Study on temperature control method of injection molding machine

Foreword

Plastic is a common material, with obvious advantages, such as: light quality, good plasticity, reusable, low cost, etc., so it is widely used in packaging, medicine, cosmetics and other fields. As the key equipment to realize plastic molding, the injection molding machine can easily realize the primary molding [1,3] of plastic products such as complex shape and high-precision size. In order to ensure the molding accuracy, efficiency and aesthetics of injection molding machine, the accuracy of radial temperature control of injection molding machine cylinder must be improved. If the temperature is relatively low, it will lead to uneven plasticizer of plastic particles, causing equipment wear or damage; if the temperature is relatively high, the polymer plastic will decompose, resulting in loose tissue, carbonization and then wrapped in the inner wall of the material cylinder or screw surface, seriously affecting the product quality [4,5]. Usually, according to the process requirements, the temperature of the injection molding machine cylinder will be divided into 3~5 temperature intervals, and different plastic injection molding temperature will be different, so it is difficult to realize the radial temperature control of the injection molding machine cylinder. The traditional PID control algorithm has the characteristics of simple structure and fast response speed, which is widely used in the temperature control of injection molding machine cylinder [6,7]. For the multi-temperature interval control, the injection molding machine often adopts multiple single loop PID independent control, but the temperature control of the injection molding machine is prone to external environment, voltage fluctuation and other factors, and the adjacent temperature interval interferes with each other. To sum up, the temperature control of injection molding machine has obvious coupling and nonlinearity. If only the traditional PID control is used, its parameters need to be adjusted repeatedly, and it is difficult to achieve the ideal control effect. At present, many advanced control strategies are introduced into the material cylinder temperature control algorithm, including expert control, neural network control, fuzzy control, optimal time control, etc., but these algorithms do not solve the coupling problem well [8~11].

To solve this problem, a static decoupling algorithm based on neural network combines fuzzy PID control to improve the temperature control effect of injection molding machine.

Conductor temperature characteristics

The injection system of injection molding machine is shown in Figure 1, including 1-oil cylinder, 2-hopper, 3-material cylinder, 4-nozzle, 5-mold, 6-metering section heater, 7-compression section heater, 8-solid conveying section heater. The whole heating section can be divided into solid conveying section (section I), compression section (section II) and metering section (section III). Each section is equipped with independent heating wire, arranged along the material cylinder, and the temperature field required for the injection process is constructed by setting different temperature values. Plastic particles enter the material cylinder through the hopper, and the oil cylinder will push the screw to squeeze the plastic along the material cylinder. After preheating, plasticizing, injection, pressure retention, cooling and other processes, the mold is finally opened to get plastic parts. Considering the different heating power and the total amount of plastic in different heating sections, the temperature adjustment methods are different. In addition, there is heat exchange between the adjacent heating segments, and each segment affects each other, so the plastic temperature control needs to solve the coupling problem. At the same time, the plastic density, thermal conductivity and diffusion coefficient will also change, so the temperature control of the material cylinder is nonlinear [12~15].

Picture 1. injection system

As mentioned above, the cylinder temperature control of the injection molding machine belongs to the MIMO system. According to the law of conservation of energy, the total heat Q generated by the cylinder heating wire is equal to the sum of the heat required by the plastic melt Q1 and the heat loss Q2, and the expression is as formula (1)

Temperature controller design

Static decoupling algorithm of neural networks

The neural network can realize the mapping of multiple input and multiple output, which can better solve the problems such as nonlinearity and chronotaxis, and has the advantages of strong adaptive ability and training, so a static decoupling algorithm is proposed in this paper, in order to realize the decoupling control of the cylinder temperature. The control system combining the fuzzy PID control and the static decoupling algorithm of the neural network is shown in Figure 2. In FIG. 2, θ 1, θ 2 and θ 3 are the temperature setting values of sections I, II and III of the injection cylinder respectively; u1, u2 and u3 are the control signals of the fuzzy PID controller of cylinder sections I, II and III respectively, and U1, U2 and U3 are the control voltage of the heating wire of cylinder I, II and III respectively; T1, T2 and T3 are the actual temperature output values of sections I, II and III respectively.

The fuzzy controller adopts two-input three-output structure, where the input variable is the temperature deviation e of each section and the change rate e [6,6], the language theory domain is {NB, NM, NS, ZO, PS, PM, PM, PB}; the output variable is PID controller parameter variation Δ kp, Δ ki, Δ kd, the theory domain is [5,5], the language theory domain is {NB, NM, NS, ZO, PS, PM, PB}. The membership function adopts the trigonometric function, the reasoning method adopts the Mamdni, and the deblurring method adopts the area center of gravity method. The setting principles of the parameters of kp, ki, kd and others are as follows:Fuzzy PID controller

If the error is relatively large, in order to improve the system response speed and reduce the overshoot, a larger Δ kp, smaller Δ ki and Δ kd should be selected.

If the error and the error change rate are not large, in order to reduce the system overshoot and appropriately improve the response speed, Δ kp, Δ ki and Δ kd should be selected moderately.

If the error change rate is relatively small, a larger Δ kp, Δ ki and a smaller Δ kd should be selected. Fuzzy rulesAs shown in Table 1

Simulation and experimental studies

simulation

To verify the feasibility and effectiveness of the method, a simulation study is carried out. The temperature control system of injection molding machine based on the traditional PID algorithm and the algorithm described in the paper is established for simulation comparison. The temperature of section I of the injection molding machine is set to 180℃, the temperature of section II is set to 210℃, and the temperature of section III is set to 230℃. The simulation results are shown in Figure 4. The simulation results show that the traditional PID control the temperature overshoot of section I of cylinder is 4.7℃ and stable time consuming about 76s; the temperature overshoot of cylinder II is 19.3℃ and stable time consuming about97s; temperature overshoot of cylinder III is 15.4℃ and stable time takes about 77s. Using the control algorithm described in the paper, the temperature of sections I, II and III is almost no overadjusted, the temperature control curve is smooth, and the time required to reach the steady state will be reduced. The results show that the static decoupling of neural network can reduce the influence of temperature coupling interference well.

Further, a step interference amount of 20℃ is applied to the material cylinder section II at t=130s to verify the anti-interference ability of the system. The simulation results are shown in Figure 5. It can be seen from the simulation results that by PID control, the temperature overshoot of segments I, II and III is 9.5℃, 9.3℃, 4.2℃, about 30s, 43s and 37s, and the stable state, in the control method described in the paper, is 0.5℃, 3.2℃ and 0.4℃, about 8s, 22s and 13s. The simulation results show that the control method has good decoupling, anti-interference and robustness.

• PID控制仿真结果

• 文中所述方法仿真结果

trial

The PID control algorithm and the fuzzy PID control algorithm based on the static decoupling of the neural network were transplanted respectively for real-time inspectionMeasure the temperature of the material cylinder section III to verify the accuracy of the temperature control. The experimental device is shown in Figure 6. During the experiment, the temperature of cylinder section I is set at 180℃, the temperature of cylinder section II is set at 210℃, and the temperature of cylinder section III is set at 230℃. The test results are shown in Table 2. The test results show that the control method can improve the temperature control accuracy, which has good decoupling ability and anti-interference ability

Machine picture