Three-dimensional printer temperature control method with learning mechanism

Abstract

The invention relates to a three-dimensional printer temperature control method with a learning mechanism. The method comprises an initialization step; a nozzle temperature calculation step based on an estimation system model, external unknown interference estimation and an unknown weight estimation coefficient; judging an estimation system model; an external unknown interference estimation judgment step; an unknown weight estimation coefficient error judgment step; a spray head temperature error judgment step; and judging the number of iterative learning times and outputting the system temperature. The model-free control method with the learning mechanism is used for temperature control of the three-dimensional printer, the problem that an accurate system model is difficult to obtain can be solved, and the convergence speed and robustness of the system are improved on the basis of existing performance.

Description

Technical Field

[0001] The present invention relates to the technical field of temperature control, and more specifically, to a temperature control method of a three-dimensional printer with a learning mechanism. Background Art

[0002] With the rapid development of 3D printing technology, the demand for precision and efficiency of fused deposition modeling printers continues to rise. Figure 1 As shown in Figure 1, temperature control is the core link in the fused deposition modeling printing process, which directly determines the printing quality and process stability. The stability and accuracy of the nozzle temperature are crucial to ensure the continuity and quality of the printing process. The main challenges faced by nozzle temperature control include: (1) high temperature change rate, especially in the heating stage, the set temperature needs to be reached quickly; (2) maintaining the stable state of the melt during the printing process to avoid temperature fluctuations; (3) the traditional PID control strategy is not adaptive enough, which can easily lead to overshoot and oscillation. In response to these challenges, existing improvement methods include fuzzy PID control, experimental verification of parameter extraction, and integration of system circuits. Fuzzy PID control combines the adaptability of fuzzy control with the accuracy of PID control to improve the stability and response speed of the system. Fuzzy control can handle imprecise or uncertain information, and dynamically adjust PID parameters to adapt to changing printing conditions through reasoning through a fuzzy rule base. However, this method requires the construction of a set of fuzzy rule tables, and changes in the external environment may cause the rule table to be inaccurate or even invalid. The experimental verification parameter extraction method determines the PID parameters that are most suitable for specific printing materials and environments through experiments and simulations, but this method requires a lot of experimental verification and has a limited scope of application. The system circuit integration method integrates the temperature control system with other printer components (such as wire feeding speed and printing speed) to achieve collaborative control, thereby improving the coordination and efficiency of the entire printing process. However, this method is relatively complex in system design and has a high degree of design difficulty.

[0003] In the research field of 3D printing technology, temperature control is a core element to ensure printing quality and efficiency. Although current research focuses on integrated control strategies based on system models, the effectiveness of these strategies is highly dependent on the accuracy of the model, which is often difficult to guarantee in practical applications. The accuracy of the model is limited by many factors such as material properties, environmental conditions, and the mechanical structure of the printer. In different operating environments, the internal parameters of the control system may change significantly, affecting the accuracy of the model. At the same time, due to the nonlinear characteristics of the temperature control process, the system has low robustness to external interference (such as noise, magnetic fields, or temperature fluctuations). The combination of these factors leads to insufficient stability and reliability of the designed control system in actual operation.

[0004] The sensitivity of the nozzle resistance to temperature causes the control system to exhibit strong nonlinearity and uncertainty, making the construction of the system model extremely complex. Compared with traditional model-based 3D printer temperature controllers, model-free control methods, especially those with learning mechanisms, are more suitable for temperature control. These methods do not need to rely on accurate system models and can be dynamically adjusted according to the real-time environment of the system. Summary of the invention

[0005] The main purpose of the present invention is to overcome the above-mentioned defects in the prior art and provide a 3D printer temperature control method with a learning mechanism, which can realize the temperature control of the 3D printer without an accurate system model, thereby improving the convergence speed and robustness.

