A self-generation method and system of PLC control strategy in industrial field

Abstract

This invention discloses a method and system for automatically generating PLC control strategies in the industrial field, belonging to the field of programmable logic controller (PLC) technology. The method includes: reading physical process parameters and performing normalization processing to generate normalized process feature vectors; performing weighted similarity matching with a pre-built control strategy template library to select the optimal template as the initial control strategy framework; calculating the slope of the start-up speed curve, the slope of the stop-down speed curve, and brake delay parameters; performing dynamic simulation verification on the start-up speed curve slope parameters, automatically iteratively correcting overshoot when it exceeds the limit; filling the verified parameters into a pre-set PLC function block template, generating PLC program code, and sending it to the target PLC. This invention achieves fully automatic generation of the entire process from process parameters to PLC control programs, and has advantages such as smooth start-up and stop processes, accurate differential protection, high parameter reliability, and long-term adaptive operation.

Description

Technical Field

[0001] This invention relates to the field of programmable logic controller (PLC) technology, and more specifically to a method and system for self-generating PLC control strategies in the industrial field. Background Technology

[0002] In engineering practice, there are significant differences in the physical process parameters of each new project. For example, belt conveyors that use programmable logic controllers to implement frequency conversion start-stop control, multi-machine linkage interlocking protection and brake timing control have significant differences in physical process parameters such as belt length, inclination angle, drive power, rated flow and number of linkages. This means that each project implementation requires redesigning the PLC control program according to the specific parameters on site. The workload is large and highly dependent on the personal professional experience of engineers, resulting in long project delivery cycles and high labor costs.

[0003] In existing technologies, frequency conversion starting of belt conveyors typically uses a linear speed setpoint curve. The acceleration remains constant during the speed increase phase, but a step change occurs at the curve's endpoint, generating a significant instantaneous impact on belt tension. Under heavy loads or steep inclines, this can easily lead to speed overshoot, belt oscillation, and even material spillage. Even in projects using S-shaped speed curves, the slope parameter tuning relies heavily on engineers manually setting it based on historical experience from similar projects. When facing projects with unique process parameter combinations, these experience-based parameters often lack sufficient physical basis, making it difficult to guarantee the smoothness of the start-up and shutdown process. Regarding multi-machine linkage differential speed protection, traditional solutions generally use a fixed percentage of the rated speed as a uniform differential speed protection threshold for all adjacent equipment pairs, failing to consider the impact of material transfer point distances and current load on allowable speed differences between different equipment pairs. The fixed threshold is often too high under conditions of small equipment spacing or heavy loads, preventing timely detection of actual differential speed exceeding limits; conversely, it is often too low under conditions of large equipment spacing or light loads, leading to frequent false shutdowns caused by normal speed fluctuations, affecting production continuity. This one-size-fits-all protection strategy cannot adapt to diverse operating conditions and poses a risk of insufficient safety and reliability. Regarding the verification and maintenance of control parameters, current engineering practices typically involve directly sending manually tuned control parameters to the target PLC without an offline pre-verification step. If improper parameter settings lead to abnormal startup, downtime and rework are required, causing project delays and additional costs. Furthermore, factors such as belt wear and aging over long-term operation, changes in idler roller friction characteristics, and seasonal fluctuations in environmental conditions can all cause the actual control response to gradually deviate from the theoretical curve. Existing solutions lack the ability to automatically compensate and correct control parameters without shutting down the system, leading to a continuous accumulation of speed tracking errors over long-term operation. Control accuracy and equipment operational safety decline over time, making it difficult to meet the requirements for reliable operation throughout the entire lifecycle of industrial equipment.

[0004] Therefore, designing a self-generating method and system for PLC control strategies in the industrial field to solve the above problems is of great significance. Summary of the Invention

[0005] To address the problems existing in the background technology, this invention provides a method for self-generating PLC control strategies in the industrial field, comprising the following steps:

[0006] S1: Read the physical process parameters from the industrial equipment process configuration information, perform interval normalization on each parameter, and generate a normalized process feature vector; S2: Perform weighted similarity calculation between the normalized process feature vector and the normalized process feature vector of each template in the pre-built control strategy template library, select the template with the highest similarity, and use the initial value of the control parameters of this template as the initial control strategy framework; S3: Calculate the start-up speed curve slope parameter, stop-up speed curve slope parameter, and brake delay parameter of the drive equipment based on the physical process parameters; S4: Calculate the differential protection threshold of each adjacent equipment pair based on the linkage topology of multiple drive equipment, and generate multi-machine linkage differential protection interlocking logic; S5: Substitute the start-up speed curve slope parameter obtained in step S3 into the drive equipment dynamics simulation. The model undergoes numerical simulation to calculate the speed response overshoot. When the overshoot exceeds the preset maximum allowable overshoot, the slope parameter of the starting speed curve is iteratively corrected until the overshoot meets the constraint or reaches the maximum number of iterations. S6: The slope parameter of the starting speed curve verified or corrected in step S5, the slope parameter of the stopping speed curve obtained in step S3, the brake delay parameter, and the differential protection threshold obtained in step S4 are filled into the preset PLC function block template to generate PLC program code and send it to the target PLC. S7: During the operation phase of the target PLC, the average speed tracking error is periodically calculated in a sliding window manner. When the average error exceeds the preset error threshold, the slope parameter of the starting speed curve is adaptively corrected online, and the correction result is written to the target PLC in real time after being constrained by the parameter safety boundary.

[0007] Further, steps S1 and S2 specifically include the following sub-steps: S11: Read the physical process parameters of the drive equipment from the engineering configuration file, wherein the physical process parameters include at least the belt length. (m), belt inclination angle (°), Rated power of drive inverter (kW), rated material flow rate (t / h), maximum design speed of the belt (m / s) and number of linked equipment ;

[0008] S12: Normalize each physical process parameter read in step S11 according to the following formula, and combine all normalization results in sequence as follows: Normalized process feature vector For use in step S21:

[0009] ;

[0010] In the formula: For the first A normalized component, dimensionless; For the first The measured value of a physical process parameter, with the same dimensions as the parameter; For the first historical sample in the control strategy template library The minimum value of each parameter, with respect to the units of measurement. same; For the first historical sample in the control strategy template library The maximum value of each parameter, with respect to its units and dimensions. same;

[0011] S21: Normalize the process feature vector obtained in step S12 With the control strategy template library, the first Normalized process feature vector of each template Calculate the weighted similarity score using the following formula :

[0012] ;

[0013] In the formula: for and The similarity score is dimensionless. , The larger the value, the closer the process conditions are; for The A normalized component, dimensionless; for The A normalized component, dimensionless; For the first The dimensionless weighting coefficients of the dimension. Pre-set based on engineering experience;

[0014] S22: Traverse all templates in the control strategy template library and execute step S21, then select... The largest template is used to output the initial values ​​of its control parameters to steps S3 and S4 as the initial control strategy framework.

[0015] Further, step S3 specifically includes the following sub-steps: S31: Calculate the slope parameter of the variable frequency start-up speed curve of the drive equipment according to the following formula. :

[0016] ;

[0017] In the formula: The slope parameter (s) of the starting speed curve -1 ); Dimensional coefficient (s) -1 ·m·kW The control strategy template library is determined by least squares regression fitting on historical samples. Rated power (kW) for the drive inverter; The power exponent is dimensionless and is determined by least squares regression fitting. The belt length is in meters. This is the tilt angle correction factor, which is dimensionless and set based on engineering experience. The belt tilt angle is (°). This is a load correction factor, dimensionless, set by engineering experience; This represents the current actual material flow rate (t / h). Rated material flow rate (t / h);

[0018] Step S31 generates the slope parameter of the startup speed curve. ;

[0019] S32: Calculate the shutdown correction factor using the following formula. And the slope parameter of the stopping speed curve :

[0020] ;

[0021] ;

[0022] In the formula: This is a shutdown correction factor, dimensionless. ; This is a dimensionless correction factor for belt length. It is set by engineering experience; The belt length is in meters. Preset reference belt length (m); The slope parameter (s) of the starting speed curve obtained in step S31 -1 ); The slope parameter of the stopping speed curve (s) -1 ), ;

[0023] Step S32 generates the slope parameter of the stopping speed curve. ;

[0024] S33: Calculate the brake delay parameters using the following formula. :

[0025] ;

[0026] In the formula: The brake delay parameter (s) is the delay time after the drive equipment issues a frequency converter stop command to activate the brake. The shortest delay time (s) allowed by the brake hardware; The belt movement time coefficient is dimensionless and determined by regression analysis of historical project data. The belt length is in meters. The maximum design speed of the belt (m / s); This is the load influence coefficient, dimensionless, and set by engineering experience. This represents the current actual material flow rate (t / h). Rated material flow rate (t / h);

[0027] Step S33 generates brake delay parameters This ensures that the brake engages after the frequency converter deceleration is completed.

[0028] Further, step S4 specifically includes the following sub-steps: S41: parse the linkage topology of multiple drive devices and determine the upstream and downstream sequence numbers of each device; sequentially obtain the sequence numbers of each pair of adjacent devices (the first pair). Taiwan and the tower, Rated operating speed (m / s), rated material flow rate (t / h), current actual material flow rate (t / h) and distance between material transfer points (m); The maximum distance between all adjacent device pairs in the statistical system. (m); Step S41 generates parameter sets for each adjacent device pair;

[0029] S42: For each pair of adjacent devices obtained in step S41, calculate the differential protection threshold using the following formula. :

[0030] ;

[0031] In the formula: For the first Differential protection threshold (m / s) for adjacent equipment, when the first The measured speed of the device exceeded the first The measured speed of the equipment exceeded The differential protection is triggered at any time. For the first Rated operating speed of the equipment (m / s); It is the basic tolerance factor, dimensionless, and set by engineering experience; This is the spacing correction factor, dimensionless. It is set by engineering experience; For the first Distance between material transfer points of adjacent equipment (m); This represents the maximum distance (m) between all adjacent device pairs within the system. This is a load correction factor, dimensionless. It is set by engineering experience; For the first Current actual material flow rate of the equipment (t / h); For the first Rated material flow rate of the equipment (t / h);

[0032] Perform step S42 to traverse all adjacent device pairs and generate a differential protection threshold array. And multi-machine linkage differential protection interlocking logic.

