472272_en-US

liliconcom

technoeconomic analysis

Hi everyone-

Has anyone ever addressed questions in using JMP to help with technoeconomic evaluations? I have a customer wanting to know how JMP can help.

In the past, I have built Excel models for multi-step chemical syntheses where the input costs, yields, purities, cycle times, etc. were tied together and then would run an in silico DOE by varying the assumptions on those factors. I used this to determine where to spend R&D budget to optimize the process and make sure the step we focused on would have the most impact on overall cost.

I think this can be expanded to a broader technoeconomic analysis by building a model (Excel, Aspen, or similar) and modeling IRR, break-even point, or whatever metric they find most useful. This would lead to a profiler that I think would be much more informative than a sensitivity analysis using a traditional tornado chart. Maybe even then carrying out Monte Carlo simulations on key factors to get a probability for achieving an acceptable range or results.

Anyone else have experience with this type of analysis or can recommend another approach?

Thanks,

-Scott

 

Re: technoeconomic analysis

Hi Scott,

I did a bit of this when I worked in the chemical industry. I built some cost models in Excel with various factors including production unit occupancy, cycle times etc. And then profiled this in JMP with the Excel addin. I added Monte Carlo simulation to give an estimate of the uncertainty in the final cost estimates, given uncertainty in the inputs.

I thought this was all very clever and that it would be incredibly useful.

My audience didn't much seem to care. I don't think they were ready for a distribution of cost estimates. They didn't much care which factors were most influential. They really just wanted best case / worst case.

I have shown similar examples to customers. Again, they don't much seem to care. I am not aware of anyone ever buying JMP to enable them to do this.

Hopefully your audience is more receptive. It seems like a great application for JMP.  But I would caution spending too much time on this. 

I hope that helps.

Phil

Re: technoeconomic analysis

Hi Phil-

Thanks for your thoughts on this. I did a quick literature search and found a few references that are doing the same thing as you suggest: create models in external software and then simulate and model outcome probabilities within JMP. This seems to be the right level of response to the customer for now.

-Scott

Re: technoeconomic analysis

Hi Scott, I didn't mean to "accept as solution" but I can't seem to undo this at the moment so I'll leave it be. 

I presume the customer has already looked at the simple example case for this in our help documentation? 

https://www.jmp.com/support/help/en/16.2/#page/jmp/example-of-simulating-general-formulas.shtml#ww511345

I personally find the search term "stochastic optimization" useful for finding this and the other monte-carlo simulation content in our Online help docs. Julian did a nice demo of this  a few years back:  https://youtu.be/rfR61lJ3mhk?t=2266

-Patrick