real time, the content of which includes the number and position of the abnormal hammer head, and the maintenance personnel are notified through the sound and light alarm and the remote monitoring platform, so as to prompt timely maintenance to prevent vibration abnormalities from causing equipment damage and production safety hazards.

[0083] The preset striking frequency refers to the number of strikes per unit time of the ramming hammer, and is determined by the energy input required for the formation of the coke cake and the production rhythm, and is usually set to be between 20 times / minute and 60 times / minute, and is set to be 40 times / minute in this embodiment, which can ensure the balance between the formation quality and the equipment life.

[0084] The preset striking stroke refers to the up-down movement distance of each strike of the ramming hammer, and is determined by the coal layer thickness and the compaction characteristics of the coal type, and is usually set to be between 100 mm and 300 mm, and is set to be 200 mm in this embodiment, which can ensure uniform compaction and stable structure of the coal cake.

[0085] The preset change fluctuation threshold refers to the maximum fluctuation range of the image change value that is acceptable, and is determined by the equipment running stability requirement and the error tolerance, and is usually set to be between ±3 mm and ±10 mm, and is set to be ±5 mm in this embodiment, which can effectively identify abnormal vibration or stroke deviation.

[0086] The preset determination duration refers to the data analysis time period for judging the first temporary hammer head state, and is determined by the striking period and the data change rate, and is usually set to be between 10 seconds and 60 seconds, and is set to be 30 seconds in this embodiment, which can stably reflect the ramming state trend.

[0087] The preset correction duration refers to a data time period for collecting further characteristics of the abnormal hammer head and performing parameter correction, and is determined according to the adjustment response period and the data convergence speed. The preset correction duration is usually set to be between 10 seconds and 60 seconds, and is set to be 20 seconds in the embodiment. The preset correction duration can quickly realize dynamic optimization of the striking parameters.

[0088] The system collects key parameters in the operation process of the tamping hammer at multiple monitoring points on the coke oven tamping assembly line in real time, including image change value, vibration acceleration, image definition, coal cake texture roughness and compaction density. The system preliminarily screens out first temporary hammer heads according to a preset image change fluctuation threshold, further screens out second temporary hammer heads according to vibration acceleration and change within a certain time period, and extracts the final abnormal hammer heads in combination with the image definition and vibration characteristics of adjacent hammer heads. Subsequently, the system dynamically corrects the change fluctuation threshold according to the spatial distribution, quantity and compaction density of the abnormal hammer heads, and adjusts the striking frequency or stroke of the tamping hammer through the texture roughness data feedback, so as to realize early warning and adaptive control of faults.

[0089] The multi-dimensional parameter cross-fusion establishes a fault prediction logic chain, and improves the accuracy and response speed of tamping hammer fault identification. The image change value reflects the actual motion trajectory of the tamping hammer, and the deviation from the preset fluctuation threshold can identify mechanical motion abnormalities. The vibration acceleration further verifies the stability of the striking process and captures potential structural faults. The image definition is used to judge the force and precision of the actual contact of the hammer head with the coal bed, and to prove the credibility of the vibration and change data. The compaction density and texture roughness reflect the quality of the finished coal cake, and feedback the overall effect of the tamping process. Based on the multi-level screening and closed-loop control logic of "preliminary screening-re-screening-comprehensive determination-adaptive adjustment", three layers of data linkage from mechanical characteristics-product quality-operation parameters are realized, false positives and false negatives are effectively avoided, the continuous and stable operation of the coke oven is ensured, and the service life of the equipment is prolonged. The problem of low monitoring accuracy and slow response speed caused by excessive reliance on simulation devices for monitoring and deviation from the actual tamping environment is effectively solved.

[0090] Specifically, the process of screening a plurality of second temporary hammer heads from all the first temporary hammer heads according to the vibration acceleration and the change of the image change value of each of the first temporary hammer heads within a preset determination duration includes:

[0091] A curve of the image change value changing with time within the preset determination duration is drawn to obtain a change value change curve;

[0092] The slope of the change value change curve at each adjacent time is calculated to obtain a plurality of slopes;

[0093] When the absolute value of the slope is greater than a preset absolute value of slope threshold, or when positive and negative changes exceeding a preset number of reversals occur continuously within a preset observation period, a difference between an image change value at a final time and an image change value at an initial time within the preset determination duration is calculated to obtain a change difference;

[0094] When the change difference is greater than a preset change difference threshold, it is determined according to the change of the vibration acceleration that the first temporary hammer head is the second temporary hammer head, so as to screen out a plurality of second temporary hammer heads from all the first temporary hammer heads.