[0033] Furthermore, step S5 specifically includes the following sub-steps: S51: using the equivalent mass of the driving device (kg), resistance to movement (N) and the component of the material's gravity along the belt direction (N) is used as a fixed input parameter for the simulation model, with the slope parameter of the starting speed curve obtained in step S3 being the parameter. Determine the time series of the inverter output drive force (N) is used as the excitation input to perform numerical integration simulation of the start-up process of the drive device, and the simulated speed-time series is output. (m / s); Step S51 generates the simulation velocity time series. ;

[0034] S52: The simulated velocity time series obtained in step S51 Calculate the speed response overshoot using the following formula. :

[0035] ;

[0036] In the formula: Speed ​​response overshoot (%) for The maximum instantaneous velocity value (m / s) in the middle; Set the target speed (m / s) for the current time.

[0037] Step S52 generates the velocity response overshoot. ;

[0038] S53: The result obtained in step S52 With the preset maximum allowable overshoot Comparison: If Then, based on the current As the output parameter of the final start-up speed curve; if Then correct it according to the following formula. Then return to step S51 to re-simulate and evaluate until... Or the number of iterations reaches the preset maximum number of iterations. until:

[0039] ;

[0040] In the formula: The slope parameter (s) of the starting speed curve to be corrected in this iteration. -1 ); The step size coefficient is dimensionless and is set based on engineering experience. The overshoot of the velocity response obtained in step S52 (%) The preset maximum allowable overshoot (%);

[0041] When the number of iterations reaches At that time, in each iteration Minimum corresponding Output the value and generate an alarm; step S53 generates the slope parameter of the final start-up speed curve verified by offline simulation. and parameter safety boundaries, while verifying the passed Multiply by the preset lower margin coefficient and upper margin coefficient respectively to obtain and Store the parameters in the safety boundary storage unit.

[0042] Further, step S6 specifically includes the following sub-steps: S61: Retrieve speed curve generation function block templates, brake timing control function block templates, differential interlock protection function block templates, and auxiliary equipment start / stop timing function block templates from the preset PLC function block template library; The slope parameter of the starting speed curve output in step S5 is... The slope parameter of the stopping speed curve obtained in step S3 and brake delay parameters Fill in the corresponding parameter placeholders for the speed curve generation function block template and the brake timing control function block template respectively; and input the differential protection threshold array obtained in step S4. Fill in the corresponding parameter placeholders for the differential interlock protection function block template; after filling, generate PLC program code conforming to the standard ST language format; step S61 generates PLC program code that can be directly downloaded;

[0043] S62: Through the communication connection established with the target PLC, the PLC program code generated in step S61 is sent to the target PLC and the program update is triggered; after receiving the download confirmation signal returned by the target PLC, step S6 is completed.

[0044] Further, step S7 specifically includes the following sub-steps: S71: After the target PLC completes each full start-up or shutdown transition process, at a fixed sampling time interval... (s) Collect the measured velocity sequence of this transition process and velocity curve reference value sequence ( , (The total number of sampling points in this transition process) is used to calculate the speed tracking error of this transition process using the following formula. and recently Mean of sliding window error in the secondary transition process :

[0045] ;

[0046] In the formula: The speed tracking error (m / s) during this transition process; This represents the total number of sampling points during this transition process; For the first Measured velocity values ​​(m / s) at each sampling time; For the first The reference velocity value (m / s) output by the velocity curve generation function block at each sampling time.

[0047] ;

[0048] In the formula: The mean value of the sliding window speed tracking error (m / s); The sliding window size (in times) is a preset value; For the first in the sliding window Velocity tracking error (m / s) during the first transient process;

[0049] Step S71: Generate the mean error of the sliding window. ;

[0050] S72: The result obtained in step S71 Compared with the preset error threshold Comparison: If No corrections will be made; if Then, the slope parameter of the corrected starting speed curve is calculated using the following formula. :

[0051] ;

[0052] In the formula: The corrected starting speed curve slope parameter (s) -1 ); The slope parameter of the current startup speed curve (s) -1 ); The adaptive correction step size coefficient is dimensionless and set by engineering experience; The mean value of the sliding window error obtained in step S71 (m / s); The preset error threshold (m / s) is used.

[0053] right Apply safety boundary constraints to make ,in and The allowed lower and upper limits of the parameter safety boundary storage unit are stored in step S5; after constraint The speed curve generation function block is written to the target PLC in real time through the communication interface to complete the online parameter update without downtime; step S72 realizes the online adaptive correction of the slope parameter of the start-up speed curve.

[0054] Furthermore, after the PLC program code is issued in step S6 and the target PLC is debugged on-site, step S8 is also included: normalizing the process feature vector of this project. Step S5 finally outputs the slope parameter of the start-up speed curve. The slope parameter of the stopping speed curve obtained in step S3 and brake delay parameters and the differential protection threshold array obtained in step S4 The control strategy template library is written as a new template record through the writing interface of the control strategy template library, and the historical maximum and minimum values ​​of each physical process parameter in the control strategy template library are updated at the same time; step S8 enables the control strategy template library to continuously accumulate historical samples, thereby improving the template matching accuracy of subsequent projects.

[0055] This invention also provides a PLC control strategy self-generation system for industrial applications, comprising: a process parameter input and feature vector extraction module, which reads the physical process parameters of the drive equipment and generates normalized process feature vectors; a control strategy template library, which stores the normalized process feature vectors of historical projects and the corresponding initial values ​​of control parameters, and is equipped with a template writing interface; a template matching module, which is connected to the process parameter input and feature vector extraction module and the control strategy template library, selects the optimal matching template through weighted similarity calculation, and outputs the initial values ​​of the control parameters of the template to the speed curve parameter calculation module and the topology analysis and differential speed threshold calculation module; and a speed curve parameter calculation module, which calculates the slope parameter of the starting speed curve based on the physical process parameters. Stop speed curve slope parameter and brake delay parameters The topology parsing and differential speed threshold calculation module parses the interconnected topology of multiple drive devices, calculates the differential speed protection threshold for each pair of adjacent devices, and generates a differential speed protection threshold array. and multi-machine linkage differential protection interlocking logic; simulation verification and parameter self-calibration module, for the above Perform dynamic simulation evaluation and iterative correction when overshoot exceeds the limit. , will satisfy the constraints Output to the PLC code generation module, and and Write parameters to the safety boundary storage unit; PLC code generation module, , , and Enter the preset PLC function block template to generate PLC program code and send it to the target PLC; a bypass parameter update interface is provided, which connects to the online monitoring and adaptive correction module; the online monitoring and adaptive correction module periodically calculates the sliding window mean of the speed tracking error, and calculates the corrected value when the mean error exceeds a preset threshold. After being constrained by safety boundaries, the data is written to the target PLC in real time; the parameter safety boundary storage unit stores the data. and Boundary constraints are provided to the online monitoring and adaptive correction module.

[0056] In a preferred embodiment, the process parameter input and feature vector extraction module includes a parameter reading unit and a normalization processing unit; the normalization processing unit outputs the normalized process feature vector to the template matching module and the velocity curve parameter calculation module, and outputs the original physical process parameters to the velocity curve parameter calculation module and the topology analysis and differential speed threshold calculation module; the simulation verification and parameter self-calibration module includes a dynamic simulation unit, an overshoot evaluation unit, and a parameter correction unit; the parameter correction unit sends the corrected parameters to the dynamic simulation unit when the overshoot exceeds the limit. This will trigger a resimulation, which will eventually occur when the constraints are met. The output is sent to the PLC code generation module and the safety boundary value is written to the parameter safety boundary storage unit; the online monitoring and adaptive correction module includes a data acquisition unit, an error calculation unit, and a parameter correction output unit; the parameter correction output unit reads the boundary value from the parameter safety boundary storage unit and performs... After applying constraints, the data is written to the target PLC, and the parameter records in the PLC code generation module are updated synchronously through the bypass parameter update interface.

[0057] The beneficial effects achieved by this invention are as follows:

[0058] First, this invention constructs a control strategy template library and introduces a normalized process feature vector and a weighted similarity matching mechanism. This enables the system to automatically retrieve the control strategy framework that is closest to the current engineering process conditions from historical projects. This framework serves as the initial basis for subsequent fine-tuning of parameters, eliminating the uncertainty of engineers relying on personal experience to estimate parameters when facing unfamiliar application scenarios. This transforms the generation process of control strategies from relying on personal experience accumulation to a systematic retrieval based on engineering data, significantly reducing the difficulty of initial project configuration. Furthermore, the matching accuracy of subsequent projects is continuously improved as the sample size of the template library expands.

[0059] Secondly, based on the physical process parameters of the drive equipment, this invention comprehensively considers factors that fundamentally affect the startup process, such as belt length, inclination angle, drive power, and load rate. It automatically calculates the slope parameters of the variable frequency drive startup S-curve, the slope parameters of the shutdown speed curve, and the brake delay parameters using analytical formulas. The S-curve maintains continuous acceleration change throughout the entire start-stop transition process, avoiding the step acceleration shock generated by traditional linear curves at the moment of speed switching. This effectively reduces peak belt tension and speed overshoot during startup and prevents reverse rotation or material buildup at transfer points caused by excessive deceleration of large-angle, long-distance belts during shutdown. This significantly improves the overall start-stop smoothness of the drive equipment and provides significantly better protection for the belt's mechanical structure than empirically tuned linear control schemes.

[0060] Third, this invention adaptively derives the differential speed protection threshold for each pair of adjacent devices based on the material transfer point spacing and the current actual load of multiple interconnected devices, quantitatively incorporating the two core physical factors affecting material transfer safety—spacing and load—into the threshold calculation. This mechanism ensures that the differential speed protection threshold is precisely matched with the actual safety tolerance of each operating condition: under small-spacing, heavy-load conditions, the threshold is moderately tightened due to the combined effect of the spacing correction term and the load correction term, thereby significantly reducing the probability of protection maloperation caused by normal speed fluctuations; under large-spacing, light-load conditions, the threshold is moderately widened with the positive effect of the spacing correction term, avoiding misjudging tolerable speed differences as exceeding limits, thereby significantly reducing the probability of protection failure. Compared with fixed empirical thresholds, the adaptive differential speed protection threshold simultaneously considers protection sensitivity and reliability, overcoming the fundamental limitation of fixed thresholds in adapting to the differences in multiple operating conditions.

