A Dynamic Thermal Assisted Method for Fuzzy PID Control Algorithm Based on Time Series

Abstract

The present invention discloses a dynamic thermal assistance method based on a fuzzy PID control algorithm for time series. Temperature data in the drilling and riveting area is collected in real time through an infrared sensor. Based on a fuzzy rule base, the temperature deviation and its rate of change are mapped to the adjustment weights of the PID parameters, solving the problem of parameter mismatch of traditional PID in nonlinear systems. Process parameter database matching: An association rule base between material combinations such as metal-composite materials and preset temperature thresholds is established. Combining the thermal physical properties of materials, such as thermal conductivity and melting point, the temperature control target is dynamically set to achieve differential regulation of dissimilar materials. By constructing a temperature-vibration coupling model based on an LSTM neural network, the internal relationship between temperature and vibration is accurately coordinated. The vibration disturbance caused during the heating process is accurately obtained, and the dynamic adjustment of the temperature and vibration during the drilling and riveting process is achieved by fine-tuning the power of the active damper.

Description

Technical Field

[0001] The present invention relates to the technical field of automated processing and intelligent control, and particularly relates to a dynamic thermal assistance method based on a fuzzy PID control algorithm for time series. Background Art

[0002] Flow drill riveting technology, as an efficient process for connecting dissimilar materials, is widely used in fields such as aerospace and automotive manufacturing. However, in traditional flow drill riveting processes, the accuracy of heat source temperature control directly affects the riveting quality and material properties. In existing technologies, PID control algorithms are mostly used to adjust the electromagnetic induction heating power for thermal-assisted drill riveting, but there are the following problems:

[0003] 1. Poor temperature control adaptability: The parameters of traditional PID control are fixed, making it difficult to cope with the non-linear and time-varying temperature fluctuations caused by factors such as material heat capacity differences and environmental heat losses during drill riveting. Overshoot or response lag is likely to occur, resulting in uneven rivet forming or thermal damage to the base material, such as local ablation of composite materials.

[0004] 2. Insufficient compatibility of dissimilar materials: The thermal expansion coefficients of metals and composite materials differ significantly. Existing methods lack differential temperature control strategies for different material combinations, leading to cross-sectional stress concentration and a decrease in riveting strength.

[0005] 3. High energy consumption and reliance on manual experience: Traditional processes require setting heating power thresholds through repeated trial and error, lacking a dynamic optimization mechanism based on real-time data and unable to adaptively match process parameters, resulting in energy waste.

[0006] Furthermore, in flow drill riveting technology, a rotating drill bit penetrates the upper and lower plates to form a riveting hole, and then a carbon steel rivet is plastically deformed under mechanical force to tap and fasten between the two plates, with the screw head seated on the upper plate, thus achieving the connection of dissimilar materials. However, during the high-speed and high-pressure drill riveting process, the dynamic coupling effect of temperature and vibration factors still has a significant impact on process accuracy and material properties, specifically manifested as the following technical problems:

[0007] In the drill riveting process, the interaction between temperature and vibration directly affects the processing quality: Temperature effect: The instantaneous temperature generated by friction between the carbon steel screw head and the plate can reach 300 - 500 °C. Although the melting point of the carbon steel rivet > 1400 °C will not soften, high temperatures will cause the problem of thermal stress concentration, specifically manifested as thermal stress concentration during plate connection. At high rotational speeds, frictional heat generation will inevitably result in differences in the thermal expansion coefficients of metal substrates such as high-strength steel and carbon steel rivets, leading to a peak interfacial residual stress > 150 MPa.

