Method for operating and regulating a pipe chain conveyor for bulk material transport

Abstract

The present application relates to the technical field of conveyor regulation, and particularly relates to a pipe chain conveyor operation regulation method suitable for bulk material conveying. Time sequence data of motor working states and monitoring point operation states of the pipe chain conveyor in a preset period and a historical period are acquired, differences between the motor working states in the historical period and the preset period are compared, and a historical reference time most matched with the working conditions at each time in the preset period is screened out; the historical reference time is compared with the operation state characteristics at each time in the preset period in multiple dimensions, a blockage probability is dynamically calculated, and a blockage time is screened out; differences in blockage probabilities of each monitoring point at the blockage time and a time relationship of the blockage time are analyzed, an abnormal transmission chain is constructed, and a blockage propagation path is revealed; on this basis, a feature vector is constructed, a neural network is combined to reduce misjudgment, and a refreshed abnormal transmission chain is obtained; and finally, adaptive regulation intensity at a current time is calculated based on the refreshed abnormal transmission chain, and a regulation strategy closed-loop optimization is realized.

Description

Technical Field

[0001] This invention relates to the field of conveyor control technology, and specifically to a method for controlling the operation of a tubular chain conveyor suitable for conveying bulk materials. Background Technology

[0002] A tubular chain conveyor is a closed, continuous industrial conveying device for transporting bulk materials, typically such as cement and granular ore. It usually relies on a chain driving scrapers to transport the bulk material from the inlet to the outlet. The transport path can be complex, and depending on specific industrial transport needs, it may include horizontal, vertical, or inclined paths. Due to the small gap between the chain and the pipe during transport, blockages can easily occur, leading to abnormal discharge from the tubular chain conveyor.

[0003] Different causes of material blockage require different handling methods and control intensities. Existing technologies typically use neural networks to capture data features to attribute the causes of blockage at monitoring points, thereby determining the corresponding handling methods and control intensities. However, in actual transportation scenarios, different causes of material blockage may exhibit similar data patterns, and blockages are often caused by the coupling of multiple factors, resulting in a propagation effect between monitoring points. Therefore, relying solely on the data features of monitoring points while ignoring engineering logic can easily lead to misjudgments in the attribution of abnormal causes of blockages, affecting subsequent control results. Summary of the Invention

[0004] To address the technical problem that in real-world transportation scenarios, different causes of material blockage may exhibit similar data patterns, and blockages are often triggered by multiple coupled factors with a propagation effect between monitoring points, relying solely on the data characteristics of monitoring points while ignoring engineering logic can easily lead to misjudgments of the abnormal causes of blockages, affecting subsequent control results, this invention aims to provide a method for controlling the operation of tubular chain conveyors suitable for bulk material transportation. The specific technical solution adopted is as follows:

[0005] Acquire time-series data of various motor working statuses of the tubular chain conveyor within a preset time period and historical time periods, as well as time-series data of various operating statuses at each monitoring point of the tubular chain conveyor.

[0006] Compare the differences between the motor operating status time series data in the historical period and the preset period, and filter out the historical reference time corresponding to each moment in the preset period in the historical period; at each monitoring point, compare the differences between each moment in the preset period and the corresponding historical reference time in all operating status time series data, calculate the blockage probability at each moment, and thus filter out the blockage moment at each monitoring point.

[0007] The differences in the probability of material blockage between different monitoring points and the time characteristics between material blockage times are analyzed to determine the abnormal transmission chain at each material blockage time. After constructing the feature vector of each monitoring point on the abnormal transmission chain at the material blockage time, the abnormal transmission chain is refreshed after obtaining the abnormal attribution using a neural network.

[0008] Based on the quantitative and length characteristics of the anomaly attribution types in the refreshed anomaly propagation chain, the adaptive control intensity of each anomaly attribution at the current moment is determined.

[0009] Furthermore, the method for obtaining the historical reference time includes:

[0010] Within a preset time period, select any time as the target time. Under the time sequence data of each motor operating state, calculate the absolute value of the difference between the target time and each historical time within the historical time period, and use it as the deviation factor.

[0011] The sum of the deviation factors between the target time and each historical time within the historical period under all motor operating state time series data is negatively correlated and normalized, and then used as the state similarity value between the target time and each historical time.

[0012] Within a historical period, historical moments with state similarity values ​​greater than a preset similarity threshold are used as comparison moments corresponding to the target moment.

[0013] Within a historical period, sequentially connected comparison moments are grouped into subsequences. Among all subsequences, the comparison moment in the longest subsequence is taken as the historical reference moment corresponding to the target moment.

[0014] Furthermore, the method for obtaining the time of material blockage includes:

[0015] At each monitoring point, the differences between each moment within the preset time period and the historical reference moment are analyzed in all operational status time series data to obtain the comprehensive deviation factor corresponding to each moment within the preset time period.

[0016] At each monitoring point, for any moment within a preset time period, the mean of the comprehensive operational deviation factors between that moment and all corresponding historical reference moments is used as the mean characteristic value, and the variance of the mean characteristic values ​​of all moments within the preset time period is used as the fluctuation factor.

[0017] At each monitoring point, for any moment within a preset time period, the normalized value of the product of the mean characteristic value corresponding to that moment and the fluctuation factor of the preset time period is used as the blockage probability corresponding to that moment.

[0018] Within a preset time period, at each monitoring point, the moment when the probability of material blockage is greater than the preset blockage threshold is taken as the blockage moment at each monitoring point.

[0019] Furthermore, the method for obtaining the comprehensive deviation factor includes:

[0020] At each monitoring point, for any moment within a preset time period, under each type of time series data of the operating state, the absolute value of the difference between the operating state value of that moment and the corresponding historical reference moment is taken as the operating deviation value between that moment and the corresponding historical reference moment.