[0095] The preset absolute value of slope threshold is a parameter for judging whether the image change change rate is abnormal, depends on the normal vibration characteristics of the hammer head operation, and is usually set to be between 0.1 and 1.0 mm / s, which is set to 0.5 mm / s in the embodiment, and can effectively identify rapid and abnormal change fluctuations.

[0096] The preset number of reversals is a parameter for judging the frequency of positive and negative changes in the change curve, depends on the stability and noise level of the device vibration, and is usually set to be between 3 and 10 times, which is set to 5 times in the embodiment, and can accurately capture periodic abnormal fluctuations of the change.

[0097] The preset change difference threshold is a parameter for determining whether the cumulative change amplitude of the change exceeds the normal range, depends on the maximum change allowed by the hammer head design, and is usually set to be between 1 and 5 mm, which is set to 3 mm in the embodiment, and can effectively distinguish between normal micro-movement and abnormal deviation.

[0098] By drawing a change curve of the image change value with time, the slope of the curve at adjacent times is calculated, and then whether the absolute value of the slope exceeds the preset threshold or whether multiple positive and negative reversals occur within the preset observation period is analyzed. If the conditions are met, the start and end difference of the change within the determination duration is calculated, and the vibration acceleration change is combined to further screen the first temporary hammer head, and finally determine a plurality of second temporary hammer heads, and accurately lock the hammer head that may have abnormal vibration and change fluctuations.

[0099] Through the joint determination of multiple parameters, the trend of the change (slope and number of reversals) and the change of the vibration acceleration are combined to effectively capture the subtle and abnormal dynamic fluctuations of the hammer head in operation, and to avoid misjudgment caused by a single parameter. The preset slope threshold reflects the severity of the change change, the number of reversals reflects the instability of the change, and the change difference quantifies the overall deviation amplitude. The vibration acceleration as a dynamic response parameter assists in confirming the abnormality. The logical correlation between these parameters through time series changes constitutes a multi-level dynamic monitoring mechanism for the device state, thereby improving the sensitivity and accuracy of fault identification and ensuring the stable operation and safe maintenance of the coke oven rammer.

[0100] Please continue to readFigure 2 Fig. 2 shows a determination logic diagram for determining the second temporary hammer head according to the present embodiment;

[0101] The process of determining the first temporary hammer head as the second temporary hammer head according to the change of the vibration acceleration comprises:

[0102] The average value of all the vibration accelerations within the preset determination duration is calculated to obtain an average acceleration, and the vibration acceleration at the last moment of the preset determination duration is marked to obtain a marked acceleration;

[0103] The ratio of the difference between the average acceleration and the marked acceleration to the average acceleration is calculated to obtain a vibration acceleration decay rate;

[0104] When the vibration acceleration decay rate is greater than a preset decay rate threshold, the first temporary hammer head is determined as the second temporary hammer head.

[0105] The vibration acceleration sequence is collected within the preset determination duration of each first temporary hammer head, the average value of all the acceleration values within the duration is calculated as a reference baseline, the acceleration at the last moment of the determination duration is marked as a current response value, and the ratio of the difference between the two to the average value is the vibration acceleration decay rate. If the decay rate exceeds the preset threshold, it indicates that the vibration energy of the hammer head is significantly lost during the striking process, i.e., the striking force or the transmission is abnormal, and the hammer head is upgraded from the first temporary set to the second temporary set.

[0106] By comparing the average acceleration with the decay rate of the acceleration at the last moment, not only the overall trend of the vibration response can be quantified, but also the sudden drop of the end acceleration can be sensitively captured. The average value reflects the inertia level of normal striking, the end value reveals the instantaneous state of the current output, and the difference and ratio between the two are related to the dynamics of the hammer head and the real-time change of the mechanical load, forming a three-layer monitoring of "baseline-instantaneous response-energy decay". This effectively avoids single-point noise interference and improves the robustness and accuracy of fault identification.

[0107] Specifically, the process of screening a number of abnormal hammer heads from all the second temporary hammer heads according to the vibration acceleration and the image clarity of each second temporary hammer head and the next second temporary hammer head adjacent thereto comprises:

[0108] The absolute value of the ratio of the difference between the image clarity of each second temporary hammer head and the next second temporary hammer head adjacent thereto and the image clarity of the second temporary hammer head is calculated to obtain a clarity change rate;

[0109] The standard deviation of all the clarity change rates within a preset observation duration is calculated to obtain a clarity change fluctuation value;

[0110] When the clearness change fluctuation value is greater than the preset clearness change fluctuation threshold, a number of abnormal hammer heads are screened out from all the second temporary hammer heads according to the vibration acceleration of each second temporary hammer head and the next second temporary hammer head adjacent thereto.

