Transformer substation terminal strip fastening tool control method and system based on torque feedback

Abstract

The invention relates to the technical field of motion control, and discloses a transformer substation terminal strip fastening tool control method and system based on torque feedback, which are applied to a double-tool-bit fastening tool, and double shafts corresponding to double tool bits are independently driven. The method comprises the following steps: acquiring tightening torque corresponding to different thread types and fastening task requirements to formulate a corresponding terminal fastening strength standard, and training to obtain a multi-layer sensor capable of analyzing the progress of a fastening task by combining a deep learning strategy in a testing process and utilizing change information fed back by the torque based on the terminal fastening strength standard; the method comprises the following steps: constructing a motion constraint of double-shaft fastening of a double-tool-bit fastening tool based on a terminal strip fastening strength standard, evaluating the collaboration of double shafts based on the motion constraint, and further realizing the optimal adjustment of the speed and controlling the operation of the double-tool-bit fastening tool by using a multi-target optimization algorithm based on multi-population coevolution and an S-shaped acceleration and deceleration speed instruction. Therefore, the working efficiency of the fastening tool is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of motion control, and in particular to a method and system for controlling a transformer substation terminal block fastening tool based on torque feedback. Background Art

[0002] In order to ensure the safe operation of the power system, maintenance work on substations is crucial. Through the time statistics of each link of the substation secondary system pre-test, the terminal block tightening work takes the longest time. However, in specific applications, the existing tightening tools lack unified terminal tightening force standards and torque feedback, making it impossible to adapt to different thread types and evaluate the tightening progress, which greatly reduces work efficiency.

[0003] It can be seen that how to improve the working efficiency of existing terminal block fastening tools has become a technical problem that needs to be solved urgently by those skilled in the art. Summary of the invention

[0004] The invention provides a method and system for controlling a transformer substation terminal block fastening tool based on torque feedback, which solves the problem of how to improve the working efficiency of the existing terminal block fastening tool.

[0005] In order to solve the above technical problems, the first aspect of the present invention provides a control method for a substation terminal block tightening tool based on torque feedback, which is applied to a double-blade tightening tool, in which the double axes corresponding to the double blades in the double-blade tightening tool are independently driven; the method comprises:

[0006] Obtain and filter historical torque data of various thread diameters in the substation terminal block in several tightening tasks to establish the tightening force standards for the terminal block under various tightening tasks and various thread diameters;

[0007] Controlling the double-tip fastening tool to fasten the target thread according to the terminal block fastening force standard to obtain torque change data, and using the torque change data and the target thread as sample data to train a neural network model to obtain a multi-layer perceptron for identifying the progress of the fastening task;

[0008] Based on the terminal block tightening force standard, the motion constraint between the two axes of the double-blade tightening tool is constructed and solved to obtain the rotation speed of the two axes;

[0009] Acquire the target tightening task and the real-time torque data when the double-tip tightening tool performs the target tightening task, and input the real-time torque data into the multi-layer perceptron for processing to obtain the real-time tightening task progress;

[0010] The motion constraint is evaluated through the real-time tightening task progress, and the double cutter head in the double cutter head tightening tool is controlled to run at the dual-axis rotation speed or improved speed according to the evaluation result, so as to achieve the tightening task of the substation terminal block; the improved speed is an S-shaped acceleration and deceleration optimized by a multi-objective optimization algorithm based on multi-population co-evolution.

[0011] As one of the preferred solutions, an electric module is provided at the connection between each cutter head and the corresponding tool handle in the double-cutter head fastening tool; the electric module includes a motor so that the corresponding cutter head is independently driven by its own motor; the electric module is also used to feed back the torque data of the corresponding cutter head.

[0012] As one of the preferred solutions, the historical torque data of various diameter threads in the substation terminal block in several tightening tasks are obtained and screened to establish the terminal block tightening force standards under various tightening tasks and various diameter threads, including:

[0013] Obtain historical torque data required for various diameter threads in the substation terminal block in several tightening tasks to construct a historical torque set;

[0014] Selecting from the historical torque set the minimum historical torque data in which the tightening degree meets the progress requirement of the tightening task and satisfies the success rate threshold in several tightening tasks, so as to establish the tightening force standard of terminal blocks under various thread diameters;

[0015] The minimum historical torque data is selected by the following formula:

[0016]

[0017] In the formula, P i is the historical moment set; is the success rate threshold; n h N is the number of tightening degrees that meet the progress requirements of the hth tightening task among several tightening tasks, h = 1, 2, 3; i,j The number of historical torque data with j value for the i-th type diameter thread.

[0018] As one of the preferred solutions, the torque change data and the target thread are used as sample data to train a neural network model to obtain a multi-layer perceptron for identifying the progress of a tightening task, including:

[0019] During the training process, the torque change data and the tightening task progress of the target thread are forward propagated and back propagated to update the weights and biases of the neural network model to obtain a trained neural network model;

[0020] In the verification process, fuzzy rules are introduced to adjust the hyperparameters of the trained neural network model, and the training process and the verification process are repeated through the adjusted neural network model until a preset number of repetitions is reached, and a multilayer perceptron for identifying the progress of the tightening task is output.

[0021] As one of the preferred solutions, the method of introducing fuzzy rules to adjust the hyperparameters of the trained neural network model includes:

[0022] Taking the training error of the training process and the verification error of the verification process as input parameters, converting the input parameters into fuzzy sets, and calculating the parameter membership of each input parameter in the fuzzy sets;

[0023] Setting a fuzzy logic decision rule, and performing Mamdani reasoning on the fuzzy set and the parameter membership through the decision rule to obtain an output fuzzy set and a corresponding set membership;

[0024] The output fuzzy set and the corresponding set membership are defuzzified by the centroid method to obtain an adjustment hyperparameter, so as to adjust the parameters of the trained neural network model by the adjustment hyperparameter.

