Do you use mixture DoE (Design of Experiments*) in formulating? (I'm a formulator of solder products, and this is a particular professional interest- although recently most of my time has been taken up dealing with routine lab stuff and manufacturing problems)
*Such a clunky abbreviation.
Sometimes, yes. Perhaps I should elaborate a bit on the development process.
There are four big players in the engine oil additive world: Infineum, Lubrizol, Afton and Oronite. Between then they provide most of the additives globally (and all of them have much wider business footprints too). For many engine oils, each additive supplier will offer a package of additives - a pre-mixed concentrate of all the main additives - to which you add the appropriate base oils to get the finished oil. In truth it is often much more complex than that, but it's a simplified view.
This means that many oil suppliers, particularly smaller ones with small resource pools, can take an additive package, blend it to the right viscosity using the right base oils and then have whatever approvals or performance that package offers - no need for expensive and lengthy in-house product development teams and programs. Larger oilcos will have a more complex portfolio of products, some of which may follow the same approach (maybe the lower-tier stuff), but some will be bespoke. Either developed in collaboration with the add-co (and therefore exclusive) or using an add-co package and then modifying it with additional ingredients to enhance or differentiate it. Or even developed from first principles using whatever prime additives serve the purpose.
When developing a formulation there are a number of factors to consider. Performance is important, as well as meeting whatever OEM and industry approvals are required to underpin the product being developed. There's also cost. And supply chain - location of additive and base oil suppliers vs your own blending facilities for example, and also location of your blending facilities compared to your sales markets. Not to mention the various health and safety requirements, REACH and other registrations and the HSSE consequences of using certain chemicals.
In terms of performance, there are some constraints in terms of what you can use. Lots of oil specifications include limits on various physical and chemical parameters - for example phosphorus (which poisons catalytic converters) or metallic elements (Zn, Ca, Mg etc) which form DPF/GPF-blocking oxides. There are also some additives that don't interact well with each other and can cause instability or degradation of performance. A lot of this is learned through experience and so when designing an experiment there are places you simply don't go because you know it's off-limits. But when trying to decide if one detergent type and one anti-wear type is better than another you might form an experimental matrix to screen for the best combinations. The further you go in the development process the more expensive and time-consuming it becomes though, so you have to use quick screening tests early on to allow you to spend the time and money on the most likely candidate(s). Given what I said up-thread about the number of options for additives, a full design matrix of all base oil and additive types and dosages across the whole formulation would be impossibly large. There are some modelling options that can help predict outcomes.
