Conveyor operation state control system based on PLC program

Abstract

The present application relates to the technical field of industrial automation control and conveying equipment intelligent operation and maintenance, in particular to a conveyor running state control system based on PLC program, comprising: a data acquisition module, which acquires real-time running state data, load weight data and physical space parameters of the conveyor; an empty load resistance self-adaptive modeling module, which constructs and slide updates current self-adaptive reference data under the empty load state; a load-background decoupling module, which generates running residual sequence under the load state; a physical constraint feature extraction module, which constructs a cost matrix of dynamic time warping algorithm based on the physical space parameters, filters random disturbance and extracts real deterioration degree; and an adaptive control module, which generates intervention coefficient according to the real deterioration degree and sends it to the PLC, dynamically adjusts the output limiting amplitude of the conveyor drive frequency converter, and realizes layered separation and closed-loop intervention of running background, random disturbance and mechanical deterioration.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control and intelligent operation and maintenance technology for conveying equipment, specifically a conveyor operation status control system based on PLC programs. Background Technology

[0002] As coal mine main transportation systems develop towards long-distance, heavy-load, and continuous operation, the stable operation of conveyors under complex working conditions is receiving increasing attention. During operation, conveyors are simultaneously affected by various factors such as changes in material load, impact of falling material, changes in idler friction, and drift in the mechanical state of the belt. How to ensure the continuity of transportation while timely identifying and controlling operational anomalies has become an important problem that needs to be solved in the field of current conveying control. Traditional conveyor operation control currently relies mainly on the following methods: total current threshold monitoring, fixed reference comparison, and PLC fixed amplitude limiting protection; However, methods such as total current threshold monitoring, fixed reference comparison, and PLC fixed limit protection all have certain drawbacks. For example, total current threshold monitoring is difficult to distinguish between normal load fluctuations and real mechanical anomalies, and is prone to false alarms or missed alarms. Fixed reference comparison cannot adapt to the drift of no-load resistance with changes in temperature, aging, and tension, resulting in distorted diagnostic results. PLC fixed limit protection lacks verification of material propagation laws and mechanical cycle characteristics, making it difficult to effectively distinguish between random impacts and continuous deterioration, and thus it is difficult to meet the needs of refined control under dynamic operating conditions. Summary of the Invention

[0003] The purpose of this invention is to provide a PLC-based conveyor operation status control system to solve the following technical problems: Existing traditional conveyor operation control technologies are significantly insufficient in distinguishing between normal load fluctuations and actual mechanical anomalies under complex dynamic working conditions, as well as in effectively distinguishing between random material drop impacts and continuous mechanical degradation. There is an urgent need for a PLC-based conveyor operation status control system that can more accurately handle the conveyor's basic operating background, the layered separation of random material disturbances and actual mechanical degradation, and dynamically optimize the inverter output limiting to achieve closed-loop intervention. The purpose of this invention can be achieved through the following technical solutions: A conveyor operation status control system based on a PLC program includes: a data acquisition module for acquiring real-time operation status data and physical space parameters of the conveyor, including load weight data; and an adaptive control module for sending control commands to the programmable logic controller (PLC) to dynamically adjust the output limiting of the conveyor drive frequency converter; characterized in that the system further includes a no-load resistance adaptive modeling module, a load-background decoupling module, and a physical constraint feature extraction module connected in series via communication. The no-load resistance adaptive modeling module is communicatively connected to the data acquisition module. It is used to construct an initial benchmark model based on real-time operating status data when the conveyor is determined to be in a no-load state based on load weight data, and to update the initial benchmark model through a sliding window to generate the current adaptive benchmark data in subsequent no-load states. The load-background decoupling module is communicatively connected to the unloaded resistance adaptive modeling module. It is used to acquire real-time operating status data under load and compare the real-time operating status data under load with the current adaptive benchmark data to generate an operating residual sequence. The physical constraint feature extraction module is communicatively connected to the load-background decoupling module. It is used to construct a spatiotemporal mapping constraint based on physical space parameters, and accordingly construct a cost matrix for a dynamic time warping algorithm. The cost matrix contains the initial cost of each matching path. The running residual sequence and a preset mechanical feature sequence representing the single jamming resistance waveform of the idler roller extracted based on measured data are simultaneously input into the dynamic time warping algorithm containing the cost matrix for matching, so as to filter random disturbances and extract the true degradation degree. The true degradation degree is then output to the adaptive control module, so that the adaptive control module can generate an intervention coefficient based on the true degradation degree and generate the control command based on the intervention coefficient.

[0004] Optionally, the data acquisition module includes: The status acquisition unit is used to acquire motor torque current data, running speed data, tension force data and weighing reading as the real-time running status data, wherein the weighing reading is used as the load weight data; the parameter configuration unit is used to receive and store physical distance, belt elastic wave transmission speed and idler roller spacing as the physical space parameters, wherein the physical distance is the spatial distance from the material drop point to the drive roller.

[0005] Optionally, the no-load drag adaptive modeling module includes: A no-load determination unit is used to acquire the weighing reading. When the weighing reading is continuously below a preset dead zone threshold for a preset time, the conveyor is determined to be in the no-load state. When the weighing reading is higher than or equal to the preset dead zone threshold, or is below the preset dead zone threshold but has not reached the preset time, the conveyor is determined to be in the loaded state. The preset dead zone threshold is 2% to 5% of the rated full-load weight of the conveyor. A reference grasping unit is connected to the no-load determination unit and is used to extract the three-dimensional mapping relationship of the motor torque current data, the running speed data, and the tension force data in the no-load state to construct the initial reference model. A reference correction unit is connected to the reference grasping unit and is used to introduce a sliding window with a preset forgetting factor, use the three-dimensional mapping relationship in the current no-load state to perform weighted correction on the initial reference model, and output the current adaptive reference data containing the reference current component under different running speeds.

[0006] Optionally, the load-background decoupling module is specifically used for: Under the load condition, the reference current component corresponding to the current operating speed data in the current adaptive reference data is subtracted from the real-time collected motor torque and current data, and the operating residual sequence is output.

[0007] Optionally, the physical constraint feature extraction module includes: A spatial distance constraint unit is used to calculate the material transport delay time window based on the physical distance, the running speed data, and the belt elastic wave transmission speed; a periodic penalty unit is used to calculate the spatial frequency period based on the idler spacing and the running speed data; a matching solution unit is connected to the spatial distance constraint unit and the periodic penalty unit, and is used to fuse the material transport delay time window and the spatial frequency period into the cost matrix to perform sequence matching on the running residual sequence.

[0008] Optionally, when calculating the material transport delay time window, the spatial distance constraint unit is specifically used for: The physical distance is divided by the operating speed data to obtain the basic delay time; the physical distance is divided by the belt elastic wave transmission speed to obtain the reference tolerance time, and the reference tolerance time is multiplied by a preset tolerance coefficient to obtain the fluctuation tolerance time; the basic delay time plus the fluctuation tolerance time is used as the upper limit of the material transfer delay time window, and the basic delay time minus the fluctuation tolerance time is used as the lower limit of the material transfer delay time window.

[0009] Optionally, when the matching solution unit updates the cost matrix using the material transport delay time window, it specifically executes the following logic: Data segments in the running residual sequence whose amplitude exceeds a preset fluctuation threshold are identified as fluctuation features, and it is determined whether the occurrence time of the fluctuation feature is between the lower limit and the upper limit of the material transfer delay time window; when the occurrence time of the fluctuation feature is greater than or equal to the lower limit and less than or equal to the upper limit, the initial cost of the corresponding matching path in the dynamic time warping algorithm is retained; when the occurrence time of the fluctuation feature is less than the lower limit or greater than the upper limit, the cost of the corresponding matching path in the dynamic time warping algorithm is set to a preset maximum overflow penalty threshold. It refuses to be identified as a physical disturbance.

[0010] Optionally, the periodic penalty unit, when calculating the spatial frequency period, is specifically used for: The spatial frequency period is obtained by dividing the idler roller spacing by the running speed data. When the matching solution unit updates the cost matrix using the spatial frequency period, it specifically executes the following logic: it determines whether the absolute value obtained by performing a floating-point remainder calculation on the spatial frequency period based on the time interval of the fluctuation peak in the running residual sequence, or the absolute difference between the spatial frequency period and the floating-point remainder result, is within a preset zero-tolerance range, wherein the zero-tolerance range is a time deviation range set based on the spatial frequency period; when the absolute value of the floating-point remainder or the absolute difference is within the zero-tolerance range, it is determined that there is a mechanical jamming feature, and the cost of the corresponding matching path remains unchanged; when neither the absolute value of the floating-point remainder nor the absolute difference is within the zero-tolerance range, it is determined that there is a random disturbance feature, and the cost of the corresponding matching path is amplified by a preset penalty multiple.

[0011] Optionally, the adaptive control module includes: The degradation degree mapping unit is used to extract the amplitude of the running residual sequence retained by the matching solution unit, compare it with multiple preset abnormal threshold intervals, and map it into a discrete real degradation degree scalar. The underlying execution unit is connected to the degradation degree mapping unit and is used to convert the real degradation degree into the intervention coefficient with an integer data format, and send the intervention coefficient to the programmable logic controller by writing to the register, so as to control the programmable logic controller to dynamically adjust the output limit of the conveyor drive frequency converter according to the intervention coefficient.

