Part quality in additive manufacturing (AM) is highly dependent on the process control, but there is a lack of adequate AM control methods and standards. In particular, standard programing language (such as G-code / M-code for conventional machine tools) is not well defined, method for synchronizing laser power to scan speed does not exist. Furthermore, there is no uniform way of implementing laser scan strategies among the AM machine vendors. Therefore, the scan strategy customization is not only difficult, but even with identical strategy, due to varying implementation methods, the build quality could be very different on different machines. This creates great uncertainty in part qualification process. This project will focus on developing algorithms, methods and standard protocols for AM process control. Specifically, an open-platform modular AM software will be developed, with all control parameters accessible, and build-in modes to implement typical control methods. A portable (universal) AM controller will also be developed to execute the build commands / scan strategies developed by the software, and build quality will be studied to verify and optimize the control methods. Ultimately, control methods will be formulated based on mode and parameter settings in the software / controller, and consistent built quality should be expected on the compatible systems.
Develop algorithms, methods and standard protocols for Additive Manufacturing (AM) process control, and implement it with software and hardware tools for open control of AM systems to enable more flexible process optimization by manufacturers.
What is the Technical Idea?
Additively manufactured metal parts can have many quality issues, such as pores or cracks, dimensional errors, poor surface finish, as well as more difficult to detect qualities such as residual stresses and inhomogeneous microstructure. The part quality is primarily affected by AM process control, though there is a lack of control methods and standards. This inadequacy has been clearly identified in the AMSC Standardization Roadmap as gaps - PC1: Digital format and digital system control. PC5: Parameter Control. PC17: Motion control. D6: Software-encodable/Machine-readable Guidelines.
The NIST Additive Manufacturing Metrology Testbed (AMMT) development process had confirmed all these gaps, and many efforts were spent to overcome them. A G-code based AM machine control language was defined as the programming language for AMMT, and a corresponding G-code interpreter was developed and embedded with motion control software. A controller was also built to execute the build commands. This actually creates a framework for an open platform AM process control, though many improvements are still needed to develop it into a useful tool and reference for AM researchers and manufactures. Therefore, in this project following key products / deliverables are proposed.
What is the Research Plan?
The AMMT development has accumulated many best practices and methods in AM process control, such as power-speed-position synchronization and thermal conductivity prediction (feedforward control), which need to be further developed into standard protocols. The AM software to be developed will be based on these protocols and provide end to end capability to process CAD designs into AM build commands. The software will be open source, key functions such as STL slicing, path planning and G-code interpretation will be packaged into modules. Each module can be used separately or embedded into other applications, and interfaces will be provided for user developed plug-ins, such as various thermal conductivity models. A portable AM controller will be developed (also as part of AMMT improvement) to facilitate the technology transfer to various stake holders by interfacing the AM software with their existing AM equipment, such as galvo and laser units. Feedback control methods and protocols will also be developed based on the real-time monitoring signals, such as melt-pool size measurements and layer-wise imaging.
The AMMT improvement / maintenance will be made from following perspectives: (1) Based on the existing AMMT architecture, review the control block diagram, identify the time critical and resource critical subroutines, streamline the command and data flow, standardize the control interface, develop the fault detection mechanism, and enforce quality assurance and version control to reduce the down time. (2) Develop periodic and/or continuous feedback control based on melt-pool measurement, layer-wise imaging and/or other real-time monitoring feedbacks. (3) Develop a new controller following the standard protocols, and based on a cost-effective single system (such as PLC or FPGA) without any customized hardware. Some real-time monitoring capabilities may be limited, but the reliability and usability should be greatly enhanced. Meanwhile this also provides a sample controller design for the AM software / protocol, and makes the technology more easily to be adapted.
Unique capabilities at NIST (i.e. AMMT) will be utilized to investigate the causal relationships between scan strategy and part quality metrics. Optimization objectives will be identified with NIST and external collaborators, strategies will be developed based on the material, part design and function. Build orientation effects will also be investigated for different scan strategies. The AM software, AMMT and scan path planning hence form a closed loop for the continuous development / improvement of AM process control protocol and standards.
The control strategies developed for the laser powder bed fusion (LPBF) method (AMMT) will also be implemented for the directed energy deposition (DED) method (Optomec). Interfaces will be developed to utilize the standard modules such as STL slicer and path planning of the AM software, and such compatibilities will also help to develop more universal AM control protocols.