[0021] The average of the operational deviations between this moment and each historical reference moment across all operational time series data is used as the comprehensive operational deviation factor.

[0022] Furthermore, the method for obtaining the exception propagation chain includes:

[0023] At each monitoring point, the time-series of material blockage are combined into a subsequence to obtain all the material blockage periods, where each material blockage period contains at least two material blockage times.

[0024] Choose one monitoring point as the target point, and use other monitoring points on the same transportation route as the target point as matching points. Choose one blockage time from all the blockage periods of the target point as the time to be measured.

[0025] Analyze the differences in the probability of material blockage between the time to be tested and the time of material blockage within the blockage period at each matching point, and determine the first abnormal response factor of the time to be tested for each blockage period at the matching point;

[0026] Analyze the time difference characteristics between the time to be measured and the time of blockage within the blockage period in the matching point, and determine the second abnormal response factor of the time to be measured for each blockage period under each matching point;

[0027] The normalized value of the product of the first abnormal response factor and the second abnormal response factor between the time to be tested and each blockage period under each matching point is used as the abnormal response index of the time to be tested for each blockage period under each matching point.

[0028] Among all the blocking periods at each matching point, the blocking period with the largest abnormal response index greater than the preset response threshold between the time to be tested is taken as the blocking response period at each matching point at the time to be tested.

[0029] The time to be measured and the initial time of all blockage response periods are sorted in chronological order. The target point and the corresponding matching point are then sorted according to the time order to obtain the anomaly transmission chain corresponding to the time to be measured.

[0030] Furthermore, the method for obtaining the first abnormal response factor includes:

[0031] The absolute value of the difference between the probability of material blockage between the time to be tested and the probability of material blockage between each time blockage in each blockage period under each matching point is used as the blockage state deviation value corresponding to the time to be tested. The value obtained by negatively mapping the mean of the blockage state deviation values ​​between the time to be tested and the total number of blockage times under each blockage period under each matching point is used as the first abnormal response factor of the time to be tested for each blockage period in each matching point.

[0032] Furthermore, the method for obtaining the second abnormal response factor includes:

[0033] The absolute value of the time difference between the time to be measured and the starting time of the blockage within each blockage period at each matching point is used as the duration factor;

[0034] The preset transmission time of the material between the target point and each matching point is obtained. The absolute value of the difference between the duration factor and the preset transmission time is negatively correlated and mapped to the value of the second abnormal response factor for the time to be tested under each material blockage period at each matching point.

[0035] Furthermore, after constructing the feature vector of each monitoring point in the anomaly propagation chain at the moment of material blockage, and refreshing the anomaly propagation chain after obtaining anomaly attribution using a neural network, the process includes:

[0036] Construct the feature vector of each monitoring point in each anomaly propagation chain at the time to be tested;

[0037] For any anomaly propagation chain, the monitoring points in the anomaly propagation chain and the feature vectors of the monitoring points are used as the input of a pre-trained neural network to obtain the anomaly attribution at each monitoring point in each anomaly propagation chain.

[0038] In each anomaly propagation chain, the monitoring points whose anomaly attribution is inconsistent with the anomaly attribution of the first monitoring point are deleted, thus obtaining the refreshed anomaly propagation chain.

[0039] Furthermore, the method for obtaining the feature vector includes:

[0040] Within the blockage period of the time to be tested, the normalized value of the difference between the maximum blockage probability at all blockage times and the blockage probability at the last blockage time is used as the first blockage recovery factor.

[0041] The time difference between the time to be measured and the starting time of the blockage period is negatively correlated and normalized, and then used as the second blockage recovery factor.

[0042] The normalized value of the product of the first blockage recovery factor and the second blockage recovery factor at the time to be tested is used as the recovery index of each monitoring point on each abnormal transmission chain at the time to be tested.

[0043] The recovery index of each monitoring point, the ranking value of each monitoring point in each abnormal transmission chain, and the blockage probability of each monitoring point at the time to be tested constitute the feature vector of each monitoring point in each abnormal transmission chain.

[0044] Furthermore, the determination of the adaptive adjustment intensity of each anomaly attribution at the current moment based on the quantity and length characteristics of the anomaly attribution types in the refreshed anomaly propagation chain includes:

[0045] For any anomaly attribution, the proportion of the number of updated anomaly propagation chains belonging to that anomaly attribution type in all updated anomaly propagation chains is used as the first adjustment factor for that anomaly attribution.

[0046] The sum of the lengths of all refreshed anomaly propagation chains belonging to this type of anomaly attribution is used as the second adjustment factor corresponding to this type of anomaly attribution.

[0047] The normalized value of the product of the first adjustment factor and the second adjustment factor for this abnormal attribution is used as the dynamic adjustment coefficient;

[0048] The control intensity of the abnormal attribution at the previous time step is weighted and fused with the dynamic adjustment coefficient to obtain the adaptive control intensity corresponding to the abnormal attribution at the current time step.