[0012] Optionally, the system includes a low-level execution layer and an edge computing layer in its hardware deployment; the low-level execution unit in the adaptive control module is deployed in the low-level execution layer; the data acquisition module, the no-load resistance adaptive modeling module, the load-background decoupling module, and the physical constraint feature extraction module are all deployed in the edge computing layer, which is independent of the low-level execution layer, so as to achieve decoupling of computing power between control and diagnosis.

[0013] Compared with the prior art, the present invention has the following beneficial effects: This application presents a PLC-based conveyor operation status control system. By collecting motor torque current, running speed, tension force, weighing readings, physical distance, belt elastic wave transmission speed, and idler spacing, it constructs and slides-updates adaptive reference data under no-load conditions. Under load conditions, it separates the operating residual sequence and verifies the cost matrix of the dynamic time warping algorithm by combining the material transfer delay time window and the idler spatial frequency period. It extracts the true degradation degree and generates intervention coefficients to be written into the programmable logic controller. This provides a closed-loop control scheme that can dynamically adjust the output limit of the conveyor drive frequency converter. Compared with total current threshold monitoring, fixed reference comparison, and PLC fixed limit protection, it can better distinguish between normal load fluctuations, random material drop impacts, and true mechanical degradation. It adapts to the drift of no-load resistance caused by changes in temperature, aging, and tension. Furthermore, it improves the stability, feasibility, and precision of control by decoupling the computing power of the edge computing layer and the bottom execution layer. Attached Figure Description

[0014] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a structural diagram of the system of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0016] like Figure 1 As shown, a conveyor operation status control system based on a PLC program is implemented. The system includes: The data acquisition module is used to collect real-time operating status data and physical space parameters of the conveyor, including load weight data; The no-load resistance adaptive modeling module communicates with the data acquisition module. It is used to build an initial benchmark model based on real-time operating status data when the conveyor is determined to be in a no-load state based on load weight data. In subsequent no-load states, the initial benchmark model is updated through a sliding window to generate the current adaptive benchmark data. The load-background decoupling module communicates with the no-load resistance adaptive modeling module to obtain real-time operating status data under load, and compares the real-time operating status data under load with the current adaptive benchmark data to generate an operating residual sequence. The physical constraint feature extraction module communicates with the load-background decoupling module. It is used to construct spatiotemporal mapping constraints based on physical space parameters and construct the cost matrix of the dynamic time warping algorithm accordingly. The cost matrix contains the initial cost of each matching path. The running residual sequence and the preset mechanical feature sequence representing the single jamming resistance waveform of the idler roller extracted based on measured data are simultaneously input into the dynamic time warping algorithm containing the cost matrix for matching to filter random disturbances. Based on the waveform characteristics of the matched running residual sequence, the true degradation level scalar representing the mechanical additional load level is extracted. The adaptive control module communicates with the physical constraint feature extraction module and is used to generate intervention coefficients based on the actual degradation degree. The actual degradation degree is mapped to intervention coefficients with integer data format according to the preset degradation degree segmentation interval. The value range of the intervention coefficients is multiple discrete integers corresponding to different limiting strategies. The adaptive control module writes the intervention coefficients to the programmable logic controller (PLC) by writing to the register to dynamically adjust the output limiting of the conveyor drive frequency converter.

[0017] This embodiment provides a PLC-based conveyor operation status control mechanism. Specifically, this mechanism is applied to a long-distance heavy-duty belt conveyor in the main transport roadway of a coal mine. The upstream of the conveying line connects to the crusher's material drop point, and the downstream connects to the main shaft coal bunker. The drive end is located at the head of the conveyor, and weighing devices, speed encoders, tension sensors, and motor current acquisition interfaces are installed along the line. The system runs continuously throughout a complete production shift, first undergoing no-load inspection, then entering loaded conveying, experiencing local transient impact loads or roller jamming faults, and finally dynamically correcting the inverter output limit through the control link. Specifically, the data acquisition module continuously receives multi-source state quantities of the conveyor during operation; the real-time operating status data is essentially a set of state parameters characterizing the dynamic mechanical behavior of the conveyor: motor torque current reflects the work capacity required to overcome resistance at the drive end, running speed reflects the belt linear speed, tension reflects the stability of the belt under longitudinal force, and load weight data reflects whether there is a continuous material flow on the belt; physical space parameters correspond to the fixed structural conditions of the conveyor line itself, such as the distance from the material drop point to the drive roller, the propagation capacity of elastic disturbances in the belt, and the periodic distribution of the idler support structure. After acquiring this data, the no-load resistance adaptive modeling module does not directly regard the data at a certain moment as the normal value. Instead, it prioritizes finding the section where the conveyor is no-load and the mechanical load is mainly formed by the belt's own weight, idler friction, and the tensioning system. An initial benchmark model is established in this section. The essence of this model is to summarize what kind of no-load current will correspond to different speeds and tension conditions under the current equipment health state. Since the belt will change its resistance characteristics due to temperature, aging, and slow drift of the tensioning mechanism, the system uses a sliding window to update the benchmark in the subsequent no-load stage to make it closer to the current health state of the equipment, so as to dynamically replace the static reference threshold in the initial installation stage. During the loaded conveying phase, the load-background decoupling module compares the real-time measured operating status with the current adaptive benchmark. The resulting operating residual sequence no longer primarily represents basic friction and no-load sustaining work, but rather additional factors such as material disturbance, localized mechanical jamming, and abnormal belt friction. This sequence is essentially a time-series graph of abnormal fluctuations after filtering out the basic steady-state operating background data. For example, in an illustrative time-series scenario, three consecutive sampling segments... , , The corresponding motor torque and current data are relatively stable, but S2 suddenly spikes; if this spike occurs simultaneously when speed and tension do not change significantly, the residual sequence will... The location forms a local peak; However, the residual peak value alone cannot distinguish between random material drop impact and actual jamming. Therefore, the physical constraint feature extraction module further introduces the physical space constraint of the conveyor to correct the cost matrix of the dynamic time warping algorithm. Its core is not to simply compare waveform similarity, but to examine whether the fluctuation conforms to the actual propagation law of materials and mechanical parts on the conveyor. If the fluctuation sequence conforms to the propagation delay from the drop point to the drive end and does not have the periodic characteristics of the idler roller, it is more likely to be a material impact. If the fluctuation occurs repeatedly and the repetition interval corresponds to the spatial distribution of the idler roller, it is more likely to be idler roller jamming, local wear, or belt defects. The actual deterioration degree extracted by this process is closer to the mechanical condition deterioration itself, rather than on-site noise. The preset mechanical feature sequence is a standard waveform data sequence extracted from theoretical mechanical simulation based on historical typical fault samples or calibrated with field measured data. It is used to characterize the single resistance pulse triggered at the drive end when a single mechanical component of the conveyor jams. This sequence serves as a reference comparison benchmark for the DTW algorithm and is used to measure the morphological similarity between the actual operating residual and the typical mechanical fault waveform. The adaptive control module maps the actual degradation degree to an intervention coefficient and sends control commands to the frequency converter through the PLC. The dynamic adjustment of the output limit here is not limited to simple speed reduction. It can also be manifested as limiting the maximum output frequency, suppressing excessive acceleration, limiting the rise slope of the drive torque, etc., so that the conveyor can maintain stable and controlled operation in the stage of suspected mechanical deterioration and avoid local anomalies from being further amplified. It should be noted that if any key sensor loses connection for a short period of time, such as loss of weighing data or a sudden change in the speed encoder, the system can suspend the benchmark update and anomaly identification for the corresponding period, and only retain the original safety interlocks and fixed protection of the PLC; if the edge side cannot form a true degradation degree with a confidence level, no new intervention coefficient will be written to the PLC, and the previous stable control value will be maintained or the system will be switched to conservative limiting mode to avoid control malfunctions. At the beginning of the night shift, the conveyor runs unloaded at the inspection speed, and the system completes the initial baseline establishment for the shift. The crusher starts continuous feeding, and the conveyor enters the loaded state. Midway, a large piece of coal suddenly falls onto the belt, causing a short-term increase in the drive current. However, the timing of the fluctuation is consistent with the material propagation law, and the system filters it as random material impact, without triggering excessive intervention. After a period of time, a repetitive micro-jamming occurs at a certain idler roller, and the residual fluctuation begins to show repetitive characteristics related to the support spacing. The system recognizes that the actual deterioration degree has increased, and then adjusts the inverter output limit to a more stable range through the PLC to prevent the abnormality from further developing into belt breakage or overload shutdown. The purpose of this step is to extract the basic operating status characteristics, random material impact characteristics, and mechanical component deterioration characteristics of the conveyor system from the coupled electrical data sequence in a hierarchical manner, and to achieve closed-loop intervention in the form of a PLC control system that can be implemented in practice. In this embodiment, the data acquisition module includes: The status acquisition unit is used to acquire motor torque and current data, running speed data, tension force data, and weighing readings as real-time operating status data, of which the weighing readings are used as load weight data; the parameter configuration unit is used to receive and store physical distance, belt elastic wave transmission speed, and idler roller spacing as physical space parameters, of which the physical distance is the spatial distance from the material drop point to the drive roller.

