PLC data intelligent analysis and optimization control method

Abstract

The invention relates to the technical field of kiln sintering production, in particular to a PLC (Programmable Logic Controller) data intelligent analysis and optimization control method, which comprises the following steps of: acquiring a center coordinate and an attitude angle of a current workpiece to be operated, introducing the center coordinate and the attitude angle into a PLC control process, dynamically correcting an original mechanical arm path, and directly driving a controller to execute actions; calculating a position error and an angle error under the current execution action, fusing the position error and the angle error through a joint error energy function, generating a deviation value, and outputting an action state mark according to a comparison result of the deviation value and a preset deviation threshold value; a historical failure trend quantity is introduced, a behavior inertia function is constructed in combination with the deviation value and the action state mark, a trend state value is output through the behavior inertia function, and automatic switching of process logic in the PLC is guided according to a comparison result of the trend state value and a preset state threshold value.

Description

Technical Field

[0001] This invention relates to the field of kiln sintering production technology, and more particularly to a PLC data intelligent analysis and optimization control method. Background Technology

[0002] In modern kiln sintering production lines, the automated stacking of sandwich panels and columns directly affects the geometric accuracy of the products, the yield rate of finished products, and the stability of the production line cycle time. However, in actual production environments, the incoming material condition is not stable. Not only is there often a planar positional offset of up to 160 mm, but there are also complex factors such as angular torsion, edge damage, localized reflection, and slight tray deformation. This makes it difficult to meet the accuracy requirements using robotic arm operations that rely solely on traditional geometric templates or fixed paths. During the gripping and placement process, columns are often extremely sensitive to angular errors due to their slender structure; even slight tilting can cause misalignment in subsequent stacking layers or even column tipping. Sandwich panels are large and have significant edge irregularities; if their center position is not consistently identified, additional deviations will be introduced into the trajectory planning. In most existing systems, the robotic arm's path is usually hard-coded in the PLC with fixed coordinates, lacking the ability to adaptively adjust based on real-time deviations. This leads to problems such as adsorption deviation and inaccurate placement when material fluctuations occur.

[0003] Furthermore, existing automated equipment generally lacks a quantitative mechanism for the results after an action is completed. Whether an action is completed or not often relies on time windows or limit switches for rough judgment, failing to identify whether placement is truly stable, especially for columns where "contact is possible but instability is not," which is impossible to detect. The lack of closed-loop judgment not only prevents the system from correcting deviations in a timely manner but also causes error accumulation in multi-layer stacking, leading to serious misalignment in later layers. Under high-cycle conditions, these hidden errors are often only discovered at the finished product stage, resulting in the scrapping of the entire batch. Existing control logic also lacks the ability to adjust the process based on the success or failure trend of actions. Failures can only be addressed with simple retries or manual intervention, failing to achieve dynamic identification and strategy switching of unstable regions. Therefore, the entire industry urgently needs a closed-loop control method with data-driven capabilities from deviation identification to trajectory correction, action quality evaluation, and control logic switching, enabling PLCs to truly assume the role of intelligent decision-making. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a method for intelligent analysis and optimized control of PLC data.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A PLC data intelligent analysis and optimization control method, comprising:

[0007] The input image is acquired, and edge extraction is performed on the input image to generate an edge point set. The principal axis direction unit vector is obtained by fitting the edge point set. The center coordinates of the edge point set are calculated by using a weighted function. The center point coordinates are mapped to the workpiece center coordinates through a preset coordinate transformation matrix. The angle between the principal axis unit vector and the horizontal axis of the workpiece center coordinates is calculated to generate the attitude angle.

[0008] Introducing target path parameters, calculating the initial deviation between the workpiece center coordinates and attitude angle and the preset target path parameters, based on the initial deviation, combining the compensation function to correct the workpiece center coordinates and attitude angle, and outputting the corrected workpiece center coordinates and attitude angle, writing the corrected workpiece center coordinates and attitude angle into the control buffer through the PLC register, and calling the internal position control subroutine to pass the command to the robotic arm controller, which then drives the robotic arm to perform the action.

[0009] Calculate the position error and angle error under the current action, fuse the position error and angle error through the joint error energy function to generate the deviation value, and output the action status mark based on the comparison result of the deviation value and the preset deviation threshold.