[0008] Traditional processes adopt a separate temperature and vibration control strategy, but do not address the inherent correlation defect between the two: a single temperature control such as the PID algorithm only compensates for the substrate temperature by adjusting the heating power, but ignores the following problems: vibration causes the infrared temperature sensor to shift by 0.5 - 1 mm, and the temperature measurement error reaches ±15 °C; carbon steel rivets have a high thermal conductivity of 50 W / m·K, and heat quickly diffuses to the substrate, exacerbating the thermal damage of the composite material; fixed PID parameters cannot adapt to dynamic thermal load changes such as the friction heat power during drilling > 2 kW, and only 0.8 kW is required during riveting, resulting in a temperature overshoot > 10% or a response lag with an adjustment time > 6 s. Summary of the Invention

[0009] In view of the technical problems in the prior art, the present invention provides a flow drill riveting thermal assistance method based on a fuzzy PID control algorithm, including the following steps:

[0010] Step S1: Real-time collect temperature data of the drill riveting processing area;

[0011] Step S2: Compare the temperature data with the preset current material combination temperature threshold in the process parameter database, and calculate the temperature deviation value and its change rate;

[0012] Step S3: Construct a fuzzy inference rule base, and dynamically tune the PID parameters of the PID controller based on the fuzzy rule base according to the temperature deviation value and its change rate. The PID parameters include the proportional coefficient , integral coefficient and differential coefficient ;

[0013] Step S4: Generate a control signal according to the tuned PID parameters, and adjust the output power of the electromagnetic induction heating device, so as to perform dynamic temperature control on the drill riveting processing area;

[0014] During the heating-up process of step S4, construct a temperature-vibration coupling model based on the LSTM neural network, obtain historical temperature data and historical vibration data, obtain the heating-up sequence in the historical temperature data and the corresponding heating-up vibration sequence in the historical vibration data, calculate the cross-correlation to determine the effective time lag, use the heating-up sequence as the input, and the heating-up vibration sequence after the effective time lag as the label to train the temperature-vibration coupling model to obtain an optimized temperature-vibration coupling model;

[0015] Import the heating-up sequence in step S4 into the temperature-vibration coupling model based on the LSTM neural network to obtain the predicted heating-up vibration sequence; dynamically adjust the control force of the active damper according to the data of the heating-up vibration sequence;

[0016] Step S5: Based on Steps S1 to S4, dynamically adjust the temperature and vibration of the drilling and riveting processing area, obtain the riveting force curve during each drilling and riveting process, and optimize the preset parameters of the fuzzy inference rule base by analyzing the stability of the riveting force curve.

[0017] Further, the method for establishing the fuzzy inference rule base in Step S3 is specifically as follows:

[0018] Define the fuzzy set of the temperature deviation value as {negative large, negative small, zero, positive small, positive large}, and the fuzzy set of the change rate as {negative fast, negative slow, zero, positive slow, positive fast};

[0019] Construct a fuzzy rule table to map the adjustment weights of the temperature deviation value and its change rate to the PID parameters. The rule form is:

[0020] If e is A and Δe is B, then Δ = C, Δ = D, Δ = E, where A and B are input fuzzy sets, and C, D, and E are output fuzzy sets;

[0021] Among them, represents the temperature deviation value, Δe represents the change rate of the temperature deviation value, Δ represents the change value of the proportional coefficient, Δ represents the change value of the integral coefficient, Δ represents the change value of the differential coefficient.

[0022] Further, the process parameter database includes:

[0023] The preset temperature thresholds corresponding to different material combinations;

[0024] The material combinations include at least two categories: metal-metal and metal-composite material;

[0025] The thermophysical parameters of each material, including the thermal conductivity, melting point, and allowable maximum heat affected zone width;

[0026] The mapping relationship table between historical riveting quality data and temperature control parameters.

[0027] Further, during the process of dynamically tuning the PID parameters in Step S3:

[0028] The adjustment weight of the proportional coefficient is positively correlated with the absolute value of the temperature deviation value;

[0029] The adjustment weight of the integral coefficient is negatively correlated with the absolute value of the change rate of the temperature deviation value;

[0030] The adjustment weight of the differential coefficient is positively correlated with the product of the temperature deviation value and the change rate.

[0031] Further, the power adjustment of the electromagnetic induction heating device adopts PWM pulse width modulation technology, and the corresponding relationship between the power output and the PID control signal satisfies:

[0032] ;

[0033] where, is the heating power at time t, represents the temperature deviation value at time t.

