Wire rod fixed-length cutting control method and electronic device

Abstract

The application discloses a wire fixed-length cutting control method and electronic equipment, and comprises the following steps: acquiring the actual tension of the wire, the theoretical wire feeding length of the wire and the speed adjustment amount of the wire feeding servo motor; calculating the length error according to the set cutting length and the theoretical wire feeding length of the wire; taking the actual tension of the wire and the length error as input quantities, taking the speed adjustment amount of the wire feeding servo motor as an output quantity, and performing adaptive fuzzy control to generate an initial control signal; performing time sequence learning and correction on the initial control signal by using a long short-term memory network to obtain a corrected control signal; weighting and fusing the initial control signal and the corrected control signal to obtain a final control signal; and forming a final wire feeding length instruction according to the final control signal. The application provides a new idea for adaptive control of fixed-length cutting of elastic wires.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wire fixed-length cutting control, and in particular to a wire fixed-length cutting control method and electronic device. BACKGROUND

[0002] The existing technologies in the field of wire fixed-length cutting are mainly divided into two categories. One is mechanical positioning type, which limits the wire length by physical structures such as adjustable baffle and multi-specification mechanical components to achieve cutting. This type of method cannot compensate for the bending or accumulation deformation of the wire due to the dependence on physical structures, resulting in poor cutting accuracy, and the structure is strongly bound to the wire diameter, so the hardware needs to be adjusted when changing specifications, which has poor adaptability and low switching efficiency. The other is traditional electric control type, which adjusts the feeding and cutting actions by traditional PID algorithm, such as double closed-loop control of flying saw fixed-length system. This type of method has no self-adaptive ability to nonlinear disturbances such as wire elastic deformation and wire feeding pressure fluctuation, and the cutting timing lags behind at high-speed feeding, with large overshoot.

[0003] In the process of fixed-length cutting of elastic wire, the actual cutting length is easily affected by tension fluctuation due to the nonlinear elastic deformation characteristics of the wire, and the defects of the above-mentioned traditional control methods are particularly significant. On the one hand, even if a single fuzzy control is used to deal with nonlinear and fuzzy conditions, it is difficult to handle the time sequence deviation of length error accumulation over time. On the other hand, if the actual cutting length is directly measured by means such as laser ranging and visual detection, high-precision detection hardware needs to be additionally invested, which greatly increases the equipment cost, and detection delay is easy to occur in high-speed cutting scene, which is not suitable for engineering batch production. Therefore, there is an urgent need for an elastic wire fixed-length cutting control technology that takes into account fixed-length accuracy, cost control and engineering feasibility. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a wire fixed-length cutting control method and electronic device to solve the technical problems of insufficient accuracy and poor adaptability of elastic wire fixed-length cutting caused by nonlinear deformation, tension fluctuation and time sequence accumulation error in the related art.

[0005] According to a first aspect of the embodiments of the present application, a wire fixed-length cutting control method is provided, comprising:

[0006] obtaining actual tension of wire, theoretical wire feeding length of wire and speed adjustment amount of wire feeding servo motor;

[0007] calculating length error according to set cutting length and the theoretical wire feeding length of the wire;

[0008] performing adaptive fuzzy control by taking the actual tension of the wire and the length error as input quantities and taking the speed adjustment amount of the wire feeding servo motor as output quantity to generate an initial control signal;

[0009] The initial control signal is time-series learned and corrected using a long short-term memory network to obtain a corrected control signal.

[0010] The initial control signal and the modified control signal are weighted and fused to obtain the final control signal;

[0011] Based on the final control signal, a final wire length instruction is generated.

[0012] Optionally, the length error is calculated based on the set cutting length and the theoretical wire feeding length, including:

[0013] The length error is calculated using the following formula:

[0014] ;

[0015] In the formula, For the first The length error at each sampling time; To set the cutting length; For the first The theoretical wire feeding length of the wire feeding wheel at each sampling time;

[0016] ;

[0017] In the formula, For the first The encoder count of the feed wheel at each sampling time, where D is the diameter of the feed wheel. i This is the transmission ratio.

[0018] Optionally, the adaptive fuzzy control includes:

[0019] The input and output quantities are fuzzified and fuzzy inference is performed. After fuzzy inference, the fuzziness is defuzzified by the centroid method, and the speed adjustment of the wire feeding servo motor is output.

