Method for Synchronous Speed Control of Extrusion Blow Molding Machines

Speed Synchronization Control Methods for Extrusion Blow Molding Machines

Achieving precise speed synchronization across multiple drive axes in an extrusion blow molding machine is not a luxury — it is the backbone of consistent wall thickness, dimensional accuracy, and overall product quality. When the extruder screw, parison take-off, blow molding mold actuation, and haul-off or winder units fall out of step, the result is wasted material, rejects, and downtime. Below is a deep dive into the control strategies that manufacturers rely on to keep every axis locked in harmony.

Why Speed Matching Matters So Much

In extrusion blow molding, the molten parison exits the die and must be handled with surgical timing. The screw rotation sets the volumetric throughput. The take-off or traction unit pulls the parison downward at a speed that directly determines wall thickness distribution. The blow mold closes, air inflates, and the finished part cools. If any of these speeds drift — even by a few percent — the parison wall thins in some zones and thickens in others, leading to weak spots, uneven cooling, or outright rupture during inflation.

The coupling between extrusion rate and traction speed is governed by a simple but unforgiving relationship:

Traction speed ≈ Extrusion speed × (die cross-section / target parison cross-section) × draw balance factor (typically 0.9 to 1.1)

When this ratio stays near unity, the parison is in a natural draw state and surface quality peaks. Deviations of just 3–5% can trigger visible defects — sharkskin, longitudinal streaks, or wall eccentricity. That tight tolerance window is exactly why open-loop speed control simply does not cut it.

Classical Synchronization Architectures

Master-Slave Follower Control

The most straightforward approach designates one drive — usually the extruder motor — as the master. All other axes, including the traction motor and any haul-off rollers, run as slaves that follow the master speed through a fixed ratio or real-time command tracking.

The advantage is simplicity. A single PID loop on the master, with slave drives receiving a scaled reference, keeps the system easy to commission. In practice, this works well when load disturbances are mild and the mechanical coupling between axes is stiff. However, the moment the traction rollers encounter a sudden load change — say, the parison catches on a guide or the mold clamp shifts — the slave lags behind. The PID on the slave corrects, but there is always a transient error. During that gap, wall thickness deviates.

Equal-Speed (Parallel) Control

In this scheme, every motor receives its own independent speed command. There is no inter-axis communication. Each drive runs its own PID loop against an encoder or resolver feedback signal.

Parallel control shines in systems where axes are mechanically decoupled and disturbances on one axis should not propagate to others. But that very independence is its Achilles heel for blow molding. If the extruder speed dips under a torque spike and the traction motor does not feel the parison stretches thin. By the time the traction PID recovers, several meters of off-spec parison have already been produced. For high-precision film or bottle applications, this approach alone is insufficient.

Cross-Coupling Control

Proposed in 1980 by Y. Koern at the University of Michigan for dual-axis motion platforms, cross-coupling control builds a bridge between axes. Each drive compares its actual speed not only to its own setpoint but also to the speed of its neighbor. The speed error between axes feeds a compensation term that is injected into both controllers.

Mathematically, if motor A runs at ω_A and motor B at ω_B, the coupling controller adds a term proportional to (ω_B − ω_A) into the command for motor A, and the opposite term into motor B. The result is that both axes converge toward a common speed much faster than independent PID loops could manage.

In extrusion blow molding, cross-coupling handles the extruder-to-traction link beautifully. When the extruder slows due to a melt pressure surge, the coupling term instantly tells the traction motor to decelerate as well, keeping the parison draw ratio stable. Industrial implementations have shown sync errors shrinking from ±0.5 mm down to ±0.05 mm when EtherCAT-based distributed clocks replace traditional analog synchronization.

Intelligent and Adaptive Strategies for Modern Machines

Fuzzy Logic Control for Highly Nonlinear Systems

The blow molding process is messy. Film thickness depends on air ring cooling, die gap, melt temperature, and ambient conditions — all interacting in ways that defy clean mathematical models. The relationship between a valve position on an automatic air ring and the resulting film thickness is nonlinear, time-varying, and strongly coupled across the ring circumference. One valve adjustment affects its neighbors significantly, and there is always transport lag.

Fuzzy logic sidesteps the need for an exact plant model. A thickness probe measures the film profile, the controller computes the deviation and its rate of change, then applies linguistic rules — "if thickness is thin and getting thinner, close the valve a little more" — to drive servo motors on each air ring segment. With adaptive scaling factors that shift as process conditions drift, fuzzy controllers have proven robust where classical PID either oscillates or gives up entirely.

Deviation Coupling and Ring Coupling

Building on cross-coupling, deviation coupling control compares each axis output against a common reference trajectory and feeds the synchronization error back as a compensating signal. Ring coupling goes further: it considers both the tracking error of each axis to its setpoint and the mutual error between adjacent axes. For blow molding machines with multiple traction rollers or multi-layer die systems, ring coupling ensures that no single roller becomes the weak link in the chain.

Neural Network and Model-Predictive Approaches

More recent research layers neural network estimators on top of the control loops. A trained network predicts the parison wall thickness profile based on real-time inputs — screw speed, melt pressure, die temperature, traction speed — and adjusts the setpoints proactively rather than reactively. Model-predictive control (MPC) takes a similar philosophy, solving a short-horizon optimization at each control cycle to compute the best speed trajectory for all axes, respecting constraints like maximum motor torque and minimum gap between parison and mold.

Practical Implementation Tips

PREVIOUS：Multi-axis coordinated control method of extrusion blow molding machine NEXT：Real-time monitoring and control of pressure in the extrusion blow molding machine