[0111] The preset observation time length is a data window length for capturing state change of adjacent hammer heads, and is determined according to a tamping hammer striking period and a system response rate, and is usually set to be between 10 seconds and 60 seconds, and is set to be 30 seconds in the embodiment, and can smooth short-term fluctuations and identify abnormal trends in time.

[0112] The preset clearness change fluctuation threshold is a critical value for distinguishing normal imaging jitter from abnormal blur fluctuation, and is determined according to a camera noise level, illumination change and coal dust interference degree, and is usually set to be between 0.05 and 0.2, and is set to be 0.1 in the embodiment, and can effectively distinguish slight clearness fluctuation caused by mechanical jitter or light and shadow interference from significant image blur caused by poor hammer head contact.

[0113] For each second temporary hammer head and the next hammer head adjacent thereto, the absolute value of the ratio of the image clearness difference value corresponding to each strike in the preset observation time length to the clearness of the former is calculated to obtain a series of clearness change rates, then the standard deviation of all the change rates is calculated as the clearness fluctuation value, and when the fluctuation value exceeds the preset threshold, it indicates that the imaging quality of the adjacent hammer heads appears abnormal fluctuation, at this time, the system combines the vibration acceleration of the two hammer heads to mark the hammer head with abnormal or attenuated vibration acceleration as the final abnormal hammer head.

[0114] Through the joint determination of the clearness change rate and the standard deviation thereof, the stability of image blur and imaging jitter can be quantified, and the microscopic inconsistency of the contact between the hammer head and the coal cake is reflected, when the clearness fluctuation abnormality is triggered, the vibration acceleration is further investigated, the “visual abnormality” at the image level and the “dynamic abnormality” at the mechanical level are closely related, the three-level screening of “visual clue-fluctuation threshold-vibration verification” is formed, the multi-parameter and multi-level fusion determination can not only eliminate false alarms caused by single sensor noise, but also accurately capture underlying faults such as poor striking contact or mechanical looseness, and the reliability and timeliness of the alarm are significantly improved.

[0115] Please continue to refer to Figure 3 As shown in the figure, it is a determination logic diagram for determining abnormal hammer heads in the embodiment.

[0116] The process of screening a number of abnormal hammer heads from all the second temporary hammer heads according to the vibration acceleration of each second temporary hammer head and the next second temporary hammer head adjacent thereto includes:

[0117] The average value of the vibration acceleration of each second temporary hammer head in the preset observation time length is calculated to obtain an observation average acceleration.

[0118] calculating the average value of the vibration acceleration of each of the next second temporary hammer head in the preset observation time length, to obtain a comparative average acceleration;

[0119] calculating the ratio of the difference between the observation average acceleration and the comparative average acceleration to the observation average acceleration, to obtain a vibration acceleration attenuation rate;

[0120] when the vibration acceleration attenuation rate is greater than a preset attenuation rate threshold, determining that the corresponding two second temporary hammer heads are both the abnormal hammer heads, to screen out a plurality of abnormal hammer heads from all the second temporary hammer heads.

[0121] The preset attenuation rate threshold is a critical ratio for determining whether the vibration acceleration appears a significant drop, and depends on the normal vibration characteristics of the hammer head and the noise level, and is usually set to be between 10% and 50%, and is set to be 20% in the embodiment, so that the abnormal attenuation of the vibration energy can be accurately captured, and thus the possible mechanical failure can be timely identified.

[0122] For each pair of adjacent second temporary hammer heads, the average values of the vibration accelerations of the previous hammer head and the next hammer head in the preset observation time length are calculated respectively as the observation average acceleration and the comparative average acceleration; then the ratio of the difference between the two average values to the observation average acceleration is calculated to obtain the vibration acceleration attenuation rate; when the attenuation rate exceeds the preset attenuation rate threshold, the pair of hammer heads are both determined as abnormal hammer heads and are included in the final alarm range.

[0123] The influence of the global vibration baseline drift is effectively eliminated through the "front and back hammer head comparison": the observation average acceleration reflects the normal dynamic level, the comparative average acceleration reveals the instantaneous response difference of the adjacent hammer heads, and the attenuation rate of the two directly quantifies the drop degree of the striking energy; when the attenuation rate exceeds the threshold, it indicates that the mechanical transmission or the striking contact is abnormal. Through the dynamic monitoring of the adjacent comparison, the universal fluctuation caused by the overall shaking of the equipment or the environmental interference can be accurately distinguished from the local attenuation caused by the local hammer head failure, and the reliability and timeliness of the abnormal identification are greatly improved.