[0020] Based on the length error and the rate of change of error, the conversion coefficients between the fuzzy universe of discourse and the actual universe of discourse are updated recursively in real time.

[0021] Based on the speed adjustment amount and conversion coefficient of the output wire-feeding servo motor after defuzzification, the initial control signal of the final fuzzy control output is obtained.

[0022] Optionally, the input and output quantities are fuzzified, including:

[0023] Construction length error The fuzzy sets are respectively , express That is, the actual length Set the cutting length The number of cables to be supplied needs to be reduced; represents , error close to 0, no need to adjust; represents , that is, the actual length <Set the shear length , need to increase the wire.

[0024] The fuzzy set of the wire servo motor speed adjustment amount is constructed , represents to reduce the motor speed, reduce the wire length; represents to keep the speed, the error has been corrected; represents to increase the motor speed, increase the wire length;

[0025] According to the three fuzzy sets, the respective membership functions are constructed, and the membership mapping of the fuzzy set adopts a triangular membership function.

[0026] Optionally, the fuzzy rule base of the fuzzy reasoning is as follows:

[0027] (1) ; ; ;

[0028] (2) ; ; ;

[0029] (3) ; ; ;

[0030] The fuzzy rule base above outputs a fuzzy set through Min-Max composition method reasoning, for the rule R i : , wherein , the membership thereof satisfies , and the membership of the final output fuzzy set is the maximum value of all rule implication results, that is, .

[0031] Optionally, the wire servo motor speed adjustment amount output after defuzzification is:

[0032] ;

[0033] In the formula, represents the wire servo motor speed adjustment amount corresponding to the tth discrete point; m is the total number of discrete points of the universe of discourse of the wire servo motor speed adjustment amount ; represents the membership corresponding to the tth discrete point; This represents the speed adjustment of the wire feeding servo motor output after defuzzification at the k-th sampling time.

[0034] Optionally, based on the length error and the rate of change of error, the conversion coefficients between the fuzzy universe of discourse and the actual universe of discourse are recursively updated, including:

[0035] ;

[0036] In the formula The coefficients to be adjusted; Indicates the first Rate of change of length error at each sampling time Indicates the first The length error at each sampling time This represents the length error at the (k-1)th sampling time. Mapping the actual universe of discourse to a fuzzy universe of discourse. Mapping the fuzzy domain of discourse to the actual control domain; The sampling period.

[0037] Optionally, a long short-term memory network is used to perform time-series learning and correction on the initial control signal to obtain a corrected control signal, including:

[0038] The initial control signal of the fuzzy control output at the current moment is combined with the temporal characteristics of historical error and historical tension to construct the input vector of LSTM at the current moment.

[0039] LSTM cells learn the dynamic characteristics of the input vector through their internal gating mechanism and cell state.

[0040] The hidden state learned by the LSTM is linearly transformed through a fully connected output layer to map a new correction control signal predicted by the LSTM network.

[0041] Optionally, the final wire length command expression is as follows:

[0042] ;

[0043] In the formula, This indicates the final feed line length command at the k-th sampling time. Indicates the first k The theoretical wire feeding length of the wire feeding wheel at each sampling time; Indicates the first k The final speed adjustment of the feed servo motor at each sampling moment. This is the sampling period of the PLC; D The diameter of the wire feed wheel, i This is the transmission ratio.

[0044] According to a second aspect of the embodiments of the present application, an electronic device is provided, comprising:

[0045] one or more processors;

[0046] a memory for storing one or more programs;

[0047] When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to the first aspect.

[0048] The technical solutions provided by the embodiments of the present application can include the following beneficial effects:

[0049] The application proposes a fusion control technology of adaptive fuzzy control and LSTM time sequence learning, which is used for elastic wire fixed-length cutting. Specifically, the actual tension and length error of the wire are fuzzified, rule reasoning and defuzzification are performed, and an initial rotating speed control signal is generated. Then, an LSTM is used to learn the time sequence dependence relationship of "tension-error-control signal" to generate a correction signal. Finally, a mathematical model is used to convert the fused rotating speed adjustment amount into a wire feeding length compensation amount, forming a complete control closed loop.

[0050] The adaptive fuzzy control can adapt to the nonlinear and fuzzy characteristics of the elastic wire, the LSTM can solve the time sequence cumulative error problem, and the fixed-length accuracy is improved. At the same time, the actual cutting length does not need to be directly measured, and the control is realized only through indirect indicators such as tension and theoretical length, which reduces the hardware investment cost. The algorithm and mathematical model are clear and easy to implement on a controller such as PLC, and the control accuracy and engineering feasibility are considered, which stably improves the reliability of the elastic wire fixed-length cutting.

