Open loop anti sway crane control

[Oumesh]

Good Day,

I have the task of rerofitting an AC inverter that controls the long travel on a crane at a shipping terminal which has built in software to prevent swaying of the load automatically. It is operator driven and this control is to compensate for operator short falls. The only feedback that I have from the system is the rope length in the form of a 4-20mA signal. The rest of the signals are operator driven like the speed reference from the joystick and the stop/run signals. What would be the best way to approach this type of control?

 

[anorlunda]

That's goind to be difficult. You can hare about the theory here

http://www.clescrane.com/soft/video.html

 

[skeptic2]

Can you specify the type of crane you are talking about. I've had experience with overhead cranes of the type found in the tops of factories. The way the crane operators move a load without it swaying is, after lifting the load, they throw the throttle full speed forward. Relative to the ground, the load remains nearly motionless but relative to the crane the load swings backwards. When the load, relative to the crane, reaches its maximum swing, the crane operator stops the crane by plugging. When the load swings forward and is directly under the crane the operator again throws the throttle full forward. Now the crane and load are moving forward at full throttle with very little swing. Experienced crane operators are very good at this and it will be difficult to design a system that will be better than the operators.

Perhaps from the rope length you can calculate the period of the pendulum created by the load. This will tell you when to stop the crane after starting and when to start again after the load has caught up with the crane.

 

[jrmichler]

Perhaps from the rope length you can calculate the period of the pendulum created by the load. This will tell you when to stop the crane after starting and when to start again after the load has caught up with the crane.

Fixed that for you.

There is another way. First calculate the period of the pendulum, then program the travel motor drive to ramp from zero to full speed in a time equal to the period. There will be a very small swing during the ramp, then it will travel with zero swing. This approach works a little better than the on-off-on method if your calculated swing period is slightly wrong. The slow start is another advantage because it is easier to get the load to an exact position.

It's true that experienced crane operators are very good at this, but I've seen an inexperienced operator get out of synch and amplify the swing. And the inexperienced operator does not want to just stop, wait for it to settle down, and start over.

 

[skeptic2]

I used the word “perhaps” because the rope length would give only a rough approximation of the pendulum length. Loads vary significantly in size and shape so the center of mass may be significantly below the hook. Some crane operators continue raising and lowering the hoist as they are moving towards their destination. The ramp method would make that more complicated. A better approach would be monitor the angle of the cable and stop the crane at the maximum angle and proceed when the load is directly below the crane, but that wasn’t an option.

I think we have to define how much of an angle is “a very small swing” and the maximum desired speed which will determine the ramp time. I disagree with your statement that it will travel with zero swing after the ramp. During the ramp the load will be behind the crane and when the crane starts traveling at constant speed the load will swing forward and continue swinging back and forth. Also note that the crane will have to ramp down its speed in anticipation of reaching its destination and deal with the remaining swing at the same time. The ramp method will result in slower operation.

I used to install radio controls for cranes and operated them as part of the installation. I mastered the technique without much practice and believe most anyone else could too.

 

[sophiecentaur]

I used to install radio controls for cranes and operated them as part of the installation. I mastered the technique without much practice and believe most anyone else could too.

It is quite unbelievable how inept some 'other people' can be. You can't rely on perfect levels of concentration, even with good operators, after a boring morning's work.

Actually, this seems to be a pretty scary - safety of life - sort of business. I should have thought that you would need some seriously well qualified (expensive) consultant to be giving you the sort of information that you are hoping to get from PF. We normally give huge disclaimers about this sort of advice.

It's great fun doing this sort of stuff on a theoretical basis but the prospect of several tonnes of load swinging about in real life (I have seen YouTube movies) should be very worrying for an inadequately qualified control system designer. (I am assuming that you are or you wouldn't be asking for PF help. )

 

[sophiecentaur]

the rope length would give only a rough approximation of the pendulum length.

It's basically the period that's needed and that can be found from the rate of change of angle with the motion of the carriage. An on-going solution to the differential equation would tell you all that's required. After all, that's what the brain of a manual operator is doing on an informal basis. The rope length is basically a red herring.

 

[anorlunda]

The OP hasn't been seen on PF since 2011. So what we have so far is more than adequate answer. Thanks all who contributed.

Thread closed.