Temperature control method and control device of hot runner system for inverted mold

Abstract

The invention belongs to the technical field of temperature control, and discloses a temperature control method and device of a hot runner system for an inverted mold. The method comprises the following steps: defining a critical melt area in a physical sprue which directly influences the product quality in a sprue of a hot runner system as a virtual sprue; dynamically generating a melt viscosity target track of the virtual gate region according to the type of an injection molding material and an injection molding process stage; based on measurable injection molding process parameters and sprue area temperature, estimating a current melt viscosity value of a virtual sprue area in real time through a state observer; calculating to obtain an optimal power control sequence by adopting a multivariable model prediction controller; and driving a sprue area heater according to the optimal power control sequence, so that the actual change of the melt viscosity of the virtual sprue area tracks a melt viscosity target track. According to the invention, the defects such as salivation and wire drawing can be effectively eliminated, so that the sprue quality stability and the production efficiency are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of temperature control, in particular to a temperature control method and device for a hot runner system of an inverted mold. BACKGROUND

[0002] At present, the hot runner temperature control technology commonly used in inverted molds is essentially still based on the temperature control scheme of traditional injection molds, and its core is to measure the physical temperature of a certain point of the hot nozzle by using a thermocouple and to stabilize the temperature of the point at a set value by using a PID controller. The traditional temperature control scheme has the following inherent defects: (1) The direct object controlled by the existing technology is the physical temperature of a certain point beside the heater, but the key factors affecting the quality of the product (such as flow, stringing, and gate scar) are the actual viscosity and phase state of the critical melt region inside the gate. Stabilizing the physical temperature does not mean that the melt state is optimal, resulting in unsatisfactory control effect and reliance on the experience of operators for repeated debugging for high-quality product production.

[0003] (2) PID control is a feedback control of after-the-fact correction. When there are severe dynamic disturbances such as mold opening and injection during the injection cycle, the temperature has already fluctuated, and the PID controller responds passively, which has a natural lag. For inverted mold applications with extremely high requirements for gate quality, this lag is enough to cause defects such as flow and cold material. Although there are optimization methods such as feedforward, the control logic of the reaction formula has not been fundamentally changed.

[0004] (3) The traditional temperature control system is a single-loop system, and it fails to consider the injection speed, holding pressure, and other process parameters that have an instantaneous and significant impact on the melt of the gate as feedforward information. Therefore, for inherent and predictable disturbances during the injection cycle, the system lacks predictability and collaborative processing capability, and the anti-interference ability is weak.

[0005] In summary, the core problem of the existing hot runner temperature control technology for inverted molds is that the physical quantity (temperature of a certain point) controlled is disconnected from the quality target (melt state at the gate) ultimately pursued, and the control strategy is passive and isolated, which cannot meet the demand for precise, proactive, and forward-looking control of the gate quality in the high-speed and dynamic injection cycle of the inverted mold. The present application is proposed to overcome the deficiencies of the prior art. SUMMARY

[0006] The purpose of the present application is to provide a temperature control method and system for a hot runner system of an inverted mold to solve the problems raised in the background art.

[0007] The application provides a temperature control method for a hot runner system of an inverted mold, comprising the following method steps: step S10, defining a critical melt region in a physical gate inside a gate of the hot runner system as a virtual gate, the melt state of the virtual gate directly determining the quality of the gate.

[0008] Step S20, dynamically generating a melt viscosity target trajectory of the virtual gate region according to the type of injection material and the injection process stage, and estimating a current melt viscosity value of the virtual gate region in real time based on measurable injection process parameters and gate region temperature through a state observer; wherein the injection process parameters at least include one of injection speed, holding pressure and screw position.

[0009] Step S30, taking the melt viscosity target trajectory as a set target, taking the current melt viscosity value as a feedback amount, and calculating an optimal power control sequence by using a multivariable model predictive controller; wherein the multivariable model predictive controller is internally provided with a dynamic process model related to associated injection process parameters, heating power and melt viscosity of the virtual gate region.

[0010] Step S40, driving the gate region heater according to the optimal power control sequence, so that the actual change of the melt viscosity of the virtual gate region tracks the melt viscosity target trajectory.

[0011] The application also provides a temperature control system for a hot runner system of an inverted mold, the system comprising: a virtual gate determination unit, configured to define a critical melt region in a physical gate inside a gate of the hot runner system as a virtual gate, the melt state of the virtual gate directly determining the quality of the gate.

[0012] A generation and observation unit is configured to dynamically generate a melt viscosity target trajectory of the virtual gate region according to the type of injection material and the injection process stage, and estimate a current melt viscosity value of the virtual gate region in real time based on measurable injection process parameters and gate region temperature through a state observer; wherein the injection process parameters at least include one of injection speed, holding pressure and screw position.

[0013] An optimal solution unit is configured to take the melt viscosity target trajectory as a set target, take the current melt viscosity value as a feedback amount, and calculate an optimal power control sequence by using a multivariable model predictive controller; wherein the multivariable model predictive controller is internally provided with a dynamic process model related to associated injection process parameters, heating power and melt viscosity of the virtual gate region.

[0014] The heater control unit is used to drive the gate area heater according to the optimal power control sequence, so that the actual change of melt viscosity in the virtual gate area tracks the target trajectory of melt viscosity.