[0061] Fourth, this invention introduces an offline simulation verification and self-calibration process based on the dynamic model of the drive equipment before the parameters are sent to the target PLC. An iterative proportional correction rule automatically adjusts the slope parameter of the start-up speed curve to a range that meets the overshoot constraint, and stores the verified parameter safety boundary for subsequent online correction. After the system is put into operation, a sliding window error statistics mechanism continuously monitors the trend of speed tracking error. When the mean error exceeds a threshold, the slope parameter of the start-up speed curve is automatically fine-tuned based on the degree of error exceedance, and the correction result is constrained by the safety boundary determined in the offline verification stage. The offline verification process ensures that the parameters are pre-evaluated before being sent, effectively avoiding the risk of downtime and rework due to on-site start-up overshoot caused by parameter deviation. The online adaptive correction process automatically compensates for operating condition drift without shutting down the system, maintaining the speed tracking accuracy at a safe level during long-term operation and significantly improving the reliability throughout the entire lifecycle. Attached Figure Description

[0062] Figure 1 The graph shows a comparison of the speed response stability index between Example 1 and Comparative Example 1. (a) is a comparison of the start-up speed-time curves of Example 1 and Comparative Example 1 under the same process parameters. (b) is a biaxial bar graph comparing the speed overshoot and belt tension peak values ​​of Example 13 and Comparative Example 1 under three sets of parameters.

[0063] Figure 2 This is a comparison chart of the differential protection accuracy of Examples 1-3 and Comparative Example 2. The false start rate and failure to start rate of each method are displayed in four groups of four bar charts according to four working conditions: small spacing heavy load, small spacing light load, large spacing heavy load, and large spacing light load.

[0064] Figure 3 The different initial k in Examples 1-3 and the comparative examples accThe velocity response overshoot σ is shown in the iteration convergence curve, where (a) is the curve of σ changing with the number of iterations in Example 1, (b) is the curve of σ changing with the number of iterations in Example 2, and (c) is the curve of σ changing with the number of iterations in Example 3.

[0065] Figure 4 A flowchart of a method for self-generating PLC control strategies in the industrial field according to the present invention. Detailed Implementation

[0066] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0067] This invention provides a method for automatically generating PLC control strategies in the industrial field. Taking the physical process parameters of industrial drive equipment as input, it automatically generates PLC program code conforming to the IEC61131-3 standard and distributes it to the target PLC through steps such as feature extraction and template matching, nonlinear curve parameter calculation, multi-machine differential speed protection threshold derivation, offline simulation verification and self-calibration, and online adaptive correction. Simultaneously, it continuously optimizes control parameters during the operation phase. (See reference...) Figure 4 The following instructions will be provided in step-by-step order.

[0068] Step S1 is the process feature vector extraction step, which includes two sub-steps: step S11 and step S12. In step S11, the system reads the physical process parameters of the drive equipment from the engineering configuration file. These physical process parameters include at least the belt length. (m), belt inclination angle (°), Rated power of drive inverter (kW), rated material flow rate (t / h), maximum design speed of the belt (m / s) and number of linked equipment The parameters mentioned above describe the drive equipment from three dimensions: geometric dimensions, process load, and system configuration. Together, they constitute the complete input conditions that determine the selection of control strategies and parameter calculations. The engineering configuration file can be in XML or JSON format, and is entered by engineers during the project configuration phase using process data provided by site surveys or design drawings.

[0069] In step S12, each physical process parameter read in step S11 is subjected to interval normalization according to equation (1), and all normalization results are combined sequentially as follows: Normalized process feature vector This is for use in step S21.

[0070] ;

[0071] In equation (1), For the first A normalized component, dimensionless; For the first The measured value of a physical process parameter, with the same dimensions as the parameter; For the first historical sample in the control strategy template library The minimum value of each parameter, with respect to the units of measurement. same; For the first historical sample in the control strategy template library The maximum value of each parameter, with respect to its units and dimensions. Same. Equation (1) is the minimum-maximum normalization method, which maps parameters of different physical dimensions and magnitudes to a uniform interval. This eliminates the interference of dimensional and order-of-magnitude differences on subsequent similarity calculations, ensuring that parameters with significant differences, such as belt length and the number of linked devices, are treated equivalently in vector space.

[0072] Step S2 is the template matching step, which includes two sub-steps: step S21 and step S22. In step S21, the normalized process feature vector obtained in step S12 is... With the control strategy template library Normalized process feature vector of each template Calculate the weighted similarity score according to formula (2). .

[0073] ;

[0074] In equation (2), for and The similarity score is dimensionless. , The larger the value, the closer the process conditions are; for The A normalized component, dimensionless; for The A normalized component, dimensionless; For the first The dimensionless weighting coefficients of the dimension. Pre-set based on engineering experience; Indicates all The summation is performed on each feature dimension. Equation (2) defines similarity in the form of the reciprocal of the weighted Euclidean distance: the denominator contains... The geometric distance between two feature vectors in the weighted space is measured; the smaller the distance, the better. As the distance approaches infinity, the closer it gets to 1. Approaching 0, thus monotonically mapping spatial distance to Intervals offer intuitive physical meaning and low computational cost. Weight coefficients for each dimension. This reflects the relative importance of the corresponding process parameters to the selection of the control strategy. For example, belt length and tilt angle have a significant impact on the control curve parameters, so their corresponding weights should be set relatively high; inverter power and rated flow are secondary; the number of linked equipment has a significant impact on the multi-machine interlocking logic structure, and can also be given a high weight. The specific values ​​of the weight coefficients can be obtained through sensitivity analysis of historical project data, or they can be preset directly by experienced engineers based on domain knowledge, and then optimized through cross-validation after the system has accumulated sufficient samples.

[0075] In step S22, all templates in the control strategy template library are traversed and step S21 is executed to select... The largest template is used, and its initial control parameter values ​​are output to steps S3 and S4 as the initial control strategy framework. The initial control parameter values ​​of the selected template provide a reasonable initial estimate for subsequent fine-grained calculations. Even if subsequent formula calculations correct for them, the historical experience information contained in the initial values ​​still has reference value for judging the safe range of parameters.

[0076] Step S3 is the nonlinear curve parameter calculation step, which includes three sub-steps: S31, S32, and S33. The core of frequency converter start-stop control for belt conveyors lies in the design of the speed setpoint curve. A reasonable speed curve must ensure smooth start-up, avoid belt impact and material spillage, and also ensure safe shutdown, preventing belt reversal or material accumulation at transfer points due to excessively rapid shutdown of heavy-duty, steeply inclined belts. In engineering practice, the S-curve (modified Sigmoid function) is commonly used as the speed setpoint curve, and its mathematical description is... ,in Set the speed (m / s) for the target. The inflection point of the curve (in seconds) is typically taken as 50% of the total startup time. For the slope parameter (s) -1 The advantages of an S-curve are: acceleration approaches zero at the beginning and end of the start-up process, increases and then decreases in the middle stage, and the acceleration changes continuously throughout the process. This results in a much smaller impact on the belt's mechanical structure compared to a linear curve, and a smaller peak impact on belt tension for the same start-up duration. (Slope parameter) The larger the value, the steeper the curve, and the faster the start-up. The smaller the value, the flatter the curve, and the slower the start-up. Therefore, It is the core parameter that determines the speed and smoothness of the startup process.

[0077] In step S31, the slope parameter of the S-shaped speed curve of the variable frequency drive device is calculated according to formula (3). .

[0078] ;

[0079] In equation (3), The slope parameter (s) of the starting speed curve -1 ); Dimensional coefficient (s) -1 ·m·kW The control strategy template library was determined by least squares regression fitting of historical samples. Rated power (kW) for the drive inverter; The power exponent is dimensionless and determined by least squares regression fitting. The value is usually less than 1, reflecting the diminishing marginal effect of power on starting acceleration capability; The belt length is in meters. This is the tilt angle correction factor, which is dimensionless and set based on engineering experience. The recommended value range is 0.5 to 0.8. The belt tilt angle is (°). This is a load correction factor, dimensionless, set based on engineering experience, with a recommended value range of 0.2 to 0.4. This represents the current actual material flow rate (t / h). Rated material flow rate (t / h).

[0080] The construction logic of equation (3) is as follows: molecule This expresses the physical law that the higher the power of the frequency converter, the stronger the starting drive force and the higher the allowable starting speed; in the denominator... The longer the belt, the greater the total inertia of the system, the weaker the acceleration capability under the same driving force, and the lower the allowable starting rate; factor The larger the tilt angle, the greater the component of gravity along the belt direction, and the greater the starting resistance, requiring a smoother acceleration process; factor The heavier the load, the greater the system inertia and the more significant the impact on belt tension, requiring a further reduction in the acceleration rate. After logarithmically linearizing equation (3), it can be transformed into... As the dependent variable, with For a linear regression problem with independent variables, the least squares method is used to fit and determine the independent variables using historical project data that has already been debugged. and The value of can be determined by periodically refitting as the sample size of the template library increases to improve accuracy.

[0081] In step S32, the shutdown correction coefficient is calculated according to formula (4). Then calculate the slope parameter of the stopping speed curve according to formula (5). .

[0082] ;

[0083] ;

[0084] In equation (4), This is a shutdown correction factor, dimensionless. ; This is a dimensionless correction factor for belt length. Based on engineering experience, the recommended value range is 0.3 to 0.6. The belt length is in meters. For the preset reference belt length (m), it is recommended to take 100m. In formula (5), The slope parameter (s) of the starting speed curve obtained in step S31 -1 ); The slope parameter of the stopping speed curve (s) -1 ),Depend on It can be known That is, the shutdown curve is never steeper than the startup curve. Equation (4) expresses that as the belt length increases, the shutdown curve needs to be gentler than the startup curve to prevent elastic wave propagation caused by a sudden drop in speed when the long belt is braked, which could lead to local overload or belt misalignment. equal hour Values ,when Much larger hour Significantly increased, The corresponding reduction in speed and the extended downtime are consistent with the operating procedures for slow stops commonly required in long-distance belt conveyor engineering practice.

[0085] In step S33, the brake delay parameter is calculated according to equation (6). .

[0086] ;

[0087] In equation (6), The brake delay parameter (s) is the delay time after the drive equipment issues a frequency converter stop command to activate the brake. The minimum allowable delay time (s) for the brake hardware is provided by the brake manufacturer. The belt movement time coefficient is dimensionless and determined by regression analysis of historical project data. The belt length is in meters. The maximum design speed of the belt (m / s); This is the load influence coefficient, dimensionless, and set by engineering experience. This represents the current actual material flow rate (t / h). The rated material flow rate is (t / h). In equation (6), The order of magnitude of the time required for the belt to travel the entire distance at its maximum speed, multiplied by a coefficient. The estimated time required for the frequency converter to decelerate from rated speed is then calculated; load correction item. Considering the factors that the kinetic energy of the leather is greater under heavy loads and that deceleration takes longer; and The maximum value is taken to ensure that the delay is not lower than the lower limit of the brake hardware safety, so as to prevent the brake from engaging before the frequency converter has completed deceleration, which would cause mechanical shock and damage to the brake and reducer.

