Temperature Control of Digital Glass Forming Processes

Abstract: Digital Glass Forming (DGF) is a new manufacturing process for low-batch glass fabrication. The work zone temperature in DGF processes must be maintained in the glass's working range to ensure good fabrication. If the temperature is too low, the filament will not wet to the substrate or previously deposited material and, if the temperature is too high, the filament may disengage from the substrate or previously deposited material, or it may partially vaporize. In this work, a real-time temperature control system capable of synchronizing process parameter, thermal camera, and visual camera data for the DGF process is introduced. A process parameter map for a scan velocity of 0.5 mm/s is constructed, as is a data-driven dynamic temperature process model. A digital controller is designed to regulate the work zone temperature. The temperature controller is a closed loop tracking controller that adjusts the commanded laser power to regulate the measured temperature. Two sets of experiments are conducted to analyze the controller performance. In the first set of experiments, single tracks on a substrate are fabricated with constant laser power and with the closed loop temperature controller. It is seen that the closed loop controller is able to extend the process parameter map into regions where using a constant laser power will result in a failed build. In the second set of experiments, walls are fabricated. Using constant laser power results in a failed build (i.e., material vaporization at the corners and the filament prematurely detaching from the substrate) as the temperature process dynamics change with layer and at the corners. The closed loop controller successfully fabricated the wall without vaporization at the corners and premature filament detachment as the controller adjusts the laser power to account for the changing temperature process dynamics.

• The paper introduces a closed loop temperature regulation system for digital glass forming that extends the process envelope and improves build stability.
• The methodology employs spatially resolved thermal imaging and a first-order LTI model achieving over 70% prediction accuracy for effective real-time control.
• Experimental results demonstrate that the system mitigates defects such as filament detachment and vaporization, enabling defect-free, continuous deposition.

Temperature Regulation in Digital Glass Forming: Real-Time Control and Process Expansion

Introduction

This work addresses temperature regulation in Digital Glass Forming (DGF), a critical bottleneck in automated, low-batch, high-precision glass fabrication. The process’s success is dominated by the temperature-dependent viscoelasticity of glass, requiring work zone temperatures within specific bounds to ensure adhesion and structural integrity. The study critiques previous feedback and pseudo-control approaches for their lack of adaptability, susceptibility to misalignments and emissivity variations, and slow, trial-heavy process parameter search. It proposes an RT-Linux-based, spatially resolved thermal imaging system with closed loop digital control, claiming significant advancement in stability, process extensibility, and part quality.

Experimental System and Sensing Architecture

The hardware platform is built around a high-power 500 W Yb-fiber laser, a custom volumetric focusing system, programmable filament feeding, and high-precision motion/positioning infrastructure. The work zone temperature is captured with a long-wave IR thermal camera (PI 640i), offering up to 125 Hz frame rates and sub-0.2 K NEDT for area measurements up to 1500°C. This provides dynamic region-of-interest capability as opposed to pyrometers’ fixed, single-point sampling, crucial for tracking the true high-intensity work zone despite filament deflections or path perturbations. Average emissivity is empirically set at 0.72 for glass at the relevant temperature ranges, but the authors highlight this as a domain needing further quantitative work.

Modeling and Identification of Process Dynamics

A first-order linear time-invariant (LTI) model is adopted for the mapping from the commanded laser power to the work zone temperature, with parametric identification using PRBS, chirp, and sinusoidal input signals. The steady-state operating point is centered at 42.6 W laser power and 888°C temperature, corresponding to defect-free, stable deposition. Model parameters, derived from system identification on experimental deposition data, achieve >70% prediction accuracy under the constraints of fixed scan velocity and filament feed. The transfer function exhibits gain ($b=0.6304$) and dominant root ($a=0.8296$ in z-domain), providing a computationally efficient but sufficiently robust model for closed loop control.

The model abstracts from several nonlinearities and dynamic variations—temperature-dependent specific heat, filament diameter stochasticity, process boundary changes (substrate thickness, thermal inertia), and rapid shifts in thermal conductivity and emissivity. Nevertheless, the low order model’s tractability is critical for real-time in-situ control, and the omitted nonlinearities are flagged as the major limiting factor for theoretical generalization.