[0049] The present invention has the following beneficial effects:

[0050] This system acquires time-series data on the operating status of various motors in the tubular chain conveyor within a preset and historical time period, as well as time-series data on multiple operating statuses at various monitoring points of the conveyor. By comparing the differences in motor operating status between historical and preset time periods, it filters out the historical reference time that best matches the operating conditions at each moment within the preset time period, avoiding false anomalies caused by differences in production batches or material types. After filtering the historical reference time, it compares it with the multi-dimensional operating status characteristics of each moment within the preset time period, enabling dynamic calculation of the probability of material blockage, and allowing the selection of the blockage moment at each monitoring point based on this indicator. Furthermore, since relying solely on the data characteristics of monitoring points to utilize neural networks for anomaly attribution typically only captures data correlations while ignoring the anomaly propagation effect between monitoring points, this invention analyzes the differences in the probability of material blockage at different monitoring points at different blockage times and the temporal relationships between these blockage times. An anomaly propagation chain is constructed, revealing the propagation path of the blockage from its source downstream, providing quantitative evidence for locating the root cause. Then, based on the anomaly propagation chain, a feature vector is constructed for each monitoring point at the blockage time. The combination of feature vectors, the anomaly propagation chain, and the neural network effectively reduces misjudgments of anomaly attribution caused by multi-factor coupling and anomaly propagation, and yields a refreshed anomaly propagation chain. Finally, based on the number and length characteristics of the attribution types of monitoring points in the refreshed anomaly propagation chain, the adaptive control intensity at the current time is calculated, achieving closed-loop optimization of the control strategy and enabling the equipment to maintain stable operation under complex working conditions. Attached Figure Description

[0051] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 A flowchart illustrating a method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying, provided in an embodiment of the present invention.

[0053] Figure 2 A flowchart illustrating a method for obtaining the moment of material blockage according to an embodiment of the present invention;

[0054] Figure 3 This is a flowchart illustrating a method for obtaining an exception propagation chain, as provided in an embodiment of the present invention. Detailed Implementation

[0055] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a tubular chain conveyor operation control method suitable for bulk material conveying proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0057] The following description, in conjunction with the accompanying drawings, details a specific scheme for the operation control method of a tubular chain conveyor suitable for bulk material conveying provided by the present invention.

[0058] Please see Figure 1 The diagram illustrates a method flowchart for controlling the operation of a tubular chain conveyor suitable for bulk material conveying, according to an embodiment of the present invention. The method includes the following steps:

[0059] Step S1: Obtain the timing data of various motor working statuses of the tubular chain conveyor within the preset time period and historical time periods, as well as the timing data of various operating statuses at each monitoring point of the tubular chain conveyor.

[0060] The motor of the tubular chain conveyor is the core power source, and its working state affects the stability of material conveying. Therefore, it is necessary to acquire various motor working state timing data of the tubular chain conveyor. In this embodiment of the invention, the types of motor working state timing data may include: current (acquired by deploying a Hall current sensor or current transformer at the three-phase input terminal of the motor), torque (acquired by deploying a torque sensor at the output shaft of the motor), and speed (acquired by a magnetoelectric speed sensor not located at the output shaft of the motor). At the same time, in order to more accurately analyze the operating state of the tubular chain conveyor, multiple monitoring points are set in the tubular chain conveyor. The locations of the monitoring points can be selected from the feed inlet, discharge outlet, chain, and pipe, and are distributed at equal intervals on the chain and pipe. Then, various operating state timing data at each monitoring point are acquired. In this embodiment of the invention, the types of operating state timing data may include: pressure (acquired by a pressure sensor), vibration (acquired by a vibration sensor), temperature (acquired by a temperature sensor), and flow rate (acquired by a radar level gauge or ultrasonic level gauge).

[0061] In this embodiment of the invention, the collection period of all the aforementioned time-series data is divided into a preset period and a historical period. The preset period is 10 minutes before the current time, and the historical period is set to any 10 minutes in the historical material conveying process where no abnormal blockage behavior occurred.

[0062] It should be noted that the time series data acquisition frequency is set to once per second. The acquisition frequency and the length of the time series data can be adjusted according to the implementation scenario and are not limited here. In addition, the horizontal axis of each type of time series data is time, and the vertical axis is the normalized value (removing the influence of units). The specific normalization method is well known to those skilled in the art and will not be described or limited here.

[0063] Step S2: Compare the differences between the motor operating status time sequence data in the historical period and the preset period, and filter out the historical reference time corresponding to each moment in the preset period in the historical period; at each monitoring point, compare the differences between each moment in the preset period and the corresponding historical reference time in all operating status time sequence data, calculate the blockage probability at each moment, and thus filter out the blockage moment at each monitoring point.

[0064] The motor's operating status directly reflects load changes. If anomaly attribution analysis is performed directly on the time-series data of the monitoring points within a preset time period, misjudgments may occur due to differences in the motor operating conditions of the tubular chain conveyor. Furthermore, parameters such as motor efficiency and chain friction coefficient can change after long-term operation. Therefore, it is crucial to first compare the differences between the time-series data of the motor's operating status in historical periods and the preset time period. Then, by selecting the historical reference time corresponding to each moment within the preset time period from the historical periods, comparisons can be made with the motor operating conditions in subsequent processes, effectively improving the accuracy of anomaly analysis.

[0065] Preferably, in one embodiment of the present invention, the method for obtaining historical reference time includes:

[0066] Within a preset time period, a target time is randomly selected. For each motor operating state time series data, the absolute value of the difference between the target time and each historical time within the historical period is calculated as a deviation factor. A larger deviation factor indicates a more significant deviation between the target time and that historical time under that motor operating state time series data, meaning a higher degree of inconsistency in the operating conditions reflected by the motors at those two times. Then, the sum of the deviation factors between the target time and each historical time within the historical period under all motor operating state time series data is calculated. Based on the aforementioned logic, a larger sum indicates a more significant overall difference in the motor operating states between the target time and that historical time. Therefore, this sum is negatively correlated and normalized to correct the logical relationship, obtaining a state similarity value between the target time and each historical time. A larger state similarity value indicates a more consistent overall motor state between the target time and that historical time, making the two times more comparable. This negative correlation mapping and normalization can be performed using the formula... ,in, Let x represent an exponential function with the natural constant e as the base, and let x represent the independent variable.