[0018] This embodiment provides a data acquisition mechanism for conveyor state modeling. Specifically, in the coal mine main transportation scenario of the previous embodiment, if only the motor current is collected without collecting speed, tension force, weighing and spatial structure parameters, the subsequent anomaly judgment is prone to confuse load changes with mechanical resistance changes, and cannot distinguish between random material drop impacts and structural faults. Therefore, this embodiment further limits the composition and collection content of the data acquisition module. Specifically, the status acquisition unit receives at least four types of signals: the first type is motor torque and current data, used to characterize the power output level of the drive system to overcome conveying resistance; the second type is operating speed data, used to characterize the current linear speed of the belt, because under the same mechanical conditions, different speeds will bring different inertia and friction losses; the third type is tension data, used to characterize the belt preload and stress distribution, and changes in tension will change the resistance seen at the drive end; the fourth type is weighing readings, used to characterize whether the material on the belt is continuous and the load level, of which the weighing readings are directly used as load weight data; The parameter configuration unit stores data related to the geometry and mechanical propagation characteristics of the conveyor line; the physical distance corresponds to the distance along the line between the upstream drop point and the drive roller of the head, which determines the time basis for the material impact to be transmitted from the drop point to the drive end response; the belt elastic wave transmission speed reflects how fast the longitudinal disturbance of the belt propagates when it is impacted. This parameter enables the system to determine that the signal is not only transmitted through the overall movement of the material, but also through the elastic response of the belt, which may be transmitted earlier or later; the idler spacing reflects the spatial periodicity of the support system. When the idler bearing is locally jammed or the support is uneven, it will leave repeating characteristics in the current residual with a rhythm related to the spacing. For ease of explanation, the following simplified operating scenario can be used: Assume that in the continuous sampling segment, the current segment is... , , The speed segment is , , The tension segment is , , The weighing segment is , , If only observation A rise in elevation makes it impossible to directly determine whether it's due to increased material weight or roller jamming; however, if both are detected simultaneously... Significantly higher than and If W2 remains unchanged while Z2 and V2 remain stable, then an increase in I2 is retained as a mechanical anomaly. Here, we adopt... , , This is specifically used to represent tension segments to avoid confusion with the symbols used to represent time intervals in subsequent embodiments; As a supplementary protection mechanism, if the weighing device experiences a short-term jump due to coal accumulation or vibration, the system can mark that period as an untrusted load segment and suspend the no-load judgment based on the data of that segment; if the speed encoder is distorted, the current and tension force acquisition will be retained, but no speed-dependent spatial constraints will be generated; if the physical distance, elastic wave transmission speed, or idler roller spacing has not been manually configured, the system can revert to performing only basic acquisition and fixed threshold protection, without activating the intelligent identification link of physical constraints. On the same main conveyor belt in the coal mine, the status acquisition unit obtains the motor torque current from the frequency converter side, the speed pulse from the tail speed measuring roller, the tension force from the hydraulic sensor of the tensioning station, and the weighing reading from the weighing bridge; the parameter configuration unit pre-stores the line distance from the material drop point to the head roller, the belt elastic wave transmission speed corresponding to the current belt material, and the standard idler frame spacing; thus, the system can not only determine the cause of the current change, but also verify whether the change conforms to the spatial propagation law; The purpose of this step is to provide a complete and engineering-interpretive data foundation for subsequent benchmark modeling, residual stripping, and physical constraint verification, thereby enabling the traceability and feasibility of anomaly identification. In this embodiment, the no-load drag adaptive modeling module includes: The no-load determination unit is used to acquire the weighing reading. The conveyor is determined to be in a no-load state only when the following combined constraints are met simultaneously: the weighing load is lower than the no-load dead zone threshold and continues to reach the preset time; the conveyor is in a stable and continuous operating range, without start-stop or inching; and there are no drastic step fluctuations in the tension force and the running speed. If the weighing reading is higher than or equal to the preset dead zone threshold, or is lower than the preset dead zone threshold but has not reached the preset time, or does not meet the above conditions of stable operating range and no step fluctuations, the conveyor is determined to be in a loaded state or an unhealthy filtering state. The reference capture unit, connected to the no-load determination unit, is used to extract the three-dimensional mapping relationship of motor torque current data, running speed data and tension force data in the no-load state to construct an initial reference model. The reference correction unit, connected to the reference capture unit, is used to introduce a sliding window with a preset forgetting factor, use the three-dimensional mapping relationship in the current no-load state to perform weighted correction on the initial reference model, and output the current adaptive reference data containing the reference current component under different running speeds.

[0019] This embodiment provides an adaptive benchmark establishment and correction mechanism. Specifically, based on the data collected in the previous embodiment, if any low-amplitude current segment is directly set as the initial benchmark, the following judgment errors are likely to occur: First, when the conveyor belt carries a small amount of loose material, the weighing signal does not reach the threshold but actually deviates from the pure no-load condition; Second, the stiffness changes caused by belt creep, environmental temperature difference, and tensioning device displacement will cause low-frequency slow drift of the no-load holding current, resulting in the static fixed benchmark losing its characterization accuracy. Therefore, this embodiment introduces no-load judgment, benchmark grabbing, and sliding update with a forgetting factor. Specifically, the no-load determination unit prioritizes judging the presence of a stable material load based on the weighing reading. The dead-zone threshold is set to filter out minor fluctuations caused by zero-point drift of the weighing bridge, mechanical vibration, and belt rebound. To prevent misjudgments caused by slow sensor drift and material adhesion residue, the no-load determination unit also introduces a dynamic zero-point compensation mechanism and variance verification logic: after each transition from load to no-load, the variance of the weighing reading within the sliding time window is calculated in real time. ;like If the mean is less than the preset steady-state variance threshold and is within the original preset dead zone range, then zero-point drift self-calibration is triggered, and the current mean is set as the new relative zero. Only when the relative calibrated weighing reading remains below the preset dead zone threshold for a preset time, and the current weighing change rate is limited to within the permissible range, is it finally determined to be in an unloaded state, thus strictly shielding the interference of dynamic fluctuations and residual materials; the preset dead zone threshold is a specifically set weight value or load percentage, usually set to 2% to 5% of the rated full load weight of the conveyor; the specific value is adjusted based on the static calibration error of the weighing bridge on site and the maximum fluctuation amplitude during historical unloaded operation, to ensure that neither slight load is missed nor pure noise is mistakenly judged as load; Only when the weighing reading remains below the dead zone threshold for a predetermined time, to avoid recording abnormal no-load data from four unhealthy no-load conditions—standby, low-speed jogging, maintenance operation, and belt slippage—into the benchmark, causing benchmark drift and distortion that contradicts the goal of resisting aging and temperature drift, the no-load determination unit adds a joint constraint mechanism; the system must simultaneously meet the following requirements. The following conditions are required for the baseline to be updated: the weighing load is below the no-load dead zone threshold and remains below it for a specified duration; the conveyor is in a stable and continuous operating range, without starting, stopping, or jogging; and there are no drastic step fluctuations in tension or running speed. Only when all of the above combined constraints are met will the system determine that there is no material flow on the belt that forms a continuous driving burden, thus allowing this period to participate in baseline modeling. Once the weighing reading is higher than or equal to the threshold, the system switches to a loaded state and no longer uses it to update the no-load baseline. The reference capture unit extracts the relationship between motor torque current, operating speed, and tension force under no-load conditions. Its engineering implications are: even under no-load conditions, increasing speed alters drive power due to friction and inertia; increasing tension changes belt adhesion and support resistance. Therefore, no-load current must be described in conjunction with both speed and tension force, rather than retaining a single value. A simplified operating condition description can be used as follows: the no-load segment is divided into... , There are two sections. Section A corresponds to the first preset speed range and the first preset tension range. Section B corresponds to the second preset speed range and the tension value that overlaps with the first preset tension range. Both are healthy no-load, but the reference current will not be the same. Therefore, the reference model needs to record this correspondence. The benchmark correction unit further employs a sliding window to slowly correct the initial benchmark. The forgetting factor allows the latest idle data to gradually influence the benchmark while preserving the continuity of historical health states. Specifically, the preset forgetting factor is a constant within the interval (0,1), preferably ranging from 0.01 to 0.1. In this embodiment, the preset forgetting factor is specifically set to... Set the window size of the sliding window to Each sampling period, with a sliding step size of [number] times. Each sampling period; during each sliding window update, the benchmark correction unit employs a first-order hysteresis filter or exponential smoothing algorithm to make the current adaptive benchmark data equal to the historical benchmark model multiplied by . Add the currently extracted 3D mapping relationship multiplied by Its specific update formula based on the exponential smoothing algorithm is as follows: in, For the first The current adaptive baseline data after the second sliding window update. For the first The historical benchmark model of the previous iteration cycle. For the first The newly extracted 3D mapping relationship under the current no-load state. For a preset forgetting factor and satisfying This allows for quantitative control over the weighting of new state data. To avoid the problem that traditional fixed forgetting factors cannot simultaneously account for normal slow aging and sudden changes in equipment condition, this system uses a preset forgetting factor. Designed as an adaptive variable based on time series variance, its mathematical expression is: ,in Based on the steady-state forgetting base, For the currently extracted in front Local deviation variance within each sliding cycle It is the gain constant; When the conveyor experiences a drastic change in ambient temperature or after the replacement of some idler rollers, the foundation resistance experiences a step change. If the forgetting factor increases significantly, the system will automatically amplify it. This forces the baseline model to converge rapidly to the new physical state; while during steady-state operation, Degenerate into a minimum value This adaptive forgetting mechanism effectively filters out high-frequency sensor noise to the greatest extent possible, significantly improving the robustness of benchmark modeling under complex dynamic operating conditions and effectively avoiding the generation of pseudo-residuals. The aforementioned weighted correction mechanism uses smoothing filtering to adapt to low-frequency, slowly varying parameters such as belt aging, lubrication degradation, and temperature fluctuations, avoiding the introduction of benchmark calculation errors by single transient abnormal sampling; the slow drift following algorithm aims to make the adaptive benchmark model fit the long-term state trend of the equipment, while effectively filtering out high-frequency random disturbance signals. As a supplementary protection mechanism, if there is no idle window for a long period of time, such as continuous upstream material supply during a production shift, the system can continue to use the current adaptive reference data formed during the most recent reliable idle period and add a time-sensitive identifier to the reference. If there are drastic speed fluctuations or tensioning station maintenance during the idle period, the period will not be used for reference updates even if the weighing conditions are met, so as to avoid erroneously absorbing unsteady conditions into the healthy reference. If the weighing reading is stuck near the threshold for a long time and switches repeatedly, a duration criterion or hysteresis interval can be added to prevent frequent jitter between idle and loaded states. After the night shift handover, the conveyor is kept running for inspection after the coal flow is emptied; the system detects that the weighing bridge reading is continuously lower than the dead zone range, and determines that it has entered the no-load section; it captures the relationship between no-load current and tension at different inspection speeds at this time to form the initial benchmark for the shift; in the early morning period, the ambient temperature in the roadway drops, the rigidity of the rubber belt increases, and the no-load current rises slightly compared to the initial start-up. The system does not regard this as a fault, but slowly corrects the current benchmark through a sliding window to ensure that the benchmark always matches the current healthy mechanical background of this conveyor. The purpose of this step is to make the baseline model reflect the resistance level that the current equipment should have under no-load and healthy conditions, thereby providing a reliable reference for subsequent load-bearing anomaly stripping, and thus achieving the separation of chronic drift and sudden anomalies. In this embodiment, the load-background decoupling module is specifically used for: Under load, the reference current component corresponding to the current operating speed data in the current adaptive reference data is subtracted from the real-time collected motor torque and current data, and the operating residual sequence is output.

