Generative AI for Lightweighting and Cost Reduction of Washing Machine Parts

Generative AI that explores lightweight, high-stiffness designs for washing machine structural components under strict performance targets

Generative AI for Lightweighting and Cost Reduction of Washing Machine Parts — Narnia Labs case study (originally: 세탁기 부품의 경량화 및 원가 절감을 위한 생성형 AI)

Project Overview

Developed a generative AI workflow to lightweight and cost-reduce structural components in large washing machines while preserving their performance targets. From a small set of core geometric criteria, the system produces a large variety of candidate designs and automatically evaluates their structural response, enabling exploration of the safety-vs-cost trade-off far beyond what a human engineer can realistically try.

My Role

AI Researcher, Narnia Labs

  • Individual contributor on the implementation code for the parametric-design and structural-verification pipeline
  • Built and tuned the generation–evaluation loop that surfaces the mass-vs-stiffness trade-off

The Challenge

Reducing part cost while preserving structural performance is constrained by both data and process limits:

  • Data-limited design exploration: Innovative structural changes are desirable for cost reduction, but limited reference data restricts how broadly the initial design space can be explored.
  • Cost vs. durability trade-off: Large-drum washing machines subject parts to high rotational and dynamic loads. Reducing part weight and cost while retaining safety margins is a challenging multi-objective problem.

Technical Approach

Delivered an intelligent design–verify pipeline for structural components:

  • Auto-generation from a minimal seed: Given core geometric criteria for a part, the system combinatorially generates a large number of candidate designs, breaking free from the initial data limitation.
  • Automated structural verification: Instead of engineers running individual analyses, the system evaluates each candidate’s structural response — vibration, load-bearing — automatically.
  • Intelligent optimization (roadmap): A next step is to close the loop with an AI that inverts the evaluation results into a proposed “minimum weight, maximum stiffness” design.

Impact

  • Unbounded structural-design exploration: Frees the search from the limits of legacy data and human intuition, systematically covering a much broader design space.
  • Cost-quality trade-off surfaced explicitly: Automatically analyzed candidates reveal the optimal frontier that reduces mass (cost) while retaining strength (quality).
  • R&D efficiency through pipeline automation: Replaces manual test cycles with an automated pipeline, accelerating early-stage decisions and demonstrating tangible cost-reduction potential.

Client: Electronics manufacturer (undisclosed) Organization: Narnia Labs Category: Electronics

Source: Narnia Labs Case Study #48