[0034] Further, the specific process of obtaining the temperature rise sequence in the historical temperature data and the corresponding vibration rise sequence in the historical vibration data, and calculating the cross-correlation to determine the effective time delay is as follows:

[0035] Obtain the cross-correlation coefficient between the temperature rise sequence and the vibration rise sequence, expressed as:

[0036]

[0037] where, represents the cross-correlation coefficient between the temperature rise sequence and the vibration rise sequence at time lag , is the temperature of the temperature rise sequence at time, is the amplitude of the vibration rise sequence at time, is the mean value of the temperature rise sequence, represents the mean value of the vibration rise sequence, is the length of the time series, is the time lag, representing the number of delay steps of the vibration rise sequence relative to the temperature rise sequence;

[0038] Take the time lag corresponding to the value greater than the preset threshold as the effective time delay, that is .

[0039] Further, step S3 is specifically:

[0040] Input variable fuzzification:

[0041] Obtain the temperature deviation value Determine the corresponding universe of discourse, and use the triangular membership function to divide the corresponding fuzzy set into {negative large , negative small , zero , positive small , positive large };

[0042] Obtain the change rate of the temperature deviation value Determine the corresponding universe of discourse, and use the Gaussian membership function to divide the corresponding fuzzy set into {negative fast , negative slow , zero , positive slow , positive fast };

[0043] Construct a fuzzy inference library based on the fuzzy inference table corresponding to the temperature deviation value and its change rate;

[0044] Input the real-time temperature deviation value and the change rate of the temperature deviation value , and obtain the fuzzy output of the corresponding PID parameters based on the fuzzy inference library;

[0045] Defuzzification: Use the center of gravity method COG to convert the fuzzy output into an accurate PID parameter adjustment amount, thereby completing the tuning of the PID parameters of the dynamic tuning PID controller.

[0046] Furthermore, the process flow of drilling and riveting successively includes a drilling period, a riveting period, and a tool retraction period, and the change rate of the temperature deviation value in the processing area during the drilling period, the riveting period, and the tool retraction period Determine that the corresponding universes of discourse are different;

[0047] And construct a process-oriented fuzzy rule library based on this;

[0048] Temperature deviation value Determine the corresponding universe of discourse based on the melting point of the drilling and riveting material.

[0049] The positive progressive effect of the present invention lies in:

[0050] 1) Real-time collect the temperature data of the drilling and riveting area through an infrared sensor, and map the temperature deviation and its change rate to the adjustment weight of the PID parameters based on the fuzzy rule library, so as to solve the problem of parameter mismatch of traditional PID in non-linear systems.

[0051] 2) Process parameter database matching: Establish an association rule library between material combinations such as metal-composite materials and preset temperature thresholds, and dynamically set the temperature control target in combination with the thermal physical properties of the materials, such as thermal conductivity and melting point, to achieve differential control of dissimilar materials.

[0052] 3) The present invention accurately coordinates and controls the internal relationship between temperature and vibration by constructing a temperature-vibration coupling model based on the LSTM neural network, accurately obtains the vibration disturbance brought during the heating process, and realizes the dynamic adjustment of the temperature and vibration during the drilling and riveting process by fine-tuning the power of the active damper. Description of the Drawings

[0053] Figure 1 is the step flow chart of a flow drilling and riveting thermal assistance method based on the fuzzy PID control algorithm of the present invention. Detailed Embodiment

[0054] The following specific examples illustrate the implementation manners of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0055] A flow drill riveting thermal assistance method based on a fuzzy PID control algorithm includes the following steps:

[0056] Step S1: Collect temperature data of the drill riveting processing area in real time; specifically, temperature data collection and preprocessing collect temperature signals of the drill riveting area in real time through an infrared temperature sensor, and the sampling frequency is set to 100 Hz; perform Kalman filtering on the original temperature data to eliminate noise interference and ensure that the temperature error input to the controller ≤ ±0.5 °C;

[0057] Step S2: Compare the temperature data with the preset current material combination temperature threshold in the process parameter database, and calculate the temperature deviation value and its change rate;

