top of page
  • Writer's pictureBrad Cortlandt

Machines Deciphering Their Own Blueprints

Lately, I've been thinking a lot about the future of machine learning and the endless possibilities it holds. I'm particularly fascinated by the idea of machine learning fully comprehending a system and its possibilities. I believe that we're laying the foundations for such advancements today, and one of the key components to this is understanding systems through drawings and schematics.

Drawings and schematics are the building blocks upon which systems and structures are built. They provide a clear insight into the design, functionality, and interrelations of various components. Unfortunately, the biggest challenge we face is the inconsistency across these drawings. Symbol conventions, design notations, and representation methods vary considerably, making it difficult to link drawings to real-time data from PLC tags and datasheets.

It would be ideal if every symbol, relationship, and data point in the documentation aligned seamlessly, but the reality is far from that. Manual interventions to reconcile these inconsistencies are labour-intensive and prone to human errors. It's nearly impossible to manually link every drawing for systems given how rapidly they change.

But why bother? We have drawings, monitoring systems that display sensor values, and datasheets. What advantage would machine learning bring to this if it could understand the drawings and schematics of a system?

Let's take a moment and imagine a scenario where machine learning models are trained to recognize and understand symbols, determine their relationships, link them to relevant datasheets, and then connect them to real-time values. The transformative potential of such an application is vast and largely untapped. It could provide the foundation for limitless possibilities, revolutionizing how we maintain and design systems and equipment.

Here are a few examples of how this advancement could drastically change how we operate:

Automated Design Optimization: The system will determine the most efficient design based on criteria like energy consumption, cost, or performance, proposing optimized system designs.

Adaptive System Evolution: The model will suggest incremental design changes over time, ensuring the system continuously evolves and improves even post-deployment.

Component Suggestion and Replacement: The system will suggest the most appropriate components or materials to use. If a component is discontinued or unavailable, it will propose alternatives, ensuring design continuity.

Environmental Impact Reduction: The system will suggest design changes or operational tweaks that reduce ecological impact.

Enhanced Drawing Review: The process of reviewing drawings will become much more efficient. Models could instantly spot inconsistencies, suggest changes, or even predict potential challenges that might arise during the implementation phase.

Troubleshooting & Maintenance: The automated system could quickly pinpoint the issue, provide potential solutions, and recommend preventative measures for the future.

Innovative Designs: The designing of new systems or components that we still need to conceptualize. The innovation potential is boundless.

The list is endless, but you get the idea. Integrating machine learning into system design and problem-solving could revolutionize how industries operate, making processes more efficient, cost-effective, and responsive to changing conditions or requirements. This advancement would be revolutionary, especially in sectors reliant on complex systems and operations, such as manufacturing, energy, aerospace, marine, medical etc.

As we stand at the intersection of intricate engineering designs and advanced machine learning capabilities, it's evident that we're on the brink of a transformative era. Fusing these two domains promises to reconfigure the landscape of system design, maintenance, and optimization. It's no longer necessary to go through the tedious and manual processes of deciphering schematics or cross-referencing datasheets. Instead, we're ushering in a future where machine learning not only understands these systems at a granular level but also proactively offers insights and solutions.

For me, this evolution signifies liberation from the constraints of traditional system analysis. With machine learning models bearing the brunt of system understanding, engineers and technicians can channel their expertise toward enhancing efficiency and fostering innovation.

Sure, there will be challenges along the way, but the potential rewards of this technological evolution outweigh the risks. We're on the cusp of a new era, and I'm excited for the future of machine learning and the applications of good we can design it to do.

Brad Cortlandt

Founder/CEO - fixaible

12 views0 comments


Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page