[0025] As one of the preferred solutions, the motion constraint is expressed by the following formula:

[0026]

[0027]

[0028] Where Ω is the motion constraint; T1 and T2 are the time required for the two tool heads to complete the corresponding fastening tasks; ω is the constraint variable; D1 and D2 are the movement process of the threads of the two tool heads in the corresponding fastening tasks; P1 and P2 are the leads of the threads of the two tool heads in the corresponding fastening tasks; N RPS1 ,N RPS2 are the rotation speeds of the two cutter heads in the corresponding tightening tasks.

[0029] As one of the preferred solutions, the motion constraint is evaluated by the real-time fastening task progress, and the double cutter head in the double cutter head fastening tool is controlled to run at the dual-axis rotation speed or improved speed according to the evaluation result, including:

[0030] When the real-time fastening task progress satisfies the motion constraint, controlling the double cutter heads in the double cutter head fastening tool to run at the dual-axis rotation speed;

[0031] When the real-time fastening task progress does not satisfy the motion constraint, controlling the double cutter head in the double cutter head fastening tool to run at the improved speed;

[0032] Wherein, the improvement rate is expressed by the following formula:

[0033]

[0034] Where a(t) is the expected acceleration; a m is the acceleration threshold; t s and t a It is a time-varying node.

[0035] As one of the preferred solutions, the S-shaped acceleration and deceleration optimized by the multi-objective optimization algorithm based on multi-population co-evolution includes:

[0036] Using the motion constraint as a constraint condition of the multi-objective optimization algorithm and constructing an evaluation function of the multi-objective optimization algorithm;

[0037] The time-varying node and the acceleration threshold are used as optimization variables, and the first multi-objective optimization algorithm, the second multi-objective optimization algorithm, and the third multi-objective optimization algorithm are combined to solve the evaluation function to obtain a non-dominated solution set;

[0038] Based on the target tightening task, a target time variation node and a target acceleration threshold are selected from the non-dominated solution set to optimize the S-type acceleration and deceleration formula to obtain the improved speed.

[0039] As one of the preferred solutions, the first multi-objective optimization algorithm, the second multi-objective optimization algorithm and the third multi-objective optimization algorithm are combined to solve the evaluation function to obtain a non-dominated solution set, including:

[0040] Setting initialization parameters to generate an initial population, uniformly and randomly dividing the initial population into three sub-populations, and respectively allocating them to the first multi-objective optimization algorithm, the second multi-objective optimization algorithm, and the third multi-objective optimization algorithm for iteration to generate three corresponding sub-generation solution sets;

[0041] The subpopulation and the offspring solution set are merged and non-dominated sorted, solutions are selected therefrom to update the subpopulation, and the proportion of the subpopulation is adjusted according to the offspring survival rate;

[0042] Based on the updated sub-population, the offspring solution set solving step and the non-dominated sorting step are iteratively executed until a preset number of iterations is reached, and the final non-dominated solution set is output.

[0043] A second aspect of the present invention provides a substation terminal block tightening tool control system based on torque feedback, which is applied to a double-blade tightening tool, in which the double axes corresponding to the double blades in the double-blade tightening tool are independently driven; the system comprises:

[0044] The standard establishment module is used to obtain and filter the historical torque data of various diameter threads in the substation terminal block in several tightening tasks, so as to establish the tightening force standards of the terminal block under various tightening tasks and various diameter threads;

[0045] A model training module, for controlling the double-tip fastening tool to fasten the target thread according to the terminal block fastening force standard, obtaining torque change data, and using the torque change data and the target thread as sample data to train a neural network model, thereby obtaining a multi-layer perceptron for identifying the progress of a fastening task;

[0046] A constraint solving module, used for constructing and solving the motion constraints between the two axes of the double-blade fastening tool based on the terminal block fastening force standard to obtain the rotation speed of the two axes;

[0047] A progress acquisition module, used to acquire the target tightening task and the real-time torque data when the double-tip tightening tool performs the target tightening task, and input the real-time torque data into the multi-layer perceptron for processing to obtain the real-time tightening task progress;

[0048] A speed control module is used to evaluate the motion constraint through the real-time tightening task progress, and control the double cutter head in the double cutter head tightening tool to run at the dual-axis rotation speed or improved speed according to the evaluation result, so as to achieve the tightening task of the substation terminal block; the improved speed is an S-shaped acceleration and deceleration optimized by a multi-objective optimization algorithm based on multi-population co-evolution.

[0049] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0050] (1) The corresponding tightening torque is set based on different thread types and tightening task requirements, and the corresponding terminal tightening force standards are formulated to facilitate standardized and unified operations. At the same time, in the test process, the deep learning strategy is combined with the change information of torque feedback to obtain a multi-layer perceptron that can analyze the work process through training, thereby achieving the adjustment of the appropriate tightening torque during the work process and real-time monitoring of the task process;

[0051] (2) Based on the terminal block tightening force standard, the motion constraint of the dual-axis tightening of the dual-head tightening tool is constructed, and the coordination of the dual axes is evaluated based on the motion constraint. Then, the multi-objective optimization algorithm based on multi-population co-evolution and the S-shaped acceleration and deceleration speed instructions are used to optimize the speed of the dual heads of the dual-head tightening tool, so as to improve its dual-axis coordination and further improve the working efficiency of the tightening tool. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the implementation mode will be briefly introduced below. Obviously, the drawings described below are only some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Figure 1 It is a flow chart of a method for controlling a substation terminal block tightening tool based on torque feedback provided by a certain embodiment of the present invention;

[0054] Figure 2 is a schematic diagram of a neural network model training process provided by a certain embodiment of the present invention;

[0055] Figure 3 is a schematic diagram of a speed adjustment process of a double-tip fastening tool provided by an embodiment of the present invention;

[0056] Figure 4 It is a structural diagram of a substation terminal block tightening tool control system based on torque feedback provided by a certain embodiment of the present invention. DETAILED DESCRIPTION

[0057] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] In the description of this application, the terms "first", "second", "third", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of the feature. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0059] In the description of the present application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two components. The terms "vertical", "horizontal", "left", "right", "upper", "lower" and similar expressions used herein are for illustrative purposes only, and do not indicate or imply that the system or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0060] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those commonly understood by those skilled in the art. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood by specific circumstances.