[0088] Step S4 is the differential speed protection threshold derivation step, which includes two sub-steps, S41 and S42. When multiple belt conveyors are connected in series, if the speed of the upstream belt conveyor is higher than that of the downstream belt conveyor by a certain value, material will accumulate or be thrown out at the transfer point, causing spillage, blockage, or even equipment damage. In engineering, the differential speed protection threshold is usually used as the trigger criterion for emergency stop of the linkage. However, traditional fixed empirical thresholds cannot adapt to different equipment spacing and load conditions, resulting in missed detections under heavy load conditions with small spacing or false alarms under light load conditions with large spacing. This method dynamically derives the differential speed protection threshold for each pair of adjacent equipment to solve the above problems.

[0089] In step S41, the linkage topology of multiple drive devices is parsed, the upstream and downstream sequence numbers of each device are determined, and each pair of adjacent devices (the first pair) is sequentially obtained. Taiwan and the tower, Rated operating speed (m / s), rated material flow rate (t / h), current actual material flow rate (t / h) and distance between material transfer points (m), and calculate the maximum distance between all adjacent device pairs in the system. (m), for use in the normalization calculation of step S42.

[0090] In step S42, for each pair of adjacent devices obtained in step S41, the differential protection threshold is calculated according to formula (7). .

[0091] ;

[0092] In equation (7), For the first Differential protection threshold (m / s) for adjacent equipment, when the first The measured speed of the device exceeded the first The measured speed of the equipment exceeded The differential protection is triggered at any time. For the first Rated operating speed of the equipment (m / s); This is the basic tolerance factor, dimensionless, set by engineering experience, with a recommended value range of 0.015 to 0.020; This is the spacing correction factor, dimensionless. Based on engineering experience, the recommended value range is 0.008 to 0.015. For the first Distance between material transfer points of adjacent equipment (m); This represents the maximum distance (m) between all adjacent device pairs within the system. This is a load correction factor, dimensionless. Based on engineering experience, the recommended value range is 0.005 to 0.010. For the first Current actual material flow rate of the equipment (t / h); For the first Rated material flow rate (t / h) of the equipment.

[0093] The construction logic of equation (7) is as follows: basic tolerance term Ensures a minimum tolerance margin under all operating conditions to prevent accidental activation of protection due to normal speed fluctuations; Spacing correction item. This reflects the physical laws that the greater the distance between transfer points, the longer the material's flight time in the air, and the greater the permissible speed difference. As a normalized spacing ratio, it avoids the absolute spacing value from directly affecting the threshold and causing dimensional coupling.

[0094] Load Correction Item Adjust the differential protection threshold according to the current actual load rate. The increase is correspondingly due to the improvement. With a smaller value, this item only makes minor adjustments to the threshold without significantly changing the main threshold.

[0095] Perform step S42 to traverse all adjacent device pairs and generate a differential protection threshold array. And multi-machine linkage differential protection interlocking logic.

[0096] Step S5 is the simulation verification and self-calibration step, which includes three sub-steps: S51, S52, and S53. If the control parameters are directly sent to the PLC without verification after calculation, significant parameter deviations may lead to problems such as start-up overshoot and belt oscillation, requiring downtime for repairs and causing project delays. Therefore, this method introduces an offline simulation verification step before parameter sending, using a simplified dynamic model of the driving equipment to pre-evaluate and automatically correct the generated parameters.

[0097] The dynamic equations for the start-up process of the drive equipment are derived based on Newton's second law. All rotational and translational masses of the belt conveyor system (including the motor rotor, reducer, drive roller, belt, and material) are uniformly converted into the system's equivalent mass. (kg), establish the following first-order linear ordinary differential equation: ;

[0098] In the above equation, The system's equivalent mass (kg) is calculated from the belt mass, material mass, and the moment of inertia of the drive system. The formula for calculating the moment of inertia is: ( The moment of inertia is (kg·m²). (where the radius of the drive roller is in meters). This is the derivative of linear velocity with respect to time, i.e., acceleration (m / s²). The output drive force (N) of the frequency converter is derived from the S-shaped speed setpoint curve through the frequency converter torque control loop. The shape of the curve changing over time and The determined S-shaped curve is consistent; The motion resistance (N) includes belt friction resistance, idler roller resistance, and air resistance, etc. It is approximately a constant in the simulation and can be estimated based on the belt conveyor design specifications or actual no-load start-up data. This represents the component of the material's gravity along the belt direction (N). ,in The total mass (kg) of the material on the conveyor belt. The acceleration due to gravity (m / s²) The belt inclination angle (°) is for the upward-moving belt conveyor. As resistance, the downward belt conveyor As an auxiliary force.

[0099] In step S51, with , and As a fixed input parameter, the result obtained in step S3 Definite Using the time series as excitation input, numerical integration simulation is performed on the startup process of the driven device, and the simulated speed time series is output. (m / s). Numerical integration employs the fourth-order Runge-Kutta method to discretize and solve the aforementioned first-order ordinary differential equation. A step size of 0.1s to 0.5s is recommended to balance computational accuracy and efficiency. The fourth-order Runge-Kutta method is a classic numerical method for solving initial value problems of ordinary differential equations. In each integration step, the slope is estimated four times and weighted averaged. Under the same step size, its truncation error is on the order of the fourth power of the step size, resulting in computational accuracy far exceeding that of the Euler method. Furthermore, it does not require increasing the step frequency, making it suitable for implementation in the configuration software or host computer of industrial control systems.

[0100] In step S52, the result obtained in step S51 is... Calculate the speed response overshoot using equation (8). .

[0101] ;

[0102] In equation (8), Speed ​​response overshoot (%) for The maximum instantaneous velocity value (m / s) in the middle; Set the current target speed (m / s). Speed ​​overshoot is a key indicator for measuring the smoothness of the startup process. Excessive overshoot means that the belt speed momentarily exceeds the set value during startup, which has adverse effects on belt tension, material condition, and mechanical structure. In engineering, it is usually required that... No more than 5%.

[0103] In step S53, the result obtained in step S52 is... With the preset maximum allowable overshoot Compare. If Then, based on the current As the output parameter of the final start-up speed curve; if Then, it should be corrected according to equation (9). Then return to step S51 to re-simulate and evaluate until... Or the number of iterations reaches the preset maximum number of iterations. until.

[0104] ;

[0105] In equation (9), The slope parameter (s) of the starting speed curve to be corrected in this iteration. -1 ); To correct the step size coefficient, which is dimensionless and set based on engineering experience, a value range of 0.1 to 0.3 is recommended. The current overshoot (%) obtained in step S52; The maximum allowable overshoot (%) is preset. Equation (9) is the proportional correction rule: when the overshoot exceeds the upper limit, it is adjusted proportionally to the overshoot. Double reduction This flattens the start-up curve to reduce overshoot. The larger the value, the greater the correction magnitude per step and the fewer the convergence iterations, but it may cause the parameters to oscillate back and forth. A smaller value results in more stable convergence, but may require more iterations; a value of 0.15-0.25 is recommended. When the number of iterations reaches... At that time, in each iteration Minimum corresponding The value is output, and an alarm is sent to the engineer, prompting them to manually verify the rationality of the parameters under this operating condition. Step S53 will also verify the passed values. Multiply by the preset lower limit margin coefficient to obtain Multiply by the preset upper limit margin coefficient to obtain The parameters are stored in the parameter safety boundary storage unit for use in step S7 during online adaptive correction.

[0106] Step S6 is the PLC code generation step, which includes two sub-steps: step S61 and step S62. In step S61, the speed curve generation function block template, brake timing control function block template, differential interlock protection function block template, and auxiliary equipment start / stop timing function block template are retrieved from the preset PLC function block template library. The slope parameter of the starting speed curve output in step S5 is used. The slope parameter of the stopping speed curve obtained in step S3 and brake delay parameters Fill in the corresponding parameter placeholders for the speed curve generation function block template and the brake timing control function block template respectively, and input the differential protection threshold array obtained in step S4. Enter the corresponding parameter placeholders for the differential interlock protection function block template. After completion, PLC program code conforming to the IEC 61131-3 standard Structured Text (ST) language format will be generated. IEC 61131-3 is an internationally recognized standard for programmable logic controller (PLC) programming languages. Its ST language is similar to a high-level programming language, with strong readability and good portability. The generated code can be deployed and run on mainstream industrial PLC platforms without requiring engineers to manually rewrite the program logic for different platforms.

[0107] In step S62, the PLC program code generated in step S61 is sent to the target PLC through a communication connection (such as Ethernet or industrial bus) established with the target PLC and the program update is triggered. After receiving the download confirmation signal returned by the target PLC, step S6 is completed.

[0108] Step S7 is the online adaptive correction step, which includes two sub-steps: S71 and S72. After the system is put into operation, factors such as changes in belt tension over time, idler wear, and seasonal changes in ambient temperature can cause deviations between the actual control response and the theoretical speed curve. If not corrected, the tracking error will continue to accumulate during long-term operation, affecting control accuracy and equipment safety. To ensure that the control parameters continuously match the current state of the equipment, this method designs an online adaptive correction mechanism based on sliding window error statistics, which corrects the error without stopping the machine. Perform automatic correction.

[0109] In step S71, after the target PLC completes each full start-up or shutdown transition process, a fixed sampling time interval is used. (s) Collect the measured velocity sequence of this transition process and velocity curve reference value sequence ( , (where the total number of sampling points in this transition process is 10), calculate the speed tracking error of this transition process according to formula (10). Then calculate the nearest according to formula (11). Mean of sliding window error in the secondary transition process .

[0110] ;

[0111] In equation (10), The speed tracking error (m / s) during this transition process; This represents the total number of sampling points during this transition process; For the first Measured velocity values ​​(m / s) at each sampling time; For the first The reference velocity value (m / s) output by the velocity curve generation function block at each sampling time. Equation (10) uses the mean absolute error (MAE) to measure the tracking quality of a single transition process. The mean of the absolute value of the error can avoid the evaluation inaccuracy caused by the mutual cancellation of positive and negative errors. In equation (11), The mean value of the sliding window speed tracking error (m / s); This is the sliding window size (times), a preset value, and it is recommended to use 10 to 30 times; For the first in the sliding window The speed tracking error (m / s) of the transition process. Equation (11) takes the moving average over multiple cycles to filter out single measurement noise and reflect the trend change of the system's tracking performance. Correction is only triggered when the average error exceeds the limit after multiple cycles, thereby avoiding frequent corrections caused by occasional interference.

[0112] In step S72, the result obtained in step S71 is... Compared with the preset error threshold Comparison: If No corrections will be made; if Then, the slope parameter of the corrected starting speed curve is calculated according to equation (12). .

