Are you considering an energy system optimisation tool? These questions help you evaluate your options. Optimisers all promise the best design. They differ in whether they find it, how fast, for which problems, and who helps when a result needs explaining. The questions work for any option: a commercial platform, an open-source framework, or a model you build yourself with an AI assistant. They are ordered by importance, with price last. After each one, we also say how Sympheny answers it, so you can compare.
An optimiser earns its keep when there are too many combinations to compare by hand: several energy carriers, several buildings, cost against CO₂, a large investment. If you are sizing a single technology, the design is already fixed, or you are modelling national grids and markets, a simpler or more specialised tool will serve you better.
Many planning tools simulate scenarios you define and report which of them is cheapest. An optimiser searches the whole design space for you. Even then, a solver stops when it is close enough to the proven optimum, and a model can be internally wrong while still producing a clean result.
"Optimal" with no method named. No stated tolerance. Different results from the same inputs with no explanation.
A solver can find the exact optimum of a model that is too coarse. If the model averages away the hours that matter, ignores temperature levels or treats every machine as running at full load, the optimum describes a system that doesn't exist.
Annual or monthly energy balances only. Storage sized on daily averages. No clear answer on what is simplified.
A good study compares many scenarios: prices, demand, technology options, CO₂ targets. If one run takes a night, you will compare three variants instead of thirty, and you are back to guessing.
Timings shown only on the demo case. Solver licence costs that appear after you sign.
Tools are built around particular projects. One designed for a single country's district heating workflow may be excellent there and awkward everywhere else. The right tool fits the projects you will actually do over the next few years.
A fixed workflow that only fits some of your projects, with no way to extend it.
Results are only as good as the costs, efficiencies, weather, demand profiles, tariffs and emission factors behind them. Collecting and updating these is often more work than building the model.
Default values with no source or date. Data you can't override.
An energy concept gets challenged by clients, boards, funders and sometimes auditors. You need to explain why the optimiser chose this design, and show what each tonne of CO₂ saved costs.
A black box. Results you can view but not export. No way to show the alternatives that were considered.
Sooner or later a result will look odd the day before a client meeting. What matters then is whether someone who understands energy optimisation answers, and how fast.
Support only through a forum. Response times nobody will put in writing.
A powerful tool that only one person can drive is a risk. So is a tool used twice a year, if everyone has to relearn it every time.
Only one person in the company can run it. Every project starts with relearning the tool.
Every hour spent retyping data between GIS, spreadsheets and the model is an hour not spent on the engineering, and a chance to introduce errors.
Manual re-entry at every step. No programmatic access at all.
Energy concepts are revisited for years, and your models and know-how build up in the tool. If it stops being maintained, so does your ability to update your work.
No export path for your models. A security claim that turns out to mean neither certified nor documented.
Price comes last because it is the smallest number in the decision. A licence costs a fraction of the engineering time it saves, and both are small next to the investment the study shapes: on a CHF 20 million energy system, a design that is just 1% better is worth CHF 200,000. The expensive mistake is not paying more for the right tool. It is paying less for one that gets the design wrong, then living with that system for 20 to 30 years. Decide what you need, make sure the tool can do it, and only then compare prices between the tools that can.
A cheap licence that needs expensive staff time or consultants to produce a usable result.
A simulation tool calculates how a system you have designed will perform. An optimiser chooses the design: which technologies, what sizes and how to run them, so that cost, CO₂ or both are as low as possible within your constraints. Many projects use both: an optimiser to find the concept, then detailed simulation to engineer it.
It is possible, and the same 11 questions apply. The ones a self-built model has to answer on its own are correctness, data, support and longevity: who checks the results, who keeps the data current, and who maintains the code when its author moves on.
It can be. Many are rigorous and transparent. Look hardest at questions 7, 8 and 10: who supports you, who on your team can use it, and who maintains it. Include staff time in the total cost.
Build the same reference case in each, ideally a past project whose outcome you know. Time it, run a few consistency checks on the results, and see how long it took your own team to get there.
Start a free trial, or book a demo and we will build a case that mirrors yours.