[0020] This embodiment provides an operational residual separation mechanism. Specifically, after the no-load baseline adaptive correction of the previous embodiment, the system has determined the approximate amount of drive current required for the conveyor to maintain its own operation under the current speed and tension conditions. However, in actual production, the total current during loaded conveying is simultaneously superimposed with factors such as material weight, material drop impact, local friction deterioration, and mechanical jamming. If separation is not performed, subsequent diagnosis will still be affected by the background interference. Therefore, this embodiment generates an operational residual sequence under loaded conditions. Specifically, the load-background decoupling module searches for the reference current component corresponding to the current operating speed during each load sampling period and separates it from the real-time motor torque current. The reference current operating speed is crucial here, as the no-load basic power consumption of the same device differs between low-speed cruising and rated speed operation. This embodiment emphasizes retrieving the corresponding reference data based on the current operating speed. The separated residual eliminates the steady-state load current required to maintain the basic operation of the system, effectively highlighting the additional current distortion characteristics caused by transient impacts or unsteady-state friction. In engineering, these increments may originate from material accumulation, material drop impact, idler jamming, belt misalignment friction, abnormal roller coating, etc. It should be noted that the reference current component selected from the current adaptive reference data corresponding to the current operating speed data in this embodiment does not mean that the tension force dimension is deleted from the reference model; its data source is still the three-dimensional mapping relationship of motor torque current data, operating speed data and tension force data in the previous embodiment; during engineering execution, the system first determines the effective reference surface or effective reference band under the current tension state in the three-dimensional mapping relationship based on the current tension force data, and then reads or interpolates the reference current component according to the current operating speed data in the effective reference surface or effective reference band; in other words, the embodiment emphasizes that the deduction object corresponding to the current operating speed data is the speed-related reference current component after being constrained by the tension force condition, which is used to ensure that the residual separation maintains the calculation form of the embodiment and is consistent with the three-dimensional mapping relationship in the embodiment; The following simplified data flow scenario can be used as an example: Assume the real-time current segment at a certain load moment is... The reference current segment corresponding to the current speed is The output residual fragment is If in continuous segments , , In China, only The formation of a bulge indicates that additional resistance beyond the healthy baseline has occurred at that moment; if , , A gradual and gradual increase is more likely to indicate a continuous increase in material load; if multiple peaks repeat in a certain rhythm, it provides clues for subsequent mechanical cycle verification. Unlike the traditional approach of directly checking whether the total current exceeds the limit, this embodiment preserves a purer abnormal response trajectory. The engineering significance of this is that when the conveyor is running at full load and in good health, the total current can reach a high rated value, but this does not indicate a fault. On the contrary, some minor roller jamming may not be obvious in the total current, but it will form an identifiable local residual after the reference is removed. As a supplementary protection mechanism, if the current speed is between two known reference speed ranges, the corresponding reference current component can be selected by using the nearest range or interpolation method; if the current tension is between two known tension states, the tension dimension can be selected by nearest range or interpolation first, and then the reference current component can be determined by combining the speed dimension; if the speed acquisition is abnormal, fine residual separation is paused, and only the original load current is retained as the observation quantity; if the current adaptive reference data is not timely due to the lack of a reliable no-load window for a long time, the system can reduce the residual interpretation weight and mark the diagnostic result as a low confidence state to avoid using the expired reference for strong control actions; During the stable coal flow transportation process, a transient surge in the head current occurred at a certain moment. The system verified that the belt speed was within the rated operating range at this time, and confirmed that the current tension did not deviate from the usable range of the benchmark model. Then, the current benchmark current at the corresponding speed was called for stripping. After stripping, it was found that the current increase did not come entirely from the normal transportation background, but left a significant pulse in the residual sequence. This pulse was sent to the subsequent motion calibration stage to determine whether it originated from random coal block impact or from a real mechanical anomaly formed by a certain idler support point. The purpose of this step is to decompose the total current during operation under load into a healthy background component and an abnormal additional component, thereby enhancing the expression of fault clues and reducing the masking caused by differences in basic operating conditions. In this embodiment, the physical constraint feature extraction module includes: The spatial distance constraint unit is used to calculate the material transfer delay time window based on physical distance, running speed data, and belt elastic wave transmission speed; the period penalty unit is used to calculate the spatial frequency period based on idler spacing and running speed data; the matching solution unit is connected to the spatial distance constraint unit and the period penalty unit, and is used to integrate the material transfer delay time window and the spatial frequency period into the cost matrix to perform sequence matching on the running residual sequence.

[0021] This embodiment provides a verification mechanism based on conveyor kinematic constraints. Specifically, residual sequences alone are insufficient for reliable judgment because the impact of large materials and minor mechanical jamming may both manifest as peak values ​​in the residuals. If a standard sequence matching algorithm without physical background is still used, the system may mismatch occasional material drop pulses as fault characteristics. Therefore, this embodiment further integrates spatial distance constraints and periodic penalty mechanisms into the matching solution process. Specifically, the spatial distance constraint unit provides a preset time window in the drive response where material impact may occur, based on the physical distance from the material drop point to the drive roller, the current belt speed, and the elastic wave propagation capability of the belt. The spatiotemporal mapping constraint is constructed by building a dynamic two-dimensional spatiotemporal feasible domain constraint matrix based on the material transmission delay time window and the spatial frequency period. Its dimension is the same as the cost matrix of the dynamic time warping algorithm. Consistent; for the cost matrix Any path node to be matched in ,in Represents the time step of running the residual sequence. The time step representing the mechanical feature sequence, and its corresponding spatiotemporal mapping constraint factor. Defined as: in, The exact time of occurrence of the current residual characteristic fluctuation. These are the upper and lower limits of the material transfer delay time window. The time interval between adjacent characteristic peaks. For spatial frequency period The zero-tolerance range set for the baseline; This module is based on the following mechanical constraints: after the material is loaded onto the conveyor belt, the dynamic load step sensed by the drive end cannot break through the hysteresis limit of the physical transmission of the system, and its timing lag is strictly controlled by the longitudinal kinematics of the belt and the elastic wave delay transmission theoretical model; therefore, any fluctuations that occur in time that are significantly outside this reasonable window should not be interpreted as disturbances caused by normal material feeding. The periodic penalty unit forms a spatial frequency period based on the idler spacing and running speed. Its engineering significance is that if a certain idler or several adjacent idlers jam, when the belt passes through these support points continuously, it will leave a repetitive rhythm related to the support spacing at the drive end. This rhythm is not ordinary random noise, but a projection of the mechanical structure period in the time dimension. After writing this periodic information into the matching rules, the algorithm not only evaluates the waveform morphological similarity, but also verifies whether its repetitive rhythm conforms to the mechanical jamming characteristics. The matching solution unit fuses the above two types of information into the cost matrix; the specific transformation logic is as follows: assuming the running residual sequence at time... The The data points are The first mechanical characteristic sequence characterizing the single-time jamming resistance waveform of the idler roller The data points are The initial base cost of its matching path is calculated as follows: The system is based on physical distance. Running speed and belt elastic wave transmission speed Constructing a material transport delay time window And generate time window constraint weight terms. The logic is as follows: Meanwhile, the system is based on the roller spacing. and running speed Calculate the spatial frequency period And for the time interval of adjacent fluctuation characteristics. Constructing the periodic penalty coefficient Define time interval For spatial frequency period The floating-point modulo result is The logic behind the punishment is as follows: in, greater than The preset penalty multiplier; physical space parameters are explicitly transformed into the algorithm's penalty term and weight matrix, integrated into the dynamic time warping algorithm, and used to update the node costs of the cost matrix. Its complete construction formula is expressed as: Through the above formula, the physical space parameters are quantitatively fused and mapped to the DTW cost matrix; To overcome the mismatch defect of the traditional DTW algorithm under nonlinear tension of the conveyor belt, this scheme further introduces a wave velocity dynamic compensation mechanism based on real-time tension in the cost matrix update; under actual heavy load or large start-stop conditions, the elastic wave transmission speed of the belt... It is not a constant value, but rather varies with the local tension. The change in undergoes a nonlinear drift; Therefore, the system collects the current tension value in real time and uses the formula Real-time correction of elastic wave propagation velocity, wherein The fundamental wave velocity under zero tension. The elastic coupling coefficient of the belt material; the corrected dynamic wave velocity. Material transfer delay time window is implemented in real time In the boundary calculation, the time window constraint weight term in the cost matrix is ​​made... It possesses the adaptability of the physical boundary of tension; The following simplified matching scenario can be used: Assume there are two candidate fluctuation segments in the residual sequence. and ;like If it occurs within a reasonable conveying time window but is inconsistent with the idler roller cycle, it is closer to material impact; if Although it may not be very large, if it appears repeatedly with an approximately fixed rhythm, it is closer to roller jamming. During the calculation, the system assigns different costs to different candidate paths, so that the interpretation path that conforms to the physical law maintains a low cost, while the interpretation path that violates the physical law is increased in cost or even eliminated. As a supplementary protection mechanism, if the current speed fluctuation is too large, making it difficult to stably define the transmission window and period within the same time period, the system can mark that time period as a transitional state, only output the observation results, and not generate strong intervention commands; if there are local non-standard arrangements in the idler spacing, such as maintenance sections or buffer idler sections, the corresponding parameters can be configured separately for that section to avoid misjudgment of the uniform period; if the elastic wave propagation parameters are missing, the system will only enable period verification and will not enable strong time window constraints to maintain the basic availability of the diagnostic link; During the middle of the night shift, a large piece of coal occasionally fell from the upstream crusher, resulting in a single peak in the residual sequence. Spatial distance constraints indicate that the timing of this peak is consistent with the time it takes for the material to travel from the drop point to the drive end. However, the periodic penalty reveals that there are no subsequent repetitive fluctuations consistent with the support rhythm of the idler rollers, so the system treats it as a random impact. Near the machine head, an idler roller bearing begins to show an abnormal increase in frictional resistance, resulting in several small but repetitive peaks in the residual. At this point, the matching solution unit retains the relevant path due to its obvious spatial periodic characteristics and increases the true degradation degree. The purpose of this step is to directly embed the spatial structure and propagation law of the conveyor into the time-series matching process, thereby achieving differentiated identification of random material disturbances and real mechanical degradation. In this embodiment, the spatial distance constraint unit is specifically used for calculating the material transport delay time window as follows: The basic delay time is obtained by dividing the physical distance by the running speed data; the baseline tolerance time is obtained by dividing the physical distance by the belt elastic wave transmission speed; the fluctuation tolerance time is multiplied by the preset tolerance coefficient to obtain the fluctuation tolerance time; the basic delay time plus the fluctuation tolerance time is used as the upper limit of the material transfer delay time window, and the basic delay time minus the fluctuation tolerance time is used as the lower limit of the material transfer delay time window.

