Grantee Research Project Results

Final Report: Enhanced Efficiency of UBC Recycling Powered by Artificial Intelligence

EPA Contract Number: 68HERC24C0032
Title: Enhanced Efficiency of UBC Recycling Powered by Artificial Intelligence
Investigators: Molstad, Emily
Small Business: Valis Insights, Inc.
EPA Contact: Richards, April
Phase: I
Project Period: December 1, 2023 through May 30, 2024
Project Amount: $100,000
RFA: Small Business Innovation Research (SBIR) - Phase I (2024) RFA Text |  Recipients Lists
Research Category: Small Business Innovation Research (SBIR)

Description:

VALIS is dedicated to delivering software and data solutions that enhance metal recycling by empowering recycling facilities to make data-driven process decisions that increase material recovery and profitability. VALIS has developed its flagship product, VALI-Sort, a software application that integrates with existing processing equipment, to enable metal recyclers to tap into data that is already being generated throughout their plant but is inaccessible or unrealized in its current form. To date, VALIS’s technology has been built for and deployed within the automotive recycling market, but it presents an opportunity to increase the recycling rate of used beverage containers (UBC), which currently stands at a mere 45%, and support the recovery of the roughly $1 billion worth of cans that end up in US landfills each year. The purpose of this research was to validate integration with processing equipment and data streams relevant to UBC recycling facilities, to adapt the core data models, and to buildout novel insights features, with the ultimate goal of demonstrating VALI-Sort’s potential to drive improvements within the UBC market.

During the SBIR Phase 1, three objectives were accomplished. These objectives served to improve data capture, quality transparency, and process performance visibility for UBC recycling facilities, enabling them to make better, data driven decisions.

• Objective 1: Hardware Integration – Develop and deploy integrations with the equipment of interest at a partnering UBC recycling facility. This includes connection to the feeding and sortation equipment.
• Objective 2: Software Integration – Deployment of necessary integrations to capture operational data, managed externally from the VALI-Sort application. This objective encompasses backend implementations and model adaptations.
• Objective 3: Application Adaptation – Adaptation of data models and features from their current applicability at automotive recycling facilities to align with UBC processing facilities.  This objective is addressed through novel implementation of backend data models/services and frontend buildout.

Summary/Accomplishments (Outputs/Outcomes):

Objective 1 - Hardware Integration: The VALIS team, in coordination with its piloting partner, and the partner’s equipment providers and system integrators, carried out development efforts needed to integrate with the equipment of interest in the UBC recycling line installed at the selected facility, including sortation technology and upstream feeding systems. The work completed under Objective 1 resulted in the de-risking of the critical integration points within the UBC recovery line. The VALIS team was able to capture all the variable processing parameters associated with the recovery of UBCs from a mixed feedstock material. The performance of the UBC recovery process was boiled down into KPIs, measurable by the data that VALIS is collecting and analyzing. By connecting with and capturing all the parameters of the recovery line, the VALI-Sort product enables insights into the relationship between those parameters and the identified performance metrics. The VALIS team is now collecting all system and equipment parameters of interest across the entire UBC recovery line, with the exception of two sorting systems due to a delay in the commissioning of the equipment. Data capture from these systems has been fully de-risked and is expected to be complete shortly after the conclusion of the reporting period.

Objective 2 - Software Integration: The VALIS team established, through conversations with the partnering facility, that non-automated data collection of procurement, production, and sales material data is acceptable as an initial implementation of VALI-Sort for their UBC recovery process. The partnering facility demonstrated their existing data management system along with their other supplemental data platforms. These conversations have informed the development team on the production insights of interest, shortcomings of existing material data management, and critical data points that the VALI-Sort model will need to manage.

The major results of the work completed under Objective 2 include the de-risking of integration with the partnering facility’s existing software; updates to the material sourcing and exports data model in the VALI-Sort application; the buildout of a UBC production dashboard feature  with functionality to import existing inventory and production tools; and support of a UBC recovery process sampling methodology, to provide more accurate supplier KPIs and process performance auditing capabilities. The VALI-Sort application supports various material container types, supplier management methodology, customer management methodology, and provides the partnering facility with a breakdown of production KPI’s.

Objective 3 - Application Adaptation: The VALIS team built out a set of insight features in parallel with the integration efforts completed in Objective 1. The major results of the application adaptations include the de-risking of the association of system generated data and material in inventory; the completion of the sourcing insights feature which provides the user with automated, detailed reports on supplier quality; a multitude of performance metric presentations based on KPIs identified through conversations with production stakeholders; and the completion of the facility insights feature, which provides the user with a high-level health check on their facility’s performance. Conversations with the partnering team towards the end of the contract identified additional KPIs of interest that the VALIS team will be pursuing.

Conclusions:

The successful completion of the previously described objectives has resulted in key findings that serve as proof of concept for VALI-Sort’s applicability in UBC recycling and to inform future development activities to deliver a production-ready version for the applicable customer segment. These development activities will enable the operational utilization of the VALI-Sort application at the partnering UBC recovery facility. Additionally, these improvements to the VALI-Sort application will enable VALIS to serve the product to a larger segment of metals recovery facilities, including those with multi-step processes, and will extend to meet the specified needs of UBC recovery facilities, as compared to automobile recycling facilities. Additional hardware integration and software development work will be required for subsequent integrations and customer segment expansions. The fundamental backbone of the product is more robust and will be able to support more out-of-the-box integrations with new metals recovery facilities.

The VALIS team carried out a multitude of commercialization activities throughout the reporting period including showcasing of capabilities at the Recycled Material Association (ReMA, formerly ISRI) Annual Tradeshow and Convention. With over 6,000 attendees from across the industry this allowed for significant sales pipeline development and product validation. Additionally, as a part of the Technical and Business Assistance (TABA) program Foresight Science and Technology conducted a market survey and delivered a commercial readiness assessment report to VALIS. During the survey, potential customer interviews were conducted which showed strong support for VALIS products such as VALI-Sort. The feedback on VALI-Sort's technology was summarized as providing high value to customers and would be highly competitive in the market. Foresight stated:

The current software in this industry does not adequately address industry needs. It is described as collecting very limited data, prone to inconsistencies, and with very little to no AI or machine learning capabilities. The key advantage of the proposed software tool, VALI-Sort, is its capability to integrate directly with sensor-based sorting infrastructure to provide high value insights into the quality and value of materials recycling streams. Finally, stakeholders have expressed a serious level of interest in the technology.

Based on preliminary data, the VALIS team estimates that the partnering demonstration facility can leverage VALI-Sort’s capabilities to increase monthly profit while increasing its recovery rate at the facility studied, which is good for their bottom line and for the planet.  Interactions with the partnering facility and other industry leaders indicates strong commercial interest in VALI-Sort. The applications of VALI-Sort expand beyond integration with sorting systems and into a full facility optimization and integrated management platform for the recycling market.

SBIR Phase II:

Enhanced Efficiency of UBC Recycling Powered by Artificial Intelligence