Hybrid Modeling for Energy Efficient CNC Grinding
Project Lead: West Virginia University
Partners: Indiana Technology and Manufacturing Companies (ITAMCO), University of Buffalo
Member % Cost Share: 50.08%
CESMII % Cost Share: 49.92%
Duration: 18 Months
Problem Statement
Grinding is the machining process with the highest specific energy consumption (SEC). Grinding is omnipresent in industry and essential for precise processing of hard and brittle parts. The total grinding time for ITAMCO’s End Gear 170 (see right) accumulates to ~11 hours and energy is responsible for 33% of the total manufacturing cost of a gear 170.
Project Goal
The principal goal is to reduce the extremely high energy consumption of grinding processes for gear manufacturing by at least 15% through hybrid modeling of the grinding system holistically.
Technical Approach
To achieve this goal, we will develop novel hybrid modeling methods that combine multi-physics equation-based models with data-driven machine learning models. The hybrid model input includes both machine tool parameters and sensor data as well as data from ERP and tool management systems. The hybrid model’s output provides grinding process parameters (wheel speed, depth of cut, infeed duration) as well as grinding tool reconditioning schedule and parameters (dressing and sharpening) that reduce the overall grinding system’s SEC. The model will be implemented in a Smart Manufacturing App on the CESMII SM platform. The industrial testbed will be located on-premise at ITAMCO, a leading US gear manufacturer and SME.
Deliverables/Outcomes/SM Marketplace
- Delivered complete data model template for machine tool messages
- Collected and validated grinding manufacturing process sample data set
- Documented generalized hybrid model for grinding
- Documented development framework for hybrid model
- Hybrid model Smart Manufacturing App source code submitted to CESMII GitHub repository
- Functional Hybrid model Smart Manufacturing App implemented at ITAMCO
- Smart Manufacturing profile for CNC grinding made available for reuse by the broader CNC grinding community.
- Data-driven and physics-based predictive models for CNC grinding
Potential Impact
- The US demand for gears is expected to grow by 6.4% to $40 billion in sales. Grinding will remain the core technology to produce large-scale, high-quality gear components. The novel, scalable, and generalizable hybrid modeling approach and its deployment in the CESMII SM platform environment will provide a blueprint for other use cases to reduce the energy consumption of the US grinding industry.
Benefits
The project showcases rapid recovery of Smart Manufacturing adoption cost through energy savings and productivity increases in an industry with energy intensive processes. The project creates an opportunity to scale its impact for other interested CESMII members across industries (automotive, aerospace, medical, etc.) and applications (milling, turning, etc.) within the larger CESMII network.