[0113] ;

[0114] In equation (12), The corrected starting speed curve slope parameter (s) -1 ); The slope parameter of the current startup speed curve (s) -1 ); The adaptive correction step size coefficient is dimensionless and set based on engineering experience; a recommended value range is 0.01 to 0.05. The mean value of the sliding window error obtained in step S71 (m / s); The preset error threshold (m / s) is used. Equation (12) and Equation (9) have similar structures, both being proportional correction rules. When the mean error exceeds the threshold, the error is corrected proportionally. Double reduction This flattens the curve to improve tracking performance. The smaller values ​​(0.01 to 0.05) are used to ensure a smooth online correction and avoid disturbances caused by sudden parameter changes.

[0115] right Apply safety boundary constraints to make ,in and The allowed lower and upper limits of the parameter safety boundary storage unit are stored in step S5. After constraint... The speed curve of the target PLC is written in real time through the communication interface to generate function blocks, enabling online parameter updates without downtime. Safety boundary constraints are a necessary guarantee for online correction: without constraints, the adaptive algorithm may correct parameters to physically unreasonable ranges under abnormal data interference, leading to start-up failure or control anomalies; with the safety boundaries determined in the offline simulation stage, online correction always adjusts within the verified parameter range, ensuring the operational safety of the system.

[0116] Step S8 is the template library update step. After the PLC program code is issued in step S6 and the target PLC is debugged on-site, the normalized process feature vector of this project will be updated. The final output of step S5 The result of step S3 and and the differential protection threshold array obtained in step S4 The control strategy template library is written as a new template record through its write interface, and the historical maximum and minimum values ​​of each physical process parameter in the control strategy template library are updated simultaneously to dynamically maintain the normalization used in step S12. and Step S8 allows the control strategy template library to continuously accumulate historical samples. As the sample size increases, the similarity retrieval candidate range in step S21 becomes wider, and the regression coefficients in step S31... and The fitting accuracy continued to improve, and the template matching accuracy and parameter calculation accuracy of subsequent projects also improved accordingly, realizing the system's self-learning and continuous optimization.

[0117] This invention also provides an industrial PLC control strategy self-generation system for implementing the above method. The system includes a process parameter input and feature vector extraction module, a control strategy template library, a template matching module, a speed curve parameter calculation module, a topology analysis and differential speed threshold calculation module, a simulation verification and parameter self-calibration module, a PLC code generation module, an online monitoring and adaptive correction module, and a parameter safety boundary storage unit. The modules are connected by signals to form a complete closed-loop workflow from parameter input to strategy generation, verification and correction, code output, and online optimization.

[0118] The process parameter input and feature vector extraction module includes a parameter reading unit and a normalization processing unit. The parameter reading unit reads the physical process parameters through the engineering configuration file interface and transmits them unidirectionally to the normalization processing unit as parameter signals. The normalization processing unit has built-in maximum and minimum values ​​of each physical process parameter in historical samples. After performing interval normalization on each parameter, it outputs the normalized process feature vectors unidirectionally to the template matching module and the speed curve parameter calculation module, respectively. Simultaneously, it outputs the original physical process parameters unidirectionally to the speed curve parameter calculation module and the topology analysis and differential speed threshold calculation module. After the template writing in step S8 is completed, the normalization processing unit reads the updated historical sample maximum and minimum values ​​from the control strategy template library to update its built-in normalization range, maintaining consistency between the normalization calculation and the sample range in the template library.

[0119] The control strategy template library stores the normalized process feature vectors and corresponding initial values ​​of control parameters for historical projects in a structured manner, and includes a template writing interface for writing new project data. Each historical record contains three fields: normalized feature vector, control parameter value, and basic project information, facilitating efficient retrieval by the template matching module. The template matching module is signal-connected to the process parameter input and feature vector extraction module, receiving the normalized process feature vector; it is also signal-connected to the control strategy template library, reading the normalized process feature vectors and weight coefficients of all templates; it selects the optimal matching template through weighted similarity calculation, and outputs the initial values ​​of the control parameters of this template unidirectionally to the speed curve parameter calculation module and the topology analysis and differential speed threshold calculation module, respectively, as the initial estimation basis for the parameter calculations of the two modules. The speed curve parameter calculation module is signal-connected to the process parameter input and feature vector extraction module, receiving physical process parameters; it is also signal-connected to the template matching module, receiving the initial values ​​of control parameters; and it calculates the parameters based on the physical process parameters. , and ;Will The parameter signal is output unidirectionally to the simulation verification and parameter self-calibration module. and The parameter signal is output unidirectionally to the PLC code generation module. The topology parsing and differential speed threshold calculation module is signal-connected to the process parameter input and feature vector extraction module to receive the linkage topology and process parameters of each device; it is also signal-connected to the template matching module to receive the initial values ​​of control parameters; it parses the linkage topology and calculates the differential speed protection threshold for each pair of adjacent devices, generating a differential speed protection threshold array. The system includes multi-machine linkage differential speed protection interlocking logic, which outputs parameter signals unidirectionally to the PLC code generation module. The simulation verification and parameter self-calibration module includes a dynamic simulation unit, an overshoot evaluation unit, and a parameter correction unit. The dynamic simulation unit is signal-connected to the speed curve parameter calculation module and receives... After performing numerical integration simulation, the simulated velocity time series is unidirectionally transmitted to the overshoot evaluation unit. The overshoot evaluation unit calculates the overshoot. Later The comparison result is unidirectionally transmitted to the parameter correction unit. The parameter correction unit... Send the corrected data to the time-to-time dynamics simulation unit The re-simulated control signal, in Or the number of iterations reaches The time will eventually come Unidirectional output to the PLC code generation module, and and One-way write parameter safety boundary storage unit. The PLC code generation module is signal-connected to the simulation verification and parameter self-calibration module, receiving the verified parameters. ; Connects to the speed curve parameter calculation module for signal reception. and It is connected to the topology analysis and differential threshold calculation module to receive the differential protection threshold array. The system includes interlocking logic; the parameters are filled into a preset PLC function block template to generate PLC program code; the PLC program code is sent to the target PLC via the downlink communication interface, and the download confirmation signal from the target PLC is received via the uplink communication interface; a bypass parameter update interface is provided, which is connected to the online monitoring and adaptive correction module signal for receiving the corrected parameters online. To maintain consistency between the recorded parameters inside the module and the actual operating parameters of the PLC.

[0120] The online monitoring and adaptive correction module includes a data acquisition unit, an error calculation unit, and a parameter correction output unit. The data acquisition unit is unidirectionally connected to the speed measurement signal output and speed curve reference value output of the target PLC, respectively, acquiring the measured speed value and the reference speed value at a fixed sampling period. At the end of each transition process, it unidirectionally transmits the speed data sequence to the error calculation unit. The error calculation unit then sequentially calculates... and ,exist When and Unidirectional transmission to the parameter correction output unit. The parameter correction output unit reads from the parameter safety boundary storage unit. and Calculate and constrain Then, the speed curve of the function block parameter register is unidirectionally written to the target PLC via the parameter writing interface using a write command signal, achieving online parameter updates without downtime; simultaneously, The parameter update is unidirectionally transmitted to the PLC code generation module via the bypass parameter update interface, maintaining consistency between the two parameter records. The parameter safety boundary storage unit is bidirectionally connected to the simulation verification and parameter self-calibration module, receiving data written during the offline verification phase. and It is also connected to the online monitoring and adaptive correction module via a unidirectional signal, outputting to it. and It is used to apply boundary constraints to the online correction results. The parameter safety boundary storage unit acts as a bridge for transferring safety constraint information between the offline verification stage and the online operation stage, keeping the safety limits of the two stages synchronized and preventing online corrections from exceeding the parameter safety range verified offline.

[0121] Example 1: This example uses a long-distance belt conveyor project with two tandem belt conveyors in a mine as an application example. The industrial PLC control strategy self-generation system in this example consists of the following components: The parameter reading unit in the process parameter input and feature vector extraction module reads physical process parameters through the project configuration file interface. The normalization processing unit performs interval normalization and outputs normalized process feature vectors unidirectionally to the template matching module and the speed curve parameter calculation module, while simultaneously outputting the original physical process parameters unidirectionally to the speed curve parameter calculation module and the topology analysis and differential speed threshold calculation module. The control strategy template library stores the normalized process feature vectors and corresponding initial control parameter values ​​of historical projects in a structured manner and has a template writing interface. After receiving the normalized process feature vectors, the template matching module selects the optimal template through weighted similarity calculation and outputs the initial control parameter values ​​of the template to the speed curve parameter calculation module and the topology analysis and differential speed threshold calculation module, respectively, as the initial estimation basis for parameter calculation in both modules. The speed curve parameter calculation module calculates based on the physical process parameters. , and Afterwards, The unidirectional output is sent to the simulation verification and parameter self-calibration module. and Unidirectional output to the PLC code generation module. The topology analysis and differential speed threshold calculation module analyzes the linkage topology and calculates... After the array and multi-machine linkage differential speed protection interlocking logic, the output is unidirectionally sent to the PLC code generation module. The simulation verification and parameter self-calibration module includes a dynamic simulation unit, an overshoot evaluation unit, and a parameter correction unit, which work together to complete the task. Offline iterative verification will satisfy the constraints. Unidirectional output to the PLC code generation module, and and The parameters are written to the safety boundary storage unit. The PLC code generation module fills the verified control parameters into the preset function block template, generates PLC program code conforming to the IEC61131-3 standard, and sends it to the target PLC through the downlink communication interface; it has a bypass parameter update interface, which is connected to the signal of the online monitoring and adaptive correction module for receiving the corrected parameters online. To maintain consistent parameter records, the online monitoring and adaptive correction module includes a data acquisition unit, an error calculation unit, and a parameter correction output unit. The parameter correction output unit reads the safety boundary value from the parameter safety boundary storage unit, applies constraints to the calculation results, and then... The parameters are written to the target PLC via the parameter writing interface, and simultaneously updated in the PLC code generation module via the bypass parameter update interface, ensuring consistency between the two. The parameter safety boundary storage unit is written by the simulation verification and parameter self-calibration module and outputs unidirectionally to the online monitoring and adaptive correction module, serving as a bridge for the transmission of safety constraint information between the offline verification stage and the online operation stage.