[0010] Historical failure trend data is introduced, and a behavior inertia function is constructed by combining deviation value and action status marker. The trend status value is output through the behavior inertia function, and the automatic switching of process logic inside the PLC is guided by the comparison result of the trend status value and the preset status threshold.

[0011] Preferably, the step of guiding the PLC to automatically switch process logic based on the comparison result between the trend state value and the preset state threshold specifically includes:

[0012] When the path is determined to be normal, the workpiece center coordinates corresponding to the normal path are sent to the robotic arm controller and the next action is initiated.

[0013] When a path is determined to be a retry path, a small disturbance compensation is applied to the original workpiece center coordinates.

[0014] When a path is determined to be a failure or an abnormal path, the current action status flag is recorded and the target workstation is skipped.

[0015] Preferably, after constructing a behavior inertia function by combining deviation values ​​and action state markers, and outputting trend state values ​​through the behavior inertia function, the method further includes:

[0016] When an action is marked as failed, the historical failure trend is updated using an exponential decay mechanism.

[0017] When the action status is marked as successful, the historical failure trend is corrected downwards.

[0018] Preferably, the parameters of the exponential decay function include the current historical failure trend amount and the historical decay coefficient.

[0019] Preferably, the parameters of the weighting function include the directional weights corresponding to the edge points, the image coordinates of the edge points, and the number of edge points. The directional weights corresponding to the edge points are calculated and generated by the unit vectors of the edge points in the image gradient direction, the unit vectors in the principal axis direction, and the smoothing coefficient.

[0020] Preferably, the parameters of the compensation function include a position deviation compensation weight coefficient, a historical stacking deviation trend weight factor, and the average deviation residual value observed during the historical path execution process.

[0021] Preferably, the parameters of the joint error energy function include position error, angle error, weighting factor, height deviation, and structural stability adjustment term. The weighting factor is set by the fixture type and workpiece geometry. The height deviation is used to assist in judging the abnormality of the object's invisible structure. The height deviation is obtained by a laser displacement sensor.

[0022] Preferably, the parameters of the behavior inertia function include an adjustment coefficient, a deviation value, an action status flag, and an action execution sequence number, wherein the action execution sequence number is used to indicate the order of the current workpiece in the stacking task.

[0023] Preferably, the historical failure trend is stored in a register inside the PLC and updated after each action is executed.

[0024] Preferably, the target path parameters include preset workpiece center coordinates and preset attitude angles.

[0025] The beneficial effects of this invention are as follows:

[0026] This invention establishes a dynamic path adjustment mechanism based on workpiece spatial deviation and posture information acquired by a vision system, ensuring the robotic arm accurately aligns with the true position and orientation of the batch of materials before each action. A robust deviation identification mechanism is established based on the center point and spindle angle extracted from edge structure consistency, providing stable input for subsequent control. A compensation term for historical stacking offset trends is introduced during path correction, enabling the system to automatically suppress deviation accumulation over long periods, eliminating the need for manual recalibration. After action execution, this invention designs an energy function that simultaneously considers positional deviation, posture deviation, and height anomalies to quantify placement quality, identifying structural instability that is difficult to detect with traditional visual judgment. Combining this quantification index, this invention further constructs a logical switching mechanism that determines process branches based on historical failure trends and current action results. This allows the PLC to automatically select a strategy based on the operating status: continue execution, fine-tune and retry, or skip the current workstation, thereby avoiding cycle interruptions caused by repeated failures. Compared with traditional control methods that rely on fixed paths, lack judgment mechanisms, and have no historical learning capabilities, this invention forms a complete closed-loop chain from deviation identification, path correction, quality judgment to process switching, enabling the robotic arm's movements to truly have adaptive capabilities. This effectively improves stacking accuracy, reduces manual intervention, and lowers the risks caused by material fluctuations, enabling stable and reliable operation in high-precision, high-cycle, and highly variable sintering stacking scenarios. Attached Figure Description

[0027] Figure 1 This is a flowchart of a PLC data intelligent analysis and optimization control method in a specific embodiment of the present invention; Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] refer to Figure 1 As shown, this application proposes a PLC data intelligent analysis and optimization control method, including:

[0030] Step 1: Acquire the input image, extract edges from the input image to generate an edge point set, fit the edge point set to obtain a unit vector along the principal axis, calculate the center coordinates of the edge point set using a weighted function, map the center point coordinates to the workpiece center coordinates using a preset coordinate transformation matrix, and calculate the angle between the principal axis unit vector and the horizontal axis of the workpiece center coordinates to generate the attitude angle. Specifically, this includes:

[0031] This step is used to obtain the spatial position and rotation angle of the workpiece to be operated, which serves as the sole input data for subsequent path correction. In actual working conditions, the types of stacked workpieces include sandwich panels and columns, with inconsistent dimensions and irregularities in their edges, such as reflections, chipped corners, and irregularities. This causes traditional geometric template matching methods to drift during actual recognition. To adapt to scenarios with unstable stacked material shapes and a high probability of occlusion, a center positioning method with edge tolerance is needed. This step uses a fixed industrial camera to acquire images, establishes an edge point set after edge detection, and introduces a weighted mechanism based on principal direction consistency to calculate the center coordinates. Simultaneously, to determine the orientation of the workpiece, the principal axis direction of the overall contour needs to be estimated and output as an angle.

[0032] The system uses a fixed-mount industrial camera to acquire images. The camera is connected to the image processing module via an industrial interface. A PLC triggers the camera to capture a single frame after the robotic arm reaches the predetermined detection position, with each frame containing a single object to be manipulated. After image acquisition, the image is sent to the processing module for edge extraction, using the Canny operator to obtain a set of edge points. ,in For the first Image coordinates of 10 edge points. This set contains all detected high-gradient edge points, covering the outer contour of the object and part of the internal disturbance boundaries.

[0033] To reduce the interference of irregular edges on the calculation of center coordinates, the following weighted function is used to calculate the center coordinates of the edges:

[0034] ;

[0035] in, For the first The orientation weights of the edge points are defined as follows:

[0036] ;

[0037] in Represents edge points The unit vector in the gradient direction of the image. This represents the unit vector along the principal axis obtained by fitting the entire set of edge points. The smoothing coefficient is empirically set to a constant. The above expression measures the consistency between the direction of the edge point and the principal axis direction; points with higher consistency have higher weights. Therefore, it accounts for a larger proportion in the calculation of the center coordinate.

[0038] Calculated center coordinates The image-machine coordinate transformation matrix established during the system initialization phase will be mapped to the workpiece center coordinates in the robot arm coordinate system. And attitude angle Then by The angle between the workpiece and the horizontal axis of the image coordinate system is calculated, representing the rotational difference between the current workpiece and the standard placement orientation.

[0039] Step Two: Introduce the target path parameters, calculate the initial deviation between the workpiece center coordinates and attitude angles and the preset target path parameters. Based on the initial deviation, use a compensation function to correct the workpiece center coordinates and attitude angles, and output the corrected workpiece center coordinates and attitude angles. Write the corrected workpiece center coordinates and attitude angles into the control buffer through the PLC register, and call the internal position control subroutine to pass the command to the robot arm controller. The robot arm controller then drives the robot arm to perform actions, specifically including:

[0040] This step is used to retrieve the workpiece center coordinates obtained in the previous step. With attitude angle By introducing a PLC control process, the original robotic arm path is dynamically corrected, and the actuator is directly driven to complete the gripping or placing action. The goal of this path correction mechanism is to ensure that the gripping point and the placement point are precisely aligned even when there are positional or angular deviations in the material. The angle of the end effector is adjusted synchronously through the attitude angle to avoid failures such as misalignment, poor fit, or misplacement.

[0041] Considering the following significant characteristics of the stacking scenario addressed by this invention: (1) There is a systematic shift between batches of material stacking, and the deviation direction of each batch of stacking has a certain regularity, for example, one side is always biased or the pallet often has rotational error due to platform accuracy; (2) The objects to be clamped include two different structures, columns and plates, with large differences in the shape of the adsorption surface, resulting in significant differences in the tolerance range of the end effector for posture alignment during the clamping and placement stages; (3) The path compensation accuracy needs to be controlled within ±3mm and ±2°, and the compensation time should not exceed 50ms, otherwise it will slow down the overall production line cycle time. Therefore, the path correction strategy needs to take into account both the flexibility of compensation and the low latency of execution.

