AI Generated FEA Results
Pressure Vessel, Heat Exchanger, Risk Management, FEA, AI randy bogard Pressure Vessel, Heat Exchanger, Risk Management, FEA, AI randy bogard

AI Generated FEA Results

The era of AI generated FEA results is here!

I like comparing FEA results to closed-form solutions because they help validate each other. 

I ๐ฅ๐จ๐ฏ๐ž comparing AI generated FEA results to closed-form solutions because they give me a good laugh. 

Thankfully the trend of getting AI generated FEA results through regulators appears to be over. In Canada Annex J of CSA B51 spells out the exact requirements regarding using FEA in equipment design. A couple of effective requirements stand out to me.

J.2) FEA may be used to support equipment design where the configuration is not covered by the available rules.

Many times AI generated FEA is submitted for simple designs already covered by ASME Division 1 and Division 2. This rule prevents FEA-only submissions for simple designs.

J.3) FEA report shall be certified by a Professional Engineer

The entire point of creating AI generated FEA results is to circumvent engineering costs. This rule prevents submissions from unqualified actors putting the public at undue risk. 

While your regulatory jurisdiction may not have any requirements regarding the use of FEA, I strongly support considering these two requirements for internal rules. Please be careful when selecting a fly-by-night unlicensed FEA firm with pricing that is too good to be true. 

Read More
Move Data from U-Forms into Spreadsheets for AI
Mechanical Engineering, Pressure Vessel, Heat Exchanger, Risk Management, COMPRESS, AI randy bogard Mechanical Engineering, Pressure Vessel, Heat Exchanger, Risk Management, COMPRESS, AI randy bogard

Move Data from U-Forms into Spreadsheets for AI

Are you ready for AI? Your data is not.

One of the big themes of Reuters Downstream 2025 was the use of AI in maintenance and inspection. I found the concept of using AI for predictive analytics in maintenance incredibly intriguing. However, letโ€™s not pretend that your important data is in a format that AI can actually use. You need to digitize your data.

Are your U-Forms sitting in a filing cabinet somewhere?

How about the historical measured thickness data from internal inspections? Are those values just floating around in PDFs?

What about the interesting information found during external inspections? Repairs? Rerates? Derates? FFS assessments? Wouldnโ€™t you want all of that data digitized in a way that makes it easy for AI to pull from?

We all know that AI is the future but AI is only as good as the data you are feeding it. Consider yourself behind if the items listed above are not in a usable, digital format.

Read More