When you are awarded a job with a given accuracy spec to reach at X confidence level for each main control point, what is your approach to plan your field work?
I'm a Starnet user and use the pre-analysis program which allows me to enter approximate coordinates, instrument specs and # of observations which I can use to either add observations, geometry or swap out instruments to hit spec and I find it works quite well.
If a simple closed traverse with the instrument we have on hand doesn't meet spec I can start adding cross ties, extra common points, extra rounds of observations and if at a certain point the amount of extra observations becomes too great I can swap out instruments and rerun it. And of course after I process the actual observations I can verify whether it did or didn't hit spec & reobserve if necessary but ideally I'm after a solution that has enough of a buffer that it doesn't require an additional site visit.
Does anyone else have a different solution for planning their field visit?
Well, I'm with you on the preanalysis routine, but I think that 98% of folks just wing it.
Some state manuals have guidance on procedures for achieving a particular specification, but a lot of state manuals were written a decade or two ago...
If you have a specific accuracy spec to meet (rare), and if the specifier actually knows WTF they are asking for (rarer still), then preanalysis is your friend.
OUAT an architect/project manager was asking for points to be monitored to an accuracy of ±0.01'. We used preanalysis to show that we would have to use a 1 second gun set up over concrete piers and turn several wraps of angles to glass targets in order to achieve the desired spec, and that the monitoring program would therefore cost north of $150,000. They gave the job to someone else who just reported their results to two decimal places.
For the record, ±0.02' is much, much, easier to achieve than ±0.01'. And a soldier pile wall moves more than ±0.02' between the cool of the morning and the heat of the afternoon.
To be completely honest I'm part of the 98% most of the time... But then again my "standard" methods equate to hitting flies with sledgehammers usually so I don't worry.
If the client had intense tolerance specs then yeah I'd probably play around with preanalysis. If I were to provide any advice I'd need to see what the project looks like on Google Maps or something and know what the control is for and what the spec actually is...
Location is going to determine your means and methods more often than not. Are you in an area where you can shore yourself up with GNSS? Can you see multiple control points in each setup? Can you resection off 3-5 control points at a time? Yada yada...
You said a nasty word. PLAN. That is a no no word in surveying from what I hear. Just go get it done lol. Just kidding but from what i have seen last few years its not far off.
I think this scene sums up what you're laying down Ole River.... 😉
OUAT an architect/project manager was asking for points to be monitored to an accuracy of ±0.01′. We used preanalysis to show that we would have to use a 1 second gun set up over concrete piers and turn several wraps of angles to glass targets in order to achieve the desired spec, and that the monitoring program would therefore cost north of $150,000. They gave the job to someone else who just reported their results to two decimal places.
Hell, this happens to me at least a few times per year.
Fine by me. Not my problem if they can't understand accuracy, and I don't need to take on liability for a spec that I can't meet.
😂
Fine by me. Not my problem if they can’t understand accuracy, and I
don’t need to take on liability for a spec that I can’t meet.
OK as long as you have plenty of other work. My major point is that sometimes the preanalysis shows you that the sought specification is unrealistic.
@Mark Mayer
Agreed. I wasn't trying to be flip about it...when I submit monitoring proposals, there's an explanation of the methodology we use to calculate time & effort (COE deformation manual), exhibits of the monitoring and control point locations, reports from StarNET preanalysis, and a narrative that discusses why the specifications will take so much work to achieve. It's not uncommon for the client to contact us to discuss it further, and we'll talk about how much less work it would take to loosen the specs even just a hair, and/or cut back on the number of monitoring sessions.
Most of the actual clients (cities, counties, whoever) understand what we are saying and are happy to try and reach a compromise. But it's ultimately up to some PE to decide the specs, and they tend to not want to budge on any of it because they know better. So in those cases I don't mind not getting the work.
A good approach, when either the person who wrote the scope doesn't understand accuracy, or when you know others will quote low, is to provide options in your quotation (with price and duration):
1. Compliant with (if possible) or
2. As accurate as practically achievable (schedule of rates)
3. Non compliant, but with typical equipment and procedures.
Using a potential property I may survey as a test subject, I took some time to get familiar with the preanalysis feature today. Wow, It's pretty neat.
Regarding standard error settings for the instrument library; I feel a bit silly for asking this, but for traverse, I always do BS-F1-F2-BS for what I call an angle set. Does starnet consider this one angle set or two angle sets? https://support.microsurvey.com/convert/instrumentsettings.html
Thanks.
That's one set.
I would consider that one set. But you can used the online calculator to compute your instrument settings in Starnet. https://support.microsurvey.com/convert/instrumentsettings.html <
That calculator is a little misleading as they would define a set. Pay attention to the number of BS and FS they use to define a set and adjust accordingly. For example, we turn 3 sets, so BS FS FR BR is one set do that three times and you have 6 BS and 6 FS so i would use the 6 option for starnet. I acutally copy and paste the results of the calculator into my data file with a note as to what I used as input.