[0061] In one embodiment, if Figure 1 As shown, the first aspect of the present invention provides a control method for a substation terminal block tightening tool based on torque feedback, which is applied to a double-blade tightening tool, in which the double axes corresponding to the double blade heads in the double-blade tightening tool are independently driven; an electric module is provided at the connection between each blade head and the corresponding tool handle in the double-blade tightening tool; the electric module includes a motor, so that the corresponding blade head is independently driven by its own motor; the electric module is also used to feedback the torque data of the corresponding blade head.

[0062] Specifically, the present invention takes into account the different spacings of the terminal blocks in the substation, and the double-blade fastening tool used for this purpose has the characteristic of adjustable dual-axis spacing; in order to realize the electric function of the fastening tool, an electric module containing a motor is installed at the connection between each blade of the double-blade fastening tool and its corresponding handle, so that each blade is driven by its corresponding independent motor to realize the motion control of any blade. In addition, in order to better realize the simultaneous fastening of the two heads of the fastening tool, a torque feedback unit is set in the electric module to realize variable speed tooth finding; wherein, the torque feedback unit is used to monitor the torque change during the fastening process in real time, and a torque sensor is installed on each fastening blade, and the torque data during the fastening process is fed back to the controller of the electric module in real time, so that the controller adjusts the speed and direction of the motor according to the change of the torque data, so that the fastening blade can more accurately find and adapt to the tooth shape of the fastener, and at the same time lay the foundation for evaluating the synergy of the dual props and the work progress. Furthermore, according to the main terminal block types in the substation, a standardized double-headed tool of the fastening tool is designed to form a double-headed fastening tool, and the main head spacings are 1.5 cm, 2.5 cm, and 4.5 cm respectively.

[0063] The present invention is adapted to different terminal row types by adjustable dual-axis spacing and customized tool spacing, and achieves better control of dual tools and improves work efficiency by means of dual-axis independent driving and torque feedback.

[0064] The method comprises:

[0065] S1. Obtain and filter historical torque data of various diameter threads in a terminal block of a substation in several tightening tasks, so as to establish a tightening force standard for the terminal block under various tightening tasks and various diameter threads;

[0066] In one embodiment, step S1 includes:

[0067] The historical torque data required for various diameter threads in the substation terminal block in several tightening tasks are obtained to construct a historical torque set; specifically, in order to adapt to the differences caused by different terminal block types and screw strengths, the present invention tests different diameter threads in terminal tightening work by using corresponding double-head tightening tools, sets different torque values ​​to tighten the threads, and samples and collects the torque data of different diameter threads and corresponding double-head tightening tools in various tasks to obtain a historical torque set.

[0068] Selecting from the historical torque set the minimum historical torque data in which the tightening degree meets the progress requirement of the tightening task and satisfies the success rate threshold in several tightening tasks, so as to establish the tightening force standard of terminal blocks under various thread diameters;

[0069] The minimum historical torque data is selected by the following formula:

[0070]

[0071] In the formula, P i is the historical moment set; is the success rate threshold; n h N is the number of tightening degrees that meet the progress requirements of the hth tightening task among several tightening tasks, h = 1, 2, 3; i,j The number of historical torque data with j value for the i-th type diameter thread.

[0072] Among them, when h=1, it is applicable to headless screws that do not protrude from the hole when tightened and other screws that cannot be tightened with a screwdriver whose blade width is greater than the root diameter of the screw; when h=2, it is applicable to screws and nuts tightened with a screwdriver; when h=3, it is applicable to screws and nuts that cannot be tightened with a screwdriver. The present invention selects the tightening degree in the tightening task from the historical torque set to meet the progress requirements of the tightening task, and then records its success rate under different torques, divides the combination that meets the success rate threshold into sets, and selects the minimum torque that meets the success rate threshold, so as to establish a unified terminal block tightening force standard, that is, the corresponding tightening torque is set based on different thread types and tightening task requirements, and the corresponding terminal tightening force standard is formulated to facilitate standardized and unified operations.

[0073] S2, controlling the double-blade fastening tool to fasten the target thread according to the terminal block fastening force standard to obtain torque change data, and using the torque change data and the target thread as sample data to train a neural network model to obtain a multi-layer perceptron for identifying the progress of the fastening task;

[0074] Specifically, according to the terminal block tightening force standard, a suitable tightening torque is selected to tighten the selected target type of thread, and the torque change information during the tightening process is recorded through torque feedback. At the same time, the working progress (i.e., the tightening degree of the bolt) is recorded to conduct a large number of tests and collect corresponding samples. It should be noted that the type of target thread is not limited to one type, and can be multiple types to enrich the sample data.

