SBIR: Cooperative Heuristics for Additive Manufacturing Processing (CHAMP)

OUSD (R&E) critical technology area(s): Human-Machine Interfaces

Objective: Design, develop, and demonstrate a proof-of-concept slicer, tool path algorithm, and production system which demonstrates non-linear increases in production rates for directed energy additive manufacturing (AM) for N>1 deposition heads.

Description: The Department of Defense (DoD) has identified additive manufacturing (AM) as a key enabler for supply chain agility and enhanced warfighter capability [1]. To date, multiple proofs-of-concept have been demonstrated, however significant infiltration into the industrial base and supply chain remains elusive. While agile in component design and near-net shape fabrication, AM remains a much slower process than conventional manufacturing approaches (e.g. casting, forging, machining), particularly when volumes >10 or sizes >40cm are needed. CHAMP seeks to break this paradigm and enable a step change (>10X rate) in metallic production capacity for large >50cm and medium volume >25 DoD critical components.

Directed energy deposition (DED) offers the greatest opportunity for high deposition rate and large parts due to the unconstrained volume, variety of feedstock available (e.g. wire and powder), and readily available energy sources (e.g. laser and arc). Currently, most DED systems contain one deposition head, however experimentation to increase rate with additional deposition heads is underway (e.g. ORNL MedUSA). In both cases, a major rate limiting step is the ability to efficiently structure tool paths to enable the most time efficient production route [2]. While adaptive strategies have been investigated [3], the complexity increases significantly in 3D geometries due to thermal build up and stresses, and even further when contending with multi-head spatial awareness.

DARPA is seeking innovative approaches to solving the tool path problem for multi-deposition head DED AM leveraging heuristic informed cooperative learning. Applying cooperative and collaborative learning to robotics is a relatively new field but has shown significant promise in optimizing multi-system interactions [4,5]. Additional work on reduced order modeling of processing [6,7] can provide a heuristic framework for decision making, while requiring minimal computational power. CHAMP performers will combine these approaches to demonstrate >10X production rate on components >50cm. Proposers should consider the full implication of existing and future modeling and in situ monitoring capabilities to inform heuristics and provide a solution that can adapt and improve in future materials and manufacturing improvements. This effort should focus on developing collaborative approaches to multi-head manufacturing while leveraging heuristic knowledge of the manufacturing process to inform tool path. Proposers should include background in the areas of interest, testing capabilities (virtual and experimental), and proposed path to demonstration. Developing new additive manufacturing approaches or new DED deposition technology is not in scope for this effort. Software only proposal for phase 2 is not in scope for this effort. Iterative computational trial and error is not in scope for this effort.

Phase I

This topic is soliciting both Phase I and Direct to Phase II (DP2).

DP2 Proposals:
Phase I feasibility must be demonstrated through evidence of: a completed proof of concept/principal or basic prototype; definition and characterization of properties/capabilities desirable for DoD/government and civilian/commercial use; and capability/performance comparisons with existing state-of-the-art technologies/methodologies (competing approaches).

Entities interested in submitting a DP2 proposal must provide documentation to substantiate that the scientific/technical merit and feasibility described above has been achieved and describe the potential commercial applications. DP2 Phase I feasibility documentation should include, at a minimum:

• technical reports describing results and conclusions of existing work, particularly regarding the commercial opportunity or DoD insertion opportunity, risks/mitigations, and technology assessments
• presentation materials and/or white papers/technical papers
• test and measurement data
• prototype designs/models
• performance projections, goals, or results in various use cases

The collection of Phase I feasibility material will verify mastery of the required content for DP2 consideration.

Phase I proposals:
Phase I consists of a base period of 8 months that will result in the development and demonstration of a computational framework capable of leveraging N>=2 DED deposition heads and demonstrating >2X increase in production rate over state of the art (SOA) single head production.

Successful proposals for this SBIR must offer significant arguments supporting the ability to rapidly iterate and execute to meet the timelines laid out in this solicitation, while addressing three key aspects of the program goals: (1) how multiple deposition heads work collaboratively to increase total production rate, (2) how to efficiently use of heuristic data to rapidly make decisions on efficient production pathways, (3) how the system will be integrated and tested in Phase II. Successful proposals will also demonstrate an in-depth knowledge of manufacturing optimization and should illustrate how their method might be expected to meet the envisioned metrics.

The Phase I effort is expected to center on building the computational foundation for further exploration and demonstration in Phase II. Emerging AI and ML methods are suitable for investigation. Integration of commercial off the shelf (COTS) software may be applicable, however successful proposers should demonstrate how their approach will significantly advance SOA and satisfy Phase I metrics. A successful Phase I program should clearly identify a path to successful completion of Phase II metrics.

Phase I fixed payable milestones for this program should include:

Phase I Base Period (required): 8 months

• Month 2: Concept Design Review (CoDR) on computational approach, selection of baseline test geometry, and baseline production rate
• Month 4: Preliminary Design Review (PDR). Initial report on feasibility to meet Phase I metrics
• Month 8: Critical Design Review (CDR). Interim report on TRL status, status of Phase I metrics, and path to successful demonstration of Phase II metrics.