[0058] Step S3: Construct a fuzzy inference rule base, and dynamically tune the PID parameters of the PID controller based on the fuzzy rule base according to the temperature deviation value and its change rate. The PID parameters include the proportional coefficient , integral coefficient , and differential coefficient ;

[0059] Step S4: Generate a control signal according to the tuned PID parameters, and adjust the output power of the electromagnetic induction heating device, so as to perform dynamic temperature control on the drill riveting processing area;

[0060] During the heating-up process of step S4, construct a temperature-vibration coupling model based on an LSTM neural network, obtain historical temperature data and historical vibration data, obtain the heating-up sequence in the historical temperature data and the corresponding heating-up vibration sequence in the historical vibration data, calculate the cross-correlation to determine the effective time lag, use the heating-up sequence as the input, and use the heating-up vibration sequence after the effective time lag as the label to train the temperature-vibration coupling model to obtain an optimized temperature-vibration coupling model;

[0061] Import the heating-up sequence in step S4 into the temperature-vibration coupling model based on the LSTM neural network to obtain the predicted heating-up vibration sequence; dynamically adjust the control force of the active damper according to the data of the heating-up vibration sequence;

[0062] Step S5: Based on Steps S1 to S4, dynamically adjust the temperature and vibration of the drilling and riveting processing area, obtain the riveting force curve during each drilling and riveting process, and optimize the preset parameters of the fuzzy inference rule base by analyzing the stability of the riveting force curve.

[0063] Further, the method for establishing the fuzzy inference rule base in Step S3 is specifically as follows:

[0064] Define the fuzzy set of the temperature deviation value as {negative large, negative small, zero, positive small, positive large}, and the fuzzy set of the change rate as {negative fast, negative slow, zero, positive slow, positive fast};

[0065] Construct a fuzzy rule table to map the adjustment weights of the temperature deviation value and its change rate to the PID parameters. The rule form is:

[0066] If e is A and Δe is B, then Δ = C, Δ = D, Δ = E, where A and B are input fuzzy sets, and C, D, and E are output fuzzy sets;

[0067] Among them, represents the temperature deviation value, Δe represents the change rate of the temperature deviation value, Δ represents the change value of the proportional coefficient, Δ represents the change value of the integral coefficient, Δ represents the change value of the differential coefficient.

[0068] Further, the process parameter database includes:

[0069] The preset temperature thresholds corresponding to different material combinations;

[0070] The material combinations include at least two categories: metal-metal and metal-composite material;

[0071] The thermophysical parameters of each material, including thermal conductivity, melting point, and allowable maximum heat-affected zone width;

[0072] The mapping relationship table between historical riveting quality data and temperature control parameters.

[0073] Further, during the process of dynamically tuning the PID parameters in Step S3:

[0074] The adjustment weight of the proportional coefficient is positively correlated with the absolute value of the temperature deviation value;

[0075] The adjustment weight of the integral coefficient is negatively correlated with the absolute value of the change rate of the temperature deviation value;

[0076] The adjustment weight of the differential coefficient is positively correlated with the product of the temperature deviation value and the change rate.

[0077] Further, the power regulation of the electromagnetic induction heating device adopts PWM pulse width modulation technology, and the correspondence between the power output and the PID control signal satisfies:

[0078] ;

[0079] where, is the heating power at time t, represents the temperature deviation value at time t.

[0080] Further, the specific process of obtaining the heating-up sequence in the historical temperature data and the corresponding vibration-up sequence in the historical vibration data and calculating the cross-correlation to determine the effective time lag is as follows:

[0081] Obtain the cross-correlation coefficient between the heating-up sequence and the vibration-up sequence, expressed as:

[0082]

[0083] where, represents the cross-correlation coefficient between the heating-up sequence and the vibration-up sequence at time lag , is the temperature of the heating-up sequence at time, is the amplitude of the vibration-up sequence at time, is the mean value of the heating-up sequence, represents the mean value of the vibration-up sequence, is the length of the time series, is the time lag, representing the number of delay steps of the vibration-up sequence relative to the heating-up sequence;

[0084] Take the time lag corresponding to greater than the preset threshold as the effective time lag, that is, .