[0015] Addressing the issues of disconnect between the controlled object and the quality objective, passive and lagging control strategies, and single variables mentioned in the background art, this invention achieves precise control of gate quality by defining a virtual gate and using melt viscosity as the direct control objective. Furthermore, by employing a model predictive controller and its built-in dynamic process model, it achieves forward-looking and collaborative control of the injection molding process, effectively overcoming the lag of traditional PID control. Ultimately, this improves gate quality stability and production efficiency, and effectively eliminates defects such as drooling and stringing. Attached Figure Description

[0016] Figure 1 This is a schematic flowchart of a temperature control method for a hot runner system for an inverted mold, as disclosed in an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram of the output sequence of the predicted melt viscosity using the dynamic process model disclosed in an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of the temperature control system for a hot runner system for an inverted mold, as disclosed in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 This invention provides a temperature control method 100 for a hot runner system used in inverted molds, comprising the following steps: Step S10, defining a critical melt region inside the physical gate of the hot runner system gate that directly affects the quality of the product as a virtual gate, wherein the melt state of the virtual gate directly determines the gate quality; in this step, during actual injection molding, the physical temperature at a certain point near the heater controlled by traditional methods differs significantly from the actual melt state at the gate that determines product quality. For example, even if the temperature at the measuring point remains stable, the melt at the gate may be in a non-ideal state due to injection impact, mold opening heat dissipation, etc., leading to defects such as drooling and stringing.

[0021] To address this, the present invention proposes the concept of a virtual gate. Specifically, a virtual gate is not an actual added physical component, but rather a critical melt region defined within the physical gate using an algorithm. The melt viscosity, temperature, and phase (molten or solidified) of this region directly determine the final gate quality. For example, at the moment of mold opening, if the melt viscosity in this region is too low, drooling is likely to occur; if the viscosity is too high or it has already solidified, cold material may be generated or it may affect the next injection.

[0022] Step S20: Based on the injection molding material type and injection molding process stage, dynamically generate the melt viscosity target trajectory of the virtual gate area; and, based on measurable injection molding process parameters and gate area temperature, estimate the current melt viscosity value of the virtual gate area in real time through a state observer; wherein, the injection molding process parameters include at least one of injection speed, holding pressure, and screw position; in this step, unlike the traditional fixed temperature setting mode, this invention dynamically generates the melt viscosity target trajectory of the virtual gate area for different injection molding materials (such as ABS, PC, etc., which have different rheological properties) and different injection molding process stages (such as requiring lower viscosity during injection to ensure filling fluidity, and requiring higher viscosity after holding pressure and before mold opening to prevent drooling). For example, for ABS material, lower viscosity is required during the injection stage to ensure filling fluidity, while higher viscosity is required after holding pressure and before mold opening to prevent drooling, and thus the corresponding melt viscosity target trajectory is generated in real time according to these requirements.

[0023] Meanwhile, since the virtual gate is a conceptual area, its melt viscosity cannot be directly measured. Therefore, this invention uses a state observer (such as a Kalman filter) to estimate the current viscosity value in real time. This state observer receives measurable injection process parameters (including injection speed, holding pressure, and screw position) and gate area temperature, and calculates the current melt viscosity value of the virtual gate accurately and in real time based on an internally established dynamic model.

[0024] Step S30: Using the target melt viscosity trajectory as the set target and the current melt viscosity value as the feedback quantity, a multivariable model predictive controller is used to calculate the optimal power control sequence. The multivariable model predictive controller has a built-in dynamic process model relating the injection molding process parameters, heating power, and melt viscosity in the virtual gate region. In this step, the invention uses a multivariable model predictive controller to replace the traditional PID controller, overcoming its hysteresis and isolation defects. The core of this controller is its built-in dynamic process model, which can accurately describe the dynamic relationship between injection molding process parameters, heating power, and melt viscosity in the virtual gate region.

[0025] In each control cycle, the multivariate model predictive controller predicts the trend of virtual gate viscosity change over a future period based on the current melt viscosity value obtained from the state observer, and calculates the optimal power control sequence through a rolling optimization algorithm, so that the predicted viscosity trajectory can best track the melt viscosity target trajectory generated in step S20.

[0026] Step S40: Drive the heater in the gate area according to the optimal power control sequence, so that the actual change in the melt viscosity in the virtual gate area tracks the target trajectory of the melt viscosity.

[0027] In this step, the optimal power control sequence calculated in step S30 is output to the physical heater in the gate area. By precisely adjusting the heating power, the actual state of the melt in the virtual gate area is controlled. Ultimately, the actual change in the melt viscosity in the virtual gate area can be accurately tracked to the target trajectory of the melt viscosity, thereby ensuring gate quality and stably producing high-quality products without defects such as drooling and stringing.

[0028] Addressing the issues of disconnect between the controlled object and the quality objective, passive and lagging control strategies, and single variables mentioned in the background art, this invention achieves precise control of gate quality by defining a virtual gate and using melt viscosity as the direct control objective. Furthermore, by employing a model predictive controller and its built-in dynamic process model, it achieves forward-looking and collaborative control of the injection molding process, effectively overcoming the lag of traditional PID control. Ultimately, this improves gate quality stability and production efficiency, and effectively eliminates defects such as drooling and stringing.

[0029] As an example, a critical melt region inside the physical gate that directly affects the quality of the product in the hot runner system gate is defined as a virtual gate, including: step S101, based on the transient thermodynamic field simulation during the injection molding cycle, taking the gate center axis as a reference, defining the region where the melt viscosity is between a first viscosity threshold and a second viscosity threshold as the initial physical boundary of the virtual gate; wherein, the first viscosity threshold corresponds to the critical viscosity at which the material drools, and the second viscosity threshold corresponds to the lower limit critical viscosity of the material's filling fluidity.