Performers may perform experimental or virtual testing to validate Phase 1 metrics. Performers will work with DARPA to identify potential transition partners for demonstration in Phase II. Performers will present plans to design and manufacture prototypes in Phase II.

Phase I Metrics:

• >2X increase in production rate for N=2 deposition heads
• Collaborative approaches to 0% chance of head collision during processing
• Optimized tool path based on >=1 heuristic input (e.g. Stress, thermal buildup, resolution)

Phase II

The Phase II effort consists of a base period of 18 months and an Option period of 8 months. The base period should focus on rapid integration of the computational framework developed in Phase I into testable hardware. Experimental validation of virtual testing will be required. While demonstrating experimental validation, performers should continue improving the computational framework for cooperative heuristic tool pathing, with a focus on introducing additional heads (N>=3) and variable feedstocks or energy sources (N>=2). Introducing varying feedstocks and energy sources will introduce additional heuristic analysis (N>=2) but provides a path to increased resolution and rate in future system manifestations. Successful proposals should consider how to efficiently incorporate the growing number of deposition technologies.

Phase II fixed payable milestones for this program should include:

Base Period: 18 months

• Month 3: Bench Testing Review (BTR). First demonstration of >2X increase in production rate for N=2 deposition heads using selected baseline test geometry from Phase II. Identify, 3 additional components with increasing geometric complexity for further validation ((1)first >50 cm in single direction, (2)second >60 cm in two directions, (3)third, >70 cm in three directions)
• Month 6: Initial report on incorporation of >=2 heuristic inputs in computational framework (virtual testing is acceptable)
• Month 9: Initial report on incorporation of N>=3 deposition heads and N>=2 heuristic inputs in computational framework. Experimental validation to newly selected component (1)
• Month 12: Initial report on experimental validation of selected component (2) utilizing N>=3 deposition heads and N>=2 heuristic inputs.
• Month 15: Initial report on experimental validation of selected component (3) utilizing N>=3 deposition heads, N>=2 feedstock or energy sources, and N>=2 heuristic inputs.
• Month 18: Interim report evaluating path to >10X increase in production rate for N>=2 deposition heads.

Option 1 (Component Testing): 8 Months

• Month 19: Selection of DoD relevant component for production and testing
• Month 20: N=1 deposition head production for baseline validation (3 parts) (external geometry, 5 tensile coupons in each X, Y, and Z direction)
• Month 24: N>=3 deposition head production for cross validation of performance versus baseline (3 parts) (external geometry, 5 tensile coupons in each X, Y, and Z direction)
• Month 26: Update Phase 2 report documenting comparative results and future work required to close performance gap (if present). Present future commercialization plans.

Phase III dual use applications

Increased production rate of AM has wide ranging applications for commercial and DoD applications via reduced cost and supply chain agility. Multiple applications are envisioned after successful demonstration of a cooperative heuristic informed, multi-head DED system:

1. DoD use for deployable manufacturing. The ability to produce components as needed, in the field for both repair and mission enhancement will provide significant warfighter advantage. Deployable manufacturing as thus far been limited to small components; however, this ability will enable distributed, large-scale repair and manufacturing.
2. Commercial use by US Oil and Gas industry, and other large processing industries. Large pumps typically require extensive casting facilities or machining of very large billets which can create lead times of months to years. Increasing the capability of large-scale AM can break the cost paradigm to enable expanded industrial use, enabling more distributed production capacity.

References

1. “Department of Defense Additive Manufacturing Strategy”, January 2021

2. M. Murua, et al. “Tool-path problem in direct energy deposition metal additive manufacturing: sequence strategy generation”, IEEE Access, May 2020

3. F. Kaji, et al. “Robotic laser directed energy deposition-based additive manufacturing of tubular components with variable overhang angles: Adaptive trajectory planning and characterization”, Additive Manufacturing, Vol. 61, 2023

4. H.M. La, et al., “Multirobot cooperative learning for predator avoidance”, IEEE transactions on control systems technology, Vol 23. No 1. Jan 2015

5. L. Ferranti, et al., “Distributed nonlinear trajectory optimization for multi-robot motion planning”, IEEE transactions on control systems technology, Vol 31. No 2. March 2023

6. V. Perumal, et al., “Temporal convolutional networks for data driven thermal modeling of directed energy deposition”, Journal of Manufacturing Processes, Vol 85, January 2023 7

7. D.S. Ertay, et al., “Thermomechanical and geometry model for directed energy deposition with 2D/3D toolpaths”, Additive Manufacturing, Volume 35, October 2020

Keywords

Additive Manufacturing, fabrication process, manufacturing efficiency, adaptive control, artificial intelligence, computer-aided manufacturing, decision theory, intelligent manufacturing

TPOC-1

DARPA BAA Help Desk

Email

SBIR_BAA@darpa.mil