[0067] Then, within the historical time period, the historical moments with state similarity values ​​greater than a preset similarity threshold are used as the comparison moments corresponding to the target moment. At the same time, since the working state of the motor often has time continuity, the comparison moments that are sequentially connected are formed into subsequences within the historical time period, and the comparison moment in the longest subsequence is used as the historical reference moment corresponding to the target moment.

[0068] It should be noted that the preset similarity threshold in this embodiment of the present invention is 0.7, and the specific value can be adjusted according to the implementation scenario, which is not limited here.

[0069] Tubular chain conveyors pose a risk of material blockage when transporting bulk materials, affecting normal transport efficiency and even causing equipment damage. When blockage occurs, the transport status of the material conveyed by the tubular chain conveyor changes significantly (in the case of local blockage, the material transport volume and conveying resistance at that monitoring point change, causing abnormal fluctuations in the operating status at that monitoring point). Furthermore, unlike the fluctuations in operating status caused by uneven material distribution and periodic scraper movement during normal transport, the abnormality in the operating status of each monitoring point due to material transport status during blockage is greater and the index fluctuations are more drastic. For example, the material compression during blockage may cause abnormal pressure increases at adjacent monitoring points; the impact of accumulated material with the scraper and chain causes abnormally intensified vibration and a sudden increase in chain tension. Therefore, based on the aforementioned historical reference times, at each monitoring point, the differences between each moment within a preset time period and the historical reference time period in the multi-dimensional operating status time series data are compared to obtain the blockage probability at each moment within the preset time period. This probability reflects the material transport status, and based on this index, the blockage moment at each monitoring point within the preset time period is selected.

[0070] Preferably, in one embodiment of the present invention, the method for obtaining the time of material blockage includes:

[0071] Please see Figure 2 The diagram illustrates a method flowchart for obtaining the time of material blockage in one embodiment of the present invention. The method includes the following steps:

[0072] Step S201: At each monitoring point, analyze the differences between each moment within the preset time period and the historical reference moment in all operating status time series data, and obtain the comprehensive deviation factor corresponding to each moment within the preset time period.

[0073] At each monitoring point, for any moment within a preset time period, under each type of operating state time series data, the absolute value of the difference between the operating state value at that moment and the corresponding historical reference moment is calculated, which is taken as the operating deviation value between that moment and the corresponding historical reference moment. Since the operating state of the tubular chain conveyor did not show any abnormalities at the historical reference moments, the larger the operating deviation value at this moment, the more it indicates that an abnormality has occurred in the tubular chain conveyor at that monitoring point at that moment within the preset time period.

[0074] Then, the average of the operating deviation values ​​between this moment and each historical reference moment under all operating state time series data is used as the comprehensive operating deviation factor between this moment and each historical reference moment. Based on the aforementioned logic, it can be seen that the comprehensive operating deviation factor integrates the deviation characteristics between this moment and each historical moment under all operating states. Therefore, the larger the value, the more it indicates that the operating state of the tubular chain conveyor at this monitoring point is abnormal at this moment, which is very likely to be a moment of material blockage.

[0075] Step S202: At each monitoring point, based on the fluctuation characteristics and numerical characteristics of the comprehensive deviation factor corresponding to each moment within the preset time period, determine the probability of material blockage at each moment.

[0076] When material blockage occurs, the tubular chain conveyor at the monitoring point often experiences significant and drastic fluctuations in its operating status due to unstable working conditions. Therefore, at each monitoring point, for any moment within a preset time period, the mean of the comprehensive operating deviation factor between that moment and all corresponding historical reference moments is used as the mean characteristic value. The larger the mean characteristic value, the worse the stability of the operating conditions at that moment. Then, the variance of the mean characteristic values ​​of all moments within the preset time period is used as the fluctuation factor. The larger the fluctuation factor, the more significant and drastic the fluctuations in the operating status of the tubular chain conveyor are considered to have occurred within the preset time period, and the greater the possibility of material blockage.

[0077] Given that both the mean characteristic value and the fluctuation factor are positively correlated with the probability of material blockage, in this embodiment of the invention, at each monitoring point, for any moment within a preset time period, the mean characteristic value corresponding to that moment is multiplied by the fluctuation factor of the preset time period, and the normalized value of the resulting product is used as the probability of material blockage at that moment. The higher the probability of material blockage, the higher the probability of material blockage occurring at that monitoring point of the tubular chain conveyor at that moment. Normalization is a technique well-known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0078] Step S203: At each monitoring point, within a preset time period, filter out the time of material blockage based on the probability of material blockage.

[0079] Based on the logic in step S202, it is known that the greater the probability of material blockage, the higher the possibility of material blockage abnormality. Therefore, within the preset time period, at each monitoring point of the tubular chain conveyor, the moment when the probability of material blockage is greater than the preset material blockage threshold is taken as the material blockage moment at each monitoring point.

[0080] It should be noted that the preset blockage threshold in this embodiment of the present invention is 0.65. The specific value can be adjusted according to the implementation scenario and is not limited here.

[0081] Thus, the above process can be used to filter out the time of material blockage at each monitoring point within a preset time period.

[0082] Step S3: Analyze the differences in the probability of material blockage between different monitoring points and the time characteristics between material blockage times to determine the abnormal transmission chain at each material blockage time; after constructing the feature vector of each monitoring point on the abnormal transmission chain at the material blockage time, refresh the abnormal transmission chain after obtaining the abnormal attribution using a neural network.