[0006] The present invention adopts the following technical solution:

[0007] A temperature control method for a three-dimensional printer with a learning mechanism, comprising:

[0008] S1, initialization step;

[0009] Set the 3D printer temperature control system to output the ideal temperature value n to the nozzle at time t d (t), t={0,1,2,3,…,T m}, T m is the maximum running time; detect the nozzle temperature value m(t) at time t, set m(t) = m1(t) = m0(t), m0(t) represents the nozzle temperature value at time t in the initial iteration; m1(t) represents the nozzle temperature value at time t in the first iteration; set the maximum number of iterative learning to j max ; Set the number of iterations j to 2;

[0010] S2, nozzle temperature calculation step;

[0011] Calculate the nozzle temperature m at time t during the jth iteration learning j (t), as follows:

[0012]

[0013] Among them, m j-1 (t) represents the nozzle temperature at time t during the j-1th iteration learning; Λ represents the weight factor of the 3D printer temperature control system input, Λ>0; δ t is the search step length, and 0<δ t <1; is the system model for estimating the unknown 3D printer temperature control system θ(t); n d (t+1) represents the ideal nozzle temperature value at time t+1; n j-1(t+1) represents the system output temperature value detected at time t+1 during the j-1th iterative learning; and They represent the estimation of the external unknown disturbance s(t) at time t during the j-1th and jth iterative learning respectively; w j (t) represents the unknown weight estimation coefficient of the external unknown interference s(t) at time t during the j-th iterative learning;

[0014] At time t during the j-th iteration of learning It is expressed as follows:

[0015]

[0016]

[0017] in, represents the estimated system model at time t in the j-1th iteration; ε t is the search step length and 0<ε t <1; η represents the weight factor of the input estimation of the 3D printer temperature control system, η>0; m j-2 (t) represents the nozzle temperature at time t during the j-2th iteration learning; w j-1 (t) represents the unknown weight estimation coefficient of the external unknown interference s(t) at time t during the j-1th iterative learning; n j-2 (t+1) is the system output temperature value at time t+1 during the j-2th iterative learning; It represents the estimation of the external unknown interference s(t) at time t during the j-2th iterative learning;

[0018] At time t during the j-th iteration of learning It is expressed as follows:

[0019]

[0020] Among them, γ t is the search step size and 0<γ t <1; w j-2 (t) represents the unknown weight estimation coefficient of the external unknown disturbance s(t) at time t during the j-2th iterative learning; σ represents the weight factor of the unknown weight estimation of the 3D printer temperature control system, σ>0;

[0021] w at time t during the j-th iteration of learning j (t), expressed as follows:

[0022]

[0023] Among them, β t is the search step length and 0<β t<1; α represents the weight factor of the external unknown disturbance estimation of the 3D printer temperature control system, α>0;

[0024] S3, estimation system model judgment step;

[0025] judge or Is it true? If so, execute Otherwise, return to S2; where κ is the system estimated reset coefficient, sign(·) is the sign function;

[0026] S4, external unknown interference estimation and judgment step;

[0027] judge or Is it true? If so, execute Otherwise, return to S2; where ξ is the external interference reset coefficient;

[0028] S5, a step of judging the error of the unknown weight estimation coefficient of the external unknown interference;

[0029] Judgment|w j (t)-w j-1 (t)|≤ξ2 holds true. If so, execute w j (t) = w1(t), otherwise, return to S2; where ξ2 is the external interference weight reset coefficient;

[0030] S6, nozzle temperature error judgment step;

[0031] Judgement|m j (t)-m j-1 (t)|≤κ2 is true, if true, execute m j (t) = m1(t), otherwise, return to S2; where κ2 is the system input reset coefficient;

[0032] S7, iterative learning times determination and system temperature output step;

[0033] Execute j=j+1, and determine j<j max Is it true? If so, repeat S2 to S7; otherwise, stop the operation and output the jth detected max The temperature value output by the 3D printer temperature control system at time t during the iterative learning

[0034] It can be seen from the above description of the present invention that, compared with the prior art, the present invention has the following beneficial effects:

[0035] (1) The present invention provides a 3D printer temperature control method with a learning mechanism, based on a system model for estimating an unknown 3D printer temperature control system. Estimation of external unknown interference And the unknown weight estimation coefficient w of the external unknown interference j (t), iteratively calculate the system input, i.e. the nozzle temperature m j (t), and then detect the system output temperature n corresponding to the system input temperature j (t+1), and when the number of iterations is reached, the temperature value output by the temperature control system of the 3D printer is detected to achieve temperature control of the 3D printer;

[0036] (2) The method of the present invention enters the next iteration only when the estimated system model, the external unknown interference estimation and the unknown weight estimation coefficients of the external unknown interference all meet the preset conditions, and has good convergence speed and robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the structure of the temperature control of the 3D printer;

[0038] Figure 2 It is a structural block diagram of the temperature control of a three-dimensional printer according to an embodiment of the present invention;

[0039] Figure 3 The figure is a flow chart of a temperature control method of a 3D printer with a learning mechanism according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The present invention is further described below through specific implementation modes.