[0124] Please continue to refer to Figure 4 as shown in the drawing, which is a determination logic diagram for adjusting the preset variation fluctuation threshold in the embodiment;

[0125] The preset variation fluctuation threshold is adjusted according to the compaction density, the arrangement position and the number of the abnormal hammer heads, and the process of adjusting the variation fluctuation threshold comprises:

[0126] When the number of the abnormal hammer heads is greater than a preset number threshold, the distance between every two abnormal hammer heads is obtained according to the arrangement position, to obtain a plurality of hammer head distances;

[0127] The standard deviation of all the hammer head distances is calculated to obtain a hammer head distribution degree;

[0128] When the hammer head distribution is less than a preset distribution threshold, the preset fluctuation threshold is adjusted according to the number of abnormal hammer heads and the compaction density, to obtain an adjusted fluctuation threshold.

[0129] The preset distribution threshold is a standard deviation threshold for determining the spatial aggregation degree of abnormal hammer heads, and is determined according to a hammer head arrangement spacing tolerance and a process allowable deviation. The preset distribution threshold is usually set to be between 2 mm and 10 mm, and is set to be 5 mm in the embodiment. The preset distribution threshold can effectively distinguish between a scattered fault and a concentrated fault, and ensures the pertinence of threshold adjustment.

[0130] When the number of identified abnormal hammer heads exceeds a preset threshold, the relative positions of the abnormal hammer heads on the pipeline are first read, and the spacing sequence between any two adjacent abnormal hammer heads is calculated. Then, the standard deviation of the spacing sequence is calculated to quantify the uniformity of the hammer head distribution. If the distribution is lower than the preset threshold, it indicates that the fault hammer heads are concentrated or locally failed. In order to avoid false positives caused by position aggregation, the original fluctuation threshold is dynamically increased or decreased in combination with the number of abnormal hammer heads and the deviation degree of the coal briquette compaction density, so as to generate a new judgment threshold that is more suitable for the current fault distribution characteristics.

[0131] Through the closed-loop self-adaptation of the three parameters of “abnormal number-spatial distribution-compaction density”, single-point accidental faults and regional concentrated failures can be distinguished. The number of abnormal hammer heads reflects the overall fault degree, and the distribution quantifies whether the fault is clustered. When the fault is concentrated, local vibration or fluctuation may overlap, and the threshold needs to be widened to reduce false positives. At the same time, the compaction density is closely related to the impact quality, and the deviation thereof is used to participate in the threshold adjustment to maintain the sensitivity to the output quality. Through the bottom-up parameter linkage and closed-loop correction, intelligent optimization of the fault judgment threshold is realized, and the discrimination accuracy and robustness of abnormal hammer heads under complex working conditions are greatly improved.

[0132] Specifically, the process of adjusting the preset fluctuation threshold according to the number of abnormal hammer heads and the compaction density to obtain an adjusted fluctuation threshold includes:

[0133] When the compaction density is less than a preset density threshold, a relative deviation between the compaction density and the preset density threshold is calculated to obtain a first adjustment factor.

[0134] A ratio of the number of abnormal hammer heads to the number of all the tamping hammers is calculated to obtain a second adjustment factor.

[0135] The preset change fluctuation threshold is adjusted according to the first adjustment factor, a preset first adjustment weight, the second adjustment factor and a preset second adjustment weight, to obtain an adjusted change fluctuation threshold, wherein T' = T x [1 + a x (D - D0) / D0] x (1 - b x n / N), T' is the adjusted change fluctuation threshold, T is the preset change fluctuation threshold, D is the compaction density, D0 is the preset density threshold, (D - D0) / D0 is the first adjustment factor, a is the preset first adjustment weight, b is the preset second adjustment weight, n is the number of abnormal hammer heads, N is the total number of the tamping hammers, n / N is the second adjustment factor.

[0136] The preset density threshold is a limit value for determining whether the compaction of the coal cake meets the standard, and is usually set to be between 0.8 g / cm3 and 1.2 g / cm3, and is set to be 1.0 g / cm3 in the embodiment, so that the working condition of insufficient compaction can be accurately identified.

[0137] The preset first adjustment weight is a coefficient for measuring the influence degree of the compaction density deviation on the change threshold adjustment, and is usually set to be between 0.3 and 0.7, and is set to be 0.6 in the embodiment, so that the quality feedback effect can be reasonably highlighted.

[0138] The preset second adjustment weight is a coefficient for measuring the influence degree of the number of abnormal hammer heads on the change threshold adjustment, and is usually set to be between 0.3 and 0.7, and is set to be 0.4 in the embodiment, so that the balance between the quantity feedback and the quality feedback can be ensured.

