Despite the applicability of fancy 3D modeling and simulation to mechanical engineering, working with documents and spreadsheets is often just as important. It’s researching the properties of new materials. It’s finding, reusing and merging existing standards into a specification. It’s managing hundreds of requirements in a new project.
That conjecture always rang true to me. Yet, it wasn’t until I conducted the Hardware Design Engineer study that I found some statistical proof that backed it up. Fifty-eight percent (58%) of design decision makers use CAD applications very frequently and very consistently. Only ten percent (10%) use simulation at the same frequency and consistency. Yet, fifty-three percent (53%) use documents and spreadsheets at the same rates. Knowledge, often captured as text and numbers, is just as important as the form and fit of designs.
Finding information in documents, spreadsheets, standards, patents, journals, articles, and enterprise systems like ERP and PLM can be tremendously painful and time-consuming. Today’s engineers are on extremely tight schedules, often making up time by staying late and working weekends. Improving their capacity to find, consume and reuse information promises great potential.
So it’s in that context that we’ll focus on a solution from IHS Markit in this post. Engineering Workbench is an unusual mix of search tool and knowledge management solution. First, we’ll dive into the capabilities it provides, Second, I’ll offer some commentary and analysis. Ready? Let’s dive in.
Capabilities of Engineering Workbench
Let me start off by saying that there is a lot of deeply technical capabilities involved with this solution. I’ll only go as deep as needed to get the concept across. If you want to dive deeper, there’s plenty of scientific details available on the IHS Markit site. So, with that said, let’s start at the beginning.
Connecting to Knowledge Sources
Engineers use many different sources of knowledge during design. There are standards published, journals, articles and other industry content by professional organizations. Government bodies provide regulatory guidance and standards. Patents and related are yet another source. In addition to those sources, there is a myriad of internal information that might need to be referenced as well. That includes the company’s standards, specifications, requirements and a lot more. On top of that, there is a sea of information in enterprise systems like drawings with part numbers, vaulted manufacturing instructions, and serial numbers associated with product configurations. All of these sources are relevant. All of them are important.
So the appropriate place to start is by connecting to these sources of data. Out of the box, Engineering Workbench has connected to IHS’ broad set of standards. It can be pointed to publicly available regulatory sites. It can hook up to internal sources of information like masses of documents on shared drives. It can connect to enterprise systems using 3rd party interfaces like SDKs, web services and more. The information could be unstructured, as in the case of many documents, or structured, as is the case in many enterprise systems. Furthermore, Engineering Workbench understands all of this information across six languages: English, German, French, Japanese, Chinese and Russian. Overall, it essentially connects to almost any source that is needed. It understands almost any kind of information.
Indexing Information Knowledge Sources
The next step here happens behind the scenes. This isn’t something any user has to do. Engineering Workbench does it on its own.
To serve up the right results to an engineer’s query, it needs to understand the information in those knowledge sources. To do that, it analyzes the source of information comprehensively. It uses advanced techniques like identification of document entities including titles and paragraphs, sentence and word segmentation, word part-of-speech tagging, the syntactic analysis of sentence structures, semantic indexing of sources, causal relations, and more.
All of these capabilities allow Engineering Workbench to understand the relative importance of every single sentence. In fact, this capability allows it to automatically generate a summary using the most important sentences in the document. Furthermore, in aggregate, that allows Engineering Workbench to understand how important that document is to a specific topic. That affects how relevant that document is to a specific search.
Engineering Search vs. Enterprise Search
Here’s where we need to touch on an important issue: engineering search is not enterprise search. In technical and engineering contexts, there are specific definitions and relationships between words. That is why Engineering Workbench looks to be unusual. All of the capabilities described have been honed and refined by working against IHS Markit’s technical standards. So the baseline is there for technical and engineering search.
With that done, you have a system ready to get you your information. So let’s get to that.
Searching for the Right Information
With all that behind-the-scenes work done, we can now focus on how engineers find the right information.
First, you are presented with a home page. It can be customized according to your company’s needs. As you start your search, you tick off which knowledge sources, both inside and outside your organization, that you want to use. From there, you enter what you are trying to find and are presented with results. It returns clusters of knowledge focused on specific topics relevant to your search. Of course, there are many different tools you can use to refine your search and interrogate the results, which is called a Research Assistant, including:
- Filters: You can refine the bounds within which the search can be conducted, all of which is based on your query.
- Automated Summary: Remember that summary of documents that was automatically generated? That is presented here. You can actually use a slider to include more or less in that summary. If increased, more sentences that were ranked as important to that document are shown. If less, then the least important sentences are removed.
- Alternative Searches: There are synonym and ontology suggestions to rephrase your question.
- On-the-Fly Localization: Just as Engineering Workbench can understand technical knowledge in six different languages, it can also deliver results in those same languages by translating them in real time.
- Annotating Results and Content: Right within the search, you can add a note to results or even within the content. This becomes a URL that can you revisit anytime.
In all, there are lots of tools that help you find the right information and mark it for later use.
Commentary and Analysis
Why is this a big deal?
Information for Problem Solvers
As mentioned before, engineers spend a significant amount of time trying to find the right information or the right person so they can make a well-informed design decision. This can range from ‘I know I saw that stuff somewhere’ to ‘I remember that we have an expert in the company, but I can’t remember their name.’ It also is as simple as ‘what can I use to fulfill this requirement’ or ‘are there any existing patents that address this need?’ Ultimately, engineers are problem solvers. To solve problems, they need the right information.
The Chaotic Nature of Acquisitions
While Engineering Workbench solves this kind of broader problem, it solves some pretty other specific ones as well. If you are a big manufacturer and have acquired a few other companies, then you are extremely likely to have many different PDM, PLM, ERP and other enterprise systems with information relevant to engineers. Consolidation is often an ugly picture, disrupting engineers that have to change and costing big IT monies for consultants. Instead of consolidating to fewer systems, Engineering Workbench can be the front end to finding the right stuff at the right time. But furthermore, it would let engineers in one division of a company find the hard-to-find subject matter expert in a different division of the company.
Enabling Long Term Archival
Another area of application for this kind of technology is long-term archived data. As hard as it is to get different groups within a company to use the same systems currently, it is practically impossible to get all their older released data converted. Frankly speaking, it just doesn’t happen. The best you can hope for is to bulk export it to a neutral format. In this scenario, Engineering Workbench can index that existing information in the format in which it already exists.
Overall, there is a lot of buzz about 3D modeling and simulation in mechanical design. However, finding the right information is an underserved need for engineers, right alongside the lack of a modern digital engineering notebook.
So that’s my take folks. I’m interested in getting your take. Sound off in the comments and I’m more than happy to discuss.
Take care. Talk soon.