[0030] This step uses numerical simulation to determine the initial range of the virtual gate. Specifically, a refined finite element model including the gate region is first established, which accurately reflects the geometric characteristics of the hot runner system. Then, based on actual injection molding process conditions, transient thermo-mechanical coupling simulation analysis is performed to simulate the melt flow, heat transfer, and phase change processes throughout the entire injection molding cycle.

[0031] During the simulation, the focus is on monitoring the melt viscosity distribution along the central axis of the gate. The first viscosity threshold is determined based on material drooling experiments and is set as the minimum viscosity at which drooling does not occur. The second viscosity threshold is determined based on the material's rheological properties and is set as the maximum permissible viscosity to ensure complete filling. Furthermore, by extracting all elements in the simulation results whose viscosity values ​​fall between these two thresholds, the initial physical boundary of the virtual gate can be determined.

[0032] Step S102: Within the initial physical boundary, the melt region that undergoes a complete phase change cycle from solid glass state to fully molten state and then to the first viscosity threshold in each injection molding cycle is determined as the virtual gate core region, thereby obtaining the virtual gate.

[0033] In this step, after obtaining the initial physical boundary, the core region is further determined. Specifically, by post-processing the transient simulation data, the physical state change history of each point within the initial boundary during the injection molding cycle is tracked. The focus is on identifying regions that undergo a complete phase transition cycle in each cycle. In these regions, the melt completely solidifies to a glassy state during the cooling phase, remelts during the heating phase, and its viscosity is precisely maintained above the critical flow value at mold opening. Understandably, in actual operation, a pre-written post-processing program can automatically identify the set of units that meet these conditions, ultimately determining the spatial extent of the virtual gate core area.

[0034] It should be noted that the above determination process can be completed during the mold design stage, using specialized simulation analysis software. The determined virtual gate parameters can be stored in the control device, providing accurate zone definitions for subsequent temperature control.

[0035] This implementation precisely defines the dynamic boundary of the virtual gate using dual viscosity thresholds, ensuring a direct correlation between the defined region and the material's flow and filling performance. Furthermore, by identifying the core region undergoing a complete phase change cycle, it accurately pinpoints the key melt that is most sensitive to the thermal history and plays a decisive role in gate quality. Thus, this implementation transforms the concept of a virtual gate from a theoretical idea into quantifiable and calculable engineering parameters, providing precise spatial positioning for subsequent temperature control strategies.

[0036] As an example, the step of dynamically generating the melt viscosity target trajectory of the virtual gate area according to the injection molding material type and injection molding process stage includes: step S201, determining the theoretical optimal viscosity value of the melt at the virtual gate during the injection, holding pressure, cooling and mold opening stages based on the standard rheological curve of the injection molding material and the quality requirements of the target product.

[0037] In this step, existing target generation methods based on fixed curves or simple functions treat the target trajectory as a static, passive setpoint. These methods cannot respond to the instantaneous impacts on the system caused by critical actions such as injection and mold opening, nor can they adapt to changes in the equipment's own state, resulting in poor actual control performance. To solve these problems, the melt viscosity target trajectory generation method of this invention is proposed.

[0038] In practice, the first step is to obtain the standard rheological curve of the injection molding material, which fully reflects the viscosity change of the material under different temperatures and shear rates. At the same time, the optimal viscosity range required for each process stage is determined by combining the appearance quality requirements of the target product (such as no flow marks, silver streaks, etc.) and dimensional accuracy requirements (such as shrinkage control).

[0039] For example, during the injection stage, to ensure sufficient melt filling and avoid flow defects, a low viscosity range (typically 10) needs to be maintained. 2 -10 3 During the holding phase, in order to effectively transmit pressure and control volume shrinkage, the viscosity needs to be appropriately increased to a medium grade (10 Pa·s). 3 -10 4 During the mold-making stage, to prevent drooling defects, the viscosity needs to be increased to above the safe threshold (>10 Pa·s). 4 Pa·s).

[0040] Therefore, through the comprehensive analysis of the material properties and process requirements described above, a precise theoretical optimal viscosity value is established for each process stage.

[0041] Step S202: Based on the real-time status parameters of the injection molding machine, the theoretical optimal viscosity value is dynamically corrected to generate an achievable viscosity reference trajectory that matches the current machine capacity; wherein, the real-time status parameters include at least one of the current maximum effective power of the heater and the system heat loss coefficient.

[0042] In this step, the operating status of the injection molding machine changes dynamically during actual production. Specifically, the heater may experience power decay due to long-term use, the heat loss of the insulation material will increase with the length of time it is used, and fluctuations in ambient temperature will also affect the heat dissipation characteristics of the system.

[0043] To address this, a real-time status monitoring mechanism is introduced in this step to continuously collect key parameters such as the heater's current maximum effective power and the system's heat loss coefficient. Specifically, when a decrease in heater power output is detected, the viscosity target value is automatically lowered to avoid control overshoot due to insufficient power; when increased system heat loss is detected, the viscosity benchmark is correspondingly increased to compensate for additional heat loss. This adaptive correction ensures that the generated viscosity benchmark trajectory meets process quality requirements while precisely matching the equipment's current actual capacity.