[0051] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the application. BRIEF DESCRIPTION OF DRAWINGS

[0052] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present application and, together with the specification, serve to explain the principles of the application.

[0053] Figure 1 A flowchart of a wire fixed-length cutting control method provided by the application.

[0054] Figure 2 A control structure diagram of a wire fixed-length cutting control method provided by the application.

[0055] Figure 3 A membership function image of the actual tension of the wire at the kth sampling time provided by the application. DETAILED DESCRIPTION

[0056] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application.

[0057] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0058] Figure 1 This is a flowchart illustrating a wire length cutting control method according to an exemplary embodiment. Figure 2 This is a schematic diagram of the control structure of a wire length cutting control method provided by the present invention, with reference to... Figure 1 and Figure 2 As shown, the method may include the following steps:

[0059] S1: Obtain the actual tension of the wire, the theoretical wire feeding length, and the speed adjustment of the wire feeding servo motor;

[0060] Specifically, the actual tension of the wire, the theoretical wire feeding length, and the speed adjustment of the wire feeding servo motor are obtained. The time-domain discrete sequence of each physical quantity is defined as follows (sampling period is...). , No. (Sampling time): Actual tension of the wire is The tension is directly measured by a tension sensor; the theoretical wire feed length is... ; For the first The encoder count of the feed wheel at each sampling time, where D is the diameter of the feed wheel. i The transmission ratio is [value]. The speed of the wire feeding servo motor is [value]. The speed adjustment of the wire feeding servo motor is as follows: .

[0061] S2: Calculate the length error based on the set cutting length and the theoretical wire feeding length;

[0062] Specifically, the length error is calculated using the following formula:

[0063] ;

[0064] In the formula, For the first The length error at each sampling time; To set the cutting length; For the first The theoretical feed length of the feed wheel at each sampling time.

[0065] S3: Using the actual tension and length error of the wire as input, and the speed adjustment of the wire feeding servo motor as output, adaptive fuzzy control is performed to generate an initial control signal;

[0066] Specifically, adaptive fuzzy control generates the initial control signal through three stages: fuzzification, fuzzy inference, and defuzzification. The adaptive fuzzy control includes:

[0067] A1: The input and output quantities are fuzzified and fuzzy inference is performed. After fuzzy inference, the fuzziness is defuzzified by the centroid method, and the speed adjustment of the wire feeding servo motor is output.

[0068] Specifically, the first step is to blur the input / output quantities, with the input quantities including the actual tension of the wire. and length error The output quantity, i.e., the initial control signal, is the speed adjustment amount of the wire feeding servo motor. .

[0069] Furthermore, the input and output quantities are fuzzified, including:

[0070] B2: Construction length error The fuzzy sets are respectively , express That is, the actual length Set the cutting length The number of cables to be supplied needs to be reduced; express The error is close to 0, and no adjustment is needed. express That is, the actual length Set cutting length Additional power lines are needed.

[0071] B3: Constructing the speed adjustment amount of the wire feeding servo motor Fuzzy sets , This indicates a reduction in motor speed and a decrease in wire length; This indicates that the rotation speed has been maintained and the error has been corrected. This indicates increasing the motor speed and increasing the wire length;

[0072] Based on the three fuzzy sets, their respective membership functions are constructed, and the membership mapping of the fuzzy sets adopts the triangular membership function.

[0073] Reference Figure 3 The membership mapping of the fuzzy set adopts a triangular membership function, and the actual tension of the wire The membership function is as follows:

[0074]

[0075]

[0076]

[0077] In the formula is a setting parameter of the membership function, which is calibrated according to the characteristics of the wire to be cut. Similarly, the membership of the length error is The membership function of the speed adjustment amount of the wire feeding servo motor is The universe and the parameter are calibrated according to the actual wire to be cut.

[0078] After the above fuzzy processing is completed, the next step of fuzzy reasoning is performed, and the fuzzy rule base of the fuzzy reasoning is as follows:

[0079] (1) When the tension is small, the wire is contracted, the actual length is prone to be short, and “feeding wire” needs to be considered first: when the tension is small and the actual length is short, the wire is accelerated, that is, ; when the tension is small and the error is extremely small, the current speed is maintained, that is, ; when the tension is small and the actual length is long, the wire is slightly decelerated, that is, .