[0083] In tubular chain conveyors, material blockage is often caused by the coupling of multiple factors and has a propagation effect between monitoring points. A single node failure can lead to a global shutdown. Therefore, relying solely on the blockage probability and various data features of monitoring points can only capture the correlation between monitoring point data, ignoring the engineering logic, making attribution prone to misjudgment and missed detection. Therefore, in this embodiment of the invention, by analyzing the differences in blockage probability between different monitoring points at the time of blockage and the time characteristics between blockage times, the abnormal transmission chain at each blockage time is determined. The abnormal transmission chain characterizes the chain reaction generated in the entire tubular chain conveyor when a blockage occurs at a certain monitoring point.

[0084] Preferably, in one embodiment of the present invention, the method for obtaining the exception propagation chain includes:

[0085] Please see Figure 3 The diagram illustrates a method flowchart for obtaining an exception propagation chain according to an embodiment of the present invention. The method includes the following steps:

[0086] Step S301: Based on the temporal continuity characteristics of the material blockage time corresponding to each monitoring point, determine the material blockage period corresponding to each monitoring point.

[0087] Since material blockage is a changing state, theoretically, individual moments usually do not exist. Therefore, at each monitoring point, the time-series blockage moments are combined into subsequences to obtain all blockage periods. Each blockage period contains at least two blockage moments, meaning that individual blockage moments may be anomalous outliers and will not be analyzed further here.

[0088] Step S302: Select any monitoring point as the target point, and select other monitoring points on the same transportation route as the target point as matching points. Select any blockage time from all blockage periods of the target point as the time to be measured.

[0089] Due to the large number of monitoring points and the numerous times of material blockage, for the sake of subsequent explanation and clarification, we will select one monitoring point as the target point and other monitoring points on the same transportation route as the target point as matching points (i.e., monitoring points that may be affected by the abnormal propagation of the target point). At the same time, we will select one blockage time from all the blockage periods of the target point as the time to be measured.

[0090] In the subsequent process, the acquisition process of the anomaly propagation chain is explained by analyzing the target point, the time to be tested, and the matching point.

[0091] Step S303: Analyze the differences in the probability of material blockage between the time to be tested and the time blockage within the blockage period at each matching point, and determine the first abnormal response factor of the time to be tested for each blockage period at the matching point.

[0092] The absolute value of the difference between the probability of blockage and the probability of blockage at each blockage time within each blockage period at each matching point is calculated and used as the blockage state deviation value corresponding to the test time. The mean of the blockage state deviation values ​​between the test time and the blockage time within each blockage period at each matching point is recorded as the mean state deviation. The smaller the mean state deviation, the more consistent the operating state of the target point and the matching point of the tubular chain conveyor is between the blockage time at the test time and the blockage time within that blockage period at the matching point. Therefore, these two are more likely to be affected by the same abnormal cause. Thus, the mean state deviation is negatively correlated to correct the logical relationship and obtain the first abnormal response factor for the test time in each blockage period at each matching point. Based on the above logic, the larger the first abnormal response factor, the more likely the abnormal characteristics of the matching point and the abnormal characteristics of the target point at the test time are to originate from the same source in that blockage period. The negative correlation mapping here can be expressed by the formula... ,in, Let x represent an exponential function with the natural constant e as the base, and let x represent the independent variable.

[0093] Step S304: Analyze the time difference characteristics between the time to be tested and the time of blockage within the blockage period in the matching point, and determine the second abnormal response factor of the time to be tested for each blockage period under each matching point.

[0094] Given that the transport of materials from one monitoring point to another takes time and is not instantaneous—meaning that after an anomaly occurs at an upstream monitoring point, a certain anomaly will only appear at a downstream monitoring point after a period of time—if the temporal difference between the time of the blockage at the target point and the time of the blockage at the matching point matches the transport time of materials at a normal flow rate between the target point and the matching point, it can also be considered that the abnormal characteristics of the matching point during that blockage period are very likely to originate from the same source as the abnormal characteristics of the target point at the time of the target point.

[0095] Therefore, the absolute value of the time difference between the time to be measured and the starting time of the material blockage within each blockage period at each matching point is first obtained. This value can be used as a duration factor, which characterizes the temporal difference between the time to be measured and each blockage period at the matching point. Then, the preset transmission time is obtained based on the route length between the target point and each matching point and the material transmission speed (preset transmission time = ...). The absolute value of the difference between the duration factor and the preset transmission time is calculated. The smaller the absolute value of the difference, the more likely the target point and the matching point are affected by the same abnormal cause during the blockage period between the time to be tested and the time to be matched. Therefore, the absolute value of the difference is negatively correlated to correct the logical relationship and obtain the second abnormal response factor for the time to be tested for each blockage period in each matching point. Based on the above logic, the larger the second abnormal response factor, the more likely the abnormal characteristics of the matching point and the abnormal characteristics of the target point at the time to be tested are of the same origin during the blockage period in that matching point. The negative correlation mapping here can be expressed by the formula... ,in, Let represent an exponential function with base e, and x represent the independent variable. It should be noted that when performing a negative correlation mapping between the absolute value of the difference between the duration factor and the preset transmission time, the absolute value of the difference can be pre-standardized by dividing it by a characteristic time scale (such as the preset transmission time itself) to achieve dimensionlessness, resulting in a dimensionless value. Then, based on this dimensionless value, a method can be implemented... The negative correlation mapping, in this step, the independent variable is the dimensionless value of the absolute value of the difference between the duration factor and the preset transmission time.

[0096] Step S305: Integrate the first and second abnormal response factors between the time to be tested and each blockage period at each matching point, thereby filtering the blockage response period of the time to be tested from all blockage periods at each matching point.

