complex time study question

[pavingrates]

Hello,a little background about myself will help explain the complexity of my question.I am a senior engineer with well over two hundred thousand time,motion and production studies under my belt using all methods of obtaining the data to arrive at a relative conclusion.I have studied all sorts of procedures from the most simple to very difficult with emphasis in establishing standard rates of production for everything from widgets to jet plane engines.
PROBLEMO:setting a production quota for asphalt paving for residential and small commercial driveways and parking lots where the size of the jobs vary from 500 square feet to as much as 20,000 square feet with established methods and machine manning requirements in place [all written and established by yours truly].I have found the varaibles to be overwhelming and almost impossible to eliminate or incorperate to a point within reason.The fatique factor alone took months of study to calculate due to finding the break even point of optimum asphalt temperature versus the optimum ambient temperature for comfort for the employees.
I have finally arrived at 345f asphalt @ 83f ambient for a rate of 5% ascending an additional 5% for each additinal 4 degrees in the ambient temperature.IE83 degrees F rate of 5%,@87 degrees F rate of 10%,@91 degrees F rate of 15%.The Fatique rate is not affected by a rise or fall of the asphalt temperature because the quality of the asphalt becomes unacceptiable long before resulting in a drop of efficiency,however the temperature of the asphalt does have an affect on the production in that the asphalt becomes hard to "work" when below 325 degrees.AH,but yet another varaible to deal with.
I think by now you can see that the task of establishing a production rate for asphalt paving can be somewhat of a harrowing experiance.
OTHER VARIBLES 1. quality of the preceeding preperation grading
2. speed of the paving machine
3. incline of the surface to be paved
4. number of other laborers assisting
5. length of each strip of asphalt layed
6. width of each strip
7 number of "set downs to pull a strip"
Thank you for any and all contributing information.I know all of the arguments against attempting to establish such a production rate such as, labor is a small piece of the total cost that it might be unjustified to set rates,or the old argument that the longer the cycle time the less accurate the rate and hence not a viable option to set a standard.etc. etc.

 

[dan21703]

Did you factor in the range of variability for the asphalt material? I have been under the impression refiners make it a little heavier in the summer. Especially if the customer is not a governmental agency.

 

[pavingrates]

Dan21703,you are exactly correct and yes the different wieghts and grades of asphalt have been factored into the equasions.I have concluded that for the topping or residential types of asphalt the standard rate of application would most likely fall in the range of .080 to 1.50 second per square foot,while the binder or first heavier asphalt would be in the range of .058 to 1.08 second per square foot.The difference being that the heavier binder doesnt require the handwork to finish as does the topping coat.
My problem has more to do with factoring in or eliminating the variables mentioned in my original post.The calculations are so obscure that I have hit a brick wall while all the time knowing that the numbers must be included to make the standard quotas correct enough to withstand critique.Thank you for your help and if you have more to offer please do so.

 

[Joewski]

Pavingrates Hi,

You have had lots of experience setting Stds, I have done some complex stuff and my method is to model the problem on a spreedsheet by breaking the problem up into smaller simpler problems. You have probably done the same for this problem, but the answer your seeking may acutually be hampering you finding a satisfactory result.

Generally, My clients is usually trying to cost something so so that they can accurately forcast labour, materials, resources requirements etc. So my clients real aims are usually never to underquote their jobs ie. loss money on a job. (so I'm going to take the liberty to reformulate the problem hopefully to illustrate how you can get to the same results but in a more meaningful way.)

From reading your approach, it seems you want to capture all the variables, plug em into a formula and get a time standard.

The bit I don't understand is how are your standards actually going to be used? ie At quoting or On the morning of the of laying the Ahphalt or as a post mortum, or in PDMTS to compare two methods.

So not understanding how exactly the TS are to be used is a problem.

May I humbly suggest, that it may be easier to first focus on the costs in the model, as your driving motivator for your study, given that everything can be so variable.

So for example you would have a spreedsheet that does this,
- First calculate the total amount of asphalt required
-- by calculating the total volume of materials
-- now use a lookup table to select a paving machine
-- This could contain all the factors related to machine.
---- ie speed, width.
--- Number of strips (This is a reset so treat it)

- labour is calculated in multiples assigned to a machine.
etc..

Use one columb for sub $ totals, another for $ totals, repeant the process for time required.

To deal with rates from 0.080 to 1.50 use lookup tables or rule tables to make a decision.

Now once you've got a basic model together, cost a job using it. Then go out and see how good your model was. Look at it from a perato point of view, focus on the biggest costs (and times ) first. Add to the model if your think its not accurate enough. Where you over or under? by what percentage?
Then go back and refine the model, add factors and rules of thumb, until it behaves just as you want it.

Finally to polish the model some things that I've found useful to ask myself.
Do I you really understand the Job? Is the business really an expert at laying Asphalt? Have all the risk been taken into account and forseen? ie have all the unexpected been taken into account, are the tools ready, are there enough tools. Has the preperation and maintenance been costed properly. Because here I've found were time is lost on a job or not properly costed.

Understanding risk is the single most missunderstood thing in costing and therefor timestudies involving work than is has many variable suffers from it. Even after all the work
I done, things can still go wrong with the time standards, because the risk hasn't been understood or known.

So after nearly everly day, I make the managers ask the simple question. "Did things go to plan?" if not what was overlook or not plan for?

And this is my point, the model is the plan for the work. By constantly checking the model after each job, the model becomes more refined. It's quite easy to add a couple of additional columbs to the spreedsheet as actual values, then some percentage accuracy can be gained. Also by carefully designing the spreedsheet it is possible to add new factors in quickly by simply inserting a line.

Ok, lets take an example of how I would deal with the data you have.
Data
Standard rate 0.080 to 1.50 seconds
while binder or first heavier range is from 0.058 to 1.08 seconds.

Now from a quoting point of view I automatically would assume 1.08 seconds for the binder and then 1.5seconds per sq foot for subsequent layers. For the most simplest of models! I would only lower the time as more factors are known. For example what will be the ambient temperature in that month that the job is scheduled. Again going by the worst cases.
Now lets say that you won the contract and the workers did the times at 0.08 and 0.058 per squar foot on the day., so thats 0.138 as the best time as opposed to 2.58 seconds. or 2.58-0.138=1.442 seconds per sq foot. Thats almost a factor of 10.5 times variablity. (Or for the layperson reading this post the job could be done in an hour or take all day with overtime)

So the better you refine the model the smaller that number 10.5 will become.

(You can refine your model more course, I'm stoping here because I just want to illustrate the method.)

Another thing you may want to do, is perform a sensitivity analysis on your model. Try and understand which variables affect that 10.5 factor the most.

Now, hopefully you can see the weakness from approaching the problem from a purely time study point of view. The methodology creates a greater risk for the business in getting it wrong by underquoting the job.

In summary,
By approaching the problem differently from a cost point of view, then generating the time standardard from the quote reduces the risk to the business. Hence by then studying and understanding the risks, a process of refining the model is possible, thereby decreasing cost, and being more compedative. The approach taken is to alway protect the business from financial loss, because financial loss is unacceptable. Ultimately the quote becomes the plan, which can be refined. The plan is then enacted, done and then can be compared to the actual outcome. (Following the old PLAN, DO, CHECK, ACT circle)

I hope this approach is of help.

Good luck
Joewski
Melbourne Australia.