Deterministic method for cost estimation

[LT Judd]

TL;DR

I am reading through some company literature for cost estimation of large projects and they say for feasibility estimates use "deterministic method". What does that mean?

This is for my Engineering master project. I am doing a feasibility study on a couple of different designs of small gas fuelled power stations for remote areas. A big part of it is the CAPEX costs. Most companies hold that kind of info very closely. Does anyone have any insights on this kind of thing?

 

[ChemAir]

I have not done anything I'd define as a "deterministic estimate". That term doesn't seem to apply quantitative measures, i.e.--Estimate is accurate +15%/-5%. I normally quantify the accuracy of my estimates, and my experience is that the words "budget estimate" mean very different things at different companies. As an engineer and a construction contractor, I have done MANY estimates that were based on prior experience, historical costs, and used inflation escalation to adjust them.

Can you tell us what company literature you read this in, so we are sure of context, and can understand their definition?

Regarding cost estimation in this case, I'd look for a vendor that sells the major equipment. Tell them you are a student, and see if you can get a budget estimate for the major equipment. They may be able to give you one based on recent sales. Most additional work (piping, electrical, civil, can be factored, based on using different costing guides, along with regional corrections. If you don't know a vendor, let us know what you are looking for, and we may be able to guide you.

Realize I am looking at this from a process engineering standpoint, and normally, for expensive OEM equipment, the process drives the rest of the costs.

What kind of project is this?

 

[LT Judd]

Hi, I read it in some internal company documents. ( oil and gas company). It is from the "projects" department of the company. I am in maintenance and I am doing project for my masters of engineering.The project is for a hypothetcial Power generation barge. However when it comes to money and costs, the company is quite secretive . Like wise I have tried approaching a few equipment vendors but mostly got no replies.
I am thinking that the deterministic method may mean build up the cost by its component parts , ( e,g. so much pipe , so many pumps , so many metres of cable ),whereas other methods may be based on bulk factors , i.e power plant 10 MW = expected cost "X"$million
I know there are different levels of accuracy and they are called different things in different places.
I have obtained some company guidelines and they have the levels;
Class 0 "assess" +40/-25%
Class1 "study/ feasibility" +30/-20%
Class 2 "Pre feed", and so on to class 4 with narrowing bands of accuracy.
And as they move higher up they switch from "deterministic" to 'probalistic" methods, whatever that means .
But I can't find much more data than that. As I said they keep their data pretty tight- I guess its all IP and trade secrets etc.

I am looking for a "feasibility" level, and I am looking for lists of major equipment like , air receivers , compressors, pumps, ,heat exchangers , Plus what you might call complexity or quality factors. For example a simple sea water fresh water heat exchanger may cost "x" $/W but a cryogenic heat exchanger ( glycol/lng) may cost "3x" or "4X" $/W.

 

[anorlunda]

However when it comes to money and costs, the company is quite secretive .

But there should be no objection if you ask the right way.
"This company document says deterministic estimate. What does that mean?"
If you're allowed to read the document, you must be allowed to understand the words.

 

[pbuk]

Here is a deterministic estimate for a capital project: "X will cost $200 million". Here is a probabilistic estimate: "There is a 90% chance that a budget of $200 million will be sufficient for X".

Note that "X will cost $190 million ± $10 million", "X will cost $200 million allowing for a contingency of $20 million" and "X will cost $180 million plus a contingency of $20 million" are all also deterministic estimates - in fact they and the first estimate above are all different ways of expressing the same deterministic estimate.

Common probability levels for probabilistic estimates are 50% (often referred to as the "most likely outcome" - note this is NOT the expected value of the outcome which is a deterministic estimate) and 90% (the worst likely case).

 

[pbuk]

Regarding cost estimation in this case, I'd look for a vendor that sells the major equipment. Tell them you are a student, and see if you can get a budget estimate for the major equipment. They may be able to give you one based on recent sales. Most additional work (piping, electrical, civil, can be factored, based on using different costing guides, along with regional corrections. If you don't know a vendor, let us know what you are looking for, and we may be able to guide you.

Realize I am looking at this from a process engineering standpoint, and normally, for expensive OEM equipment, the process drives the rest of the costs.

You should probably discuss scope with your supervisor, but I wouldn't have thought that building a bottom-up estimate is appropriate. There is plenty of literature giving parametric estimates - search for "power station capital costs".

 

[ChemAir]

You should probably discuss scope with your supervisor, but I wouldn't have thought that building a bottom-up estimate is appropriate. There is plenty of literature giving parametric estimates - search for "power station capital costs".

I'm not the OP. I was just wanting to read the context of where he read the "deterministic estimate", because without that, my comment was likely unhelpful.

Corporate jargon can differ.

 

[LT Judd]

Hi, The context came from internal company guidelines, I can't copy the internal document because if i even send an email with attachment from my work computer to my home computer , it triggers warnings and alarms.
But I can retype it in
re reading it now I noticed that the probabilistic / deterministic thing comes from the section estimating "contingency". There is a table which has;
cat IV small scale <50M USD - screening, study feasibility- standard deterministic
- BOD/pre feed, Budget,Control - probabilistic CSRA ( plus scenario analysis)

Looking through this a bit deeper I think you are right- its corporate jargon and may mean different things in different places.
I think the previous poster was right , For my purposes i am getting into too deep, better off doing a parametric estimate.
I got a good manual from the library - "Compass International - front end conceptual cost estimation/ year book" . It as a few case studies and examples. Unfortunately its 2016 version but I will just have to adjust for inflation (of which there hasn't been much)