[0139] When the compaction density of the coal cake is lower than the preset density threshold, the system first calculates the relative deviation of the compaction density from the threshold to obtain the first adjustment factor, and simultaneously calculates the ratio of the number of abnormal hammer heads to the total number of hammer heads to obtain the second adjustment factor, and then according to the preset first and second adjustment weights, the two are weighted and synthesized into a comprehensive adjustment coefficient, and the original change fluctuation threshold is added or subtracted to generate a new adjusted change fluctuation threshold.

[0140] The finished product quality (compaction density) and the fault severity (abnormal number) are combined, the potential influence of insufficient compaction on the system stability is quantified through the first adjustment factor, the fault diffusion degree is reflected through the second adjustment factor, and then the weight is flexibly integrated to construct a closed-loop adaptive mechanism of the three of “quality-fault-threshold”. In this way, when the coal cake is too loose or the faults are concentrated, the change fluctuation threshold can be dynamically relaxed or tightened to accurately match the current working condition, so that high sensitivity to real faults and high robustness to environmental interference can be realized.

[0141] Specifically, the process of adjusting the preset striking frequency or the preset striking stroke according to the texture roughness collected from the abnormally selected hammer head within the preset correction duration based on the adjusted change fluctuation threshold value comprises:

[0142] calculating the average value of all the texture roughness within the preset correction duration to obtain an average roughness;

[0143] when the average roughness is greater than the maximum value of the preset roughness range, reducing the preset striking frequency according to the relative deviation of the average roughness and the maximum value of the preset roughness range and a preset correction coefficient, wherein P' = P x [1 - e x (y' - ymax) / ymax], P' is the reduced preset striking frequency, P is the preset striking frequency before reduction, e is the preset correction coefficient, y' is the average roughness, and ymax is the maximum value of the preset roughness range;

[0144] when the average roughness is less than the minimum value of the preset roughness range, increasing the preset striking stroke according to the relative deviation of the minimum value of the preset roughness range and the average roughness and the preset correction coefficient, wherein L' = L x [1 + e x (ymin - y') / y'], L' is the increased preset striking stroke, L is the preset striking stroke before increase, and ymin is the minimum value of the preset roughness range.

[0145] The preset roughness range is a target interval for measuring the texture quality of the compacted surface, which depends on the type of the target compacted layer material, the construction process requirement, and the site acceptance standard, and is usually set between 0.4 mm and 1.2 mm, and is set as [0.5 mm, 1.0 mm] in the embodiment, which can provide a basis for judging whether the compaction effect is qualified and provide a standard reference for subsequent parameter adjustment.

[0146] The preset correction coefficient is a control factor for adjusting the change range of the striking frequency or the striking stroke, which depends on the equipment adjustment response capability, the construction tolerance error range, and the roughness control precision requirement, and is usually set between 0.1 and 0.5, and is set as 0.2 in the embodiment, which can realize accurate adjustment based on the roughness deviation, avoid excessive adjustment or delayed response, and improve the sensitivity and stability of the adjustment.

[0147] By monitoring the texture roughness data collected from the abnormally selected hammer head within the preset correction duration, the average value of all the texture roughness is calculated first to determine whether it is out of the preset roughness range. When the average roughness is higher than the maximum value, it indicates excessive striking, and the system reduces the striking frequency according to the relative deviation of the average roughness and the maximum value and a preset correction coefficient. Conversely, when the average roughness is lower than the minimum value, it indicates insufficient compaction, and the system increases the striking stroke according to the relative deviation of the average roughness and the minimum value and a correction coefficient, so as to realize dynamic adjustment of the compaction uniformity and quality control.

[0148] By establishing a direct correlation between roughness feedback and device parameter adjustment through the relative deviation between average roughness and preset upper and lower limits of roughness, the correction coefficient is used as the core variable of adjustment sensitivity, and quantitative coupling between roughness feedback and striking frequency / travel is realized. The underlying logic is to use surface texture as an indirect representation of compaction effect, combined with the trend of roughness exceeding or being insufficient, to dynamically optimize the striking parameters and improve the system's adaptive adjustment capability and compaction quality consistency.

[0149] Specifically, the process of screening a plurality of first temporary hammer heads from all the tamping hammers according to the image change value and the preset change fluctuation threshold value comprises:

[0150] Calculate the standard deviation of the image change value within a preset first temporary time period to obtain an image change fluctuation value;

[0151] When the image change fluctuation value is greater than the preset change fluctuation threshold value, the tamping hammer is determined to be the first temporary hammer head, so as to screen a plurality of first temporary hammer heads.