[0075] In one embodiment, the training of a neural network model using the torque change data and the target thread as sample data to obtain a multi-layer perceptron for identifying the progress of a tightening task includes:

[0076] During the training process, the torque change data and the tightening task progress of the target thread are forward propagated and back propagated to update the weights and biases of the neural network model to obtain a trained neural network model;

[0077] In the verification process, fuzzy rules are introduced to adjust the hyperparameters of the trained neural network model, and the training process and the verification process are repeated through the adjusted neural network model until a preset number of repetitions is reached, and a multilayer perceptron for identifying the progress of the tightening task is output.

[0078] Specifically, in order to find the relationship between torque change data, thread type and tightening task progress in the sample, the present invention uses a multi-task deep learning strategy for training. The neural network model training process is as follows: Figure 2 As shown, it mainly adopts the multi-layer perceptron (MLP) method that introduces the artificial neural network architecture. The MLP consists of an input layer, multiple hidden layers, and an output layer. The input layer receives input data, the hidden layer represents the learning features, and the output layer produces the final result. Each neuron in the hidden layer and the output layer has an activation function for introducing nonlinear mapping. During the training process, the neural network model uses the torque change data in the sample data as the training input, and the thread model and the tightening task progress as the output. The forward propagation and back propagation in the MLP method are used to continuously update the weights and biases of the connections of each layer to obtain the trained neural network model; during the verification process, fuzzy rule adjustment is introduced to adjust the learning rate, batch size, and number of hidden layers in the trained neural network model to avoid overfitting and underfitting, and the training process and verification process are repeated through the neural network model after parameter adjustment until the preset number of repetitions is reached to obtain a multi-layer perceptron.

[0079] The multi-layer perceptron obtained by the present invention through the multi-task deep learning strategy mainly includes: a classification function of distinguishing thread types according to torque change data by training a classification model, and a regression function of monitoring the progress of the tightening task according to the torque change data by training a regression model. The present invention combines the deep learning strategy and uses the change information of torque feedback to judge different thread types and analyze the work progress, so as to adjust the appropriate tightening torque during the work process and monitor the task progress in real time, so as to improve the work efficiency of the double-head fastening tool.

[0080] In one embodiment, the introducing fuzzy rules to adjust the hyperparameters of the trained neural network model includes:

[0081] Taking the training error of the training process and the verification error of the verification process as input parameters, converting the input parameters into fuzzy sets, and calculating the parameter membership of each input parameter in the fuzzy sets;

[0082] Setting a fuzzy logic decision rule, and performing Mamdani reasoning on the fuzzy set and the parameter membership through the decision rule to obtain an output fuzzy set and a corresponding set membership;

[0083] The output fuzzy set and the corresponding set membership are defuzzified by the centroid method to obtain an adjustment hyperparameter, so as to adjust the parameters of the trained neural network model by the adjustment hyperparameter.

[0084] Specifically, the present invention collects the training error in the model training process and the verification error in the verification process as input parameters, the error can be represented by a loss value, and the input parameters are divided into multiple fuzzy sets (low, lower, medium, higher, high), the variation range of the training error and the verification error is defined, and the membership of each input parameter in different fuzzy sets is calculated using a membership function (such as a Gaussian function, a trapezoidal function, etc.); based on historical task requirements and their corresponding moment data, a fuzzy logic "if-then" decision rule is set, such as "if the training error is high and the verification error is high, then adjust the hyperparameters to reduce overfitting", and the Mamdani reasoning method is used to calculate the output fuzzy set and its corresponding membership according to the membership of the input parameters and the defined fuzzy rules; the output fuzzy set is converted into a specific value using the centroid method, that is, one or more values ​​for adjusting the hyperparameters of the neural network model are obtained, and the specific values ​​obtained by defuzzification are applied to the parameters of the MLP model to train the MLP model and evaluate its performance; according to the model performance, the fuzzy rules and fuzzy sets are further adjusted to optimize the parameter adjustment process.

[0085] The present invention dynamically adjusts hyperparameters according to real-time training errors and verification errors, thereby improving the flexibility and adaptability of model adjustment; processes the uncertainty of input parameters through fuzzy logic, making the system more robust when facing complex and uncertain environments; dynamically adjusting hyperparameters helps to avoid overfitting or underfitting, and improves the generalization ability and overall performance of the model; the entire adjustment process is automated, reducing manual intervention, improving efficiency, and having a certain level of intelligence.

[0086] S3, constructing and solving the motion constraints between the two axes of the double-blade fastening tool based on the terminal block fastening force standard to obtain the two-axis rotation speed;

[0087] Specifically, considering that the independent drive of the dual tools needs to be synergistic, the present invention is based on the terminal block tightening force standard, and for the dual-head tightening tools used in the tightening task, combined with different thread leads and thread lengths, constructs a motion constraint between the dual axes, and then solves the preset dual-axis rotation speed, and presets it to ensure that the bolts can be tightened simultaneously without disturbance. For this purpose, a motion constraint is constructed, which is expressed by the following formula:

[0088]

[0089]

[0090] Where Ω is the motion constraint; T1 and T2 are the time required for the two tool heads to complete the corresponding fastening tasks; ω is the constraint variable to limit the relationship between T1 and T2; D1 and D2 are the movement process of the threads of the two tool heads in the corresponding fastening tasks; P1 and P2 are the leads of the threads of the two tool heads in the corresponding fastening tasks; N RPS1 ,N RPS2 are the rotation speeds of the two cutter heads in the corresponding tightening tasks.

[0091] By solving the motion constraint according to T1=T2, the preset rotation speeds of the two axes corresponding to the two cutters of the double-cutting head fastening tool can be obtained. The present invention improves the working efficiency of the fastening tool by constructing the motion constraint between the two axes of the double-cutting head fastening tool and evaluating the synergy of the two axes based on the motion constraint.

