In the not too distant future, machines – not humans – will be the primary consumers of engineering information.
This is the fourth article in a series about the principles of the Transform Manifesto. If you want to start at the beginning, go here.
Today, a growing number of industrial machines and software applications rely on data from specifications and standards to run their processes. Humans are the bridge – and the bottleneck – between the data and the machines. Most of the data is stored in PDF which necessitates hard coding data into machines, a time-consuming and error-prone task that is not scalable. In the not too distant future, machines – not humans – will be the primary consumers of engineering information. To that end, engineering documents and data should be structured for a machine to read, interpret, and execute.
Many years ago, a major aerospace company built an app wizard which enabled a mechanic to input information about a bolt and receive the correct torque setting for the installation of the bolt. The torque specs came from a National Aerospace Standard (NAS) and was a complex set of data, organized by the type of bolt (material, thread type, thread class, etc), the substrate, and other criteria. Because all the source data was stored in a PDF file, building this app was a tedious job that required hardcoding data, and any updates also had to be updated manually. Despite the tedious manual labor required, this was a monumental achievement at the time, but it was only the seed of what was to come.
In order for manufacturers to leverage the full benefits of the digital thread and autonomous manufacturing, we need to rethink how machines interact with engineering data.
The information in engineering documents must be readable, interpretable, and executable by machines without human intervention. Documents, from their general structure down to the most granular data elements (equations, tables, etc), must be quantified in standardized formats that machines can understand and act upon. Furthermore, the content must be written and tagged in such a way that machines can analyze the full scope and context of information, combine it with data from other sources, and automatically make decisions and take actions based on the machine’s interpretation.
The SWISS platform has already developed machine-readable and machine-interpretable data. We are not far from a time when these tools are leveraged by a majority of engineering-intensive industries and machines read, interpret, and execute autonomously. We must prepare today for that inevitability.
What do you think? Will machines do more engineering in the future? If you want to automate more of your processes with machine-readable data, let’s talk.
Ready for the next in the series? Coming soon!
If you want to read all seven principles of the Transform Manifesto, start here.