Continuously Cycling Double Cylinder Control

There are such relays you could use (time delay and basic latch) but I think you may want to use an intelligent controller like an Arduino or PLC when one considers the variety of things that can go wrong along with added useful data and control. For instance; did the door open and close fully each time, cycle count, STOP at selected cycle count, time to open, time to close, safety stops, delay time to motion, failure to complete motion time limit.

That was my 1st thought. I've had to make in the past a mechanism to operate a hinged joint 10,000 times for BS approval. I used a crank & wheel and just let it run for the appropriate time.

https://www.directindustry.com/prod/airwork-pneumatic-equipment/product-69633-2377939.html

https://www.measuremonitorcontrol.com/valves/solenoid/five-way

https://tameson.com/52-way-and-42-way-pneumatic-valve.html

https://www.grainger.com/category/pneumatics/pneumatic-valves/solenoid-air-control-valves?attrs=Item%7CSolenoid+Air+Control+Valve~~Valve+Ways+and+Positions%7C5-Way%2F2-Position&filters=attrs

All you need is an on/off switch and a timer with a "symmetrical recycling" function. Google Broyce Controls™ type LMMT to see what you need but most industrial timer manufacturers will make a similar product. Typical price is £25/€30/$35. The Broyce has a universal voltage input 12-230volt AC/DC but if you can't get one like that go for a timer with the same voltage as you solenoid valve coils. The top dial switch on this timer selects the function which you will set to RF. The middle dial switch sets the timing range. The lower dial is a pot resistor that allows you to fine tune your timer setting. Wiring: Connect your power supply live to one side of the on/off switch. The other side of the on/off switch to timer terminals A1 and 15. Timer terminal 16 to the live of one of the 5/2 valve solenoid coils. Time terminal 18 to the live of the other 5/2 solenoid coil. Timer terminal B1 and the neutrals of the two solenoid coils to your power supply neutral. I would also suggest a mechanical counter that increments each time the door opens and hits the counter button/lever.

Set up the air cylinder and mount two air change over switches one in open position and one in closed position. Connect the air lines so when the door is open the air relay supplies air to the close side of the cylinder and the opposite for the close cylinder.

A speed regulating valve will be needed on the open and close connections to the cylinder. Festo have the hardware to do the task as I set up a similar splitter to take a sample from a material stream. If you go to the Festo site or any other pneumatic control supplier you will find all the equipment you need.

Check this out https://www.festo.com/net/supportportal/files/10102/sa231742.

https://www.festo.com/us/en/c/products/industrial-automation/pneumatic-valves-and-valve-manifolds-id_pim23/

As you are testing a steel door there is the possibility that if it hits someone while in motion it could cause serious injury. The best option is to conduct your tests inside a cage with interlocked access so that the door is stopped when anyone opens the cage access and a key operated reset with the key held by a responsible person.

A 5/3 valve with a "closed center" position that halts motion immediately when an Emergency Stop is activated would be my preferred option. If you use a 5/3 open center valve the door can move when the valve is in the center off position and it's inertia could still cause injury. A 5/2 will push the door to one end position when power is removed which could cause someone to be trapped between the door and jamb. Forcing the door to the open position on loss of power is better but still not completely safe.

In an earlier post I suggested using a mark space timer to control the cylinder. You could use a relay set on at one end by a N/O contact on one limit switch and latched on with a N/O relay contact then tripped off at the other end by a N/C contact on a second limit switch. This has the advantage of not having to match the mark space ratio with the cylinder speed setting but limits the open and closed positions to where the limits are mounted. It also introduces two extra potential failure points which could be significant in such a long test. Pneumatic valves are prone to sticking and a driving cylinder coupled with door inertia into a fixed limit switch would be enough to smash it off it's mountings.

Yes safety last...haha