[0152] By using the standard deviation of the image change value as a quantitative indicator of running stability, the position fluctuation during the hammering process can be objectively reflected. The larger the standard deviation, the more unstable the hammering behavior, which may indicate problems such as structural loosening, obstruction or abnormal energy transmission. The preset change fluctuation threshold value is used as a screening reference, combined with the preset first temporary time period to ensure that the data sampling has time continuity and statistical representativeness, thereby avoiding false positives caused by short-term disturbances and establishing an initial screening mechanism based on image recognition and statistical analysis. This provides a data basis and target hammer range for subsequent higher-level fault identification, and has the advantages of fast response, efficient screening and low cost.

[0153] On the other hand, the present application also provides a fault prediction and alarm method for coke oven tamping hammers, comprising:

[0154] The acquisition module is used to acquire the image change value, vibration acceleration, image clarity, texture roughness of the output coal cake and compaction density of each tamping hammer during the running process at a preset striking frequency and a preset striking travel at a plurality of monitoring points arranged on a coke oven tamping flow line comprising a plurality of tamping hammers arranged side by side;

[0155] The first temporary screening module is connected with the acquisition module and is used to screen a plurality of first temporary hammer heads from all the tamping hammers according to the image change value and the preset change fluctuation threshold value;

[0156] a second temporary screening module, connected with the collecting module and the first temporary screening module respectively, for screening a plurality of second temporary hammer heads from all the first temporary hammer heads according to the vibration acceleration and the change of the image change value of each of the first temporary hammer heads within a preset judgment time;

[0157] an abnormality screening module, connected with the collecting module and the second temporary screening module respectively, for screening a plurality of abnormal hammer heads from all the second temporary hammer heads according to the vibration acceleration and the image definition of each of the second temporary hammer heads and the next second temporary hammer head adjacent thereto;

[0158] a first adjusting module, connected with the collecting module and the abnormality screening module respectively, for adjusting the preset change fluctuation threshold according to the compaction density, the arrangement position and the number of the abnormal hammer heads to obtain an adjusted change fluctuation threshold;

[0159] a second adjusting module, connected with the collecting module, for adjusting the preset striking frequency or the preset striking stroke according to the texture roughness of the abnormal hammer heads re-screened based on the adjusted change fluctuation threshold within a preset correction time;

[0160] an alarm module, connected with the abnormality screening module, for issuing an alarm for the abnormal hammer heads re-screened based on the adjusted preset striking frequency or the adjusted preset striking stroke.

[0161] The image change value, the vibration acceleration, the image definition, the texture roughness and the compaction density of each tamping hammer in operation are monitored in real time by the collecting module. The system first screens the first temporary hammer heads according to the image change fluctuation by the first temporary screening module, and then further screens the second temporary hammer heads according to the vibration acceleration and the change of the image change value by the second temporary screening module. Then the abnormality screening module further confirms the abnormal hammer heads by analyzing the vibration acceleration and the image definition of the adjacent second temporary hammer heads. The preset change fluctuation threshold is dynamically adjusted by the first adjusting module according to the number, the distribution and the corresponding compaction density of the abnormal hammer heads. Then the abnormal hammer heads are re-identified according to the adjusted threshold, and the striking frequency or the stroke is adjusted according to the texture roughness feedback by the second adjusting module. Finally, the alarm module issues a fault warning for the hammer heads still abnormal after adjustment.

[0162] Based on the logical association and step-by-step screening mechanism among multiple monitoring parameters, a refined fault identification process is constructed. The standard deviation of image change value reflects the stability of hammer head operation, the change of vibration acceleration reveals the structural response characteristics, and the image definition further verifies whether there is physical disturbance or energy anomaly in the striking process. The compaction density is used as an indicator of macroscopic compaction effect, and is linked to the adjustment of the preset change fluctuation threshold, while the texture roughness, as an important embodiment of the coal briquette forming quality, drives the fine correction of the striking frequency or stroke. By constructing a closed-loop control chain based on "monitoring-screening-analysis-adjustment-feedback-warning", the intelligent linkage between the equipment operation state and the output quality is realized, the timeliness and accuracy of fault identification are improved, and the self-adaptability and practical value of the system under complex working conditions are enhanced.