[0122] The implementation process of the PLC control strategy self-generation method in the industrial field of this embodiment includes steps S1-S8, as follows: In step S11, the parameter reading unit of the process parameter input and feature vector extraction module reads the physical process parameters of belt conveyor 1 and belt conveyor 2 from the engineering configuration file. Belt conveyor 1 is the upstream belt conveyor, and the belt length is... m, belt inclination angle Rated power of drive frequency converter kW, rated material flow rate t / h, maximum design speed of the belt m / s, number of linked devices Distance between material transfer points of the two machines m. Belt conveyor 2 is the downstream belt conveyor, belt length... m, , kW, t / h, m / s, current actual material flow rate .

[0123] In step S12, the normalization processing unit performs interval normalization on each physical process parameter, using the minimum and maximum values ​​of each parameter in the historical samples of the control strategy template library as endpoints to generate a 6-dimensional normalized process feature vector. .

[0124] In step S21, the template matching module will The weighted similarity score is calculated using Formula 2 with all templates in the control strategy template library. Weight coefficient vector This demonstrates that belt length and tilt angle have the greatest impact on the selection of control strategy.

[0125] In step S22, the historical engineering case corresponding to the template with the highest similarity is retrieved, with a belt length of 750m and an inclination angle. Power 560kW, similarity score The initial values ​​of the control parameters of the template are output to the speed curve parameter calculation module and the topology analysis and differential threshold calculation module, respectively, as the initial control strategy framework.

[0126] In step S31, the speed curve parameter calculation module calculates the slope parameter of the starting speed curve of belt conveyor 1 according to formula 3. The regression coefficient is then used. , (Determined by least squares fitting of historical samples), using engineering experience coefficients. , Current load rate Substituting into formula 3, we get s -1 The transition time for the 10% to 90% speed of the corresponding S-shaped speed curve is approximately 720 seconds, with the inflection point located at the midpoint of the transition process.

[0127] In step S32, the belt length correction factor is taken. Preset reference belt length m, calculate the shutdown correction factor according to formula 4. Then calculate the slope parameter of the stopping speed curve according to Formula 5. s -1 The shutdown transition time is approximately five times the startup transition time, meeting the engineering safety requirements for slow stopping of long-distance belt conveyors with large inclination angles.

[0128] In step S33, the belt movement time coefficient is taken. Load Influence Coefficient Shortest hardware latency s, calculate the brake delay parameters To ensure that the brake can be engaged only after the frequency converter has completed the deceleration curve, mechanical impact damage to the brake and the reducer must be prevented.

[0129] In step S41, the topology parsing and differential speed threshold calculation module parses the series linkage topology of the two belt conveyors and obtains the parameters of the first pair of adjacent devices: rated speed. m / s, distance between material transfer points m, based on the system reference maximum spacing m is used as the normalization benchmark, and the current actual flow rate. .

[0130] In step S42, the basic tolerance coefficient is taken. Spacing correction factor Load correction factor Calculate the differential protection threshold according to Formula 7. When the measured speed of the upstream conveyor belt 1 exceeds the measured speed of the downstream conveyor belt 2 by more than 0.111 m / s, an emergency stop is triggered.

[0131] In step S51, the dynamic simulation unit of the simulation verification and parameter self-calibration module uses the system's equivalent mass... kg, resistance to motion kN, gravitational component kN is a fixed parameter, obtained from step S31. s -1 Determined S-shaped driving force time series To provide stimulation, a fourth-order Runge-Kutta method is used for numerical integration simulation, with an integration step size of [missing information]. s, output simulation velocity time series In step S52, the overshoot evaluation unit calculates the velocity response overshoot according to Formula 8. peak and set target speed m / s calculated In step S53, the parameter correction unit will... With preset Comparison, because No need for iterative correction, directly using s -1 The slope parameter of the final start-up speed curve is output to the PLC code generation module, and s -1 and s -1 Write parameters to the safety boundary storage unit.

[0132] In step S61, the PLC code generation module retrieves four types of function block templates from the preset PLC function block template library, and... s -1 , s -1 , Fill in the parameter placeholders for the speed curve generation function block template and the brake timing control function block template, respectively, with 's'. The speed is filled into the differential interlock protection function block template. After filling, PLC program code conforming to the IEC61131-3 standard ST language format is generated. In step S62, the code is sent to the target PLC in the field via Ethernet. After receiving the download confirmation signal, step S6 is completed.

[0133] In step S71, the data acquisition unit of the online monitoring and adaptive correction module collects data at fixed sampling time intervals after the target PLC completes each full start-stop transition process. The system collects the measured velocity sequence and the velocity curve reference value sequence, and the error calculation unit calculates the velocity tracking error of this transient process according to Formula 10. Then, according to formula (11) The mean error of the second sliding window calculation In step S72, the parameter correction output unit adjusts the output to a preset error threshold. m / s and adaptive correction step size coefficient Comparison and correction: when Then, calculate the corrected result according to Formula 12. ,through and After constraining the safety boundaries, the speed curve of the target PLC is written to the function block parameter register in real time through the parameter writing interface, realizing online parameter updates without downtime. In step S8, after on-site commissioning is completed, the normalized process feature vector of this project is... , s -1 , s -1 , s and m / s is written as a new template into the control strategy template library, and the historical maximum and minimum values ​​of each parameter are updated at the same time, so that the control strategy template library can continuously accumulate samples and improve the matching accuracy of subsequent projects.

[0134] Example 2: This example uses a port bulk material conveying system with three belt conveyors connected in series as the application object. The parts that are the same as in Example 1 will not be described again. It demonstrates the automatic derivation of the differential speed threshold of the three machines and the influence of higher flow parameters on the control curve.

[0135] The main physical and technological parameters are as follows: The rated parameters of belt conveyor 1 are... m、 , kW t / h, m / s, the rated parameters of belt conveyor 2 are m、 , kW, the rated parameters of the belt conveyor 3 are m、 , kW; Distance between material transfer points of belt conveyor 1 and belt conveyor 2 m, the distance between belt conveyor 2 and belt conveyor 3 m, system reference maximum spacing m; Current actual traffic .

[0136] Step S2, template matching stage, using historical case examples of belt length 480m and inclination angle... The three-machine linkage is the optimal matching template, and the similarity score is... .

[0137] Step S31: Calculate the slope parameter of the starting speed curve of belt conveyor 1. s -1 (Pick , , , , The corresponding transition time for the S-shaped speed curve from 10% to 90% is approximately 511 seconds. Step S32 calculates the shutdown correction coefficient. Stopping slope parameter s -1 Step S33: Calculate the brake delay parameters. s.

[0138] Step S42 calculates the differential speed protection threshold for the first pair of adjacent equipment, belt conveyor 1 and belt conveyor 2. m / s (2.92% of rated speed); calculate the differential speed protection threshold for the second pair of adjacent equipment, belt conveyors 2 and 3. m / s. Generate multi-machine linkage interlocking logic containing two sets of adaptive differential protection thresholds.

[0139] Step S53 Verification result: No iterative corrections are needed. The system ultimately sends complete PLC program code to the target PLC, containing start / stop control for three belt conveyors, differential interlock protection, and timing of auxiliary equipment, allowing for online correction of the step size. Safety boundary s -1 , s -1 .

[0140] Example 3: This example uses two short-distance light-load belt conveyors connected in series in the raw material storage system of a steel enterprise as the object, to demonstrate the automatic generation results of control parameters under the parameter range of short belt, small inclination angle and low power. The parts that are the same as in Example 1 will not be described again.

[0141] Main physical and process parameters: The parameters for belt conveyor 1 are as follows: m、 , kW t / h, m / s, the parameters of belt conveyor 2 are: m、 , kW; Distance between material transfer points of the two machines m, system reference maximum spacing m; Current actual traffic .

[0142] Step S2 Template matching stage, similarity score Historical case studies were selected, with a belt length of 130m and an inclination angle. The initial framework is based on the dual-machine linkage.

[0143] Step S31 Calculation s -1Using the same regression coefficient and empirical coefficient, the transition time for the S-shaped velocity curve from 10% to 90% is approximately 287 seconds. Step S32 calculation , s -1 Step S33 Calculation s.

[0144] Step S42: Calculate the differential protection threshold. m / s. Step S53 Simulation Verification No iterations required. Step size adjusted online. Safety boundary s -1 , s -1 This embodiment covers At 100200m, Within the parameter range of 50-100kW.

[0145] Comparative Example 1 uses a traditional manual experience-based tuning method: the engineer directly sets a fixed linear acceleration time based on experience, without using the automatic calculation formula for the S-shaped speed curve parameters in step S3 of this invention, without performing the offline simulation verification and self-calibration in step S5, without performing the online adaptive correction in step S7, and with the differential protection threshold fixed at 2.0% of the rated speed, without performing the adaptive derivation in step S4. The remaining process parameters are the same as in each embodiment. The linear acceleration times set by the engineer based on experience under each parameter condition are as follows: Under the parameter conditions of Example 1 s, under the parameter conditions of Example 2 s, under the parameter conditions of Example 3 s. The corresponding fixed differential protection thresholds are as follows: under the parameters of Example 1 m / s, under the parameters of Example 2 m / s, under the parameters of Example 3 m / s.

[0146] Comparative Example 2 does not perform the adaptive differential speed threshold derivation of the present invention in step S4. Instead, it uniformly applies a fixed differential speed protection threshold to all adjacent devices. Steps S1 to S3 and steps S5 to S8 are the same as in each embodiment.

[0147] Comparative Example 3 skips steps S1 to S4 and directly jumps to step S6 to generate and distribute PLC code, skipping the simulation verification and parameter self-calibration in step S5 and not performing the online adaptive correction in step S7. The rest is the same as in Example 1.

[0148] Experimental Example 1 uses the S-shaped velocity curves obtained from Examples 1 to 3. Step S3 automatically calculates the parameters and compares them with the method in Comparative Example 1. The simulation uses the fourth-order Runge-Kutta numerical integration method described in step S51, with an integration step size of... s, the dynamic parameters are taken from the corresponding values ​​of each embodiment.

[0149] Evaluation metrics include speed response overshoot. (Percentage, measuring start-up smoothness), peak belt tension (kN, derived from the dynamic equation, reflecting the degree of impact on the belt's mechanical structure) and the time to reach a stable speed. (s, defined as velocity entering) of (Time required for stabilization). Overshoot Calculated using Formula 8, Driven by the simulation process Peak value is calculated. From simulated velocity sequence Analysis yielded the following results. Experimental results are as follows: Figure 1 As shown, Figure 1 In the figure, (a) shows a comparison of the start-up speed-time curves of Example 1 and Comparative Example 1, with the vertical axis representing the belt speed. (m / s), with time on the horizontal axis. (s), in the figure, the S-shaped curve response of Example 1 is represented by a solid line, and the linear curve response of Comparative Example 1 is represented by a dashed line, and labeled with... Limit line. Figure 1 In the table, (b) shows all embodiments and comparative examples under various parameter conditions. Value and A bar chart comparing values, with the left vertical axis representing... (%), with the right vertical axis being (kN), the examples in the figure are marked with blue-green columns, the comparative examples are marked with red columns, and are labeled with horizontal dashed lines. Limit.