[0080] (2) When the tension is medium, the deformation is stable, and the error is mainly caused by the fluctuation of the wire feeding accuracy, and the adjustment needs to be moderate: when the tension is medium and the actual length is short, the wire is moderately accelerated, that is, ; when the tension is medium and the error is extremely small, the current speed is maintained, that is, ; when the tension is medium and the actual length is long, the wire is moderately decelerated, that is, .

[0081] (3) When the tension is large, the wire is stretched, the actual length is prone to be long, and “decreasing the wire” needs to be considered first: when the tension is large and the actual length is short, it is not necessary to accelerate (avoid excessive wire feeding) due to the sudden contraction after stretching, and the tension is waited to be stable, that is, ; when the tension is large and the error is extremely small, the wire is slightly decelerated to prevent the actual length from being long due to subsequent stretching, that is, ; when the tension is large and the actual length is long, the wire is greatly decelerated to reduce the length of the wire and offset the error of the overlong length caused by stretching, that is, .

[0082] The fuzzy rule base above is inferred by the Min-Max composition method to output a fuzzy set, and for the rule Ri : ,in Its membership degree satisfies The final output fuzzy set membership degree is the maximum value of the implication results of all rules, that is... .

[0083] After completing the fuzzy inference, the precise compensation amount is calculated and defuzzified using the center of gravity (COG) method. The speed adjustment amount of the wire feeding servo motor is then calculated. The universe of discourse Y is discrete into m finite points, denoted as... The defuzzy output shows the speed adjustment of the wire feeding servo motor:

[0084] ;

[0085] In the formula This represents the speed adjustment amount of the wire feeding servo motor corresponding to the t-th discrete point; m is the speed adjustment amount of the wire feeding servo motor. The total number of discrete points in the universe of discourse; This represents the membership degree corresponding to the t-th discrete point; This represents the speed adjustment of the wire feeding servo motor output after defuzzification at the k-th sampling time.

[0086] A2: Based on the length error and the rate of change of error, recursively update the conversion coefficients between the fuzzy universe of discourse and the actual universe of discourse in real time;

[0087] Specifically, the conversion coefficients between the fuzzy universe of discourse and the actual universe of discourse are adjusted in real time based on the length error and the rate of change of the error. If the error is large or changes rapidly, K in and K out It will adaptively correct, making fuzzy control more sensitive to changes in "tension and error", adapting to the elastic characteristics of different wires, and achieving self-adaptation.

[0088] ;

[0089] In the formula The coefficients to be adjusted; Indicates the first Rate of change of length error at each sampling time; Mapping the actual domain of discourse to a fuzzy domain of discourse. Mapping the fuzzy domain to the actual control domain.

[0090] A3: Based on the speed adjustment amount and conversion coefficient of the output wire-feeding servo motor after defuzzification, the initial control signal of the final fuzzy control output is obtained.

[0091] Specifically, the initial control signal of the final fuzzy control output is... .

[0092] S4: The initial control signal is time-series learned and corrected using a long short-term memory network to obtain a corrected control signal; this step includes the following sub-steps:

[0093] S41: Combine the initial control signal of the fuzzy control output at the current moment with the temporal characteristics of historical error and historical tension to construct the input vector of LSTM at the current moment;

[0094] Specifically, no. k Input vector of LSTM at time step It includes the temporal characteristics of the fuzzy control initial signal, historical error, and historical tension, namely... In the formula, This indicates the initial speed adjustment amount of the fuzzy control output at the current moment; and They represent the first The length error at any given moment and the actual tension of the wire; d is the feature dimension, which is determined by the length of the historical time series.

[0095] S42: LSTM cells learn the dynamic characteristics of the input vector through their internal gating mechanism and cell state;

[0096] Specifically, LSTM cell operations capture temporal dependencies through input gates, forget gates, cell states, and output gates. Input gates This determines how much of the current input is stored in the cell state. The Sigmoid activation function outputs 0 to 1. The weights and biases of the input gate, The hidden state at time k-1 stores historical time sequence information. Forget gate. The symbol has the same meaning as the input gate; cell state update With cell state , It is a hyperbolic tangent activation function, with an output of -1 to 1, providing a nonlinear transformation for cell state; Element-wise multiplication involves the forget gate controlling the retention ratio of historical states and the input gate controlling the addition ratio of new states; the output gate... and hidden state .