[0042] The input variable comes from the previous step and is the workpiece center coordinate. and actual attitude angle The preset target path parameters are and This parameter comes from the standard placement position set in the process station or product settings. The initial deviation is obtained by directly calculating the difference. ,in:

[0043] ;

[0044] Positional deviation It is obtained from the difference between two coordinates. To prevent simple difference compensation from introducing regular error accumulation in repetitive structural scenarios, a novel compensation function that introduces a "stacking correction potential energy term" is proposed. The specific expression of the compensation function is as follows:

[0045] ;

[0046] In this formula, The position deviation compensation weighting coefficient has a range of values. This is used to adapt to the gripping tolerance range of different fixtures. The historical stacking deviation trend weighting factor is automatically learned and updated by the system. For the front The average deviation residual value observed during the execution of the secondary path. This "historical potential term" is used to suppress the control path periodic error caused by systematic deviations in repeated batch stacking, and has a dynamic correction capability that is significantly different from that of conventional linear compensation paths, especially showing long-term stability advantages in large batch stacking. For example, if the first dozens of batches of stacking are as follows... If the directional bias is 3mm, then The system will automatically add the offset to the path, without needing to reset the target value.

[0047] For attitude angle adjustment, the control system compensates... At the same time, a safety range check is added to the end-rotation mechanism of the fixture to avoid overtravel or collision caused by attitude angle adjustment. The maximum allowable angle is set in the system. If the compensation angle exceeds this range, the fixture rotation forward shift operation is triggered to move the workpiece to a relatively safe area before angle compensation is performed.

[0048] All corrected workpiece center coordinates With attitude angle The command is written to the control buffer via the PLC register and the internal position control subroutine is called to pass the command to the end effector trajectory planning module of the six-axis robotic arm controller. This process can be processed in parallel with the data transmission from the vision system, and the entire compensation-control flow is completed within 50ms.

[0049] Step 3: Calculate the position and angle errors under the current action. Fuse the position and angle errors using a joint error energy function to generate a deviation value. Output the action status flag based on the comparison between the deviation value and a preset deviation threshold. Specifically, this includes:

[0050] This step follows the output from the path correction and action execution stages. Its purpose is to provide a structured assessment of the completion status of the gripping or placing action, thus offering clear and executable success or failure signals to subsequent process control modules. Unlike traditional industrial control systems that rely on indirect judgments of action results (such as limit switch triggering or default completion within a time window), this step introduces a state determination mechanism based on visual feedback data. It directly quantifies the accuracy of action execution by measuring the spatial error between the actual action result and the target instruction. The output from the path correction step... The desired position and posture for the current operation are calculated by the PLC based on the deviation identified by the vision system. This step compares the corrected target with the actual action result, and finally determines whether the action meets the standard through error structure analysis.

[0051] After the material handling is completed, the vision system takes another picture of the stacking area, and the image is processed in the same way to extract the new position of the current action. and attitude angle To maintain consistency with the variable structure in the path correction step, the position error is calculated based on this data. The expression is:

[0052] ;

[0053] This formula is used to represent the spatial offset of an object relative to its target position on the actual stacking surface, in millimeters. In the industrial context of this invention, sandwich panels have a large surface contact area, and their allowable error is generally ±3mm; while the column structure is relatively slender and is extremely prone to tipping over when offset, requiring higher placement accuracy, generally not exceeding ±1.5mm.

[0054] Regarding attitude angle error judgment, considering the slight deviation of the fixture during attitude rotation and the fact that the geometric features of the target workpiece affect the accuracy of the vision system's angle recognition, this step does not directly rely on... Instead of using it as the sole criterion, an angle matching index based on "consistency of expected direction" is introduced, specifically the angle error. The formula for calculation is:

[0055] ;

[0056] This expression has two advantages: first, when the angular deviation is close to 0, the error value approaches 0, which aligns with physical intuition; second, it can naturally compress the angular error. The interval facilitates subsequent evaluation in conjunction with positional errors. Furthermore, based on the complex stacking structure characteristics of the invention scenario, this step proposes a joint error energy function, expressed as:

[0057] ;

[0058] in, This indicates the deviation value of the current action result, used to comprehensively assess whether it is in an "acceptable" state; This is the weighting factor of the angle error in the joint energy, which is set by the fixture type and workpiece geometry. For example, it can be set to 0.5 for plate fixtures and 1.2 for column scenarios. This is a structural stability adjustment term. Indicates the current workpiece is in The height deviation in the direction is used to help determine whether an object has experienced invisible structural anomalies such as warping or tilting. This can be obtained through a laser displacement sensor or a 3D vision module. This energy function allows the judgment to extend beyond position and angle deviations in two-dimensional images, incorporating the stability of the stacked structure into the judgment system, which is the most valuable innovation in this step. For example, if the vision system identifies a column with a placement deviation of only 1.1mm and an angle error of 1.5°, but its top height is 6mm higher than the theoretical value (i.e., it is tilted), then even if the position and angle are within the threshold, It is still possible to exceed the set upper limit, and the system will determine the action as a failure. This mechanism significantly improves the system's ability to identify incorrect placement behavior where "the structure is unstable but the image is normal". The system sets a deviation threshold. ,like If the action is successful, the status is marked as 1; otherwise, it is marked as 0. This status mark is passed as the sole output to the next control logic.