[0150] Depend on Figure 1 As can be observed in (a), the S-shaped velocity curve response of Example 1 shows that the acceleration approaches zero in the initial and final stages of startup, increases and then decreases in the middle stage, and the velocity curve is smooth and continuous throughout the process, finally stabilizing at approximately 742 s. Around m / s Significantly lower The limit. The linear velocity curve of Comparative Example 1 is at... At time s, the acceleration undergoes a step change, causing the velocity to instantaneously exceed the set value. Exceeding It meets engineering safety limits and reaches a stable speed in a shorter time, but the control quality is poorer. Figure 1 In (b), it can be observed that Examples 1 to 3 The figures all did not exceed 3.2%; while Comparative Example 1, under the parameters of Example 1, showed a figure of less than 3.2%. Exceeding the limit by 70%, corresponding to kN, which is 168% of that in Example 1. The above comparative conclusions indicate that the S-shaped velocity curve parameters are automatically calculated in step S3 of this invention. It can accurately quantify the combined impact of belt length, tilt angle, power, and load on the starting rate, ensuring that the generated S-curve meets the overshoot constraint under any combination of parameters. Compared with linear curves set manually based on experience, it has a significant advantage in protecting the belt structure. The continuous change in acceleration of the S-curve avoids the step acceleration impact generated by the linear curve at the moment of start-stop, thereby reducing the peak belt tension and speed overshoot.

[0151] Experiment Example 2: This experiment uses the Monte Carlo simulation method. m、 m、 200 sets of operating condition parameters were randomly sampled uniformly within the range. The protection accuracy of the adaptive differential speed threshold method used in Examples 1 to 3 and the fixed 2% rated speed threshold method used in Comparative Example 2 under different operating conditions was compared. The 200 sets of operating conditions were divided into four subcategories according to equipment spacing and load size: small spacing heavy load ( m and 48 groups in total), small-pitch light load ( m and 46 groups in total), large spacing heavy load ( m and 52 groups in total), large spacing and light load ( m and (A total of 54 groups).

[0152] For each set of operating conditions, 400 random speed disturbance events were simulated, including 200 actual differential speed limit exceedance events that should trigger an emergency stop, and 200 potential false triggering events caused by normal speed fluctuations. The false trigger rate (the probability of a normal fluctuation being erroneously stopped) and the failure to trigger rate (the probability of an actual speed limit exceedance not being detected) were statistically analyzed for each subclass. A high false trigger rate indicates excessive protection sensitivity, while a high failure to trigger rate indicates insufficient protection reliability; both should be kept as low as possible.

[0153] Experimental results are as follows Figure 2 As shown, Figure 2 The chart is a four-cluster bar chart. The vertical axis represents the protection false trip rate or failure rate (%), and the horizontal axis represents the four operating condition subclasses. The chart uses four different colors to represent the false trip rate of the adaptive method (dark blue), the failure rate of the adaptive method (dark green), the false trip rate of the fixed 2% method (red), and the failure rate of the fixed 2% method (orange).

[0154] Depend on Figure 2 It can be observed that in the small-pitch heavy-load subclass, the false trigger rate of the fixed 2% threshold method is as high as 22.5%, while that of the adaptive method is only 3.2%, a reduction of 85.8%. This is because: under small-pitch heavy-load conditions, the equipment spacing is short, the material travel time is short, and the speed fluctuation during normal operation is relatively small. The fixed 2% threshold is too low, resulting in an excessively high false trigger rate for normal speed fluctuations; while the adaptive method, through formula 7... Item spacing correction will The threshold corresponding to smaller, closer spacing conditions should be appropriately tightened, through The threshold is further tightened under heavy load conditions, ensuring that the threshold accurately matches the safety margin of the current operating condition, thereby significantly reducing the false start rate. In the large-spacing light-load subclass, the rejection rate of the fixed 2% threshold method is as high as 18.3%, while the adaptive method is only 2.4%, a reduction of 86.9%. The reason is that under the large-spacing light-load condition, the equipment spacing is large and the material travel time is long, allowing for a larger speed difference without causing material accumulation risk. The fixed 2% threshold is too low, causing a large number of tolerable speed differences to be identified as exceeding the limit, while the adaptive method, through... The positive correction of the term increases the upper limit of the threshold for large-spacing conditions and reduces the failure rate. Based on 200 sets of data, the adaptive method significantly outperforms the fixed 2% method in both overall false alarm rate and failure rate, demonstrating the effectiveness of the adaptive differential speed threshold derivation technique in step S4. Formula 7 quantitatively incorporates the physical factors affecting material handling safety—equipment spacing and load—into the threshold calculation, ensuring that the differential speed protection threshold matches the actual safety tolerance of each operating condition. This approach balances protection sensitivity and reliability, overcoming the fundamental limitation of fixed thresholds in adapting to the differences in multiple operating conditions.

[0155] Experiment 3 verifies the convergence of parameter iteration in step S5, which involves simulation verification and self-calibration. Five initial parameter sets are configured for each of Examples 1 to 3. The values ​​cover three categories: smaller than the nominal value, and larger than the nominal value. To establish a convergence criterion, the step size coefficient is adjusted. Maximum number of iterations Next, perform iterative correction according to formula (9) and record the results for each group. The process of change with the number of iterations.

[0156] Experimental results are as follows Figure 3 As shown, Figure 3 It contains three sub-graphs, (a), (b), and (c) corresponding to Examples 1, 2, and 3, respectively. The vertical axis of each sub-graph represents the velocity response overshoot. (%), the horizontal axis represents the number of iterations, and each curve in the graph represents a set of initial... The values ​​are represented by five curves, each with a different shade of color to distinguish between three categories of initial values: smaller (lighter), nominal value (medium), and larger (darker). The convergence criterion is marked with a horizontal dashed line.

[0157] Depend on Figure 3 It can be observed that there are 3 initial curves in each subgraph. The curve with a smaller or nominal value satisfies the condition from the 0th iteration. , The initial values ​​are between 0.4% and 3.2%, requiring no iterative correction. This directly proves that the nominal value automatically calculated by Formula 3 in step S31 based on the physical process parameters is correct. The values ​​are highly accurate, and the corresponding S-shaped velocity curves satisfy the overshoot constraint from the outset, even with relatively large initial values. ~9.5%, and convergence to the target value within a maximum of 6 iterations is achieved under the parameter conditions of all three embodiments. During the iteration process Monotonically decreasing without oscillation, indicating that the proportional correction rule in Formula 9 is combined with... The adjusted step size exhibits stable convergence characteristics. The maximum number of convergence iterations (6) is much smaller than the preset value. This demonstrates that the simulation verification and self-calibration closed loop in step S5 have high computational efficiency within typical operating conditions. By comparing the convergence behavior of different initial value sets, the effectiveness of the formula calculation in step S3 as an initial value estimate is verified: the correct initial value allows step S5 to pass verification without iteration, and even if the initial value deviates, the iterative correction mechanism in step S5 can reliably converge the parameters to the range that satisfies the constraints.

[0158] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An industrial field PLC control strategy self-generation method, characterized in that, Includes the following steps: S1: Read the physical process parameters from the industrial equipment process configuration information, perform interval normalization on each parameter, and generate a normalized process feature vector; S2: Perform a weighted similarity calculation between the normalized process feature vector and the normalized process feature vector of each template in the pre-built control strategy template library, select the template with the highest similarity, and use the initial value of the control parameters of the template as the initial control strategy framework. S3: Based on the physical process parameters, calculate the slope parameters of the start-up speed curve, the slope parameters of the stop speed curve, and the brake delay parameters of the drive equipment; S4: Based on the linkage topology of multiple drive devices, calculate the differential protection threshold of each adjacent device pair and generate multi-machine linkage differential protection interlocking logic; S5: Substitute the slope parameter of the starting speed curve obtained in step S3 into the dynamic simulation model of the drive equipment for numerical simulation and calculate the speed response overshoot. When the overshoot exceeds the preset maximum allowable overshoot, iteratively correct the slope parameter of the starting speed curve until the overshoot meets the constraint or reaches the maximum number of iterations. S6: Fill the preset PLC function block template with the slope parameters of the start-up speed curve verified or corrected in step S5, the slope parameters of the stop-up speed curve obtained in step S3, the brake delay parameters, and the differential protection threshold obtained in step S4, generate PLC program code and send it to the target PLC. S7: During the operation phase of the target PLC, the average speed tracking error is periodically calculated using a sliding window method; When the average error exceeds the preset error threshold, the slope parameter of the start-up speed curve is adaptively corrected online, and the correction result is written to the target PLC in real time after being constrained by the parameter safety boundary.

2. The method of claim 1, wherein, Steps S1 and S2 specifically include the following sub-steps: S11: Read the physical process parameters of the drive equipment from the project configuration file, the physical process parameters including at least the belt length. (m), belt inclination angle (°), Rated power of drive inverter (kW), rated material flow rate (t / h), maximum design speed of the belt (m / s) and number of linked equipment ; S12: Normalize each physical process parameter read in step S11 according to the following formula, and combine all normalization results in sequence as follows: Normalized process feature vector For use in step S21: ； In the formula: For the first A normalized component, dimensionless; For the first The measured value of a physical process parameter, with the same dimensions as the parameter; For the first historical sample in the control strategy template library The minimum value of each parameter, with respect to the units of measurement. same; For the first historical sample in the control strategy template library The maximum value of each parameter, with respect to its units and dimensions. same; S21: normalizing the process feature vector obtained in step S12 with the normalized process feature vector of the i-th template in the control strategy template library The weighted similarity score is calculated as follows: :​ ； In the formula: for and The similarity score is dimensionless. , The larger the value, the closer the process conditions are; for The A normalized component, dimensionless; for The A normalized component, dimensionless; For the first The dimensionless weighting coefficients of the dimension. Pre-set based on engineering experience; S22: Traverse all templates in the control strategy template library and execute step S21, then select... The largest template is used to output the initial values ​​of its control parameters to steps S3 and S4 as the initial control strategy framework.

