Many problems with baby example trying to get workers to work.
Hi all forum users,
There are several issues here, and related questions. As they all relate to a single model (I made a baby form of this model to attach), I'll start by posting them all here - maybe later split off sub threads if the discussion gets detailed.
Issues (in order of importance)
1. Workers do not work, processes that need workers are blocked
2.1 Assembly station cannot be configured to take 2 parts of main MU and 3 of sub MUs.
2.2 Can workers be despatched into sub-frames from the frame where they are defined?
(Edited in : As above, if I move the assy station into a frame, to me it seems logical that I should be able to route a footpath to the sub-frame and pass the workers into a frame too, however I doubt if this is provisioned in PlantSim). So how do common worker resources get shared with all the sub-frames of a model?
3. Tecnomatix runs incredibly slowly. Two instances in memory, main model's process is suspended - no CPU load there.
Future related stuff
4. When attaching a model SPP file, is all my info secure (stuff like licence server IP address, personal information etc.)?
5. I'd prefer my modified copies of Worker1 and Worker2 to live in ClassLibrary.Models.<MyModel>, not in the ClassLibrary.Resources. Is this possible>
6. I've minimized the footpath widths, even so they are very intrusive on a busy model. Can the paths be hidden and unhidden? (FYI With too low a width the path is visible but cannot be click-selected, need to select on the footpath name).
7. At some stage all the things I do manually - creating workplaces and footpaths and joining these to workstations and to workerpool - I will want to do with SimTalk. Is there a general way to find the SimTalk equivalents to manual editing procedures without always having to come to the forum to ask for help?
So the above is the brief version. Don't want to, or have the time to, explain this in huge long question - those that know, will know from the model and the screen shots.
Maybe just a few brief pointers -
Models.Sim_worker_cases has the original intended model
frame= _2_workers (assembly station using 2x MU1 and 3x MU2 and using workers)
frame= NoWorkers (as above but no workers) - fails as assy station produces no output
and the disabled versions of the assembly stations
frame= _2_workers_11 (assembly station using 1x MU1 and 1x MU2 and using workers)
frame= NoWorkers_11 (as above but no workers) - this is the only model producing output.
Regarding the very sluggish behaviour, laptop has recently been rebooted after a windows update and CPU indicates low loading.
Have two instances of Tecnomatix in memory - original instance with main model is suspended and CPU load thereof is 0%.
Yet doing only 1/2 dozen object moves and renames in the current model, using 1 sec of User Time required 9 sec of Kernel Time - there is something very wrong somewhere.
Thanks for all ideas and feedback
Image : No assembly table, shows worker not used at all
Version with assembly table = 2x MU1 and 3x MU2 --> Always reverts to 1x Main Mu in spite of switching to MU Type option for Assembly Table
Screenshots of process CPU utilisation, very minimal model editing changes only
PlantSim Profess Ver 220.127.116.114 & Windows 7 Ent. SP1
~~~~ 1st post in thread ~~~~
Solved! Go to Solution.
freelance simulation specialist
Resolved points 1, 2.1 and 2.2 as indicated in screen shot below.
Multiple workplaces point into the processes within the frame. Path lengths of the FP's may be altered with F8 properties or terminal points spaced out along path between WS_3 and drain (reflecting frame content taking space on main model floorplan)
Further points' solutions forthcoming
Final solution attached (resolving issues 1, 2.1 and 2.2) for those who might benefit from it.
I'm not going to resolve any other issues here (in particular the SimTalk solution deferred to real plant model).
I am still concerned that with all the worker overheads --
(2 workers, each busy for at least 82% of the total workstation activity time for each of the 28 MUs produced, a bit more in fact since the BOM checking in WS_3 has zero activity time but still worker travel time)
-- the total number of MUs produced in 5 hours in worker and workerless plants is identical. That sounds a bit fishy to me, my expectation is that there should be worker bottlenecks in this baby model at 82%+ worker occupancy.
I did experiment with dropping average worker efficiency to 61%, then 5 hr production dropped from 28 to 11 MUs (would have expected more like 17 MUs, so apparently here worker bottlenecks do have an effect).
For now I am leaving the worker bottleneck issues, a pity because it could cause me pain in the future with a more realistic model.
Summarising screen shot and model attached