[0044] Step S203: Before the critical process switching point of the viscosity reference trajectory, a preset viscosity precursor signal is injected to form the final melt viscosity target trajectory; wherein, a negative step precursor signal is injected before the start of the injection stage to reduce the viscosity target in advance with a preset amplitude; and a positive pulse precursor signal is injected before the start of the mold opening stage to increase the viscosity target in advance with a preset amplitude.

[0045] In this step, the injection start and mold opening actions will cause significant instantaneous disturbances to the system during the injection molding process. Specifically, when the screw starts to advance, it will generate a violent shear impact, which will cause the viscosity to rise instantaneously; when the mold opens, the sudden change in heat dissipation conditions may cause the viscosity to drop rapidly.

[0046] Therefore, this step employs a viscosity precursor signal injection strategy. Specifically, at a specific time before the injection phase begins (e.g., 0.1-0.5 seconds), a negative step precursor signal is injected to preemptively reduce the viscosity target by a preset amplitude (e.g., 5%-15% of the theoretical value). This pre-compensation mechanism allows the system to enter a low-viscosity preparation state in advance, effectively offsetting the viscosity peak generated during screw startup.

[0047] Accordingly, a positive pulse precursor signal is injected before the mold opening stage begins (e.g., 0.2-0.8 seconds in advance) to pre-increase the viscosity target with a preset amplitude (e.g., 10%-20% of the theoretical value). This advance control allows the melt in the virtual gate area to reach the safe viscosity level required to prevent drooling in advance, thereby solving the problem of viscosity drop that may occur due to sudden changes in heat dissipation conditions at the moment of mold opening.

[0048] This implementation method generates a dynamic viscosity target trajectory that is ideal, feasible, and anti-interference by establishing a theoretical optimal benchmark, introducing adaptive correction of equipment state, and injecting forward-looking precursor signals. This effectively overcomes the shortcomings of traditional static setting methods that cannot adapt to changes in equipment state and instantaneous process disturbances.

[0049] As an example, based on measurable injection process parameters and gate area temperature, the current melt viscosity value of the virtual gate area is estimated in real time by a state observer, including: step S204, constructing a nonlinear state-space model with the melt viscosity, temperature and thermal capacity state of the hot runner system in the virtual gate area as state variables, the heater power and injection process parameters as input variables, and the measurable gate area temperature as output variable.

[0050] In this step, the melt viscosity, temperature, and thermal capacity of the hot runner system in the virtual gate region are used as state variables. These three variables reflect the thermodynamic and rheological states of the system. Heater power and injection process parameters (including injection speed, holding pressure, and screw position) are used as input variables, directly determining the system's energy input and process conditions. The measurable temperature in the gate region is used as the output variable, representing a key observation signal that can be directly obtained from the system.

[0051] Understandably, this state-space model adopts a nonlinear form, which can more accurately describe the complex thermodynamic relationships that exist in the injection molding process, including the nonlinear changes in material viscosity, the nonlinear characteristics of heat conduction, and the nonlinear behavior of phase transition processes.

[0052] Step S205: Design an extended Kalman filter based on the nonlinear state-space model, and introduce an online model parameter identification mechanism to correct model mismatch caused by material batch differences and system aging in real time.

[0053] In this step, based on the aforementioned nonlinear state-space model, an extended Kalman filter is designed for state estimation. The extended Kalman filter effectively handles the nonlinear estimation problem in the injection molding process by linearizing the nonlinear system. The specific process is roughly as follows: The state vector of the filter is set. Input vector and observation vector : ,in, The melt viscosity in the virtual gate region. The melting temperature is... The system's thermal capacity state, For heater power, For injection speed, To maintain pressure, This is the temperature measurement value of the gating area.

[0054] Establish a discrete-time nonlinear state-space model: ,in, It is a nonlinear state transition function, established based on the thermodynamic and rheological principles of the injection molding process; It is a nonlinear observation function; The vector of model parameters to be identified; This is process noise; To observe noise.

[0055] Within the EKF (Extended Kalman Filter) framework, the nonlinear system is linearized using a first-order Taylor expansion: Prediction step: ,in, Let be the state transition Jacobian matrix.

[0056] Update steps: ,in, To observe the Jacobian matrix, This is a new information sequence.

[0057] An adaptive estimation method based on innovation is used to adjust the model parameters in real time. The parameter update law is: .

[0058] in, To update the gain matrix for the parameters, calculate using either gradient descent or stochastic approximation: ,in, For learning rate, This is the error covariance matrix for parameter estimation.

[0059] Through the above design, the extended Kalman filter can simultaneously perform state estimation and parameter identification, effectively addressing nonlinearity and model uncertainty in the injection molding process, and achieving accurate estimation of melt viscosity in the virtual gate region.

[0060] Furthermore, this invention introduces an online model parameter identification mechanism. This mechanism dynamically adjusts key parameters in the nonlinear state-space model, such as thermal conductivity and heat capacity, by comparing the differences between the filter's predicted output and the actual measured values ​​in real time. When batch changes in materials or system performance degradation are detected, the filter can automatically correct the model parameters to maintain the accuracy of the state estimation. This ensures that viscosity estimation remains highly accurate even under long-term operating conditions.

[0061] Step S206: Process temperature sensor data at a high sampling rate, update injection process parameters at the injection cycle rate, fuse multi-rate measurement data through the extended Kalman filter, and output the current melt viscosity value of the virtual gate region.