[0041] like Figure 2 and 3 As shown, the present invention provides a temperature control method for a three-dimensional printer with a learning mechanism, comprising the following steps.

[0042] S1, initialization step.

[0043] Set the 3D printer temperature control system to output the ideal temperature value n to the nozzle at time t d (t), t={0,1,2,3,…,T m}, T m is the maximum running time; detect the nozzle temperature value m(t) at time t, set m(t) = m1(t) = m0(t), m0(t) represents the nozzle temperature value at time t in the initial iteration; m1(t) represents the nozzle temperature value at time t in the first iteration; set the maximum number of iterative learning to j max ; Set the number of iterations j to 2.

[0044] S2, nozzle temperature calculation step.

[0045] Calculate the nozzle temperature m at time t during the jth iteration learning j (t), as follows:

[0046]

[0047] Among them, m j-1 (t) represents the nozzle temperature at time t during the j-1th iterative learning, and Λ (Λ>0) represents the weight factor of the system input of the 3D printer temperature control system, which controls the system input change and can solve the steady-state error of the system. t is the search step size, and 0<δ t <1, It is a system model for estimating the temperature control system θ(t) of an unknown 3D printer.

[0048] n d (t+1) is the ideal nozzle temperature value at time t+1, n j-1 (t+1) is the system output temperature value at time t+1 during the j-1th iterative learning. and They represent the estimation of the external unknown disturbance s(t) at time t during the j-1th and jth iterative learning respectively, and w j (t) represents the unknown weight estimation coefficient for the external unknown interference s(t) at time t during the j-th iterative learning.

[0049] Specifically, at time t during the j-th iteration of learning It is expressed as follows:

[0050]

[0051]

[0052] in, represents the estimated system model at time t in the j-1th iteration, ε t is the search step size and 0<ε t <1, η (η>0) represents the weight factor of the input estimation of the 3D printer temperature control system, which can ensure that the estimated system smoothly approaches the actual temperature control system as the input and output data change, without any jump changes. j-2 (t) represents the nozzle temperature at time t during the j-2th iteration learning, w j-1 (t) represents the unknown weight estimation coefficient of the external unknown interference s(t) at time t during the j-1th iterative learning, n j-2 (t+1) is the system output temperature value at time t+1 during the j-2th iterative learning. Represents the estimate of the external unknown interference s(t) at time t during the j-2th iterative learning.

[0053] Specifically, at time t during the j-th iteration of learning It is expressed as follows:

[0054]

[0055] Among them, γ t is the search step size and 0<γ t <1, w j-2 (t) represents the unknown weight estimation coefficient for the external unknown disturbance s(t) at time t during the j-2th iterative learning, and σ (σ>0) represents the weight factor of the unknown weight estimation of the 3D printer temperature control system.

[0056] Specifically, w at time t during the j-th iteration of learning j (t), expressed as follows:

[0057]

[0058]

[0059] Among them, β t is the search step size and 0<β t <1, α (α>0) represents the weight factor of the external unknown disturbance estimation of the temperature control system of the 3D printer.

[0060] S3, estimation system model judgment step.

[0061] judge or Is it true? If so, execute Otherwise, return to S2; where κ is the system estimated reset coefficient, which is a constant and relatively small value, in the order of 1e-3; sign(·) is the sign function.

[0062] S4, external unknown interference estimation and judgment step.

[0063] judge or Is it true? If so, execute Otherwise, return to S2; where ξ is the external interference reset coefficient, which is a constant value and is relatively small, at the order of 1e-3;

[0064] S5, step of judging the error of unknown weight estimation coefficient of external unknown interference.

[0065] Judgment|w j (t)-w j-1(t)|≤ξ2 holds true. If so, execute w j (t) = w1(t), otherwise, return to S2; wherein ξ2 is the external interference weight reset coefficient, which is a constant and relatively small value, on the order of 1e-3.

[0066] S6, nozzle temperature error judgment step.