[0163] The above only describes the preferred embodiments of the present application and is not intended to limit the present application; for those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method of failure prediction alarm for coke oven stoking hammer, characterized in that, The application relates to a coke oven ramming flow line comprising a plurality of ramming hammers arranged in parallel. A plurality of monitoring points are arranged on the coke oven ramming flow line comprising a plurality of ramming hammers arranged in parallel; Real-time acquisition of image change values, vibration accelerations, image definition, texture roughness of output coal briquettes and compaction densities of each ramming hammer during operation at a preset striking frequency and a preset striking stroke; Screening of a plurality of first temporary hammer heads from all the ramming hammers according to the image change values and a preset change fluctuation threshold value; Screening of a plurality of second temporary hammer heads from all the first temporary hammer heads according to the vibration accelerations and the change of the image change values of each first temporary hammer head within a preset judgment duration; Screening of a plurality of abnormal hammer heads from all the second temporary hammer heads according to the vibration accelerations and the image definition of each second temporary hammer head and the next second temporary hammer head adjacent thereto; Adjustment of the preset change fluctuation threshold value according to the compaction densities, arrangement positions and quantity of the abnormal hammer heads, so as to obtain an adjusted change fluctuation threshold value; Adjustment of the preset striking frequency or the preset striking stroke according to the texture roughness of the abnormal hammer heads re-screened based on the adjusted change fluctuation threshold value within a preset correction duration; Issuing of an alarm for the abnormal hammer heads re-screened based on the adjusted preset striking frequency or the adjusted preset striking stroke.

2. The fault prediction alarm method for coke oven stamping hammers according to claim 1, characterized in that, The process of screening a plurality of second temporary hammer heads from all the first temporary hammer heads according to the vibration accelerations and the change of the image change values of each first temporary hammer head within a preset judgment duration comprises: Drawing of a curve of the image change values changing with time within the preset judgment duration, so as to obtain a change value change curve; Calculation of the slope of the change value change curve at each adjacent time, so as to obtain a plurality of slopes; When the absolute value of the slope is greater than a preset slope absolute value threshold value or when positive and negative changes exceeding a preset overturning number continuously occur within a preset observation period, the difference between the image change value at a final time within the preset judgment duration and the image change value at an initial time is calculated, so as to obtain a change difference; When the change difference is greater than a preset change difference threshold value, the first temporary hammer head is determined to be the second temporary hammer head according to the change of the vibration acceleration, so as to screen a plurality of second temporary hammer heads from all the first temporary hammer heads.

3. The fault prediction alarm method of a coke oven stamping ram according to claim 2, characterized by, The process of determining the first temporary hammer head to be the second temporary hammer head according to the change of the vibration acceleration comprises: Calculation of the average value of all the vibration accelerations within the preset judgment duration, so as to obtain an average acceleration, and marking of the vibration acceleration at the last time of the preset judgment duration, so as to obtain a marked acceleration; Calculation of the ratio of the difference between the average acceleration and the marked acceleration to the average acceleration, so as to obtain a vibration acceleration attenuation rate; When the vibration acceleration attenuation rate is greater than a preset attenuation rate threshold value, the first temporary hammer head is determined to be the second temporary hammer head.

4. The fault prediction alarm method for coke oven stamper hammers according to claim 3, characterized in that, The process of screening a plurality of abnormal hammer heads from all the second temporary hammer heads according to the vibration accelerations and the image definition of each second temporary hammer head and the next second temporary hammer head adjacent thereto comprises: calculating a difference value of the image definition of each of the second temporary hammer heads and the next second temporary hammer head adjacent thereto and an absolute value of a ratio of the image definition of the second temporary hammer head, to obtain a definition change rate; calculating a standard deviation of all of the definition change rates within a preset observation time length, to obtain a definition change fluctuation value; when the definition change fluctuation value is greater than a preset definition change fluctuation threshold value, screening a plurality of abnormal hammer heads from all of the second temporary hammer heads according to the vibration acceleration of each of the second temporary hammer heads and the next second temporary hammer head adjacent thereto.

5. The fault prediction alarm method for coke oven stamping hammers according to claim 4, characterized in that, The process of screening a plurality of abnormal hammer heads from all of the second temporary hammer heads according to the vibration acceleration of each of the second temporary hammer heads and the next second temporary hammer head adjacent thereto includes: calculating an average value of the vibration acceleration of each of the second temporary hammer heads within the preset observation time length, to obtain an observation average acceleration; calculating an average value of the vibration acceleration of each of the next second temporary hammer heads within the preset observation time length, to obtain a comparison average acceleration; calculating a ratio of a difference value of the observation average acceleration and the comparison average acceleration to the observation average acceleration, to obtain a vibration acceleration attenuation rate; when the vibration acceleration attenuation rate is greater than a preset attenuation rate threshold value, determining that the corresponding two second temporary hammer heads are both the abnormal hammer heads, to screen a plurality of abnormal hammer heads from all of the second temporary hammer heads.