[0085] Further, step S3 is specifically as follows:

[0086] Input variable fuzzification:

[0087] Obtain the temperature deviation value Determine the corresponding universe of discourse, and divide the corresponding fuzzy set into {negative large , negative small , zero , positive small , positive large } by using the triangular membership function;

[0088] Obtain the change rate of the temperature deviation value Determine the corresponding universe of discourse, and divide the corresponding fuzzy set into {negative fast , negative slow , zero , positive slow , positive fast };

[0089] Construct a fuzzy inference library based on the fuzzy inference table corresponding to the temperature deviation value and its change rate; expressed as:

[0090] Table 1 Example of some rules

[0091]

[0092] Input the real-time temperature deviation value and the change rate of the temperature deviation value , and obtain the fuzzy output of the corresponding PID parameters based on the fuzzy inference library;

[0093] Defuzzification: Use the center of gravity method COG to convert the fuzzy output into an accurate PID parameter adjustment amount, so as to complete the tuning of the PID parameters of the dynamic tuning PID controller.

[0094] Furthermore, the drilling and riveting process flow successively includes a drilling period, a riveting period, and a tool withdrawal period, and the change rate of the temperature deviation value in the processing area during the drilling period, the riveting period, and the tool withdrawal period Determine that the corresponding universe of discourse is different;

[0095] Refer to Table 2 and construct a fuzzy rule library based on process orientation based on this;

[0096] Table 2

[0097]

[0098] Temperature deviation value Determine the corresponding universe of discourse based on the melting point of the drilling and riveting material.

[0099] The above has described the present invention in detail in conjunction with the accompanying drawings and embodiments. Those of ordinary skill in the art can make various variations to the present invention according to the above description. Therefore, certain details in the embodiments should not constitute a limitation to the present invention, and the present invention will take the scope defined by the appended claims as the protection scope.

Claims

1. A dynamic thermal assistance method for a fuzzy PID control algorithm based on time series, characterized in that It includes the following steps: Step S1: Collect the temperature data of the drilling and riveting processing area in real time; Step S2: Compare the temperature data with the preset temperature threshold of the current material combination in the process parameter database, and calculate the temperature deviation value and its change rate; Step S3: Construct a fuzzy inference rule base, and dynamically tune the PID parameters of the PID controller based on the temperature deviation value and its change rate according to the fuzzy rule base. The PID parameters include the proportional coefficient , the integral coefficient , and the derivative coefficient ; Step S4: Generate a control signal according to the tuned PID parameters, and adjust the output power of the electromagnetic induction heating device, so as to perform dynamic temperature control on the drilling and riveting processing area; In the heating-up process of step S4, construct a temperature-vibration coupling model based on the LSTM neural network, obtain historical temperature data and historical vibration data, obtain the heating-up sequence in the historical temperature data and the corresponding heating-up vibration sequence in the historical vibration data, calculate the cross-correlation to determine the effective time delay, use the heating-up sequence as the input, and use the heating-up vibration sequence after the effective time delay as the label to train the temperature-vibration coupling model to obtain an optimized temperature-vibration coupling model; Import the heating-up sequence in step S4 into the temperature-vibration coupling model based on the LSTM neural network to obtain the predicted heating-up vibration sequence; dynamically adjust the control force of the active damper according to the data of the heating-up vibration sequence; Step S5: Based on steps S1 to S4, perform dynamic adjustment of temperature and vibration on the drilling and riveting processing area, obtain the riveting force curve in each drilling and riveting process, and optimize the preset parameters of the fuzzy inference rule base by analyzing the stability of the riveting force curve.

2. A dynamic thermal assistance method for a fuzzy PID control algorithm based on time series according to claim 1, characterized in that The specific method for establishing the fuzzy inference rule base in step S3 is as follows: Define the fuzzy set of the temperature deviation value as {negative large, negative small, zero, positive small, positive large}, and the fuzzy set of the change rate as {negative fast, negative slow, zero, positive slow, positive fast}; Construct a fuzzy rule table to map the adjustment weights of the temperature deviation value and its change rate to the PID parameters, and the rule form is: If e is A and Δe is B, then Δ = C, Δ = D, Δ = E, where A and B are input fuzzy sets, and C, D, and E are output fuzzy sets; Among them, represents the temperature deviation value, Δe represents the change rate of the temperature deviation value, Δ represents the change value of the proportionality coefficient, Δ represents the change value of the integral coefficient, Δ represents the change value of the differential coefficient.