[0022] This embodiment provides a mechanism for determining the material transfer delay time window. Specifically, although the concept of spatial distance constraint has been introduced in the previous embodiment, it is difficult to uniformly configure and reproduce the system in the engineering site if the specific formation method of the time window is not given. Therefore, this embodiment further clarifies that the time window is composed of the basic delay time and the fluctuation tolerance time. Specifically, the basic delay time reflects the main transportation time required for the material to move from the drop point to the drive end along the belt. It embodies the macroscopic propagation law of the material moving forward with the belt as a whole. The fluctuation tolerance time reflects the advance or lag margin brought about by the longitudinal propagation of the tension disturbance inside the belt after the impact enters the belt. In actual field, the current fluctuation sensed by the drive end is often affected by both the overall displacement of the material and the elastic coupling of the belt. Therefore, the reasonable range of material disturbance occurrence is not a single moment, but a window that fluctuates up and down around the basic delay time. In other words, the upper limit corresponds to the case where the material conveying time is longer than the basic delay time and the disturbance propagation is delayed, while the lower limit corresponds to the case where the elastic disturbance of the belt is transmitted to the drive end ahead of time. The time window constructed in this way is more in line with the real response characteristics of long-distance flexible conveying systems, rather than the approximately fixed time delay in rigid machinery. The following illustrative timing diagram can be used: Assume that a material dropping event is recorded as event at the source. In the residual sequence during the time period Fluctuations occurred nearby; the system does not require... It must fall precisely on a single theoretical point, but rather examine... Whether it is located between the upper and lower boundaries extended by the basic transport time delay; if it is located within the boundary, it can continue to be regarded as satisfying the material impact propagation condition; if it is significantly earlier than the lower boundary or later than the upper boundary, its physical interpretation is no longer preferentially attributed to the falling material disturbance. Among them, the above-mentioned events This term is used only to refer to a single recorded material dropping event, within the aforementioned time period. The terms "time period" and "period period" are used only to refer to the periods in which candidate fluctuations occur in the residual sequence. Both are local markers in this section and are not used as control variable symbols in subsequent embodiments. Specifically, "time period" is... This indicates the time location or analysis segment, but not the intervention coefficient or register write value, to avoid confusion with the symbolic representation in the control link below; Additionally, if the operating speed is below the preset lower threshold or is in the start-stop transition zone, the basic delay time will be significantly extended. In this case, the time window can be increased by a preset tolerance on the safety side, but it will not participate in high-confidence fault stripping. If the value of the belt elastic wave propagation speed changes due to belt replacement, the parameters should be reconfigured by maintenance personnel. The system can use a conservative window before the parameters are confirmed. If there are multiple material drop points, distances can be configured for different drop points to form multiple time windows. When matching, only one needs to be satisfied. On the coal mine's conveyor line, during a certain period of the night shift, the crusher experienced uneven feeding, resulting in a significant impact at the material drop point. Based on the distance from the material drop point to the head roller and the belt speed at that time, the system calculated the main time delay of the impact's transmission to the drive response. Simultaneously, considering that the belt tension wave could transmit part of the impact before the main material, a time window with front and rear envelopes was formed. When the pulse in the residual motor torque current data fell within this window, the system preferentially interpreted it as a material transmission disturbance rather than a roller jam. It should be noted that the method of using the basic delay time ± fluctuation tolerance time here is aimed at the engineering envelope of when the driving current residual enters the key review interval, rather than simplifying the driving end response to either the arrival time of pure material or the arrival time of pure elastic wave. That is, the lower limit value represents the earliest review boundary that can be paid attention to in advance relative to the main transportation delay, and the upper limit value represents the latest review boundary that can be paid attention to later. The two together serve the path selection of the subsequent cost matrix, rather than being used alone to assert that a certain impact must occur instantaneously at this boundary point. Furthermore, when the lower limit value obtained in the above manner is less than the current sampling starting point or has no discernible significance in implementation, the lower limit value can be truncated as the starting time of the current analysis window during engineering execution to maintain the feasibility of path filtering; this processing is only a refinement of the implementation of the time boundary and does not change the core logic of constructing the time window together with the basic delay time and fluctuation tolerance time. The purpose of this step is to provide a time boundary for material disturbance that conforms to the flexible propagation mechanism of the conveyor, thereby achieving an engineered constraint on when it can be attributed to the impact of falling material. In this embodiment, when the matching solution unit updates the cost matrix using the material transport delay time window, it specifically executes the following logic: Data segments in the residual sequence whose amplitude exceeds a preset fluctuation threshold are identified as fluctuation features. The occurrence time of these fluctuation features is then determined to be between the lower and upper limits of the material transport delay time window. When the occurrence time of the fluctuation feature is greater than or equal to the lower limit and less than or equal to the upper limit, the initial cost of the corresponding matching path in the dynamic time warping algorithm is retained. When the occurrence time of the fluctuation feature is less than the lower limit or greater than the upper limit, the cost of the corresponding matching path in the dynamic time warping algorithm is set to a preset maximum overflow penalty threshold. It refuses to be identified as a physical disturbance.