[0097] Based on the analysis of the preceding steps, it is known that both the first and second abnormal response factors between the time to be measured and the blocking period at the matching point are positively correlated with the probability of the common source of anomaly attribution between the target point and the matching point at the time to be measured. Therefore, the normalized value of the product of the first and second abnormal response factors between the time to be measured and each blocking period at each matching point is used as the abnormal response index of the time to be measured for each blocking period at each matching point. The larger the abnormal response index, the higher the probability of the common source of anomaly attribution between the target point and the matching point at the time to be measured. Normalization is a technique well-known to those skilled in the art. The choice of normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here.

[0098] Finally, among all the blocking periods at each matching point, the blocking period with the largest abnormal response index greater than the preset response threshold between the time to be tested is taken as the blocking response period at each matching point at the time to be tested.

[0099] It should be noted that the preset response threshold in this embodiment of the present invention is 0.7, and the specific value can be adjusted according to the implementation scenario, which is not limited here.

[0100] Step S306: Compare the time to be tested with the time in the blockage response period under all matching points to determine the abnormal transmission chain corresponding to the time to be tested.

[0101] Based on the aforementioned process, the blockage response period of the target point in the matching point at the time to be tested can be screened out. This means that at the time to be tested, the matching point with the blockage response period and the target point may have the same source of blockage attribution, and they are more likely to influence each other. Therefore, the time to be tested and the initial time of all blockage response periods are sorted in chronological order. Then, the matching points corresponding to the target point and the blockage response period are sorted according to the sorting of time, so as to obtain the abnormal transmission chain corresponding to the time to be tested. The position order of the monitoring points in the transmission chain represents the order of abnormal transmission.

[0102] Thus, through the above steps, the abnormal transmission chain at each blocking moment in each blocking period at each monitoring point can be obtained.

[0103] After obtaining the anomaly propagation chain, we can continue to construct the feature vector of each monitoring point on the anomaly propagation chain to represent the data characteristics. The anomaly propagation chain is used to represent the engineering logic. By combining the two, we can use a neural network to obtain the anomaly attribution and refresh the anomaly propagation chain to obtain the refreshed anomaly propagation chain.

[0104] Preferably, in one embodiment of the present invention, the method for obtaining the refreshed exception propagation chain includes:

[0105] Construct the feature vector of each monitoring point in the anomaly propagation chain at the time of material blockage: Within the blockage period at the time to be measured (consistent with the time to be measured in step S302), normalize the difference between the maximum blockage probability at all blockage times and the blockage probability at the last blockage time, and use this normalized value as the first blockage recovery factor. The larger the first blockage recovery factor, the more the severity of blockage has decreased over time, and the better the anomaly recovery capability. Since the difference here may be positive or negative, normalization can be performed using... function.

[0106] Calculate the time difference between the time to be measured and the start time of the blockage period. The smaller the time difference, the shorter the duration of the anomaly at the time to be measured, and thus the less severe it is, indirectly reflecting better recovery capability. Therefore, the time difference is negatively correlated and normalized, and used as the second blockage recovery factor. The larger the second blockage recovery factor, the better the anomaly recovery capability. The negative correlation mapping and normalization can be performed using the formula... ,in, Let represent an exponential function with base e, and x represent the independent variable. It should be noted that when performing a negative correlation mapping between the time difference between the time to be measured and the start time of the blockage period, the time difference can be pre-standardized by dividing it by a characteristic time scale (such as the preset total duration of the period) to achieve dimensionlessness, resulting in a dimensionless value. Then, based on this dimensionless value, a mapping based on... The negative correlation mapping, in this step, the independent variable is the dimensionless value of the time difference between the time to be measured and the starting time of the blockage period.

[0107] Based on the aforementioned logic, it is known that both the first and second blockage recovery factors at the time of the test are positively correlated with the anomaly recovery capability. Therefore, the normalized product of the first and second blockage recovery factors at the time of the test is used as the recovery index of each monitoring point on each anomaly transmission chain at the time of the test. The recovery index characterizes the recovery capability characteristics that the monitoring points on the anomaly transmission chain may possess at the time of the test. Normalization is a technique well-known to those skilled in the art, and the choice of normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0108] Then, on the anomaly transmission chain corresponding to the time to be tested, the recovery index of each monitoring point, the ranking value of each monitoring point in each anomaly transmission chain, and the blockage probability of each monitoring point at the time to be tested are used to construct the feature vector of each monitoring point on each anomaly transmission chain. The recovery index in the feature vector reflects the dynamics of anomaly recovery, the blockage probability reflects the anomaly characteristics of the monitoring point, and the ranking value reflects the position of the monitoring point in the anomaly transmission chain. At this time, the feature vector contains both data features and considers the characteristics that affect transmission.

[0109] Therefore, for any abnormal transmission chain, the feature vectors of the monitoring points in the abnormal transmission chain and the time of blockage at the monitoring points are used as the input of a pre-trained neural network to obtain the abnormal attribution at each monitoring point in each abnormal transmission chain.

[0110] Finally, in each anomaly propagation chain, the monitoring points whose anomaly attribution is inconsistent with the anomaly attribution of the first monitoring point are deleted, thus obtaining a refreshed anomaly propagation chain. At this time, the refreshed anomaly propagation chain and the anomaly attribution of each monitoring point in the chain will be more accurate.

[0111] It should be noted that the training process of neural networks is a well-known technique, and the specific process will not be described in detail here.

[0112] At this point, we can obtain the refreshed exception propagation chains corresponding to all exception propagation chains.

[0113] Step S4: Based on the quantity and length characteristics of the anomaly attribution types in the refreshed anomaly propagation chain, determine the adaptive control intensity of each anomaly attribution at the current moment.

[0114] The quantity and length characteristics of the anomalous transmission chain under the anomalous attribution type directly reflect the magnitude of the anomalous intensity. Therefore, in this step, the adaptive control intensity of each anomalous attribution at the current moment can be determined based on the aforementioned characteristics.