[0092] S4, obtaining the target tightening task and the real-time torque data of the double-tip tightening tool when performing the target tightening task, and inputting the real-time torque data into the multi-layer perceptron for processing to obtain the real-time tightening task progress;

[0093] Specifically, the specific requirements of the tightening task, such as the tightening position, the tightening object, the target torque value, the tightening sequence, etc., are determined and obtained, and the torque data of the tightening process is collected in real time through the sensor installed on the double-head tightening tool; the real-time torque data is cleaned, noise and outliers are removed, and standardized or normalized to improve the processing efficiency of the multi-layer perceptron; during the actual tightening task execution, the real-time torque data is input into the trained multi-layer perceptron model for processing to output the real-time tightening task progress.

[0094] S5, evaluating the motion constraint through the real-time fastening task progress, and controlling the double cutter head in the double cutter head fastening tool to run at the dual-axis rotation speed or improved speed according to the evaluation result, so as to achieve the fastening task of the substation terminal block; the improved speed is an S-shaped acceleration and deceleration optimized by a multi-objective optimization algorithm based on multi-population co-evolution;

[0095] In one embodiment, step S5 includes:

[0096] When the real-time fastening task progress satisfies the motion constraint, controlling the double cutter heads in the double cutter head fastening tool to run at the dual-axis rotation speed;

[0097] When the real-time fastening task progress does not satisfy the motion constraint, controlling the double cutter head in the double cutter head fastening tool to run at the improved speed;

[0098] Wherein, the improvement rate is expressed by the following formula:

[0099]

[0100] Where a(t) is the expected acceleration; a m is the acceleration threshold; t s and t a It is a time-varying node.

[0101] Specifically, the speed adjustment process of the double-head fastening tool is as follows: Figure 3 As shown in the figure, in the actual tightening process, the initial speed of the double-head tightening tool is the double-axis rotation speed. In order to avoid the double-axis from working in coordination due to errors and disturbances, the real-time change information of the torque is used to detect the progress of the double-axis tightening task to evaluate whether the motion constraint is satisfied. If the real-time tightening task progress satisfies the motion constraint, the double-head tightening tool is controlled to continue to run at the double-axis rotation speed obtained by solving the running constraint to perform the target tightening work; if the real-time tightening task progress does not meet the motion constraint, it is determined that the double-head tightening tool is difficult to achieve double-axis coordination. Therefore, it is necessary to speed up the slower motor in the electric module, that is, to control the double-head tightening tool to run at an improved speed to perform the target tightening work.

[0102] The present invention dynamically adjusts the running speed of the double cutter heads by evaluating the relationship between the progress of the fastening task and the motion constraints in real time. When the task progress does not meet the constraints, it switches to the improved speed operation, which helps to avoid potential conflicts or errors and ensures that the fastening task can be carried out smoothly, thereby improving the flexibility and adaptability of the fastening task, making more effective use of time and resources to improve work efficiency. At the same time, the task progress can be evaluated in real time and corresponding adjustments can be made, which helps to improve the safety of the fastening task and reduce potential risks caused by improper operation or equipment failure.

[0103] In one embodiment, the S-shaped acceleration and deceleration optimized by the multi-objective optimization algorithm based on multi-population co-evolution includes:

[0104] Using the motion constraint as a constraint condition of the multi-objective optimization algorithm and constructing an evaluation function of the multi-objective optimization algorithm;

[0105] The time-varying node and the acceleration threshold are used as optimization variables, and the first multi-objective optimization algorithm, the second multi-objective optimization algorithm, and the third multi-objective optimization algorithm are combined to solve the evaluation function to obtain a non-dominated solution set;

[0106] Based on the target tightening task, a target time variation node and a target acceleration threshold are selected from the non-dominated solution set to optimize the S-type acceleration and deceleration formula to obtain the improved speed.

[0107] Specifically, in order to enable the double cutter heads in the double cutter head fastening tool to finally fasten simultaneously, the present invention adopts a multi-population co-evolutionary multi-objective optimization algorithm to adjust the time change node and the acceleration threshold, so as to solve the relevant parameters of the speed instruction of the S-type acceleration and deceleration, and constructs the evaluation function of the multi-objective optimization algorithm based on the time change node and the acceleration threshold in the improved speed formula, which is expressed by the following formula:

[0108]

[0109] Where J1, J2, and J3 are multi-objective optimization functions, which are the time required for acceleration, the influence of maximum acceleration on motor torque, and the maximum value of angular velocity. m ) is the torque of the motor when it reaches the acceleration threshold; M0 is the torque of the motor before acceleration; w0 is the preset angular velocity of the motor.

[0110] At the same time, the motion constraint Ω is used as the constraint condition of the multi-objective optimization function, the time-varying node and the acceleration threshold are used as optimization variables, and three different multi-objective optimization algorithms are combined for multi-objective optimization. The multi-objective optimization algorithm based on genetic algorithm (NSGA-II), the multi-objective optimization algorithm of particle swarm (MOPSO) and the multi-objective optimization algorithm based on decomposition differential evolution (MOEA-DE) algorithm can be used. As long as it is a multi-objective optimization algorithm, the specific algorithm type is not limited here. Then, the evaluation function is solved by the combined multi-objective optimization algorithm to obtain a non-dominated solution set, and based on the specific requirements of the target fastening task (such as accuracy, speed, stability, etc.), the most suitable time-varying node and acceleration threshold are selected from the non-dominated solution set and substituted into the S-type acceleration and deceleration formula, which is optimized to obtain an improved speed curve, which should be able to better meet the needs of the target fastening task. The present invention can comprehensively consider multiple optimization targets through the combined use of multi-objective optimization algorithms, thereby obtaining a more comprehensive and excellent solution, and further improving the working efficiency of the double-head fastening tool.