[0062] In this step, the extended Kalman filter of this invention operates with a dual-rate asynchronous mechanism, specifically as follows: The extended Kalman filter cyclically executes the core algorithm at a high frequency set internally (typically 100Hz to 1kHz, matching the sampling rate of the temperature sensor). Within each high-frequency cycle, the latest temperature measurement value of the gating area is read, and the state variables (viscosity, temperature, heat capacity) are predicted and corrected, thereby achieving millisecond-level precise tracking of the system's thermal dynamics.

[0063] Meanwhile, the filter establishes a communication interface synchronized with the main controller of the injection molding machine. When the injection molding machine enters a new injection or holding pressure stage, that is, when the injection process parameters are updated, this interface is triggered. The filter then receives the latest injection speed, holding pressure and other parameters, and uses them as new and more representative system inputs to update its internal state space model.

[0064] Through this dual-rate asynchronous mechanism, the extended Kalman filter can effectively integrate high-frequency temperature details with low-frequency macroscopic process commands. It can capture dynamics by utilizing rapid changes in temperature data and correct the driving input of the model based on precise process parameters, thereby ultimately outputting a real-time and accurate virtual gate area current melt viscosity value.

[0065] As an example, the method of using a multivariable model predictive controller to calculate the optimal power control sequence includes: step S301, in the current control cycle, obtaining the current melt viscosity value of the virtual gate region estimated in real time by the state observer as the initial state for prediction; at the same time, obtaining the melt viscosity target trajectory sequence in the future prediction time domain.

[0066] In this step, at the beginning of each control cycle, the multivariate model predictive controller first obtains its latest estimated current melt viscosity value for the virtual gate region from the state observer. This current melt viscosity value reflects the latest state of the most critical controlled variable, i.e., the melt state, after the previous control action. It is set as the initial state for the dynamic process model to make predictions, ensuring that the optimization calculation is based on the real-time feedback of the system.

[0067] Simultaneously, the controller retrieves or calculates in real-time the target trajectory sequence of melt viscosity for a predicted time domain from pre-stored process specifications. This target trajectory sequence precisely defines the ideal change path that the melt viscosity at the virtual gate should follow over multiple control cycles from the current moment. For example, when the mold opening action is anticipated, this target trajectory will pre-plan the viscosity rise curve to proactively prevent drooling.

[0068] For step S302, please refer to... Figure 2The predictable injection molding process parameter sequence is used as a feedforward input and input into the dynamic process model along with a set of candidate heating power sequences. The dynamic process model associates the dynamic relationship between the injection molding process parameters, heating power and melt viscosity in the virtual gate region, and predicts the melt viscosity output sequence of the virtual gate region in the future prediction time domain based on the initial state and the dynamic relationship.

[0069] In this step, the multivariate model predictive controller performs data preparation actions in parallel: (1) obtains the sequence of predictable injection process parameters from the injection molding machine main controller, such as the known injection speed curve and holding pressure value in the next stage, and uses these parameters as feedforward inputs; (2) generates a set of candidate heating power sequences internally, that is, a set of power operation schemes that are assumed to be applied to the heater in the future control time domain.

[0070] Subsequently, both sets of input sequences are input into the dynamic process model. This dynamic process model is a pre-established mathematical model that accurately describes the dynamic coupling relationship between injection molding process parameters, heating power, and melt viscosity in the virtual gate region. Using the initial state provided in step S301 as the starting point for calculation, the model simulates the future evolution of the system state under the combined effects of feedforward disturbances and candidate power schemes, thereby predicting the melt viscosity output sequence in the virtual gate region throughout the entire prediction time domain.

[0071] Step S303: Construct an objective function, which measures the tracking error between the predicted melt viscosity output sequence and the melt viscosity target trajectory sequence, while also weighing the stability of the candidate heating power sequence's own variation; using the objective function as the optimization objective, under the constraint of satisfying the heater power limit, solve for the optimal power control sequence that optimizes the objective function through an iterative algorithm.

[0072] In this step, the multivariate model predictive controller first constructs an objective function as a comprehensive evaluation standard, which simultaneously considers two key performance indicators: (1) tracking error: the difference between the melt viscosity output sequence predicted in step S302 and the melt viscosity target trajectory sequence obtained in step S301. Optimizing this error aims to ensure control accuracy; (2) control stability: the variation amplitude of the candidate heating power sequence itself. By penalizing the power increment, it aims to avoid drastic fluctuations in heating power, protect the heater, and improve process stability. The mathematical expression of the objective function is, for example: Where J is the objective function value, and the goal of the optimization algorithm is to find the optimal power control sequence that minimizes J. The system outputs a tracking error term, which measures the difference between the predicted output and the desired target, and is the core manifestation of control accuracy.

[0073] For the future moment The predicted virtual gate melt viscosity value is derived from the output of the dynamic process model in step S302.

[0074] For the future moment The target trajectory value of the melt viscosity is derived from the set sequence obtained in step S301.

[0075] The output error weight matrix is ​​a positive semi-definite matrix, which determines the weighting of tracking error in the overall objective function. Increasing the weighting matrix... The value of will prioritize tracking accuracy for the multivariate model predictive controller.

[0076] The summation symbol indicates that within the control time domain, from the current time ( ) to the last adjustable control moment ( ), sum up all items, and .

[0077] This is a penalty term for controlling incremental changes. This term limits the magnitude of changes in control power and is crucial for ensuring smooth control and preventing actuator wear.

[0078] For at any time The heater power control increment is defined as .

[0079] The control increment weight matrix, being a positive definite matrix, determines the importance of the stability of the control action in the overall objective function. Increasing... With a given value, the multivariate model predictive controller will output a smoother, slower-changing power signal, but this may come at the cost of some tracking speed.

