How expensive are software and training in reality?

If you haven’t experienced the benefits of using the correct tool for the job or the tremendous impact targeted training can have to rapidly increase your knowledge and exploitation of the software, a phrase that is often heard is that “software and training are expensive”.

In my view, that depends in great part from the following factors:

1. Have you purchased the most suitable software for the job?
2. How efficiently do the operators/users work and how enthusiastic are they about their subject area (can make a huge difference to the output). Are they adequately supported?
3. The overall impact of training

1. “Have you purchased the most suitable software for the job?”

If you have, then great. With the right training, lasting not more than a couple of days to a week for example, you’ll be able to really make a reduction to product development times. You should therefore be able to estimate how much quicker you can get to a design solution that is worth manufacturing and implementing. Calculate the equivalent cost in terms of salaries and reduction of physical testing costs to see how quickly you can recover the software costs but also what the longer term savings will be.

For example, if you have just purchased software that cost for argument’s sake £30.000 and in the first 6 months of use you want to replace physical testing of vehicle HVAC systems in climatic wind tunnels, you will have to consider the additional costs of 1 person to set up the model on top of standard vehicle preparation and instrumentation costs to get to a robust baseline case that you can calibrate the models against. Of course if the test data already exists, you can use this to calibrate the baseline model.

The baseline model should be calibrated with minimal divergence from values of 1 for correction factors and based on models defined by equations rather than tables where possible so that we are not having to characterise models from test data and also are able to avoid extrapolation of tables when we need to go beyond the bounds of those tables.

Evaluations

If you are still in the evaluation stage of different software packages, make sure that the person doing the evaluation:

• is suited and has the right level of experience
• is supported by their manager
• does not base their conclusions on “assumptions”. Assumptions are often dangerous, particularly when making important decisions. For example, do not assume that one software’s limitations are the same as others. There are many different tools out there with differing capabilities
• has set themselves reasonable targets to achieve by the end of the evaluation phase
• makes full use of the expert advice the software supplier should have offered

If you don’t make the considerations above, there is a high likelyhood that the conclusions that are drawn from the evaluation are flawed and that you will end up making the wrong software choice. .

Example case study

Say you would like to perform comparisons of how different refrigerants and glazing affect the loads on the energy source used to drive such HVAC systems. This can now be done in a matter of hours rather than months trying to source many different types of glazing and setting up different HVAC systems and finally instrumenting the vehicle, transporting it to the test facility and performing the tests (with all the debugging involved in such physical tests).

For example, if you could replace additional “what if” physical tests which entail the following…

• 1 fitter for at least two weeks to fit part variants (e.g. A/C compressor, glazing, etc) and instrument a car ~£1800
• 1 person in charge of getting the vehicle ready for testing and sourcing the parts ~£3000-£10000 depending on prototype part costs.
• 1 person to transport the vehicle to the tests ~ £1000

Then for each test shift:

• 1-2 engineers that will set up the vehicle in the cell ~£500-£900
• 1-2 engineers that will sit in vehicle during each test ~£500-£800
• Hire of the test cell at £4500 upwards for each test shift including facility staff.
• Potentially require 2 additional re-tests to establish variability: 2x £4500
• Add extra time for sensor and logging failures and potentially void tests which in the worst case means that you’ve just wasted the cost of an entire test shift.

…your £30,000 for software and potentially ~£1,500 for training are starting to seem a very reasonable number to have spent for the benefits you would draw from that investment and being able to repeat the above testing in a virtual environment for all sorts of control strategies, boundary conditions (including atmospheric conditions and soak temperatures) and hardware characteristics variations at the push of a few buttons using just one main person to set up and request the PC to carry out those virtual tests.

2. How efficiently do the operators/users work and how enthusiastic are they about their subject area (can make a huge difference to the output). Are they adequately supported?

You might have all the right software tools available but we still need to make sure that the engineers understand the physical principles involved in the physical system so that they can put together a model architecture that captures the characteristics of the real system and therefore generates a representative model.

Non representative models and poor data will lead to unreliable and non scalable models. If CAE (Computer Aided Engineering) has bad name in the company, ask yourself why and find the reports that document what may have gone wrong in the past so that current and new staff can be made aware and not repeat the mistakes. This way you will improve the efficiency of how you put model architectures together, parameterise the components and evaluate the results using lessons learned from the past.

3. The overall impact of training

Training is often overlooked and deemed not necessary. It is thought that a professional and high end software tool should be straight forward to pick up without the help of experts to show you the way.

A high end software will inevitably give the user a wide range of functionalities and tools to make use of. For a quick and efficient uptake of the software package, it is extremely beneficial to be trained to get up to speed in the area of interest as quickly as possible and be exploiting the full potential of the software package; i.e. accelerating the return on investment of the software purchase.

Spending even 2 days training for beginners and existing users will save you weeks of self-learning and you might not be aware of best practices, tips and tricks right from the start which will save you lots of time in the long run not only for model development but also for simulation run times which may not be optimised.

Conclusion

When evaluating the “cost” of a software package, training and system development methodology, you need to consider not only how the software purchase price compares to other tools but also what the short and long term benefits are of proper use of that software package through training. Training an engineer to use a new tool may seem more time consuming than continuing to use existing tools that are less efficient. However, you need to consider the longer term benefits of using the right tool for the job even if the tool is a new one.

Only with this mindset you’ll be able to make a meaningful evaluation of what the “cost” of software and training really is.

Written by: Alessandro Picarelli – Engineering Director

Please get in touch if you have any questions or have got a topic in mind that you would like us to write about. You can submit your questions / topics via: Tech Blog Questions / Topic Suggestion.