Shutdown / Turnaround Planning

[Pulpeng]

We have a very condensed shutdown - something like 1600 work orders (with multiple tasks and resources) completed in 7 to 10 days. 24/7. Cost is about a million dollars a day to be down.

1st Question: Given all the time in the world to plan has anyone actually pulled off a nearly live actual/baseline plan so you can level as jobs come up or go off track. I think you'd have to be constantly updating it all day and night to have any value. If you have done it what type of resources for entering did you require?

2nd - Has anyone ever set-up unknown contractors as resources? I.e - Welder 1, Welder 2... Welder 50 to try and level or do you try and change it to Welder with 5000% units. What works better?

Thanks.

 

[IRstuff]

Given that a well known military adage is, "The first casuality is the battle plan," I think you can guess what my position would be. Known unknowns and unknown unknowns always happen, and anything that is that dependent on things happening precisely has risk attached to it.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[Yjoshi]

Dear Pullpen

1. I have done many projects with live/baseline and it usually works fine. My sugestion would be:
• As TTFN said, take your time planning, use only physical predecessor (don’t use resources availability as a predecessor)
• Use PERT and DELPHY to estimate times
• Because of the importance it may be wise to have some slack on your schedule
• Update every 12 hours seems reasonable (input work done, reschedule your undone work and level resources)

2. Use generic resources, (5000% or 50 (options/schedule/ decimal)). Otherwise you will not be able to level your resource which will be the key factor.

Regards

YJO

 

[ierland]

As a general rule: update about as frequently as your task base. Thus if you plan in weeks, updating about once a week might be fine. When planning in hours updating should be done in about the same period, perhaps once each half hour or maximum 2 hours. However, having a critical planning maybe even going below hours it is absolutely useless if you want to update once a day, or even a half day.
A modern approach may be to have also an add-in for risk planning, including probabilities for breaking down certain equipment. With the attached Monte Carlo simulation you probably get better and more realistic minimum and maximum times compared with the course result of the Pert Analysis as in MS Project 2003 and 2007. Although this may still be better than just relying on single values.
Another good practise may be to add Deadlines, rather than constraints such as Must Finish On or Should Not Finish Later Than. The Total Slack will take the Deadlines into account and is a very good management tool. Another advantage is that often MS Project will warn that there is somewhere a conflict. If you neglect the conflict you will never get a warning again, thus spoiling all your efforts.
All this is probably way too late for your problem but may be of any help for the future.

Why an easy solution if you can make it complicated?
Greetings from the Netherlands

 

[BernardErtl]

We plan using generic resources (BM = Boilermaker, WE = Welder) and use coding structures to further define contractors vs. in house plant personnel, etc. We schedule as logically as we can according to how we envision crews attacking the work. We don't level the schedule itself, but level manpower in analysis reporting instead. This helps maintain maximum flexibility in the schedule and buy in from the field supervision.

We update schedules just before the end of every shift in order to have updated shift schedules and progress reports ready at shift change to help facilitate work priorities and management decision making.

Bernard
InterPlan Systems
- eTaskMaker Planning Software
- ATC Professional Turnaround Management Software

 

[ShopPlanner]

Pulpeng, if downtime is so expensive, you may consider using best-of-breed scheduling tools which produce more realistic schedules with meaningful, automatic resource leveling. You may keep one or two powerful scheduling tools as alternatives to MS Project when resource-constrained scheduling is essential. One of those tools is Schedlyzer Lite (

http://www.optisol.biz/schedlite.htm

) which optimally handles multi-functional machines, multi-skilled workers, any weekly resource calendars and calendar exceptions even for large projects. The scheduler can specify, in the input data, resource requirements of tasks by functions and/or skills not necessarily by individual resource names. It facilitates fast and extensive what-if analysis of "properly resource-leveled" workload schedules and proactive capacity planning. When workers can be grouped (by skills) into non-overlapping groups, then the group names can be retained in the schedule output.

Even a resource-leveled schedule serves only as a workflow forecast because it does not handle natural variation in task durations and uncertain events in the system. This forecast is more reliable than a critical path schedule with a lot of resource overallocation instances. To deal with such variation and uncertainty, the schedule must be updated (with automatic resource leveling) as frequently as necessary. As time progresses, the baseline schedule will be increasingly invalid.