[0067] Judgement|m j (t)-m j-1 (t)|≤κ2 is true, if true, execute m j (t) = m1(t), otherwise, return to S2; where κ2 is the system input reset coefficient, which is a constant and relatively small value, on the order of 1e-3.

[0068] S7, iterative learning times determination and system temperature output step.

[0069] Execute j=j+1, and determine j<j max Is it true? If so, repeat S2 to S7; otherwise, stop the operation and output the jth detected max The temperature value output by the 3D printer temperature control system at time t during the iterative learning

[0070]

[0071] The above embodiments are only used to illustrate the present invention, and are not used to limit the present invention. As long as they are based on the technical essence of the present invention, changes and modifications to the above embodiments will fall within the scope of the claims of the present invention.

Claims

1. A temperature control method for a 3D printer with a learning mechanism, characterized in that: include: S1, initialization step; Set the 3D printer temperature control system to output the ideal temperature value n to the nozzle at time t d (t), t={0,1,2,3,…,Tm},T m is the maximum running time; detect the nozzle temperature value m(t) at time t, set m(t) = m1(t) = m0(t), m0(t) represents the nozzle temperature value at time t in the initial iteration; m1(t) represents the nozzle temperature value at time t in the first iteration; set the maximum number of iterative learning to j max ; Set the number of iterations j to 2; S2, nozzle temperature calculation step; Calculate the nozzle temperature m at time t during the jth iteration learning j (t), as follows: Among them, m j-1 (t) represents the nozzle temperature at time t during the j-1th iteration learning; Λ represents the weight factor of the 3D printer temperature control system input, Λ>0; δ t is the search step length, and 0<δ t <1; For unknown 3D printer temperature control system Estimated system model; n d (t+1) represents the ideal nozzle temperature value at time t+1; n j-1 (t+1) represents the system output temperature value detected at time t+1 during the j-1th iterative learning; and They represent the estimation of the external unknown disturbance s(t) at time t during the j-1th and jth iterative learning respectively; w j (t) represents the unknown weight estimation coefficient of the external unknown interference s(t) at time t during the j-th iterative learning; At time t during the j-th iteration of learning It is expressed as follows: in, represents the estimated system model at time t in the j-1th iteration; ε t is the search step length and 0<ε t <1; η represents the weight factor of the input estimation of the 3D printer temperature control system, η>0; m j-2 (t) represents the nozzle temperature at time t during the j-2th iteration learning; w j-1 (t) represents the unknown weight estimation coefficient of the external unknown interference s(t) at time t during the j-1th iterative learning; n j-2 (t+1) is the system output temperature value at time t+1 during the j-2th iterative learning; It represents the estimation of the external unknown interference s(t) at time t during the j-2th iterative learning; At time t during the j-th iteration of learning It is expressed as follows: Among them, γ t is the search step size and 0<γ t <1; w j-2 (t) represents the unknown weight estimation coefficient of the external unknown disturbance s(t) at time t during the j-2th iterative learning; σ represents the weight factor of the unknown weight estimation of the 3D printer temperature control system, σ>0; w at time t during the j-th iteration of learning j (t), expressed as follows: Among them, β t is the search step length and 0<β t <1; α represents the weight factor of the external unknown disturbance estimation of the 3D printer temperature control system, α>0; S3, estimation system model judgment step; judge or Is it true? If so, execute Otherwise, return to S2; where κ is the system estimated reset coefficient, sign(·) is the sign function; S4, external unknown interference estimation and judgment step; judge or Is it true? If so, execute Otherwise, return to S2; where ξ is the external interference reset coefficient; S5, a step of judging the error of the unknown weight estimation coefficient of the external unknown interference; Judgment|w j (t)-w j-1 (t)|≤ξ2 is true, if true, then execute w j (t) = w1(t), otherwise, return to S2; where ξ2 is the external interference weight reset coefficient; S6, nozzle temperature error judgment step; Judgement|m j (t)-m j-1 (t)|≤κ2 is true, if true, execute m j (t) = m1(t), otherwise, return to S2; where κ2 is the system input reset coefficient; S7, iterative learning times determination and system temperature output step; Execute j=j+1, and determine j<j max Is it true? If so, repeat S2 to S7; otherwise, stop the operation and output the jth detected max The temperature value n output by the 3D printer temperature control system at time t during the iterative learning jmax (t).

Images