[0111] In one embodiment, the first multi-objective optimization algorithm, the second multi-objective optimization algorithm, and the third multi-objective optimization algorithm are combined to solve the evaluation function to obtain a non-dominated solution set, including:

[0112] Setting initialization parameters to generate an initial population, uniformly and randomly dividing the initial population into three sub-populations, and respectively allocating them to the first multi-objective optimization algorithm, the second multi-objective optimization algorithm, and the third multi-objective optimization algorithm for iteration to generate three corresponding sub-generation solution sets;

[0113] The subpopulation and the offspring solution set are merged and non-dominated sorted, solutions are selected therefrom to update the subpopulation, and the proportion of the subpopulation is adjusted according to the offspring survival rate;

[0114] Based on the updated sub-population, the offspring solution set solving step and the non-dominated sorting step are iteratively executed until a preset number of iterations is reached, and the final non-dominated solution set is output.

[0115] Specifically, the combined algorithm (taking NSGA-II, MOPSO and MOEA-DE algorithms as examples) solves the evaluation function in the following process:

[0116] Step 1: Set the necessary initialization parameters, such as population size, number of iterations, sub-population ratio, etc., and generate an initial population Q based on these parameters. The population contains multiple candidate solutions, and the population is uniformly and randomly divided into three sub-populations Q1, Q2, and Q3, which are assigned to NSGA-II, MOPSO, and MOEA-DE algorithms respectively;

[0117] Step 2: Use these three optimization algorithms to iterate (the specific iteration process is not discussed here) to generate corresponding child solution sets Of1, Of2, Of3, each of which contains multiple optimized candidate solutions;

[0118] Step 3: Merge the subpopulation and the three offspring solution sets to form a larger solution set, and perform non-dominated sorting on the merged solution set to distinguish solutions of different levels of quality; based on the results of non-dominated sorting, select a portion of excellent solutions to update the subpopulation, and adjust the proportion of the subpopulation based on the offspring survival rate (i.e., the proportion of excellent solutions in the subpopulation) to maintain diversity; the proportion of the subpopulation is adjusted by the following formula:

[0119]

[0120] In the formula, Q' x is the adjusted subpopulation size; Q x is the number of subpopulations; Q min , Q max are the minimum and maximum values ​​of the subpopulation number respectively; Rate(Of x ) is the survival rate of the xth generation solution set after non-dominated sorting; μ min , μ max are the minimum and maximum survival rate thresholds respectively. If it is less than the minimum threshold, the algorithm will randomly assign at least Q min If the number of subpopulations is greater than the maximum threshold, the algorithm randomly assigns at most Q max The number of sub-populations; λ is the adjustment factor of the corresponding population size of each algorithm before and after iteration;

[0121] Step 4: Determine whether the preset number of iterations has been reached. If so, output the non-dominated solution set; if not, redistribute the adjusted number of subpopulations to NSGA-II, MOPSO, and MOEA-DE algorithms, and re-enter step 2 for iterative solution until the end.

[0122] The present invention can explore the solution space more effectively and increase the possibility of finding the global optimal solution by dividing the initial population into multiple sub-populations and applying different multi-objective optimization algorithms to each of them. The parallel processing strategy adopted can accelerate the convergence process and reduce the calculation time. The adjustment of non-dominant sorting and offspring survival rate helps to maintain the diversity of the solution set and avoid premature convergence. By dynamically adjusting the sub-population ratio, the relationship between exploration and utilization can be further balanced. At the same time, the multi-objective optimization algorithm based on the co-evolution of multiple populations and the S-type acceleration and deceleration speed instructions are used to achieve optimal adjustment of the speed, so as to improve the synergy of the dual axes and further ensure the working efficiency of the fastening tool.

[0123] In the embodiment of the present application, based on the problem of how to improve the working efficiency of the existing terminal block tightening tools, a control method for substation terminal block tightening tools based on torque feedback is designed and applied to double-blade tightening tools. The double axes corresponding to the double blades in the double-blade tightening tool are independently driven; it realizes the acquisition and screening of historical torque data of various diameter threads in the substation terminal block in several tightening tasks, so as to establish terminal block tightening force standards under various tightening tasks and various diameter threads; according to the terminal block tightening force standard, the double-blade tightening tool is controlled to tighten the target thread to obtain torque change data, and the torque change data and the target thread are used as sample data to train the neural network model to obtain a multi-layer perceptron for identifying the progress of the tightening task; based on the The terminal block tightening force standard is used to construct the motion constraint between the two axes of the double-blade tightening tool and solve it to obtain the double-axis rotation speed; the target tightening task and the real-time torque data of the double-blade tightening tool when performing the target tightening task are obtained, and the real-time torque data is input into the multi-layer perceptron for processing to obtain the real-time tightening task progress; the motion constraint is evaluated according to the real-time tightening task progress, and the double blade heads in the double-blade tightening tool are controlled to run at the double-axis rotation speed or the improved speed according to the evaluation result to achieve the tightening task of the substation terminal block; the improved speed is a technical solution of S-type acceleration and deceleration optimized by a multi-objective optimization algorithm based on multi-population co-evolution; the working efficiency of the existing terminal block tightening tools is effectively improved.

[0124] It should be noted that although the steps in the above flowchart are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders.