[0023] This embodiment provides a path filtering mechanism based on a delayed time window. Specifically, the previous embodiment has obtained a reasonable time range for material transportation. However, if the matching algorithm is still open to all paths, and if the time window is only used as a non-mandatory reference, it is easy to cause confusion and cross-judgment between the random material impact characteristics and the inherent mechanical fault characteristics in the algorithm matching calculation domain. Therefore, this embodiment further writes the time window as a strong constraint into the cost matrix and rejects matching paths that do not meet the window conditions. Specifically, the matching and solving unit first identifies data segments with amplitudes greater than a preset fluctuation threshold from the running residual sequence, and considers them as candidate fluctuation features. The purpose of setting the fluctuation threshold here is to avoid including sensor noise, inverter micro-adjustments, or ordinary mechanical ripples in the strong verification objects. The preset fluctuation threshold can be set dynamically or statically: when set statically, the value is a specific percentage of the reference current component at the current operating speed; when set dynamically, the preset fluctuation threshold is set to a specific multiple of the standard deviation of the current residual under a known stable and healthy load condition of the conveyor. After identifying the candidate fluctuation, the system then checks whether its occurrence time falls within the delay time window formed by the previous embodiment. In this embodiment, physical disturbance refers to disturbance that can be explained by the actual physical propagation path of the conveyor, especially material transport disturbance that is transmitted from the drop point into the belt through material transport and belt elastic response to the drive end. To keep the terminology clear, the time window screening does not reject the objective existence of the fluctuation as a current residual, but rather rejects its inclusion in the above-mentioned disturbance explanation path that conforms to the physical propagation law of material transport. Therefore, the material impact explanation, material disturbance explanation path and physical disturbance explanation path mentioned below are all based on the same type of propagation explanation constrained by the material transport delay time window, and are not confused with the periodic mechanical jamming explanation of the idler roller. If a candidate fluctuation occurs within this window, the system does not directly determine it to be a material impact. Instead, it retains the initial weight of the corresponding matching path, allowing subsequent algorithms to combine other information for further judgment. This retention means that the path is still eligible to participate in the interpretation calculation. Conversely, if the candidate fluctuation occurs earlier than the lower limit or later than the upper limit, clearly violating the physical time relationship from the material drop point to the drive end, the system will directly increase the weight of the corresponding path related to the material disturbance interpretation to the preset maximum overflow penalty threshold. Alternatively, a preset rejection threshold can be used, which is equivalent to rejecting the interpretation during the matching phase; wherein, to prevent data overflow or system crash during DTW cost matrix accumulation calculation in the embedded industrial gateway, the maximum overflow prevention penalty threshold is... The following constraint must be satisfied: in, The maximum positive integer limit supported by the current data type of the system; The following simplified time-series judgment scenario can be described: Assume there are two peaks in the residual. and The reasonable window for material impact calculated by the system falls within the time period. arrive ;like Appeared If it is compatible with the material impact explanation, the relevant path is preserved; if Appeared or If so, it should no longer be interpreted as being transmitted from the material drop point event, and the corresponding path is blocked; in this way, subsequent algorithms will not mistakenly identify a fluctuation that does not conform to physical logic in time as a material drop disturbance. As a supplementary protection mechanism, if the amplitude of a residual fluctuation is lower than the fluctuation threshold, the system will not use forced time window path filtering for it, but will retain it as a weak signal for observation; if the time window itself is too wide due to rapid speed changes, the system can reduce the intensity of infinite rejection and replace it with high-cost penalty to avoid excessive rejection during the start-up and shutdown transition phase; if multiple peaks in the residual overlap, making it difficult to define the fluctuation starting point, the window can be judged according to the peak group center time or the main peak time. During the latter half of the night shift, after a large piece of coal impacted the conveyor, two peaks appeared successively in the residual data of the motor torque current. One peak appeared exactly within the reasonable propagation window, while the other peak appeared significantly later. The system retained the former as a material impact, but directly rejected the latter as a material impact. In this way, even if the two peaks are similar in shape, they will not be confused due to mere waveform similarity. The latter is more likely to be transferred to the mechanical anomaly channel for further review. It should be added that the weight of the corresponding matching path is set to infinity here. The rejection object is the material transmission disturbance interpretation path corresponding to the time window, not the existence of the fluctuation itself. That is, once a fluctuation does not meet the material drop propagation boundary in time, the system only refuses to continue to interpret it as a material impact, but can still retain the fluctuation for subsequent mechanical cycle verification or anomaly observation link, so as to avoid mistakenly deleting real mechanical anomalies as invalid data. Furthermore, when the same wave corresponds to multiple candidate interpretation paths, the time window filtering only blocks the path branches that do not conform to the propagation law, without blocking the other path branches that still have physical rationality. By rejecting according to the interpretation path rather than deleting according to the original wave, the system can maintain the hard boundary of the time constraint and maintain the integrity of the information required for subsequent fault identification. The purpose of this step is to use whether or not it conforms to the laws of temporal propagation as a hard boundary for material disturbance identification, so as to eliminate erroneous interpretation paths that do not conform to the laws of physical propagation in advance. In this embodiment, the periodic penalty unit is specifically used to: The spatial frequency period is obtained by dividing the idler roller spacing by the running speed data. When the matching solution unit updates the cost matrix using the spatial frequency period, it specifically executes the following logic: It determines whether the absolute value obtained by taking the time interval of the fluctuation peak in the running residual sequence as a floating-point remainder of the spatial frequency period, or the absolute difference between the spatial frequency period and the floating-point remainder result, is within a preset zero-tolerance range. The zero-tolerance range is a time deviation range set based on the spatial frequency period. When the absolute value of the floating-point remainder or the absolute difference is within the zero-tolerance range, it is determined that there is a mechanical jamming feature, and the cost of the corresponding matching path remains unchanged. When neither the absolute value of the floating-point remainder nor the absolute difference is within the zero-tolerance range, it is determined that there is a random disturbance feature, and the cost of the corresponding matching path is amplified by a preset penalty multiple.

[0024] This embodiment provides a penalty mechanism based on the periodic characteristics of the idler rollers. Specifically, the strong time window constraint of the previous embodiment alone can eliminate a large number of erroneous interpretations that do not conform to the material drop propagation law. However, there is still a difficult situation: some random disturbances, although they do not violate the propagation law in time, may form multiple peaks locally. If the verification mechanism of the idler roller spatial period is not introduced, such discrete random high-frequency mutation characteristics are easily misjudged by the system as mechanical jamming anomalies. Therefore, this embodiment continues to use the spatial period formed by the idler roller spacing to perform secondary screening of the matching path. Specifically, idlers are usually arranged at approximately fixed intervals along the conveyor line. When a certain idler or a few adjacent idlers experience poor rotation, eccentricity, or local jamming, the drive end may repeatedly feel approximately rhythmic resistance disturbances every time the belt passes through a similar support unit. Since the belt is running, the fixed spatial interval will be converted into a repetitive period in time. Therefore, the periodic penalty unit forms a spatial frequency period based on the idler spacing and running speed, which serves as a mechanical periodicity reference. During the matching phase, the system verifies whether the time interval between adjacent fluctuation peaks in the residual sequence is consistent with the period. It does not require strict identical matching, but allows them to fall within a zero tolerance range, because there are belt slippage, speed fluctuations and sensor sampling errors in the field. The preset zero tolerance range is defined as the time deviation range in which the absolute value is less than or equal to a specific proportion of the current spatial frequency period, for example, set within ±10% of the spatial frequency period, to accommodate non-ideal time shifts in the field. If multiple peak intervals are basically consistent with the cycle, it indicates that these fluctuations have the characteristics of a fixed spatial structure repeatedly acting on the belt. Therefore, the cost of maintaining the corresponding matching path remains unchanged, allowing it to continue to be a candidate for mechanical jamming. If the peak intervals are not within the zero tolerance range of the cycle, it is more likely to be random material drop, uneven distribution of coal blocks, or external disturbances. The system then amplifies the cost of the corresponding path by multiplying it by a preset penalty factor greater than 1. This preset penalty factor is set to a constant between 2 and 5, thereby reducing its priority as an explanation of mechanical failure. The following simplified scenario for determining the period can be used: Assume that peak values ​​occur continuously in the residual. , , ;like arrive The interval and arrive If the intervals of the three peaks are close to the current idler cycle, then the peaks in this group exhibit mechanical repeatability. If the variance of the intervals of the three peaks exceeds the preset range, then it is more in line with the characteristics of a random impact sequence. Through this cost adjustment, the algorithm not only evaluates the peak amplitude, but also verifies whether it conforms to the periodic rhythm of the conveyor's spatial structure. As a supplementary protection mechanism, if the conveyor line uses a non-uniform idler arrangement in certain areas, such as buffer sections, transition sections, or correction sections, the cycle parameters of that section should be maintained separately; if the operating speed changes frequently in a short period of time, the peak interval determination should use the instantaneous speed or segmented speed at the corresponding moment to avoid mistaking speed changes as cycle deviations; if the number of peaks is insufficient and only a single isolated peak appears, the system will not directly determine jamming based on cycle penalties, but will retain it as an event to be observed. During continuous coal conveying, several medium-amplitude peaks appeared successively in the residual data of motor torque and current. Time window verification could not interpret them as typical material drop impacts, and further verification by periodic penalty found that the interval between these peaks was basically consistent with the rhythm of the on-site idler roller support. Therefore, it was retained as a mechanical jamming path. During a shutdown inspection, maintenance personnel found that the temperature of a certain group of idler roller bearings in the middle was abnormally high and the frictional resistance was increased, which corresponded to the system identification results. Conversely, if multiple peaks appear, but their intervals do not have a fixed rhythm, the system classifies them as random disturbances and increases the cost to prevent them from easily triggering intervention. It should be further explained that the determination of the remainder of the aforementioned time interval with respect to the spatial frequency period is not limited to comparing only the repetition relationship of exactly adjacent single periods. In the case of occasional missed detection, merging, or uneven amplitude of peak values ​​on site, two peak values ​​may also correspond to the cumulative effect of two or more idler cycles. In this case, as long as the result of the remainder of the time interval with respect to the spatial frequency period is still within the zero tolerance range, the system can still regard it as compatible with the rhythm of the mechanical structure, and will not misjudge it as a random disturbance due to cross-cycle comparison. Furthermore, the near-zero tolerance range preferentially absorbs small differences caused by sampling quantization errors, speed fluctuations, and local slippage, rather than relaxing the period constraint without limit. In engineering implementation, this tolerance can be adjusted according to the conveyor speed stability and sampling period, so that the period penalty is neither too rigid nor loses the ability to screen random peak groups due to excessive tolerance. The purpose of this step is to transform the idler structure cycle into temporal evidence for identifying mechanical jamming, thereby achieving further separation between random peaks and true periodic degradation. In this embodiment, the adaptive control module includes: The degradation degree mapping unit is used to extract the amplitude of the operating residual sequence retained by the matching solution unit, compare it with multiple preset abnormal threshold intervals, and map it into a discrete real degradation degree scalar. The underlying execution unit is connected to the degradation degree mapping unit and is used to convert the real degradation degree into an intervention coefficient with an integer data format. The intervention coefficient is sent to the programmable logic controller by writing to the register, and the programmable logic controller is controlled to dynamically adjust the output limit of the conveyor drive frequency converter according to the intervention coefficient.