3. The method of claim 2, wherein, Step S3 specifically includes the following sub-steps: S31: Calculate the slope parameter of the variable frequency start-up speed curve of the drive equipment according to the following formula. : ； In the formula: The slope parameter (s) of the starting speed curve -1 ); Dimensional coefficient (s) -1 ·m·kW -β The control strategy template library is determined by least squares regression fitting on historical samples. Rated power (kW) for the drive inverter; The power exponent is dimensionless and is determined by least squares regression fitting. The belt length is in meters. This is the tilt angle correction factor, which is dimensionless and set based on engineering experience. The belt tilt angle is (°). This is a load correction factor, dimensionless, set by engineering experience; This represents the current actual material flow rate (t / h). Rated material flow rate (t / h); Step S31 generates the start speed profile slope parameter ; S32: Calculate the stop correction coefficient according to the following formula and the stop speed curve slope parameter : ； ； In the formula: This is a shutdown correction factor, dimensionless. ; This is a dimensionless belt length correction factor. It is set by engineering experience; The belt length is in meters. Preset reference belt length (m); The slope parameter (s) of the starting speed curve obtained in step S31 -1 ); The slope parameter of the stopping speed curve (s) -1 ), ; Step S32 generates a shutdown speed curve slope parameter ; S33: Calculate the brake delay parameter according to the following formula : ； In the formula: The brake delay parameter (s) is the delay time after the drive equipment issues a frequency converter stop command to activate the brake. The shortest delay time (s) allowed by the brake hardware; The belt movement time coefficient is dimensionless and determined by regression analysis of historical project data. The belt length is in meters. The maximum design speed of the belt (m / s); This is the load influence coefficient, dimensionless, and set by engineering experience. This represents the current actual material flow rate (t / h). Rated material flow rate (t / h); Step S33 generates brake delay parameters This ensures that the brake engages after the frequency converter deceleration is completed.

4. The method of claim 3, wherein, Step S4 specifically includes the following sub-steps: S41: Analyze the linkage topology of multiple drive devices and determine the upstream and downstream sequence numbers of each device; sequentially obtain each pair of adjacent devices, i.e., the [number]th [device]. Taiwan and the tower, , rated operating speed (m / s), rated material flow rate (t / h), current actual material flow rate (t / h) and distance between material transfer points (m); Maximum value of the distance between all pairs of adjacent devices in the statistical system (m); step S41 generates a parameter set for each pair of adjacent devices; S42: For each pair of adjacent devices resulting from step S41, the differential protection threshold is calculated as follows : ； In the formula: For the first Differential protection threshold (m / s) for adjacent equipment, when the first The measured speed of the device exceeded the first The measured speed of the equipment exceeded The differential protection is triggered at any time. For the first Rated operating speed of the equipment (m / s); It is the basic tolerance factor, dimensionless, and set by engineering experience; This is the spacing correction factor, dimensionless. It is set by engineering experience; For the first Distance between material transfer points of adjacent equipment (m); This represents the maximum distance (m) between all adjacent device pairs within the system. This is a load correction factor, dimensionless. It is set by engineering experience; For the first Current actual material flow rate of the equipment (t / h); For the first Rated material flow rate of the equipment (t / h); Perform step S42 to traverse all adjacent device pairs and generate a differential protection threshold array. And multi-machine linkage differential protection interlocking logic.

5. The method of claim 4, wherein, Step S5 specifically includes the following sub-steps: S51: Based on the equivalent mass of the driving device (kg), resistance to movement (N) and the component of the material's gravity along the belt direction (N) is used as a fixed input parameter for the simulation model, with the slope parameter of the starting speed curve obtained in step S3 being the parameter. Determine the time series of the inverter output drive force (N) is used as the excitation input to perform numerical integration simulation of the start-up process of the drive device, and the simulated speed-time series is output. (m / s); Step S51 generates the simulation velocity time series. ; S52: Calculate the speed response overshoot from the speed time series obtained in step S51 The speed response overshoot is calculated as follows : ； wherein: is the speed response overshoot (%); is the maximum instantaneous speed value (m / s) in is the current set target speed (m / s);​ Step S52 generates a speed response overshoot amount ; S53: The result obtained in step S52 With the preset maximum allowable overshoot Comparison: If Then, based on the current As the output parameter of the final start-up speed curve; if Then correct it according to the following formula. Then return to step S51 to re-simulate and evaluate until... Or the number of iterations reaches the preset maximum number of iterations. until: ； In the formula: The slope parameter (s) of the starting speed curve to be corrected in this iteration. -1 ); The step size coefficient is dimensionless and is set based on engineering experience. The overshoot of the velocity response obtained in step S52 (%). The preset maximum allowable overshoot (%); When the number of iterations reaches At that time, in each iteration Minimum corresponding Output the value and generate an alarm; step S53 generates the slope parameter of the final start-up speed curve verified by offline simulation. and parameter safety boundaries, while verifying the passed Multiply by the preset lower margin coefficient and upper margin coefficient respectively to obtain and Store the parameters in the safety boundary storage unit.

6. The method of claim 5, wherein, Step S6 specifically includes the following sub-steps: S61: Retrieve speed curve generation function block templates, brake timing control function block templates, differential interlock protection function block templates, and auxiliary equipment start / stop timing function block templates from the preset PLC function block template library; Input the slope parameter of the start speed curve output in step S5. The slope parameter of the stopping speed curve obtained in step S3 and brake delay parameters Fill in the corresponding parameter placeholders for the speed curve generation function block template and the brake timing control function block template respectively; The differential protection threshold array obtained in step S4 Fill in the corresponding parameter placeholders for the differential interlock protection function block template; after filling, generate PLC program code conforming to the standard ST language format; step S61 generates PLC program code that can be directly downloaded; S62: Through the communication connection established with the target PLC, the PLC program code generated in step S61 is sent to the target PLC and the program update is triggered; after receiving the download confirmation signal returned by the target PLC, step S6 is completed.

7. The method of claim 6, wherein, Step S7 specifically includes the following sub-steps: S71: After the target PLC completes each full start-up or shutdown transition process, at a fixed sampling time interval... (s) Collect the measured velocity sequence of this transition process and velocity curve reference value sequence ( , (The total number of sampling points in this transition process) is used to calculate the speed tracking error of this transition process using the following formula. and recently Mean of sliding window error in the secondary transition process : ； In the formula: The speed tracking error (m / s) during this transition process; This represents the total number of sampling points during this transition process; For the first Measured velocity values ​​(m / s) at each sampling time; For the first The reference velocity value (m / s) output by the velocity curve generation function block at each sampling time. ； In the formula: The mean value of the sliding window speed tracking error (m / s); The sliding window size (in times) is a preset value; For the first in the sliding window Velocity tracking error (m / s) during the first transient process; Step S71 generates a sliding window error mean ; S72: The result obtained in step S71 Compared with the preset error threshold Comparison: If No corrections will be made; if Then, the slope parameter of the corrected starting speed curve is calculated using the following formula. : ； In the formula: The corrected starting speed curve slope parameter (s) -1 ); The slope parameter of the current startup speed curve (s) -1 ); The adaptive correction step size coefficient is dimensionless and set by engineering experience; The mean value of the sliding window error obtained in step S71 (m / s); The preset error threshold (m / s) is used. right Apply safety boundary constraints to make ,in and The allowed lower and upper limits of the parameter safety boundary storage unit are stored in step S5; after constraint The speed curve generation function block is written to the target PLC in real time through the communication interface to complete the online parameter update without downtime; step S72 realizes the online adaptive correction of the slope parameter of the start-up speed curve.

8. The method of claim 7, wherein, After the PLC program code is issued in step S6 and the target PLC is debugged on-site, step S8 is also included: Normalizing the process feature vector of this project. Step S5 finally outputs the slope parameter of the start-up speed curve. The slope parameter of the stopping speed curve obtained in step S3 and brake delay parameters and the differential protection threshold array obtained in step S4 The control strategy template library is written as a new template record through the write interface of the control strategy template library, and the historical maximum and minimum values ​​of each physical process parameter in the control strategy template library are updated at the same time.

9. An industrial field PLC control strategy self-generation system for implementing the method of claim 8, characterized by, include: The process parameter input and feature vector extraction module reads the physical process parameters of the driving equipment and generates normalized process feature vectors. A control strategy template library stores the normalized process feature vectors of historical projects and the corresponding initial values ​​of control parameters, and has a template writing interface. The template matching module is connected to the process parameter input and feature vector extraction module and the control strategy template library. It selects the optimal matching template through weighted similarity calculation and outputs the initial value of the control parameter of the template to the speed curve parameter calculation module and the topology analysis and differential speed threshold calculation module. a speed curve parameter calculation module, which calculates a start speed curve slope parameter, a stop speed curve slope parameter and a brake delay parameter according to physical process parameters ;​​ The topology parsing and differential speed threshold calculation module parses the interconnected topology of multiple drive devices, calculates the differential speed protection threshold for each pair of adjacent devices, and generates a differential speed protection threshold array. and multi-machine linkage differential protection interlocking logic; The simulation verification and parameter self-calibration module is used for the following: Perform dynamic simulation evaluation and iterative correction when overshoot exceeds the limit. , will satisfy the constraints Output to the PLC code generation module, and and Write parameters to the safety boundary storage unit; PLC code generation module, , , and Fill in the preset PLC function block template, generate PLC program code, and send it to the target PLC; It is equipped with a bypass parameter update interface, which is connected to the online monitoring and adaptive correction module; The online monitoring and adaptive correction module periodically calculates the sliding window mean of the speed tracking error, and calculates the corrected value when the mean error exceeds a preset threshold. After being constrained by safety boundaries, the data is written to the target PLC in real time. a parameter safety boundary storage unit storing the and providing boundary constraints to the online monitoring and adaptive correction module.

10. The system according to claim 9, characterized in that: The process parameter input and feature vector extraction module includes a parameter reading unit and a normalization processing unit; the normalization processing unit outputs the normalized process feature vector to the template matching module and the speed curve parameter calculation module, and outputs the original physical process parameters to the speed curve parameter calculation module and the topology analysis and differential speed threshold calculation module; The simulation verification and parameter self-calibration module includes a dynamics simulation unit, an overshoot evaluation unit, and a parameter correction unit; the parameter correction unit sends a corrected parameter to the dynamics simulation unit when the overshoot exceeds the limit. This will trigger a resimulation, which will eventually occur when the constraints are met. The output is sent to the PLC code generation module and the safety boundary value is written to the parameter safety boundary storage unit; The online monitoring and adaptive correction module includes a data acquisition unit, an error calculation unit, and a parameter correction output unit; the parameter correction output unit reads boundary values ​​from the parameter safety boundary storage unit and performs... After applying constraints, the data is written to the target PLC, and the parameter records in the PLC code generation module are updated synchronously through the bypass parameter update interface.

Images