Closed Loop Temperature Control Framework

A digital tracking controller is synthesized around the discrete model, employing the Internal Model Principle to guarantee offset-free reference tracking and disturbance rejection. Temperature feedback is the average of the 200 hottest pixels from the thermal camera, which confers resilience to spatial translations of the work zone and rapid perturbations. The control law achieves two designed closed loop time constants (0.1 s and 0.54 s), an order of magnitude faster than process thermal dynamics, supporting aggressive but non-oscillatory disturbance rejection. Commanded laser power is updated every 0.1 s, well within the hardware’s actuation capability.

Experimental Validation and Performance Gains

Single-Track Deposition Outside Standard Process Map

When fabricating tracks with process parameters outside the established viable window (e.g., df = 10 mm, L = 70 W), open loop (constant power) operation yielded cyclical failures—periodic wetting, lateral filament instability, and substrate detachment. The closed loop controller, with reference temperature set at 800°C, maintained stable, continuous deposition, extending the process map into previously unreachable regions.

At smaller focus distances (df = 3 mm) and low laser powers, open loop attempts resulted in non-wetting, substrate scratching, or vaporization/detachment depending on power. The closed loop controller maintained a stable work zone at 850°C using an average power of 20 W, demonstrating that real-time feedback efficiently adapts to tight parameter coupling and process sensitivity.

Multi-Layer Wall Fabrication

For multi-layer walls (16 layers, 20 mm long), open loop control at L = 40 W succeeded for the initial few layers but failed at the ninth corner due to rising work zone temperature (reduced thermal conduction as wall height increases, localized heat accumulation at corners). The closed loop control (Tr = 940°C) automatically reduced laser power as required, finishing deposition without vaporization, detachment, or excessive corner rounding, and demonstrating improved morphological fidelity.

Strong numerical evidence is shown for the closed loop system’s robustness: the controller adaptively decreases required power as thermal boundary conditions evolve, enables process regions previously associated with persistent failures, and suppresses the formation of corner defects and track discontinuities. Controlled builds feature higher repeatability and track quality, indicating improved manufacturability.

Theoretical and Practical Implications

This study demonstrates that spatially resolved thermal feedback, combined with low-order data-driven dynamic modeling and digital control, can drastically expand the operational envelope of DGF. These findings are applicable not only to DGF but also to other scenarios involving highly temperature-sensitive viscoelastic or thermoplastic materials, where parameter sensitivity and boundary evolution are severe. The decoupling of process defect mitigation from static parameter maps implies reduced need for trial-and-error tuning and improved robustness to filament variations or environmental disturbances.

Theoretically, this research reveals the limits of first-order linear models: while effective for real-time control, future advances require integration of process- and temperature-dependent nonlinearities, dynamic adaptation to boundary condition shifts, and potential data-driven (ML) augmentation for unmodeled process regimes.

On the practical engineering front, the work directly enables higher quality, lower-defect, and more complex glass structures in low-batch contexts, accelerating the deployment of DGF across scientific instrumentation, microfluidics, and bespoke optics. The framework could be generalized to full 3D closed loop microfabrication, in situ hybrid manufacturing (metal/glass), or feedback-driven systems for ceramics and high-temperature polymers.

Conclusion

This paper establishes an effective, real-time closed loop temperature regulation strategy for Digital Glass Forming, leveraging spatial thermal imaging and low-order data-driven modeling. Experimental results demonstrate that the proposed system significantly extends the viable process parameter space, adapts to thermal boundary transitions, and directly mitigates build failures such as filament detachment and vaporization. While the employed process model omits key nonlinear effects, its practical tractability and control efficacy mark an important advancement in the real-time feedback control of additive glass manufacturing. Future efforts should focus on extending modeling fidelity and controller adaptivity, with the long-term objective of generalized closed loop control for thermally sensitive digital fabrication platforms.

Reference: "Temperature Control of Digital Glass Forming Processes" (2604.00135)