Page 34 - IJAMD-1-2

International Journal of AI for
Materials and Design
AI-driven quality assurance in AM

optimization, and predictive maintenance, thereby also extends the lifespan of AM equipment and reduces
improving the reliability, efficiency, and consistency of maintenance costs. 16,18,30 Moreover, AI-driven predictive
AM operations. 21,26,28 In this section, we delve into the maintenance enables condition-based maintenance
various ways AI contributes to quality assurance in AM, strategies tailored to the specific needs and operating
highlighting its key methodologies, applications, and conditions of AM machines, optimizing resource allocation
benefits. Table 3 outlines the specific implementation and enhancing operational efficiency.
methods and details how these methods can be practically
implemented to improve the quality, reliability, and 3.4. Design optimization
efficiency of AM processes. AI-driven generative design tools enable engineers
to explore vast design spaces and generate optimized
3.1. Defect detection and classification geometries that meet performance requirements
AI algorithms, particularly deep learning models such as while minimizing material usage and manufacturing
CNNs, excel in analyzing large volumes of imaging data constraints. 30,31 By integrating AI algorithms with CAD
to detect and classify defects in AM parts. 15,16 By training software and finite element analysis tools, designers can
on labeled datasets of defective and defect-free parts, rapidly iterate and evaluate design alternatives, considering
CNNs can learn complex patterns and features indicative factors such as stress distribution, weight reduction, and
of various types of defects, including porosity, surface thermal performance. Generative design techniques
irregularities, voids, and cracks. 27,28 These AI-driven defect leverage AI to generate innovative and unconventional
detection systems can identify defects with high accuracy designs that leverage the unique capabilities of AM,
and reliability, enabling manufacturers to perform non- such as lattice structures, topology optimization, and
destructive testing and quality control during the printing organic shapes. By automating the design process and
31
process. 10,11 Moreover, AI algorithms can adapt and leveraging AI-driven optimization, engineers can create
generalize to new defect types and variations, making AM parts with improved functionality, performance, and
them versatile tools for defect detection in diverse AM manufacturability, driving innovation and accelerating
applications. product development cycles.
3.2. Process monitoring and control In summary, AI plays a pivotal role in enhancing quality
assurance in AM processes by enabling defect detection,
AI-powered monitoring systems enable real-time tracking process monitoring, predictive maintenance, and design
and analysis of key process parameters, such as temperature, optimization. 32,33 By leveraging the power of AI-driven
humidity, build speed, and laser power, to ensure analytics, manufacturers can achieve higher levels of
consistency and quality throughout the AM process. By quality, reliability, and efficiency in AM operations,
29
integrating sensor data with AI models, manufacturers unlocking new opportunities for innovation and growth.
can identify deviations from optimal process conditions, However, realizing the full potential of AI in AM quality
detect anomalies, and predict potential defects before they assurance requires addressing challenges related to data
occur. Adaptive control algorithms can dynamically adjust availability, model interpretability, and integration with
process parameters in response to changing conditions, existing manufacturing workflows. 17,33 By overcoming
minimizing the risk of defects and optimizing part these challenges and embracing AI-driven approaches,
quality. 17,22,25,28 Furthermore, AI-driven process monitoring manufacturers can realize the transformative benefits of
enables continuous improvement and optimization of AM AM and drive the next wave of industrial revolution.
processes, leading to increased productivity and reduced
production costs. By following these practical implementation steps,
organizations can effectively harness ML techniques to
3.3. Predictive maintenance enhance quality assurance in AM processes. This systematic
AI-based predictive maintenance systems leverage sensor approach not only addresses the specific challenges faced
data from AM machines to anticipate equipment failures, in AM but also ensures that AI-driven solutions are
diagnose issues, and schedule maintenance proactively. robust, scalable, and aligned with industry standards and
30
By analyzing patterns and trends in sensor readings, AI regulatory requirements.
algorithms can detect early warning signs of equipment 4. Case study
degradation or malfunction, allowing maintenance
personnel to take preemptive action before catastrophic AM processes have revolutionized manufacturing
failures occur. Predictive maintenance not only reduces industries by enabling the production of complex
31
unplanned downtime and production interruptions but geometries with reduced material waste and lead times.

Volume 1 Issue 2 (2024) 28 doi: 10.36922/ijamd.3455