[0059] Step 4: Introduce historical failure trend data and construct a behavior inertia function by combining deviation values ​​and action status markers. Output trend status values ​​through the behavior inertia function. Guide the PLC to automatically switch process logic based on the comparison results between these trend status values ​​and preset status thresholds. Specifically, this includes:

[0060] This step is used to evaluate the deviation value output in the previous step. and action status markers Based on this, the PLC's internal process logic is automatically switched to achieve dynamic response to the results of operational actions. In the entire control loop, this step receives the judgment result and makes a system decision, directly determining whether to proceed to the next gripping / placing task, retry the current action, or skip the current target position and record the anomaly. Compared to traditional state judgment logic based on whether "completed in place" or "limit triggered," this step introduces an error energy function. Historical failure trend volume and number of times the action is executed This forms a dynamic control system that can adapt to batch differences in materials, equipment fluctuations, and fault tolerance thresholds. Its purpose is not only to make single judgments on a particular action, but also to perceive the system's operational status through data trends, thereby improving the stability of the control chain and the rationality of the strategy.

[0061] The input for this step includes three variables: the deviation value output from the previous step. (Used to reflect the overall deviation of this gripping or placement action), status marker (in Indicates failure. (Indicates success), and the action execution sequence number. This indicates that the current workpiece is the [number]th [item] in this stacking task. In addition, the system internally maintains a historical failure trend. It is used to record the failure trend of recent actions and to adjust the sensitivity and response threshold of the current judgment. The value is stored in a register within the PLC and is updated after each action is executed.

[0062] Error index It is a predefined three-dimensional error energy function in the system, derived from step three, combining information from spatial deviation, attitude deviation, and height anomaly, and exhibits good stability. A typical example illustrates this: if the current workpiece is a column, and a slight offset in the actual adsorption point causes tilting, its positional error... Only 1.2mm, angular deviation It is 2.3°, but the top height deviation is 2.3°. Reaching 6mm, ultimately leading to The value is 3.1, exceeding the set threshold. Even if adsorption is "completed" at this point, it may still be judged as a failure due to unstable physical structure. (Status marker) This will trigger the control branch judgment in this step.

[0063] To achieve stable control of the system, this step introduces a behavioral inertia function. It is used to calculate the current result in conjunction with historical failure trends, specifically the behavior inertia function. The expression is:

[0064] ;

[0065] in, The adjustment coefficient is set to [value]. This function is used to control the percentage of impact from past failures. Its design principle is: even if the current execution succeeds, if there have been multiple previous failures (such as...), the impact of previous failures will be mitigated. The system should not be immediately judged as "stable," but rather enter a moderate state; conversely, if success occurs continuously, the trend weakens, and the system gradually returns to a high-tolerance state.

[0066] according to The system sets two judgment thresholds in the PLC to determine the value. General empirical values ​​are as follows: , The judgment strategy is as follows:

[0067] when The system enters the normal path and directly proceeds to the processing flow of the next target point;

[0068] when The system determines that there is a tendency for the action to be unstable and executes a retry action;

[0069] when If the system determines that there are consecutive failures or error anomalies, it will record the anomaly and skip the current target workstation.

[0070] During execution, the PLC will call preset control subroutines according to the above classifications. For a normal path, path coordinates will be sent. The robot arm controller is then notified and the next action is initiated. The retry path applies a small disturbance compensation to the original coordinates (e.g., in...). Add ±1mm random items) to overcome action failures caused by unstable clamp adsorption or occasional errors in image detection. If a path is skipped, record the current target number and failure mark, send them to the host computer or MES system as a basis for subsequent quality traceability, and automatically index the next material task.