[0025] This embodiment provides a mechanism for the implementation of diagnostic results into PLC control quantities. Specifically, in the previous embodiment, the system was able to distinguish between random disturbances and actual mechanical jamming. However, if the diagnostic results remain on the edge display interface and cannot be converted into control quantities that can be executed by the PLC, closed-loop protection cannot be formed. Therefore, this embodiment further specifies the mapping method of the actual degradation degree and the method of issuing the intervention coefficient. Specifically, the degradation degree mapping unit receives the residual results after motion calibration. At this point, the residual amplitude is no longer the unfiltered original fluctuation, but a credible abnormal response after physical constraint filtering. During mapping, the residual amplitude can be divided into several control intervals, such as slight abnormality, continuous abnormality, and significant abnormality, corresponding to different levels of true degradation. Its physical meaning is: to positively map the residual amplitude and the corresponding periodic characteristics to the additional load and mechanical state degradation degree of the drive system to overcome local abnormalities. Specifically, let the effective average amplitude of the running residual sequence after the sports school test within the time window be . The reference current component corresponding to the current operating speed is The number of repetitions of the identified mechanical jamming feature period is: The actual degradation (The range of values ​​is specified as) The mapping formula for ) is: in, For amplitude degradation gain parameters, This is the periodic frequency penalty coefficient; using this formula, complex residual waveform data can be transformed into... A clear and continuous indicator of actual degradation; The underlying execution unit converts the actual degradation level into an integer intervention coefficient. Using an integer format and sending it via register writing aligns better with common communication methods between PLCs and frequency converters in industrial settings, facilitating stable integration with existing control chains. Upon receiving the intervention coefficient, the PLC does not need to analyze complex diagnostic mechanisms; it only needs to adjust the frequency converter's output limit according to preset logic. For example, when the intervention coefficient corresponds to the first preset level, the PLC can maintain the current production capacity; when the intervention coefficient corresponds to the second preset level, the PLC can limit the maximum frequency or delay acceleration; when the intervention coefficient corresponds to the third preset level, the PLC can force the system into a conservative operating range; if the safety limit is reached, it can still trigger a shutdown in conjunction with existing hardware protection. The following specific mapping configuration instructions can be used: Assume that the system divides the actual degradation degree into... , , Level 3, corresponding to the integer intervention coefficient , , For example, K1 corresponds to an integer value of 0, indicating no intervention, with the inverter outputting at 100% of its upper limit frequency; K2 corresponds to an integer value of 1, indicating slight intervention, with the inverter's upper limit frequency output reduced to 90% and acceleration time extended; K3 corresponds to an integer value of 2, indicating severe intervention, with the inverter's upper limit frequency output reduced to 70%. To ensure that the underlying execution unit can achieve continuous and accurate closed-loop control, the actual degradation degree... To intervention coefficient The complete mapping logic for the final frequency conversion limiting follows the calculation rules below: Continuous true degradation Intervention coefficients converted to integer data format : in, This indicates a floor operation, making the intervention coefficient... Fixed as Integer values ​​within the specified range directly correspond to the values ​​in the PLC's internal registers. One control level; Unlike traditional control strategies that crudely reduce speed based solely on the total current amplitude, this scheme introduces a nonlinear flexible limiting strategy based on degradation mechanism decoupling; when the actual degradation degree is mapped... When amplitude degradation is the primary cause, the underlying execution unit triggers a hard limiting mode, directly lowering the maximum output frequency limit of the inverter; when the actual degradation degree... When the frequency of the cycle is the primary factor, the underlying execution unit triggers a soft limiting mode to maintain the current upper limit of the output frequency, but forcibly reduces the torque rise slope of the inverter output, allowing the conveyor to maintain its rated conveying capacity. At the same time, it softens mechanical shock by lengthening the system's response time to minor load fluctuations. This decoupling limiting mechanism ensures the continuity of the main shaft's conveying while minimizing transient tearing damage to slightly deteriorated components caused by strong drive. A coefficient-limiting strategy comparison table is pre-established in the PLC program; thus, only the edge side needs to be written. The PLC can automatically select a medium-level inverter limiting strategy without having to repeatedly execute complex sequence matching algorithms within the PLC. As a supplementary protection mechanism, if communication between the edge side and the PLC is interrupted, the PLC will continue to operate according to the last effective intervention coefficient, or fall back to the local conservative mode after the timeout; if the actual degradation changes too quickly, the underlying execution unit can be set to output smoothing and minimum hold time to avoid frequent back-and-forth changes to the limit; if the diagnostic result is a low confidence state, only a warning display can be performed without writing a new intervention coefficient to prevent over-control. Near the end of the night shift, the system continuously identified residual peaks consistent with the idler rhythm, indicating that the actual deterioration level had increased from a slight anomaly to a sustained anomaly. The edge gateway converted this state into an integer intervention coefficient and wrote it into the PLC register. After reading the coefficient, the PLC appropriately lowered the maximum output frequency of the inverter and tightened the acceleration slope, so that the conveyor could reduce the impact on the abnormal idler and the additional stress on the belt without affecting the basic coal conveying continuity. If the anomaly disappeared after subsequent maintenance, the intervention coefficient gradually recovered, and the PLC released the limit again. The purpose of this step is to convert complex diagnostic results into industrial control quantities that can be directly executed by the PLC, thereby achieving closed-loop linkage from status identification to drive protection. In this embodiment, the system's hardware deployment includes a bottom execution layer and an edge computing layer; the bottom execution unit in the adaptive control module is deployed in the bottom execution layer; the data acquisition module, the no-load resistance adaptive modeling module, the load-background decoupling module, and the physical constraint feature extraction module are all deployed in the edge computing layer, which is independent of the bottom execution layer, so as to achieve decoupling of computing power between control and diagnosis.

[0026] This embodiment provides a hierarchical deployment mechanism for control and diagnosis. Specifically, in the aforementioned implementation process, benchmark updates, residual separation, and sequence matching with fused spatial constraints all involve continuous data caching, window maintenance, and high-frequency computational processing. If all these functions are integrated and configured within the main control PLC's scan cycle, it will occupy the PLC's real-time resources used for interlocking and safety protection, and the limited floating-point and matrix processing capabilities of the PLC will also affect stability. Therefore, this embodiment divides the system into a bottom execution layer and an edge computing layer. Specifically, the bottom execution layer takes priority in rigid control responsibilities, including equipment start-up and shutdown, emergency stop interlocking, inverter basic control, sensor hard contact protection, and basic communication with the upper-level dispatcher. In this layer, only the bottom execution unit in the adaptive control module is retained to receive the integer intervention coefficient given by the edge side and adjust the inverter output limit accordingly. Its function is that even if the edge side is temporarily unavailable, the PLC can still rely on the existing program to ensure the safe operation of the conveyor. The edge computing layer deploys functions such as data acquisition, benchmark adaptation, residual separation, and motion calibration. This layer can use industrial gateways, industrial control computers, or integrated display and control machines. Relying on data processing capabilities that are more suitable for time series analysis, it performs mirroring, caching, and calculation on high-frequency state data. The edge layer does not directly drive the actuators, but interacts with the underlying execution layer through the intervention coefficient, a low-bandwidth, strongly constrained result variable, forming a computing power decoupling architecture where diagnosis is at the edge and execution is at the PLC. The following simplified data link scenario can be described: the field sampling stream first enters the edge side to form a sequence window, and the edge side outputs a single control result. Write to the PLC register, then press the PLC button. The corresponding strategy modifies the frequency converter's amplitude limit; in this way, from a large amount of raw data to an integer control quantity, the calculation and execution are clearly layered; the PLC does not process complex matching matrices, and the edge side does not directly overstep its authority to control the field safety loop; Among them, the above control results This is merely a symbol for a single integer intervention coefficient used in the scenario illustration in this section, representing the value written to the PLC register from the edge side; The letter symbols used in the previous text to represent time segments, fluctuation periods or other local schematic labels have different meanings. In engineering implementation, the meaning of the integer intervention coefficient should be taken first, and the individual letter itself should not be regarded as an independent technical feature. As a supplementary protection mechanism, if the edge computing layer loses power or restarts, the underlying execution layer will still maintain its original basic control and protection logic; if the communication delay between the underlying execution layer and the edge layer increases, the PLC can freeze the most recent effective intervention coefficient according to the communication timeout policy; specifically, a millisecond-level heartbeat message monitoring mechanism is established between the underlying execution layer and the edge computing layer; when the PLC detects continuous If no heartbeat refresh signal is received from the edge layer within a heartbeat cycle, the system automatically determines that communication is interrupted or there is a fault on the edge side, and then triggers the freeze logic. If the frozen state persists for more than the preset security downgrade time threshold The PLC will actively discard the frozen coefficient and force a downgrade to the most conservative hard limit protection mode to achieve completely independent control. After the edge computing layer recovers and the PLC receives stable heartbeat messages continuously again to reach the self-healing confirmation time window, the PLC will release the downgrade mode and reintroduce the new intervention coefficient issued by the edge side through the ramp function, thus forming a complete logical closed loop of edge side fault judgment, independent safe operation and communication recovery takeover. Furthermore, to address the risk of delayed confidence misjudgments in complex edge-side algorithms under sudden massive impact loads, this system innovatively constructs a control token handshake protocol based on strong coupling with the underlying operating conditions, in addition to the conventional heartbeat message mechanism; the edge computing layer sends integer intervention coefficients to the lower-level computer. At the same time, the formula must be issued simultaneously. The generated dynamic check hash value, where This represents the millisecond-level current integral value when extracting the diagnostic window from the edge side. The PLC underlying layer receives and At the same time, the latest torque and current integral values ​​collected in real time at the nanosecond level are used for local reverse verification. If the current background on which the intervention command is based is misaligned with the actual operating condition of the PLC at the moment due to the long time-series processing delay on the edge side, the hash reverse verification will immediately fail, and the PLC will directly reject and discard the limit adjustment command. This mechanism uses the underlying hard real-time field data to reverse timestamp align with the edge soft real-time command, thus blocking the risk of lag control caused by asynchronous delay of industrial Ethernet. If the edge layer diagnostic calculation load is too high, the diagnostic refresh frequency can be temporarily reduced without affecting the PLC's millisecond-level protection link. For maintenance mode, the edge layer can also be allowed to pause writing and only retain the data recording function to avoid manual operations during maintenance being mistakenly identified as faults. In the main transportation system of this coal mine, the PLC is installed in the machine head control cabinet and is responsible for hard protection such as rope pulling, deviation, smoke temperature, and overcurrent, as well as inverter start and stop control. The edge industrial control computer obtains current, speed, tension and weighing data through industrial network mirroring, completes no-load reference correction, residual extraction and physical constraint matching locally, and periodically writes integer intervention coefficients to the PLC. During a night shift, the edge industrial control computer restarted briefly. The PLC did not lose control as a result, but maintained conservative limit and continued to operate until the edge layer recovered before taking over the diagnostic output again. The purpose of this step is to structurally separate high-real-time field execution from computationally intensive state diagnostics, thereby enabling a maintainable, scalable system deployment method in industrial settings that does not interfere with safety control.