[0080] The optimization problem is then solved under the following constraints on the heater power limit (e.g., maximum power P). max and minimum power P min This is an inherent physical property of the heater, ensuring that the solution is feasible in engineering.

[0081] ,in, and These are the minimum and maximum allowable power of the heater, respectively.

[0082] Finally, the multivariate model predictive controller is solved automatically using iterative algorithms (such as gradient descent, sequential quadratic programming, etc.). This algorithm searches through all possible candidate heating power sequences that satisfy the above constraints, repeatedly evaluating and adjusting the sequences until it finds the sequence that makes the objective function value optimal (usually minimum). This sequence is then determined as the optimal power control sequence for the current cycle.

[0083] Step S304: Output the first power control quantity corresponding to the current control cycle in the optimal power control sequence to the gate area heater; after entering the next control cycle, repeat the above steps to realize rolling optimization and feedforward control.

[0084] In this step, the multivariate model predictive controller immediately outputs the first element of the optimal power control sequence obtained in step S303, which corresponds to the first power control quantity of the current control cycle, to the actuator, namely the gate area heater. It should be noted that although the multivariate model predictive controller plans an optimization path for multiple future cycles, it only executes the first step because the plan for further future cycles will be re-optimized due to new system states.

[0085] After the actions of the current control cycle are completed, the system enters the next control cycle. The multivariable model predictive controller then repeats steps S301 to S304: acquiring new state feedback, updating the target trajectory, performing a new round of prediction and optimization, and executing the first new control variable. Through this cycle, it can continuously adapt to the dynamic changes in the injection molding process, perform feedforward control on predictable disturbances, and correct unknown disturbances through feedback, ultimately achieving high-precision and robust rolling optimization control of the melt viscosity at the virtual gate.

[0086] Please see Figure 3 This invention also provides a temperature control system 200 for a hot runner system for inverted molds. The system includes a virtual gate determination unit 2010, which defines a critical melt region inside a physical gate that directly affects the quality of the product as a virtual gate. The melt state of the virtual gate directly determines the gate quality.

[0087] The generation and observation unit 2020 is used to: dynamically generate the melt viscosity target trajectory of the virtual gate area according to the injection molding material type and injection molding process stage; and estimate the current melt viscosity value of the virtual gate area in real time through a state observer based on measurable injection molding process parameters and gate area temperature; wherein the injection molding process parameters include at least one of injection speed, holding pressure and screw position.

[0088] The optimal solution unit 2030 is used to: calculate the optimal power control sequence using the melt viscosity target trajectory as the set target and the current melt viscosity value as the feedback quantity, and employ a multivariable model predictive controller; wherein, the multivariable model predictive controller has a built-in dynamic process model that relates injection molding process parameters, heating power, and melt viscosity in the virtual gate region.

[0089] The heater control unit 2040 is used to drive the gate area heater according to the optimal power control sequence, so that the actual change of melt viscosity in the virtual gate area tracks the target trajectory of melt viscosity.

[0090] As an example, the virtual gate determination unit 2010 is specifically used to: based on the transient thermodynamic field simulation during the injection molding cycle, and taking the gate center axis as a reference, define the region where the melt viscosity is between a first viscosity threshold and a second viscosity threshold as the initial physical boundary of the virtual gate; wherein, the first viscosity threshold corresponds to the critical viscosity at which the material drools, and the second viscosity threshold corresponds to the lower limit critical viscosity of the material's filling fluidity.

[0091] Within the initial physical boundary, the melt region that undergoes a complete phase change cycle from a solid glass state to a fully molten state and then to the first viscosity threshold in each injection molding cycle is defined as the virtual gate core region, thereby deriving the virtual gate.

[0092] As an example, the generation and observation unit 2020 is specifically used to: determine the theoretical optimal viscosity value of the melt at the virtual gate during the injection, holding, cooling and mold opening stages based on the standard rheological curve of the injection molding material and the quality requirements of the target product.

[0093] Based on the real-time status parameters of the injection molding machine, the theoretical optimal viscosity value is dynamically corrected to generate an achievable viscosity reference trajectory that matches the current machine capacity; wherein, the real-time status parameters include at least one of the heater's current maximum effective power and the system heat loss coefficient.

[0094] Before the critical process switching point of the viscosity reference trajectory, a preset viscosity precursor signal is injected to form the final melt viscosity target trajectory; wherein, a negative step precursor signal is injected before the injection stage to reduce the viscosity target in advance with a preset amplitude; and a positive pulse precursor signal is injected before the mold opening stage to increase the viscosity target in advance with a preset amplitude.

[0095] As an example, the generation and observation unit 2020 is also specifically used to: construct a nonlinear state-space model with the melt viscosity, temperature and thermal capacity state of the virtual gate region and the hot runner system as state variables, the heater power and injection process parameters as input variables, and the measurable gate region temperature as the output variable.

[0096] An extended Kalman filter is designed based on the aforementioned nonlinear state-space model, and an online model parameter identification mechanism is introduced to correct model mismatch caused by material batch differences and system aging in real time.

[0097] Temperature sensor data is processed at a high sampling rate, injection process parameters are updated at the injection cycle rate, and multi-rate measurement data is fused through the extended Kalman filter to output the current melt viscosity value of the virtual gate region.

[0098] As an example, the optimal solution unit 2030 is specifically used to: in the current control cycle, obtain the current melt viscosity value of the virtual gate region estimated in real time by the state observer as the initial state for prediction; at the same time, obtain the melt viscosity target trajectory sequence in the future prediction time domain.