[0115] Preferably, in one embodiment of the present invention, the method for obtaining the adaptive control intensity includes:

[0116] For any anomaly attribution, the proportion of the number of updated anomaly propagation chains belonging to that type of anomaly attribution among all updated anomaly propagation chains is used as the first adjustment factor for that anomaly attribution. The larger the first adjustment factor, the more updated anomaly propagation chains belonging to that type of anomaly attribution there are, and the greater the proportion of that type of anomaly attribution, thus the higher the severity and the greater the intensity of regulation required.

[0117] Then, the sum of the lengths of all refreshed abnormal propagation chains belonging to this type of abnormal attribution is used as the second adjustment factor corresponding to this type of abnormal attribution. The larger the second adjustment factor, the greater the expansion trend of this type of abnormal attribution, the more timely the adjustment is needed, and the greater the intensity of regulation required.

[0118] Therefore, the normalized value of the product of the first adjustment factor and the second adjustment factor for this type of anomalous attribution is used as the dynamic adjustment coefficient. The larger the dynamic adjustment coefficient, the greater the regulatory intensity required for this type of anomalous attribution at the current moment. Normalization is a technique well-known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0119] Finally, the control intensity of the abnormal attribution at the previous time step is weighted and fused with the dynamic adjustment coefficient to obtain the adaptive control intensity corresponding to the abnormal attribution at the current time step.

[0120] The specific weighted fusion method is as follows:

[0121]

[0122] In the formula, This indicates the intensity of regulation at the current time t. This represents the control intensity at the current time compared to the previous time t-1, where K represents the dynamic adjustment coefficient. Indicates the preset memory factor ( ,For example It should be noted that at t=1, this leads to... for At this time, a cold start problem exists, and the intensity of regulation will be adjusted. Set to the preset base operating strength, the The value is determined based on the no-load operating power requirement of the tubular chain conveyor, with a typical value of 0.3.

[0123] In summary, this method acquires time-series data on the operating states of various motors in the tubular chain conveyor within a preset time period and historical time periods, as well as time-series data on multiple operating states at various monitoring points of the tubular chain conveyor. By comparing the differences in motor operating states between historical and preset time periods, the method selects the historical reference time that best matches the operating conditions at each moment within the preset time period, avoiding false anomalies caused by differences in production batches or material types. After selecting the historical reference time, it compares it with the multi-dimensional operating state characteristics of each moment within the preset time period, enabling dynamic calculation of the probability of material blockage, and allowing the selection of the blockage moment at each monitoring point based on this indicator. Furthermore, since relying solely on the data characteristics of monitoring points to utilize neural networks for anomaly attribution typically only captures data correlations while ignoring the anomaly propagation effect between monitoring points, this embodiment of the invention analyzes the differences in the probability of material blockage at different monitoring points at the time of blockage and the temporal relationship of the blockage times, constructing an anomaly transmission chain. This reveals the propagation path of the blockage from the source to the downstream, providing a quantitative basis for locating the root cause. Then, based on the anomaly transmission chain, a feature vector for each monitoring point at the time of blockage is constructed. The combination of feature vectors, the anomaly transmission chain, and the neural network can effectively reduce misjudgments of anomaly attribution caused by multi-factor coupling and anomaly propagation, and can obtain a refreshed anomaly transmission chain. Finally, based on the number and length characteristics of the attribution types of monitoring points in the refreshed anomaly transmission chain, the adaptive control intensity at the current time is calculated, realizing closed-loop optimization of the control strategy, enabling the equipment to maintain stable operation under complex working conditions.

[0124] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0125] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying, characterized in that, The method includes: Acquire time-series data of various motor working statuses of the tubular chain conveyor within a preset time period and historical time periods, as well as time-series data of various operating statuses at each monitoring point of the tubular chain conveyor. Compare the differences between the motor operating status time series data in the historical period and the preset period, and filter out the historical reference time corresponding to each moment in the preset period in the historical period; at each monitoring point, compare the differences between each moment in the preset period and the corresponding historical reference time in all operating status time series data, calculate the blockage probability at each moment, and thus filter out the blockage moment at each monitoring point. The differences in the probability of material blockage between different monitoring points and the time characteristics between material blockage times are analyzed to determine the abnormal transmission chain at each material blockage time. After constructing the feature vector of each monitoring point on the abnormal transmission chain at the material blockage time, the abnormal transmission chain is refreshed after obtaining the abnormal attribution using a neural network. Based on the quantitative and length characteristics of the anomaly attribution types in the refreshed anomaly propagation chain, the adaptive control intensity of each anomaly attribution at the current moment is determined. The method for obtaining the time of material blockage includes: At each monitoring point, the differences between each moment within the preset time period and the historical reference moment are analyzed in all operational status time series data to obtain the comprehensive deviation factor corresponding to each moment within the preset time period. At each monitoring point, for any moment within a preset time period, the mean of the comprehensive operational deviation factors between that moment and all corresponding historical reference moments is used as the mean characteristic value, and the variance of the mean characteristic values ​​of all moments within the preset time period is used as the fluctuation factor. At each monitoring point, for any moment within a preset time period, the normalized value of the product of the mean characteristic value corresponding to that moment and the fluctuation factor of the preset time period is used as the blockage probability corresponding to that moment. Within a preset time period, at each monitoring point, the moment when the probability of material blockage is greater than the preset blockage threshold is taken as the blockage moment at each monitoring point.