[0097] S43: The hidden state output by the LSTM after learning is linearly transformed through a fully connected output layer to obtain a new correction control signal predicted by the LSTM network.

[0098] Specifically, the LSTM hidden state is mapped to the predicted value of the "wire feeding servo motor speed adjustment amount" through a fully connected layer, i.e. , The weight matrix and bias of the output layer, This represents the correction amount for the speed adjustment of the wire feeding servo motor predicted by LSTM, i.e., the correction control signal.

[0099] S5: The initial control signal and the modified control signal are weighted and fused to obtain the final control signal;

[0100] Specifically, the weighted fusion of the initial control signal of fuzzy control and the correction amount of LSTM yields the final control signal. , This is the weighting coefficient. If... This indicates the acceleration / deceleration magnitude of enhanced fuzzy control; if This indicates a reduction in the acceleration / deceleration magnitude of fuzzy control. It is a dynamic compensation for the initial signal of fuzzy control, which solves the problem of insufficient adaptation of fuzzy control to time-accumulated errors, and ultimately achieves... It outputs precise speed adjustment commands, enabling real-time feedback of fuzzy control and timing prediction correction of LSTM, thereby improving the fixed-length accuracy.

[0101] S6: Based on the final control signal, generate the final wire length instruction.

[0102] Specifically, the expression for the final wire length command is as follows:

[0103] ;

[0104] In the formula, This indicates the final wire length command at time k. Indicates the theoretical wire feeding length of the wire feeder; Indicates the first k The final adjustment amount of the wire feeding servo motor speed at any given moment. The PLC's acquisition cycle is the time interval between two data acquisitions. D The diameter of the wire feed wheel, i This is the transmission ratio.

[0105] To further improve the control accuracy and adaptability of fixed-length cutting of flexible wire, the optimization iteration steps are also included, specifically: updating the weights of the fuzzy control rules based on the new length error. Adjust the membership function parameters In the formula For learning rate, Let be the weight of the i-th rule. The weight parameters of the LSTM are updated using gradient descent to minimize the error prediction loss. Loss function. The LSTM weights are updated by back propagation.

[0106] From the above embodiment, the application proposes a fusion control technology of adaptive fuzzy control and LSTM time sequence learning, which is used for elastic wire fixed-length cutting. Specifically, the actual tension of the wire and the length error are fuzzified, rule reasoning and defuzzification are performed to generate an initial rotating speed control signal, then the LSTM is used to learn the time sequence dependence relationship of "tension-error-control signal" to generate a correction signal, and finally the mathematical model is used to convert the fused rotating speed adjustment amount into a wire feeding length compensation amount to form a complete control closed loop. The adaptive fuzzy control can adapt to the nonlinear and fuzzy characteristics of the elastic wire, the LSTM can solve the time sequence cumulative error problem, and the fixed-length accuracy is improved. At the same time, the actual cutting length does not need to be directly measured, and the control is realized only through indirect indicators such as tension and theoretical length which are easy to obtain, thereby reducing the hardware investment cost. The algorithm and mathematical model are clear and easy to land on the controller such as PLC, and the control accuracy and engineering feasibility are considered, thereby stably improving the reliability of the elastic wire fixed-length cutting.

[0107] Correspondingly, the application also provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the wire fixed-length cutting control method as described above.

[0108] Correspondingly, the application also provides a computer readable storage medium, which stores computer instructions, and the instructions are executed by a processor to implement the wire fixed-length cutting control method as described above.

[0109] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. The specification and examples given are considered exemplary only, and it is intended that the application be limited only by the claims.

[0110] It should be understood that the application is not limited to the precise construction that has been described above and shown in the accompanying drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the application.

Claims

1. A wire length cutting control method, characterized by, The method comprises the following steps: acquiring actual wire tension, theoretical wire feeding length and feeding motor speed adjustment amount; calculating length error according to set cutting length and the theoretical wire feeding length; performing adaptive fuzzy control on the actual wire tension and the length error as input quantities and the feeding motor speed adjustment amount as output quantity to generate an initial control signal; performing time sequence learning and correction on the initial control signal by using a long short-term memory network to obtain a corrected control signal; performing weighted fusion on the initial control signal and the corrected control signal to obtain a final control signal; forming a final wire feeding length instruction according to the final control signal.