3. A dynamic thermal assistance method based on a time series fuzzy PID control algorithm according to claim 1, characterized in that The process parameter database includes: The preset temperature threshold corresponding to different material combinations; The material combinations include at least two categories: metal-metal and metal-composite material; The thermophysical parameters of each material, including thermal conductivity, melting point and allowable maximum heat affected zone width; The mapping relationship table between historical riveting quality data and temperature control parameters.

4. A dynamic thermal assistance method based on a time series fuzzy PID control algorithm according to claim 1, characterized in that In the process of dynamically tuning the PID parameters in step S3: The adjustment weight of the proportional coefficient is positively correlated with the absolute value of the temperature deviation value; The adjustment weight of the integral coefficient is negatively correlated with the absolute value of the change rate of the temperature deviation value; The adjustment weight of the differential coefficient is positively correlated with the product of the temperature deviation value and the change rate.

5. A dynamic thermal assistance method based on a fuzzy PID control algorithm for time series according to claim 1, characterized in that The power adjustment of the electromagnetic induction heating device adopts PWM pulse width modulation technology, and the corresponding relationship between the power output and the PID control signal satisfies: ； Among them, is the heating power at time t, represents the temperature deviation value at time t.

6. A dynamic thermal assistance method based on a fuzzy PID control algorithm for time series according to claim 1, characterized in that The specific process of obtaining the heating-up sequence in the historical temperature data and the corresponding heating-up vibration sequence in the historical vibration data, and calculating the cross-correlation to determine the effective time delay is: Obtain the cross-correlation coefficient between the heating-up sequence and the heating-up vibration sequence, expressed as: Among them, represents the cross-correlation coefficient between the temperature increase sequence and the vibration increase sequence at a time lag ; is the temperature of the temperature increase sequence at moment; is the amplitude of the vibration increase sequence at moment; is the mean value of the temperature increase sequence, represents the mean value of the vibration increase sequence, is the length of the time series, is the time lag, indicating the number of steps of delay of the vibration increase sequence relative to the temperature increase sequence; Take the time lag corresponding to greater than the preset threshold as the effective time lag, that is .

7. A dynamic thermal assistance method for a fuzzy PID control algorithm based on time series according to claim 1, characterized in that Step S3 is specifically: Input variable fuzzyfication: Obtain the temperature deviation value Determine the corresponding universe of discourse, and use the triangular membership function to divide the corresponding fuzzy set into {negative large , negative small , zero , positive small , positive large }; Obtain the change rate of the temperature deviation value Determine the corresponding universe of discourse, and use a Gaussian membership function to divide the corresponding fuzzy set into {negative fast , negative slow , zero , positive slow , positive fast }; Construct a fuzzy inference library based on the fuzzy inference table corresponding to the temperature deviation value and its change rate; Input the real-time temperature deviation value and the change rate of the temperature deviation value , and obtain the fuzzy output of the corresponding PID parameters based on the fuzzy inference library; Defuzzification: The center of gravity method (COG) is used to convert the fuzzy output into an accurate PID parameter adjustment amount, thereby completing the tuning of the PID parameters of the dynamic tuning PID controller.

8. A dynamic thermal assistance method based on a time series fuzzy PID control algorithm according to claim 7, characterized in that The technological process of drilling and riveting successively includes a drilling stage, a riveting stage, and a tool withdrawal stage, and the change rate of the temperature deviation value in the processing area during the drilling stage, the riveting stage, and the tool withdrawal stage The corresponding universes of discourse are different; And based on this, a fuzzy rule base based on process orientation is constructed; Temperature deviation value Determine the corresponding universe of discourse based on the melting point of the drill riveting material.

Images