[0071] In the state of At that time, the historical failure trend volume It will be updated, and the update method uses an exponential decay function to control the gradual reduction of the impact of long-term failures. The expression for the exponential decay function is:

[0072] ;

[0073] in, The attenuation coefficient is typically set to 1. .when At that time, the system will adjust the amplitude by a smaller amount. Correct downwards to prevent it from growing indefinitely. For example, after two consecutive successes, It will drop to 49% of the previous value, restoring the system to the normal state judgment range.

[0074] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A PLC data intelligent analysis and optimization control method, characterized in that, The method comprises the following steps: Collecting an input image, performing edge extraction on the input image to generate a set of edge points, fitting a principal axis direction unit vector based on the set of edge points, calculating the center coordinates of the set of edge points by using a weighting function, mapping the center point coordinates to the workpiece center coordinates through a preset coordinate transformation matrix, and calculating the angle between the principal axis unit vector and the horizontal axis of the workpiece center coordinates to generate a pose angle; Introducing a target path parameter, calculating the initial deviation between the workpiece center coordinates and the pose angle and the preset target path parameter, and correcting the workpiece center coordinates and the pose angle based on the initial deviation and a compensation function, and outputting the corrected workpiece center coordinates and the pose angle, writing the corrected workpiece center coordinates and the pose angle into a control buffer through a PLC register, and calling an internal position control subroutine to transmit the command to a robot controller through the robot controller to drive the robot to execute the action; Calculating the position error and the angle error under the current executed action, fusing the position error and the angle error through a joint error energy function to generate a deviation value, and outputting an action state marker according to the comparison result of the deviation value and a preset deviation threshold; Introducing a historical failure trend quantity, combining the deviation value and the action state marker to construct a behavior inertia function, outputting a trend state value through the behavior inertia function, and guiding the automatic switching of the process logic inside the PLC according to the comparison result of the trend state value and a preset state threshold. 2.The PLC data intelligent analysis and optimization control method of claim 1, wherein, The automatic switching of the process logic inside the PLC according to the comparison result of the trend state value and the preset state threshold specifically comprises: When it is determined to be a normal path, the workpiece center coordinates corresponding to the normal path are sent to a mechanical control arm controller and the next action is started; When it is determined to be a retry path, a small disturbance compensation is applied on the basis of the original workpiece center coordinates; When it is determined to be a failure or an abnormal path, the current action state marker is recorded and the target station is skipped. 3.The PLC data intelligent analysis and optimization control method of claim 1, wherein, After the behavior inertia function is constructed by combining the deviation value and the action state marker and the trend state value is outputted through the behavior inertia function, it further comprises: When the action state marker is a failure, the historical failure trend quantity is updated through an exponential decay function; When the action state marker is a success, the historical failure trend quantity is downwardly corrected. 4.The PLC data intelligent analysis and optimization control method of claim 3, wherein, The parameters of the exponential decay function include the current historical failure trend quantity and a historical decay coefficient.

5. The PLC data intelligent analysis and optimization control method of claim 1, wherein, The parameters of the weighting function include the direction weight corresponding to the edge point, the image coordinates of the edge point, and the number of edge points, and the direction weight corresponding to the edge point is calculated and generated by a unit vector of the edge point in the image gradient direction, a principal axis direction unit vector, and a smoothing coefficient. 6.The PLC data intelligent analysis and optimization control method of claim 1, wherein, The parameters of the compensation function include a position deviation compensation weight coefficient, a historical stacking deviation trend weight factor, and an average deviation residual value observed in a historical path execution process.

7. The PLC data intelligent analysis and optimization control method of claim 1, wherein, The parameters of the joint error energy function include a position error, an angle error, a weighting factor, a height deviation, and a structure stability adjustment term, the weighting factor is determined by the jig type and the workpiece geometry setting, the height deviation is used to assist in judging the abnormality of the invisible structure of the object, and the height deviation is obtained through a laser displacement sensor. 8.The PLC data intelligent analysis and optimization control method of claim 1, wherein, The parameters of the behavior inertia function include an adjustment coefficient, a deviation value, a motion state mark, and a motion execution sequence number, which is used to represent the sequence of the current workpiece in the stacking task. 9.The PLC data intelligent analysis and optimization control method of claim 1, wherein, The historical failure trend quantity is stored in a register inside the PLC and is updated after each motion execution. 10.The PLC data intelligent analysis and optimization control method of claim 1, wherein, The target path parameters include preset workpiece center coordinates and a preset posture angle.

Images