[0099] The predictable injection molding process parameter sequence is used as a feedforward input and input into the dynamic process model along with a set of candidate heating power sequences. The dynamic process model associates the dynamic relationship between the injection molding process parameters, heating power and melt viscosity in the virtual gate region, and predicts the melt viscosity output sequence of the virtual gate region in the future prediction time domain based on the initial state and the dynamic relationship.

[0100] An objective function is constructed to measure the tracking error between the predicted melt viscosity output sequence and the melt viscosity target trajectory sequence, while also weighing the stability of the candidate heating power sequence itself. Using the objective function as the optimization objective, and under the constraint of satisfying the heater power limit, the optimal power control sequence that optimizes the objective function is obtained through an iterative algorithm.

[0101] The first power control quantity corresponding to the current control cycle in the optimal power control sequence is output to the gate area heater; after entering the next control cycle, the above steps are repeated to achieve rolling optimization and feedforward control.

[0102] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A temperature control method for a hot runner system for inverted molds, characterized in that, The method includes the following steps: Step S10, defining a critical melt region inside the physical gate in the hot runner system gate that directly affects the quality of the product as a virtual gate, wherein the melt state of the virtual gate directly determines the gate quality; Step S20: Dynamically generate the melt viscosity target trajectory of the virtual gate area according to the injection molding material type and injection molding process stage; and estimate the current melt viscosity value of the virtual gate area in real time through a state observer based on measurable injection molding process parameters and gate area temperature; wherein the injection molding process parameters include at least one of injection speed, holding pressure, and screw position; Step S30: Using the melt viscosity target trajectory as the set target and the current melt viscosity value as the feedback quantity, calculate the optimal power control sequence using a multivariate model predictive controller; wherein the multivariate model predictive controller has a built-in dynamic process model that correlates injection molding process parameters, heating power, and melt viscosity of the virtual gate area; Step S40: Drive the gate area heater according to the optimal power control sequence, so that the actual change of melt viscosity in the virtual gate area tracks the melt viscosity target trajectory.

2. The temperature control method for a hot runner system for an inverted mold according to claim 1, characterized in that: The virtual gate is defined as a critical melt region inside the physical gate that directly affects the quality of the product in the hot runner system. The steps are as follows: Step S101: Based on the transient thermodynamic field simulation during the injection molding cycle, the region where the melt viscosity is between a first viscosity threshold and a second viscosity threshold is defined as the initial physical boundary of the virtual gate, with the gate center axis as the reference. The first viscosity threshold corresponds to the critical viscosity at which the material drools, and the second viscosity threshold corresponds to the lower limit critical viscosity of the material's filling fluidity. Step S102: Within the initial physical boundary, the melt region that undergoes a complete phase change cycle from a solid glass state to a fully molten state and then to the first viscosity threshold in each injection molding cycle is determined as the core region of the virtual gate, thereby obtaining the virtual gate.

3. The temperature control method for a hot runner system for an inverted mold according to claim 1, characterized in that: Based on the injection molding material type and injection molding process stage, the target melt viscosity trajectory of the virtual gate area is dynamically generated, including: Step S201, determining the theoretical optimal viscosity value of the melt at the virtual gate during the injection, holding, cooling, and mold opening stages based on the standard rheological curve of the injection molding material and the quality requirements of the target product; Step S202, dynamically correcting the theoretical optimal viscosity value based on the real-time status parameters of the injection molding machine to generate an achievable viscosity reference trajectory that matches the current machine capability; wherein, the real-time status parameters include at least one of the current maximum effective power of the heater and the system heat dissipation coefficient; Step S203, injecting a preset viscosity precursor signal before the key process switching point of the viscosity reference trajectory to form the final target melt viscosity trajectory; wherein, a negative step precursor signal is injected before the start of the injection stage to reduce the viscosity target by a preset amplitude; a positive pulse precursor signal is injected before the start of the mold opening stage to increase the viscosity target by a preset amplitude.

4. The temperature control method for a hot runner system for an inverted mold according to claim 3, characterized in that: Based on measurable injection process parameters and gate area temperature, the current melt viscosity value of the virtual gate area is estimated in real time through a state observer, including: Step S204, constructing a nonlinear state-space model with the melt viscosity, temperature, and thermal capacity state of the hot runner system in the virtual gate area as state variables, heater power and injection process parameters as input variables, and measurable gate area temperature as output variable; Step S205, designing an extended Kalman filter based on the nonlinear state-space model, and introducing an online model parameter identification mechanism to correct model mismatch caused by material batch differences and system aging in real time; Step S206, processing temperature sensor data at a high sampling rate, updating injection process parameters at the injection cycle rate, and achieving fusion of multi-rate measurement data through the extended Kalman filter to output the current melt viscosity value of the virtual gate area.