6. The fault prediction alarm method of a coke oven stamping ram according to claim 5, characterized by, The process of adjusting the preset change fluctuation threshold value according to the compaction density, the arrangement position and the number of the abnormal hammer heads to obtain an adjusted change fluctuation threshold value includes: when the number of the abnormal hammer heads is greater than a preset number threshold value, obtaining a distance between every two abnormal hammer heads according to the arrangement position, to obtain a plurality of hammer head spacings; calculating a standard deviation of all of the hammer head spacings, to obtain a hammer head distribution degree; when the hammer head distribution degree is less than a preset distribution degree threshold value, adjusting the preset change fluctuation threshold value according to the number of the abnormal hammer heads and the compaction density, to obtain an adjusted change fluctuation threshold value.

7. The fault prediction alarm method of a coke oven stamping ram according to claim 6, characterized by, The process of adjusting the preset change fluctuation threshold value according to the number of the abnormal hammer heads and the compaction density to obtain an adjusted change fluctuation threshold value includes: when the compaction density is less than a preset density threshold value, calculating a relative deviation of the compaction density and the preset density threshold value, to obtain a first adjustment factor; calculating a ratio of the number of the abnormal hammer heads to the number of all of the ramming hammers, to obtain a second adjustment factor; adjusting the preset change fluctuation threshold value according to the first adjustment factor, a preset first adjustment weight, the second adjustment factor and a preset second adjustment weight, to obtain an adjusted change fluctuation threshold value.

8. The fault prediction alarm method of a coke oven stamping ram according to claim 7, characterized by, The process of adjusting the preset striking frequency or the preset striking stroke according to the texture roughness collected from the abnormal hammer heads re-screened according to the adjusted change fluctuation threshold value within a preset correction time length includes: calculating an average value of all of the texture roughness within the preset correction time length, to obtain an average roughness; when the average roughness is greater than a maximum value of a preset roughness range, reducing the preset striking frequency according to a relative deviation of the average roughness and the maximum value of the preset roughness range and a preset correction coefficient; When the average roughness is less than the minimum value of the preset roughness range, the preset striking stroke is increased according to the relative deviation of the average roughness from the minimum value of the preset roughness range and the preset correction coefficient.

9. The fault prediction alarm method of a coke oven stamping ram according to claim 8, characterized by, The process of screening a plurality of first temporary hammer heads from all the tamping hammers according to the image change value and a preset change fluctuation threshold value includes: calculating the standard deviation of the image change value in a preset first temporary time period to obtain an image change fluctuation value; When the image change fluctuation value is greater than the preset change fluctuation threshold value, the tamping hammer is determined to be the first temporary hammer head, so as to screen a plurality of first temporary hammer heads.

10. A fault prediction alarm system of a coke oven tamping hammer, which is constructed based on the fault prediction alarm method of the coke oven tamping hammer according to any one of claims 1-9, comprising: a collection module configured to collect, at a plurality of monitoring points arranged on a coke oven tamping production line including a plurality of tamping hammers arranged in parallel, the image change value, the vibration acceleration, the image definition, the texture roughness of the produced coal cake, and the compaction density of each tamping hammer during operation at a preset striking frequency and a preset striking stroke in real time; a first temporary screening module connected with the collection module and configured to screen a plurality of first temporary hammer heads from all the tamping hammers according to the image change value and a preset change fluctuation threshold value; a second temporary screening module connected with the collection module and the first temporary screening module respectively and configured to screen a plurality of second temporary hammer heads from all the first temporary hammer heads according to the change of the vibration acceleration and the image change value of each first temporary hammer head in a preset determination time period; an abnormality screening module connected with the collection module and the second temporary screening module respectively and configured to screen a plurality of abnormal hammer heads from all the second temporary hammer heads according to the vibration acceleration and the image definition of each second temporary hammer head and the next second temporary hammer head adjacent thereto; a first adjustment module connected with the collection module and the abnormality screening module respectively and configured to adjust the preset change fluctuation threshold value according to the compaction density, the arrangement position, and the number of the abnormal hammer heads to obtain an adjusted change fluctuation threshold value; a second adjustment module connected with the collection module and configured to adjust the preset striking frequency or the preset striking stroke according to the texture roughness collected when the abnormal hammer heads are re-screened based on the adjusted change fluctuation threshold value in a preset correction time period; an alarm module connected with the abnormality screening module and configured to issue an alarm for the abnormal hammer heads re-screened based on the adjusted preset striking frequency or the adjusted preset striking stroke.

Images