[0125] In another embodiment, if Figure 4 As shown, the second aspect of the present invention provides a substation terminal block tightening tool control system based on torque feedback, which is applied to a double-blade tightening tool, in which the double axes corresponding to the double blades in the double-blade tightening tool are independently driven; the system comprises:

[0126] The standard establishment module 10 is used to obtain and filter the historical torque data of various diameter threads in the terminal block of the substation in several tightening tasks, so as to establish the tightening force standards of the terminal block under various tightening tasks and various diameter threads;

[0127] A model training module 20 is used to control the double-tip fastening tool to fasten the target thread according to the terminal block fastening force standard to obtain torque change data, and use the torque change data and the target thread as sample data to train a neural network model to obtain a multi-layer perceptron for identifying the progress of the fastening task;

[0128] A constraint solving module 30 is used to construct and solve the motion constraints between the two axes of the double-blade fastening tool based on the terminal block fastening force standard to obtain the rotation speed of the two axes;

[0129] A progress acquisition module 40 is used to acquire the target tightening task and the real-time torque data when the double-tip tightening tool performs the target tightening task, and input the real-time torque data into the multi-layer perceptron for processing to obtain the real-time tightening task progress;

[0130] The speed control module 50 is used to evaluate the motion constraint through the real-time tightening task progress, and control the double cutter head in the double cutter head tightening tool to run at the dual-axis rotation speed or improved speed according to the evaluation result, so as to achieve the tightening task of the substation terminal block; the improved speed is an S-shaped acceleration and deceleration optimized by a multi-objective optimization algorithm based on multi-population co-evolution.

[0131] It should be noted that each module in the above-mentioned substation terminal block tightening tool control system based on torque feedback can be fully or partially implemented by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules. For the specific definition of a substation terminal block tightening tool control system based on torque feedback, please refer to the definition of a substation terminal block tightening tool control method based on torque feedback above. The two have the same functions and effects, which will not be repeated here.

[0132] In summary, the present invention relates to the field of motion control technology, and discloses a control method and system for a substation terminal block tightening tool based on torque feedback, which is applied to a double-blade tightening tool, wherein the double axes corresponding to the double blade heads are driven independently; the corresponding tightening torque is set based on different thread types and tightening task requirements to formulate corresponding terminal tightening force standards, and in the test process, a deep learning strategy is combined to use the change information of torque feedback based on the terminal tightening force standard to obtain a multi-layer perceptron that can analyze the progress of the tightening task through training; based on the terminal block tightening force standard, a motion constraint for the double-axis tightening of the double-blade tightening tool is constructed, and based on the motion constraint, the synergy of the double axes is evaluated, and then the multi-objective optimization algorithm based on multi-population co-evolution and the S-type acceleration and deceleration speed instruction are used to optimize the speed, so as to improve the synergy of the double axes of the double-blade tightening tool and further improve the working efficiency of the tightening tool.

[0133] Each embodiment in this specification is described in a progressive manner, and the same or similar parts of each embodiment can be directly referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. It should be noted that the technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above-mentioned embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0134] The above-mentioned embodiments only express several preferred implementation modes of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in the technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be based on the protection scope of the claims.

Claims

1. A method for controlling a substation terminal block tightening tool based on torque feedback, characterized in that: Applicable to a double-tool fastening tool, in which the double shafts corresponding to the double tools are independently driven; the method comprises: Obtain and filter historical torque data of various thread diameters in the substation terminal block in several tightening tasks to establish the tightening force standards for the terminal block under various tightening tasks and various thread diameters; Controlling the double-tip fastening tool to fasten the target thread according to the terminal block fastening force standard to obtain torque change data, and using the torque change data and the target thread as sample data to train a neural network model to obtain a multi-layer perceptron for identifying the progress of the fastening task; Based on the terminal block tightening force standard, the motion constraint between the two axes of the double-blade tightening tool is constructed and solved to obtain the rotation speed of the two axes; Acquire the target tightening task and the real-time torque data when the double-tip tightening tool performs the target tightening task, and input the real-time torque data into the multi-layer perceptron for processing to obtain the real-time tightening task progress; The motion constraint is evaluated through the real-time tightening task progress, and the double cutter head in the double cutter head tightening tool is controlled to run at the dual-axis rotation speed or improved speed according to the evaluation result, so as to achieve the tightening task of the substation terminal block; the improved speed is an S-shaped acceleration and deceleration optimized by a multi-objective optimization algorithm based on multi-population co-evolution.

2. A method for controlling a substation terminal block tightening tool based on torque feedback according to claim 1, characterized in that: An electric module is provided at the connection between each of the cutter heads and the corresponding tool handle in the double-cutter head fastening tool; the electric module includes a motor so that the corresponding cutter head is independently driven by its own motor; the electric module is also used to feed back the torque data of the corresponding cutter head.

3. A method for controlling a substation terminal block tightening tool based on torque feedback according to claim 1, characterized in that: The acquisition and screening of historical torque data of various diameter threads in the substation terminal block in several tightening tasks to establish the terminal block tightening force standards under various tightening tasks and various diameter threads include: Obtain historical torque data required for various diameter threads in the substation terminal block in several tightening tasks to construct a historical torque set; Selecting from the historical torque set the minimum historical torque data in which the tightening degree meets the progress requirement of the tightening task and satisfies the success rate threshold in several tightening tasks, so as to establish the tightening force standard of terminal blocks under various thread diameters; The minimum historical torque data is selected by the following formula: Ρi={jρi,j>θ,ρi,j=nh / Ni,j} In the formula, P i is the historical moment set; θ is the success rate threshold; n h N is the number of tightening degrees that meet the progress requirements of the hth tightening task among several tightening tasks, h = 1, 2, 3; i,j The number of historical torque data with j value for the i-th type diameter thread.