5. The temperature control method for a hot runner system for an inverted mold according to claim 4, characterized in that: The method of using a multivariate model predictive controller to calculate the optimal power control sequence includes: Step S301, in the current control cycle, obtaining the current melt viscosity value of the virtual gate region estimated in real time by the state observer, as the initial state for prediction; simultaneously, obtaining the melt viscosity target trajectory sequence within a future prediction time domain; Step S302, inputting the future predictable injection molding process parameter sequence as a feedforward input, along with a set of candidate heating power sequences, into the dynamic process model; the dynamic process model relates the injection molding process parameters, heating power, and the dynamic relationship of the melt viscosity in the virtual gate region, using the initial state as a starting point and predicting future trends based on this dynamic relationship. Step S303: Construct an objective function to measure the tracking error between the predicted melt viscosity output sequence and the target melt viscosity trajectory sequence, while also balancing the stability of the candidate heating power sequence's own variation. Using the objective function as the optimization objective, and under the constraint of satisfying the heater power limit, obtain the optimal power control sequence that optimizes the objective function through an iterative algorithm. Step S304: Output the first power control quantity corresponding to the current control cycle in the optimal power control sequence to the heater in the gate region. After entering the next control cycle, repeat the above steps to achieve rolling optimization and feedforward control.

6. A temperature control system for a hot runner system for inverted molds, characterized in that: The system includes: a virtual gate determination unit, used to: define a critical melt region inside the physical gate of the hot runner system gate that directly affects the product quality as a virtual gate, wherein the melt state of the virtual gate directly determines the gate quality; and a generation and observation unit, used to: dynamically generate a melt viscosity target trajectory of the virtual gate region according to the injection molding material type and injection molding process stage; and, based on measurable injection molding process parameters and gate region temperature, estimate the current melt viscosity value of the virtual gate region in real time through a state observer; wherein the injection molding process parameters include at least injection speed, One of the holding pressure and screw position; an optimal solution unit, used to: calculate the optimal power control sequence using the melt viscosity target trajectory as the set target and the current melt viscosity value as the feedback quantity, employing a multivariable model predictive controller; wherein, the multivariable model predictive controller has a built-in dynamic process model relating injection molding process parameters, heating power, and the melt viscosity of the virtual gate region; a heater control unit, used to: drive the gate region heater according to the optimal power control sequence, so that the actual change in melt viscosity in the virtual gate region tracks the melt viscosity target trajectory.

7. The temperature control system for a hot runner system for an inverted mold according to claim 6, characterized in that: The virtual gate determination unit is specifically used for: based on the transient thermodynamic field simulation during the injection molding cycle, taking the gate center axis as a reference, defining the region where the melt viscosity is between a first viscosity threshold and a second viscosity threshold as the initial physical boundary of the virtual gate; wherein, the first viscosity threshold corresponds to the critical viscosity at which the material drools, and the second viscosity threshold corresponds to the lower limit critical viscosity of the material's filling fluidity; within the initial physical boundary, the melt region that undergoes a complete phase change cycle from a solid glass state to a fully molten state and then to the first viscosity threshold in each injection molding cycle is determined as the core region of the virtual gate, thereby deriving the virtual gate.

8. The temperature control system for a hot runner system for an inverted mold according to claim 6, characterized in that: The generation and observation unit is specifically used for: determining the theoretical optimal viscosity value of the melt at the virtual gate during the injection, holding, cooling, and mold opening stages based on the standard rheological curve of the injection molding material and the quality requirements of the target product; dynamically correcting the theoretical optimal viscosity value based on the real-time state parameters of the injection molding machine to generate an achievable viscosity reference trajectory that matches the current machine capability; wherein the real-time state parameters include at least one of the current maximum effective power of the heater and the system heat dissipation coefficient; injecting a preset viscosity precursor signal before the key process switching point of the viscosity reference trajectory to form the final melt viscosity target trajectory; wherein a negative step precursor signal is injected before the start of the injection stage to reduce the viscosity target by a preset amplitude; and a positive pulse precursor signal is injected before the start of the mold opening stage to increase the viscosity target by a preset amplitude.

9. The temperature control system for a hot runner system for an inverted mold according to claim 8, characterized in that: The generation and observation unit is further specifically used for: constructing a nonlinear state-space model with melt viscosity, temperature, and thermal capacity of the hot runner system in the virtual gate region as state variables, heater power and injection process parameters as input variables, and measurable gate region temperature as output variable; designing an extended Kalman filter based on the nonlinear state-space model, and introducing an online model parameter identification mechanism to correct model mismatch caused by material batch differences and system aging in real time; processing temperature sensor data at a high sampling rate, updating injection process parameters at the injection cycle rate, and achieving the fusion of multi-rate measurement data through the extended Kalman filter to output the current melt viscosity value of the virtual gate region.

10. The temperature control system for a hot runner system for an inverted mold according to claim 9, characterized in that: The optimal solution unit is specifically used to: in the current control cycle, obtain the current melt viscosity value of the virtual gate region estimated in real time by the state observer, as the initial state for prediction; at the same time, obtain the melt viscosity target trajectory sequence in the future prediction time domain; and input the future predictable injection molding process parameter sequence as feedforward input, together with a set of candidate heating power sequences, into the dynamic process model. The dynamic process model correlates the injection molding process parameters, heating power, and the dynamic relationship between the melt viscosity in the virtual gate region. Starting from the initial state, it predicts the melt viscosity output sequence of the virtual gate region in the future prediction time domain based on this dynamic relationship. An objective function is constructed to measure the tracking error between the predicted melt viscosity output sequence and the target melt viscosity trajectory sequence, while also balancing the stability of the candidate heating power sequence's own changes. Using the objective function as the optimization objective, under the constraint of satisfying the heater power limit, an iterative algorithm is used to obtain the optimal power control sequence that optimizes the objective function. The first power control quantity corresponding to the current control cycle in the optimal power control sequence is output to the heater in the gate region. After entering the next control cycle, the above steps are repeated to achieve rolling optimization and feedforward control.

Images