How to use artificial intelligence: A guide for mechanical engineers

Artificial intelligence insights

• Artificial intelligence can help eliminate mundane tasks, such as customer service and lead generation.
• Machine learning helps integrate computerized data into predictions and analyses.

In mechanical engineering, the use of artificial intelligence (AI) to automate routine tasks, optimize and streamline processes and improve quality control and testing precision is on the rise. According to a McKinsey & Company report, the percentage of companies using AI jumped from 20% in 2017 to 50% by the end of 2022.

When used correctly, AI can free engineers from menial work, allowing them to focus on alternative design and other more complex and interesting tasks. The use of AI also opens the door for engineering innovations, such as the development of smart and autonomous systems. Additionally, AI plays a pivotal role in reaching targets related to sustainability, the internet of things, predictive analytics and more.

To maximize the benefits of this AI renaissance, it is important for engineers, especially mechanical engineers, to adapt by developing new skills, familiarizing themselves with the use of AI-powered tools, learning and contributing to domain-specific languages and collaborating with computer and data scientists.

Machine learning skills for mechanical engineers

Machine learning (ML), a rapidly growing field that is deeply intertwined with AI, involves training computers to learn from data to make predictions or decisions without being explicitly programmed. The following ML skills have clear applications for mechanical engineers:

• Supervised learning techniques, including regression and classification, help engineers build predictive models for mechanical systems. For example, an AI model could be trained to predict the performance of a mechanical system based on historical output data.

• Unsupervised learning techniques, such as clustering and dimensionality reduction, can identify patterns and groupings in data. They might, for example, group similar components in a mechanical system based on their characteristics.

• Reinforcement learning techniques are used to optimize decision-making processes. They can be used to boost the performance of a mechanical system by discovering and developing the optimal control strategy.

• Neural network training and design are commonly used in image and signal processing applications in mechanical engineering. For example, they can detect defects in mechanical components from images.

• Data analysis and preparation with tools such as Python, R and MATLAB, can identify patterns and correlations in data related to mechanical systems. This involves skills such as data cleaning, feature selection and data visualization.

Figure 1: Domain-specific languages (DSL) code enables mechanical engineers to perform design optimization. Courtesy: Shreedhar Murthy Hebbur Subbaraju

The rise of ML has also created a demand for advanced engineering skills, including:

• Hardware and infrastructure design: ML algorithms often require specialized hardware and infrastructure to run efficiently. Engineers need to have knowledge of hardware design, networking and cloud computing to build and deploy ML systems.

• Algorithm development and optimization: Developing and optimizing ML algorithms require advanced skills in mathematics, statistics and computer science. It’s critical for engineers to understand the underlying algorithms and techniques used in ML and how to optimize them for specific applications.

• Model deployment and maintenance: After ML models are developed, they are deployed and maintained in a production environment. This requires skills such as software engineering, testing and maintenance.

• Domain expertise: ML models are most effective when they are developed by experts with knowledge of the specific domain. It’s important for engineers to have a deeper understanding of the application area to design and train customized ML models and to help create domain-specific languages (DSLs).

Domain-specific languages in machine learning and artificial intelligence

DSLs are programming languages designed to address a specific need of a particular application domain. The use of DSLs is not yet commonplace, but adoption is becoming more widespread, with examples including OpenModelica and Simulink. Their benefits include simplified development, improved accuracy, faster deployment and customization potential. They are generally easier to use than traditional programming languages for engineers without programming experience.

Figure 1 is a sample snippet of code in a DSL that enables mechanical engineers to perform design optimization. This code enables engineers to specify the initial design, optimization constraints and objectives and the optimization algorithm to be used. By adjusting the parameters of the optimization algorithm, engineers can fine-tune the optimization process to achieve the desired results.

Here is a breakdown of the code line by line:

optimize_design: This DSL command is used to indicate that the user wants to perform a design optimization.

input: Specifies the input parameters that the optimization algorithm will use. In this example, the input parameters are the design, constraints and objectives. The design parameter represents the initial design that the algorithm will optimize, while the constraints and objectives parameters represent the optimization constraints and objectives, respectively.

algorithm: Specifies the optimization algorithm to be used. In this example, the algorithm is a genetic_algorithm. Other algorithms such as particle_swarm_optimization or simulated_annealing could also be specified here.

parameters: Specifies the parameters for the optimization algorithm. In this example, the population_size is set to 100, the mutation_rate is set to 0.1 and the generations is set to 50. These parameters can be adjusted to fine-tune the optimization algorithm for this specific application.

output: Specifies the output of the optimization algorithm. In this example, the output is the optimized design.

How can engineers develop artificial intelligence and machine learning skills?

A variety of online platforms, such as Coursera, edX and Udemy, offer educational courses in engineering, data science and AI. Industry conferences and workshops provide opportunities to network and gain hands-on experience with new tools and technologies. On-the-job training, including working on projects and attending knowledge-sharing sessions with more experienced colleagues, can be an effective way to learn. Engineers can also experiment with new technologies as part of personal projects; for example, building a prototype or working on an open-source project.

Proper use and implementation of AI in engineering require mechanical engineers to learn hard skills like data analysis, programming and machine learning. Equally important are soft skills such as communication and collaboration to work effectively with other experts. The use and development of domain-specific languages can aid in this collaboration while allowing engineers to play a more active role in developing AI for a specific purpose.

Building these skills will allow mechanical engineers to thrive as emerging technologies like AI continue to change the way engineers work.

Do you have experience and expertise with the topics mentioned in this content? You should consider contributing to our WTWH Media editorial team and getting the recognition you and your company deserve. Click here to start this process.