4. The method for controlling a substation terminal block tightening tool based on torque feedback according to claim 1, characterized in that: The method of using the torque change data and the target thread as sample data to train a neural network model to obtain a multi-layer perceptron for identifying the progress of a tightening task includes: During the training process, the torque change data and the tightening task progress of the target thread are forward propagated and back propagated to update the weights and biases of the neural network model to obtain a trained neural network model; In the verification process, fuzzy rules are introduced to adjust the hyperparameters of the trained neural network model, and the training process and the verification process are repeated through the adjusted neural network model until a preset number of repetitions is reached, and a multilayer perceptron for identifying the progress of the tightening task is output.

5. A method for controlling a substation terminal block tightening tool based on torque feedback according to claim 4, characterized in that: The introducing fuzzy rules to adjust the hyperparameters of the trained neural network model includes: Taking the training error of the training process and the verification error of the verification process as input parameters, converting the input parameters into fuzzy sets, and calculating the parameter membership of each input parameter in the fuzzy sets; Setting a fuzzy logic decision rule, and performing Mamdani reasoning on the fuzzy set and the parameter membership through the decision rule to obtain an output fuzzy set and a corresponding set membership; The output fuzzy set and the corresponding set membership are defuzzified by the centroid method to obtain an adjustment hyperparameter, so as to adjust the parameters of the trained neural network model by the adjustment hyperparameter.

6. A method for controlling a substation terminal block tightening tool based on torque feedback according to claim 1, characterized in that: The motion constraint is expressed by the following formula: Where Ω is the motion constraint; T1 and T2 are the time required for the two tool heads to complete the corresponding fastening tasks; ω is the constraint variable; D1 and D2 are the movement process of the threads of the two tool heads in the corresponding fastening tasks; P1 and P2 are the leads of the threads of the two tool heads in the corresponding fastening tasks; N RPS1 ,N RPS2 are the rotation speeds of the two cutter heads in the corresponding tightening tasks.

7. A method for controlling a substation terminal block tightening tool based on torque feedback according to claim 1, characterized in that: The step of evaluating the motion constraint by the real-time fastening task progress and controlling the double cutter head in the double cutter head fastening tool to run at the double-axis rotation speed or the improved speed according to the evaluation result comprises: When the real-time fastening task progress satisfies the motion constraint, controlling the double cutter heads in the double cutter head fastening tool to run at the dual-axis rotation speed; When the real-time fastening task progress does not satisfy the motion constraint, controlling the double cutter head in the double cutter head fastening tool to run at the improved speed; Wherein, the improvement rate is expressed by the following formula: Where a(t) is the expected acceleration; a m is the acceleration threshold; t s and t a It is a time-varying node.

8. A method for controlling a substation terminal block tightening tool based on torque feedback according to claim 7, characterized in that: The S-shaped acceleration and deceleration optimized by the multi-objective optimization algorithm based on multi-population co-evolution includes: Using the motion constraint as a constraint condition of the multi-objective optimization algorithm and constructing an evaluation function of the multi-objective optimization algorithm; The time-varying node and the acceleration threshold are used as optimization variables, and the first multi-objective optimization algorithm, the second multi-objective optimization algorithm, and the third multi-objective optimization algorithm are combined to solve the evaluation function to obtain a non-dominated solution set; Based on the target tightening task, a target time variation node and a target acceleration threshold are selected from the non-dominated solution set to optimize the S-type acceleration and deceleration formula to obtain the improved speed.

9. A method for controlling a substation terminal block tightening tool based on torque feedback according to claim 8, characterized in that: The first multi-objective optimization algorithm, the second multi-objective optimization algorithm and the third multi-objective optimization algorithm are combined to solve the evaluation function to obtain a non-dominated solution set, including: Setting initialization parameters to generate an initial population, uniformly and randomly dividing the initial population into three sub-populations, and respectively allocating them to the first multi-objective optimization algorithm, the second multi-objective optimization algorithm, and the third multi-objective optimization algorithm for iteration to generate three corresponding sub-generation solution sets; The subpopulation and the offspring solution set are merged and non-dominated sorted, solutions are selected therefrom to update the subpopulation, and the proportion of the subpopulation is adjusted according to the offspring survival rate; Based on the updated sub-population, the offspring solution set solving step and the non-dominated sorting step are iteratively executed until a preset number of iterations is reached, and the final non-dominated solution set is output.

10. A substation terminal block tightening tool control system based on torque feedback, characterized in that: Applicable to a double-blade fastening tool, in which the double axes corresponding to the double blades in the double-blade fastening tool are independently driven; the system comprises: The standard establishment module is used to obtain and filter the historical torque data of various diameter threads in the substation terminal block in several tightening tasks, so as to establish the tightening force standards of the terminal block under various tightening tasks and various diameter threads; A model training module, for controlling the double-tip fastening tool to fasten the target thread according to the terminal block fastening force standard, obtaining torque change data, and using the torque change data and the target thread as sample data to train a neural network model, thereby obtaining a multi-layer perceptron for identifying the progress of a fastening task; A constraint solving module, used for constructing and solving the motion constraints between the two axes of the double-blade fastening tool based on the terminal block fastening force standard to obtain the rotation speed of the two axes; A progress acquisition module, used to acquire the target tightening task and the real-time torque data when the double-tip tightening tool performs the target tightening task, and input the real-time torque data into the multi-layer perceptron for processing to obtain the real-time tightening task progress; A speed control module is used to evaluate the motion constraint through the real-time tightening task progress, and control the double cutter head in the double cutter head tightening tool to run at the dual-axis rotation speed or improved speed according to the evaluation result, so as to achieve the tightening task of the substation terminal block; the improved speed is an S-shaped acceleration and deceleration optimized by a multi-objective optimization algorithm based on multi-population co-evolution.

Images