2. The method of claim 1, wherein The method for calculating the length error according to the set cutting length and the theoretical wire feeding length comprises the following steps: The length error is calculated by the following formula: ； wherein is the length error at the th sampling instant; is the set cut length; is the theoretical wire feed length of the th sampling instant; ； wherein is the encoder count of the feed wheel at the is the encoder count of the feed wheel at the D is the feed wheel diameter, i is the gear ratio.

3. The method of claim 1, wherein The adaptive fuzzy control comprises the following steps: performing fuzzy processing and fuzzy reasoning on the input quantities and the output quantity, and outputting the feeding motor speed adjustment amount by defuzzification through the barycenter method after the fuzzy reasoning; recursively updating the conversion coefficient between the fuzzy domain and the actual domain according to the length error and the error change rate; obtaining the initial control signal of the final fuzzy control output according to the defuzzification output feeding motor speed adjustment amount and the conversion coefficient.

4. The method of claim 3, wherein The method for performing fuzzy processing on the input quantities and the output quantity comprises the following steps: Constructing actual tension of wire Fuzzy set of actual tension of wire , respectively, represent small, medium, large degree of actual tension of wire, is the first sampling time Construction length error The fuzzy set of the error is , The error is , i.e. the actual length < the set cutting length , the wire feeding needs to be reduced; The error is , close to 0, no adjustment is needed; The error is , i.e. the actual length > the set cutting length , the wire feeding needs to be increased; Construction of fuzzy set of speed adjustment amount of wire feeding servo motor , indicates to reduce the motor speed and reduce the wire feeding length; indicates to maintain the speed and the error has been corrected; indicates to increase the motor speed and increase the wire feeding length;​ corresponding membership functions are constructed according to three fuzzy sets, and the membership mapping of the fuzzy set adopts a triangular membership function.

5. The method of claim 4, wherein The fuzzy rule base of the fuzzy reasoning is as follows: （1） ； ； ； （2） ； ； ； （3） ； ； ； The fuzzy rule base above outputs a fuzzy set by Min-Max composition method, and for the rule R i : where whose membership satisfies The membership of the final output fuzzy set is the maximum of the membership of all rule implication results, i.e. .

6. The method of claim 3, wherein The defuzzification output feeding motor speed adjustment amount is as follows: ； In the formula represents the speed adjustment amount of the wire feeding servo motor corresponding to the kth discrete point; t m represents the speed adjustment amount of the wire feeding servo motor corresponding to the kth discrete point; represents the total number of discrete points in the domain of the speed adjustment amount of the wire feeding servo motor; represents the speed adjustment amount of the wire feeding servo motor corresponding to the kth discrete point; t represents the membership degree corresponding to the kth discrete point; represents the speed adjustment amount of the wire feeding servo motor after demisting at the kth sampling time.​ 7. The method of claim 3, wherein The method for recursively updating the conversion coefficient between the fuzzy domain and the actual domain according to the length error and the error change rate comprises the following steps: ； wherein is the coefficient to be tuned; denotes the length error rate of the kth sampling instant, denotes the length error of the kth sampling instant, denotes the length error of the k-1th sampling instant; maps the actual universe of discourse to the fuzzy universe of discourse, maps the fuzzy universe of discourse to the actual control universe of discourse; is the sampling period.

8. The method of claim 1, wherein The method for performing time sequence learning and correction on the initial control signal by using a long short-term memory network to obtain a corrected control signal comprises the following steps: combining the initial control signal output by the fuzzy control at the current time with the time sequence characteristics of the historical error and the historical tension to jointly construct an input vector of the LSTM at the current time; learning the dynamic characteristics of the input vector by the LSTM cell through its internal gating mechanism and cell state; performing linear transformation on the hidden state output by the LSTM learning through a fully connected output layer to map a new corrected control signal predicted by the LSTM network.

9. The method of claim 1, wherein The final wire feeding length instruction expression is as follows: ； In the formula, Indicates the first k The final wire length instruction at each sampling time; Indicates the first k The theoretical wire feeding length of the wire feeding wheel at each sampling time; Indicates the first k The final speed adjustment of the feed servo motor at each sampling moment. This is the sampling period of the PLC; D The diameter of the wire feed wheel, i This is the transmission ratio.

10. An electronic device, comprising: The device comprises the following components: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1-9.

Images