Tommy Cookers wrote:... let's categorise the 'leveller' approach as semi-active suspension, ie series-combined active strut and passive (spring etc)
(for intelligent road or race suspension do we need to go beyond the abilities of the semi-active ?)
the (active) strut of the semi-active system can be electromechanical
intelligently damping (recovery) at the important lower frequencies/longer travel and simplifying the mechanical damping need...
To be frank, Tommy, I don't have an issue with the above. Just two thought's: F1 designers go to greats lengths to minimize weight (even choosing to use non-adjustable dampers to save the weight of the adjuster mechanisms - possibly saving 5 grams per). Secondly, the approach does not make the suspension "completely accessible through software algorithms", one of Scarbs suggestions.
Tommy Cookers wrote:btw ..... a real-world example of in-service problems with complex systems
LSR vehicles have solid (metal) 'tyres'
the LSR holder importantly had a software-controlled suspension functionality for 'real-time' pitch attitude management transonically
the processing hardware regularly failed due to heat etc, eventually the suspension had to be locked (with a rigid strut)
this conservative pitch attitude gave downforce throughout, but this was generally excessive, costing 30 mph ?
Curiously, I know something about your real world example (LSR = Land Speed Record, I think).
In it's design phase, I was approached by Richard Noble to design an active suspension system for ThrustSSC. I was intrigued and spent many hours of my own time exploring options.
I was presented with schemes that suggested that they did not actually want an active system, the two front wheels were passively suspended using (I recall) bump rubbers with a total of 16 mm of travel, the single wheel rear bogie was required to be suspended actively to control overall downforce. The rear wheels were steered (my aircraft background suggested that a "tail-dragger" was not the most stable layout), and the bottom of the vehicle was flat.
None of this seemed to make much sense. I tried hard to make them re-think it, but was unsuccessful. I modelled the thing, using guesstimated aero characteristics (the vehicle was to weigh 8 tonnes, I recall, and was thought to generate a downforce of up to 8 tonnes/degree). In my view, the main problem was to catch a rearward shift in centre of pressure before the front wheels left the ground, after which the active system had only inertia to work with. I evolved a solution using actively controlled canard surfaces that, at least, might have prevented the vehicle flipping over. My solution was turned down, so I chose not to become involved.
Much later, I discovered that the project used 2 of "my" active suspension controllers, controllers that had previously survived an F1 practice & racing season. Why the reliability problems? I've no idea, but we didn't see them again after the F1 season had finished. I can only assume that it was a case of (inappropriate) horses for courses. I was told that the "active suspension" idea was dropped fairly early in the program in favour of a ride height scheduling strategy driven by pitot pressure (I assume).
Ultimately, the project was successful. I am sure that was mainly due to the test piloting methods and skills of Andy Green, the driver/pilot.
I can't recall ever worrying about making the active system work with rigid wheels (see my previous post), but I did think about the consequences of an exploding wheel...
