Theory of Constraints: Structuring a Maintenance Turnaround
Most manufacturers think they're short on maintenance techs, parts budget, and machine availability.
But it's almost impossible to be short on all 3. In every situation only one of these is your actual constraint. You need to know which one it is and here's how you figure it out.
Maintenance techs:
You need 1 to 2 techs per shift who can respond immediately to whatever might break on the floor. These techs should only be assigned to work that can be quickly set down like building parts lists, running a daily check of critical systems, or rebuilding pumps on the bench. If a call comes in, the tech puts the work down and responds to the call.
These techs are still using parts but the rate of spend is generally pretty low.
Every other tech should be working on the most critical work that you can afford parts for.
That means doing repairs that return machines to standard. It doesn't mean wandering the floor, wiggling wires, or helping operators push buttons.
If you are already operating at a world class level, the repair spend isn't that much. But if you've got maintenance opportunities, your deferred maintenance backlog could have a price tag in the millions of dollars.
Real repair work consumes parts. That means your tech headcount and your parts budget are tied together, whether anyone planned it that way or not.
Scenario:
Let's assume you have 10 identical machines in your factory. After performing some machine audits, the average machine needs $500K in parts so you can return to standard. That implies a deferred maintenance backlog of $5 million.
Now assume the portion of the monthly budget that you can free-up for parts is ~$200K. One machine is down each weekday for maintenance and it takes 2 weeks to cycle through the machines.
In this hypothetical you have 3 shifts of 4 techs, 5 days a week. As discussed before, you keep 1 tech per shift on responder duty. That leaves 9 techs doing the real work at about 160 hours a month each for a total of 1,440 hours.
Some of that time needs to be dedicated to PM. PMs are mostly inspections, so they're cheap: maybe $30 an hour in filters, lube, and belts. For now we will assume that it takes up a quarter of your techs' time.
The rest needs to be repair work. Look at your last machine audit and evaluate. Suppose it shows $120K in parts to bring a line back to standard, and 200 tech-hours to install those parts. That's $600 in parts for every hour of real work (this number will be different for different industries).
Here's the tally:
360 PM hours × $30 = about $11K
1,080 repair hours × $600 in parts/hour = about $648K in parts .
That's roughly $660K a month in parts purchases to entertain your team for a month (assuming nothing needs to be set aside for emergencies).
Now look at your actual budget. If your maintenance budget has $200K discretionary after contracts and obligations, that'll cover PM supplies but it's only enough parts to consume about 30% of the labor you have.
So in this case, labor is definitely not your constraint. This means 2 things:
1) Don't buy any more labor, you already have unused labor so any techs you bring in externally should be just for expertise.
2) You still have to find what the constraint is and optimize to it.
Parts Budget vs Line Time:
At this point, techs aren't a limitation. Your parts budget is definitely less than your tech availability.
But parts aren't the only thing your techs need. They also need machine time and there are 2 ways to get it:
Take the machine down longer.
Have more people on the machine while it's down.
Repair work happens when machines are down and locked out. In our scenario we are using 1 shift for planned down time and it's day shift.
If we adjust the scenario and now it takes my factory 2 people to handle the floor calls, then I effectively have 2 techs to do my PMs and correctives, 8 hours per day.
That's 320 hours per month. Using the above scenario, if those techs are doing 80 hours of PM work ($30/hour) and 240 hours of corrective work ($600/hour) to return machines to standard, that'll consume $146,400 per month in parts.
That's less than my $200K monthly parts budget, so machine time is then my restriction even before production starts cancelling work.
Can I do something about that? Yes! Remember, I have 4 techs on each shift.
3 techs could make sense on back-shifts. Perhaps you need 2 techs to use a ladder and you want a 3rd so they can take a vacation... fine. Buy why is #4 there? And couldn't someone from another department hold a ladder?
If your night shifts aren't scheduling lines down, move techs to where the downtime is.
Suppose you move from 4 techs on each shift to 6 on day shift and 3 on each back shift. If you're only getting machine availability on day shift and 2 techs are being continuously entertained by emergencies then the "repair and PM crew" for your dayshift downtime goes from 2 to 4, a 100% improvement. Even better, you still have at least 3 techs on every shift to deal with major emergencies.
This new schedule increases the number of hours available for you to hang all those expensive parts by 100%. This means you now have enough machine time to easily hang all the parts you can afford.
The Last Constraint: Knowledge
This almost goes without saying, but if you get control of the budget and you start using machine audits to return equipment to standard, you're rapidly going to run up against a knowledge constraint.
In our example, I've pulled 2 more techs to day shift to increase my repair crew to be able to handle ~$292.8K, which outpaces my parts budget by 46%... but that only allows me to waste 46% of my potential repair time before I transfer my constraint back to machine time.
That means I would have to have accurate audits on 7 of the 10 machines so that I could be buying parts and applying them in order to prevent machine time from being my constraint (which would then put the parts I'm buying on the kit shelf for months).
This gives me 2 options:
Do all those audits, but they don't age well. You want to complete audits within 1 year. $3.5 million in audits against a $200K/month budget doesn't move very fast, but it's viable.
Move 2 more people to day shift. That reduces back crews to 2 and your day shift to 8. That increases your tech parts delivery capability to $439.2K... which means you only need to capture 45.5% of that time with repair/PM work. That means you need audits and parts purchases on 5 machines instead of 7 to keep the kit shelf clear. It also better aligns your $500K per machine (x5 machines) at $2.5 million in parts cost with a $200K/month budget (x12 = 2.4 million) for completion in 1 year.
The faster you return a machine to standard, the faster it stops acting up. The more worn components you leave on a machine, the faster the new parts you put on it will be destroyed.
I recommend getting individual machines to completion as fast as possible, but within the constraints described above.
Emergencies will, inevitably, go down as you repair machines.
This balance logic has to be done to ensure you can move cleanly through a plant turnaround. Ideally, you always want to move the constraint to your parts budget by manipulating your crewing and downtime availability.
Even if you have to sacrifice some emergency response capability, you need to be able to plan the consumption of the deferred maintenance backlog with your monthly parts budget.
Conclusion:
If your parts budget is your constraint, as it is in many shops, the kit shelf should STAY CLEAN!
Depending on the size of your shop, I recommend minimizing responders on shifts where you primarily do emergency response and put the rest where the machines can be worked on.
A tech who doesn't line up with emergency coverage or planned downtime is pure waste, no matter how good he is. Put your techs where the work is done.
If you're trying to rehab a plant that's got issues, you need to know what the limit is. Do you lack labor, parts budget, machine time, or knowledge?
And could you move that constraint to help your situation?
What does that number look like in your plant?
Comments