2. The method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying according to claim 1, characterized in that, The methods for obtaining the historical reference time include: Within a preset time period, select any time as the target time. Under the time sequence data of each motor operating state, calculate the absolute value of the difference between the target time and each historical time within the historical time period, and use it as the deviation factor. The sum of the deviation factors between the target time and each historical time within the historical period under all motor operating state time series data is negatively correlated and normalized, and then used as the state similarity value between the target time and each historical time. Within a historical period, historical moments with state similarity values ​​greater than a preset similarity threshold are used as comparison moments corresponding to the target moment. Within a historical period, sequentially connected comparison moments are grouped into subsequences. Among all subsequences, the comparison moment in the longest subsequence is taken as the historical reference moment corresponding to the target moment.

3. The method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying according to claim 1, characterized in that, The method for obtaining the comprehensive deviation factor includes: At each monitoring point, for any moment within a preset time period, under each type of time series data of the operating state, the absolute value of the difference between the operating state value of that moment and the corresponding historical reference moment is taken as the operating deviation value between that moment and the corresponding historical reference moment. The average of the operational deviations between this moment and each historical reference moment across all operational time series data is used as the comprehensive operational deviation factor.

4. The method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying according to claim 1, characterized in that, The method for obtaining the exception propagation chain includes: At each monitoring point, the time-series of material blockage are combined into a subsequence to obtain all the material blockage periods, where each material blockage period contains at least two material blockage times. Choose one monitoring point as the target point, and use other monitoring points on the same transportation route as the target point as matching points. Choose one blockage time from all the blockage periods of the target point as the time to be measured. Analyze the differences in the probability of material blockage between the time to be tested and the time of material blockage within the blockage period at each matching point, and determine the first abnormal response factor of the time to be tested for each blockage period at the matching point; Analyze the time difference characteristics between the time to be measured and the time of blockage within the blockage period in the matching point, and determine the second abnormal response factor of the time to be measured for each blockage period under each matching point; The normalized value of the product of the first abnormal response factor and the second abnormal response factor between the time to be tested and each blockage period under each matching point is used as the abnormal response index of the time to be tested for each blockage period under each matching point. Among all the blocking periods at each matching point, the blocking period with the largest abnormal response index greater than the preset response threshold between the time to be tested is taken as the blocking response period at each matching point at the time to be tested. The time to be measured and the initial time of all blockage response periods are sorted in chronological order. The target point and the corresponding matching point are then sorted according to the time order to obtain the anomaly transmission chain corresponding to the time to be measured.

5. The method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying according to claim 4, characterized in that, The method for obtaining the first abnormal response factor includes: The absolute value of the difference between the probability of material blockage between the time to be tested and the probability of material blockage between each time blockage in each blockage period under each matching point is used as the blockage state deviation value corresponding to the time to be tested. The value obtained by negatively mapping the mean of the blockage state deviation values ​​between the time to be tested and the total number of blockage times under each blockage period under each matching point is used as the first abnormal response factor of the time to be tested for each blockage period in each matching point.

6. The method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying according to claim 4, characterized in that, The method for obtaining the second abnormal response factor includes: The absolute value of the time difference between the time to be measured and the starting time of the blockage within each blockage period at each matching point is used as the duration factor; The preset transmission time of the material between the target point and each matching point is obtained. The absolute value of the difference between the duration factor and the preset transmission time is negatively correlated and mapped to the value of the second abnormal response factor for the time to be tested under each material blockage period at each matching point.

7. The method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying according to claim 4, characterized in that, After constructing the feature vector of each monitoring point on the anomaly propagation chain at the moment of material blockage, and refreshing the anomaly propagation chain after obtaining anomaly attribution using a neural network, the process includes: Construct the feature vector of each monitoring point in each anomaly propagation chain at the time to be tested; For any anomaly propagation chain, the monitoring points in the anomaly propagation chain and the feature vectors of the monitoring points are used as the input of a pre-trained neural network to obtain the anomaly attribution at each monitoring point in each anomaly propagation chain. In each anomaly propagation chain, the monitoring points whose anomaly attribution is inconsistent with the anomaly attribution of the first monitoring point are deleted, thus obtaining the refreshed anomaly propagation chain.

8. The method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying according to claim 7, characterized in that, The method for obtaining the feature vector includes: Within the blockage period of the time to be tested, the normalized value of the difference between the maximum blockage probability at all blockage times and the blockage probability at the last blockage time is used as the first blockage recovery factor. The time difference between the time to be measured and the starting time of the blockage period is negatively correlated and normalized, and then used as the second blockage recovery factor. The normalized value of the product of the first blockage recovery factor and the second blockage recovery factor at the time to be tested is used as the recovery index of each monitoring point on each abnormal transmission chain at the time to be tested. The recovery index of each monitoring point, the ranking value of each monitoring point in each abnormal transmission chain, and the blockage probability of each monitoring point at the time to be tested constitute the feature vector of each monitoring point in each abnormal transmission chain.

9. A method for controlling the operation of a tubular chain conveyor suitable for bulk material conveying, as described in claim 7, is characterized in that, The quantitative and length characteristics of the anomaly attribution types based on the refreshed anomaly propagation chain. Determine the adaptive modulation strength for each anomalous attribution at the current time, including: For any anomaly attribution, the proportion of the number of updated anomaly propagation chains belonging to that anomaly attribution type in all updated anomaly propagation chains is used as the first adjustment factor for that anomaly attribution. The sum of the lengths of all refreshed anomaly propagation chains belonging to this type of anomaly attribution is used as the second adjustment factor corresponding to this type of anomaly attribution. The normalized value of the product of the first adjustment factor and the second adjustment factor for this abnormal attribution is used as the dynamic adjustment coefficient; The control intensity of the abnormal attribution at the previous time step is weighted and fused with the dynamic adjustment coefficient to obtain the adaptive control intensity corresponding to the abnormal attribution at the current time step.

Images