[0027] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A conveyor operation status control system based on a PLC program, including: The data acquisition module is used to collect real-time operating status data and physical space parameters of the conveyor, including load weight data; The system also includes an adaptive control module for sending control commands to a programmable logic controller (PLC) to dynamically adjust the output limiting of the conveyor drive frequency converter; the system further comprises a no-load resistance adaptive modeling module, a load-background decoupling module, and a physical constraint feature extraction module connected in series. The no-load resistance adaptive modeling module is communicatively connected to the data acquisition module. It is used to construct an initial benchmark model based on real-time operating status data when the conveyor is determined to be in a no-load state based on load weight data, and to update the initial benchmark model through a sliding window to generate the current adaptive benchmark data in subsequent no-load states. The load-background decoupling module is communicatively connected to the unloaded resistance adaptive modeling module. It is used to acquire real-time operating status data under load and compare the real-time operating status data under load with the current adaptive benchmark data to generate an operating residual sequence. The physical constraint feature extraction module is communicatively connected to the load-background decoupling module. It is used to construct a spatiotemporal mapping constraint based on physical space parameters, and accordingly construct a cost matrix for a dynamic time warping algorithm. The cost matrix contains the initial cost of each matching path. The running residual sequence and a preset mechanical feature sequence representing the single jamming resistance waveform of the idler roller extracted based on measured data are simultaneously input into the dynamic time warping algorithm containing the cost matrix for matching, so as to filter random disturbances and extract the true degradation degree. The true degradation degree is then output to the adaptive control module, so that the adaptive control module can generate an intervention coefficient based on the true degradation degree and generate the control command based on the intervention coefficient.

2. The conveyor operation status control system based on PLC program as described in claim 1, characterized in that, The data acquisition module includes: The status acquisition unit is used to acquire motor torque current data, running speed data, tension force data and weighing reading as the real-time running status data, wherein the weighing reading is used as the load weight data; the parameter configuration unit is used to receive and store physical distance, belt elastic wave transmission speed and idler roller spacing as the physical space parameters, wherein the physical distance is the spatial distance from the material drop point to the drive roller.

3. The conveyor operation status control system based on PLC program as described in claim 2, characterized in that, The no-load resistance adaptive modeling module includes: An unloaded determination unit is used to acquire the weighing reading. When the weighing reading is continuously below a preset dead zone threshold for a preset time, the conveyor is determined to be in the unloaded state. When the weighing reading is higher than or equal to the preset dead zone threshold, or is below the preset dead zone threshold but has not reached the preset time, the conveyor is determined to be in the loaded state. The preset dead zone threshold is 2% to 5% of the conveyor's rated full load weight. A reference capture unit, connected to the no-load determination unit, is used to extract the three-dimensional mapping relationship between the motor torque current data, the operating speed data, and the tension force data under the no-load state to construct the initial reference model; a reference correction unit, connected to the reference capture unit, is used to introduce a sliding window with a preset forgetting factor, use the three-dimensional mapping relationship under the current no-load state to perform weighted correction on the initial reference model, and output the current adaptive reference data containing the reference current component under different operating speeds.

4. The conveyor operation status control system based on PLC program as described in claim 3, characterized in that, The load-background decoupling module is specifically used for: Under the load condition, the reference current component corresponding to the current operating speed data in the current adaptive reference data is subtracted from the real-time collected motor torque and current data, and the operating residual sequence is output.

5. The conveyor operation status control system based on PLC program as described in claim 4, characterized in that, The physical constraint feature extraction module includes: A spatial distance constraint unit is used to calculate the material transport delay time window based on the physical distance, the running speed data, and the belt elastic wave transmission speed; a periodic penalty unit is used to calculate the spatial frequency period based on the idler spacing and the running speed data; a matching solution unit is connected to the spatial distance constraint unit and the periodic penalty unit, and is used to fuse the material transport delay time window and the spatial frequency period into the cost matrix to perform sequence matching on the running residual sequence.

6. The conveyor operation status control system based on PLC program as described in claim 5, characterized in that, The spatial distance constraint unit, when calculating the material transport delay time window, is specifically used for: The physical distance is divided by the operating speed data to obtain the basic delay time; the physical distance is divided by the belt elastic wave transmission speed to obtain the reference tolerance time, and the reference tolerance time is multiplied by a preset tolerance coefficient to obtain the fluctuation tolerance time; the basic delay time plus the fluctuation tolerance time is used as the upper limit of the material transfer delay time window, and the basic delay time minus the fluctuation tolerance time is used as the lower limit of the material transfer delay time window.

7. The conveyor operation status control system based on PLC program as described in claim 6, characterized in that, When the matching solution unit updates the cost matrix using the material transport delay time window, it specifically executes the following logic: Identify data segments in the running residual sequence whose amplitude exceeds a preset fluctuation threshold as fluctuation features, and determine whether the occurrence time of the fluctuation feature is between the lower limit and the upper limit of the material transfer delay time window; When the occurrence time of the fluctuation feature is greater than or equal to the lower limit and less than or equal to the upper limit, the initial cost of the corresponding matching path in the dynamic time warping algorithm is retained. When the occurrence time of the fluctuation feature is less than the lower limit or greater than the upper limit, the cost of the corresponding matching path in the dynamic time warping algorithm is set to a preset maximum overflow penalty threshold. It refuses to be identified as a physical disturbance.

8. The conveyor operation status control system based on PLC program as described in claim 7, characterized in that, The periodic penalty unit, when calculating the spatial frequency period, is specifically used for: The spatial frequency period is obtained by dividing the idler roller spacing by the running speed data. When the matching solution unit updates the cost matrix using the spatial frequency period, it specifically executes the following logic: it determines whether the absolute value obtained by performing a floating-point remainder calculation on the spatial frequency period based on the time interval of the fluctuation peak in the running residual sequence, or the absolute difference between the spatial frequency period and the floating-point remainder result, is within a preset zero-tolerance range, wherein the zero-tolerance range is a time deviation range set based on the spatial frequency period; when the absolute value of the floating-point remainder or the absolute difference is within the zero-tolerance range, it is determined that there is a mechanical jamming feature, and the cost of the corresponding matching path remains unchanged; when neither the absolute value of the floating-point remainder nor the absolute difference is within the zero-tolerance range, it is determined that there is a random disturbance feature, and the cost of the corresponding matching path is amplified by a preset penalty multiple.

9. The conveyor operation status control system based on PLC program as described in claim 8, characterized in that, The adaptive control module includes: The degradation degree mapping unit is used to extract the amplitude of the running residual sequence retained by the matching solution unit, compare it with multiple preset abnormal threshold intervals, and map it into a discrete real degradation degree scalar. The underlying execution unit is connected to the degradation degree mapping unit and is used to convert the real degradation degree into the intervention coefficient with an integer data format, and send the intervention coefficient to the programmable logic controller by writing to the register, so as to control the programmable logic controller to dynamically adjust the output limit of the conveyor drive frequency converter according to the intervention coefficient.

10. The conveyor operation status control system based on PLC program as described in claim 9, characterized in that, The system comprises a low-level execution layer and an edge computing layer in its hardware deployment; the low-level execution unit in the adaptive control module is deployed in the low-level execution layer; the data acquisition module, the no-load resistance adaptive modeling module, the load-background decoupling module, and the physical constraint feature extraction module are all deployed in the edge computing layer, which is independent of the low-level execution layer, so as to achieve decoupling of computing power between control and diagnosis.

Images