In September I was honored to speak at the California Land Surveyors Association (CLSA) Los Angeles chapter meeting and dinner on the topic of 3D data preparation and 3D machine control as it relates to surveyors.

In my June column, “Requiring Licensure for Digital Stakeout,” I wrote about California’s position that, among other things, data preparation for the purposes of use in 3D GPS equipment is within the field of engineering and surveying. This statement means that the preparation of the GPS bound data must be under the responsible charge of a licensed engineer or surveyor.

The discussion centered on several factors, starting with the fact that 3D GPS machine control is here now and is here to stay. Surveyors are the natural professionals to prepare the data for the tasks this equipment will perform, and the business model of many engineering and surveying firms needs to change to accommodate this line of work. Some surveyors are looking for ways to get into this service while others are still hesitant to jump in.

When Does Interpreting Become Re-Engineering?

A main issue the CLSA surveyors raised is the overriding question, “When do you cross the line from preparing ‘digital stakeout’ to that of ‘re-engineering the site?”’ Of course, there is no easy answer to this; we all may define what we are doing in different ways. For instance, adding vertices to existing 3D breaklines to densify the model may be construed as adding data to the engineer’s model, and as such, constitutes a re-engineering task. Conversely, another person might consider this act as simply an interpretation of existing data that has nothing to do with a re-engineering task. It could be considered as making the data “more palatable” for the task at hand.

I will offer the following response, though: “You cross the line from preparing digital stakeout data to re-engineering when you change the design away from what the engineer intended.” I submitted to the audience that surveyors are routinely making these types of decisions every day when they perform traditional stakeout. Here are some of my thoughts in response to questions raised by CLSA members:

What if the profile data and the cross sectional data don’t agree?Submit an RFI to the engineering firm with copies to the owner. Document the response. Don’t be surprised when you are instructed to follow one component of the data and to ignore another. We have often been told to follow cross section sheet data over any conflicts with the profile, but this could vary per job.

What if the surface data (contours) doesn’t agree with the annotated slopes, grades or spot shots on the plans? What if the data is simply lacking or insufficient such as in the case where too few contours are shown to depict the surface to the blade of a bulldozer? What if the design is satisfactory, but the contractor recognizes that maintenance would be difficult to impossible? What if the plans indicate features that need to be built, but the surface data and related annotations are insufficient or missing?For all of these situations, I suggest submitting an RFI to the engineering firm with copies to the owner. Begin to assess your interpretation of the information so as to make construction more effective. Be prepared to propose your suggestions in lieu of an unsatisfactory or insufficient response. Document the response.

What if the surface provided is augmented by details such that the accumulation of the data provides a complete concept; however, there is no physical replication of this information such as in a TIN file or other usable form? Does assembling this data into a 3D surface amount to engineering?I wouldn’t think so as long as no data interpretations are rendered. The person preparing the data model is simply compiling available data, as presented and submitted, into a monolithic and contiguous form. It wouldn’t hurt at all, though, to have a PE or LS check it over and agree to its accuracy and intended use. Of course, this would be performed for a fee that attributes accountability to the professional.

What if the surface is available but subsurfaces are required? Does the preparation of this data constitute engineering? Does preparing adjunct data required for construction such as utility trenching constitute engineering?For these items, as long as no design changes are instilled into the final product, I think that this is akin to computing information for stakeout with traditional stakes. If the surface data is easily identifiable and the cross sectional data is clear, it should not be too difficult to develop a surface representing exactly what is approved. However, if your state board’s position states that all data geared for GPS usage must be handled by a licensee, then perhaps it does constitute something a professional will need to be accountable for.

This 3D model illustrates the level of accuracy needed to perform and deliver a satisfactory work product to contractors using machine control.

Raising More Questions

The following questions were also raised during our discussion and indicate the curiosities that surround this new task of 3D modeling for machine control. Many of these comments were directed at non-DOT projects since it was noted that many DOTs require that roadways be engineered in 3D with an associated model as a deliverable.

What do the modern software solutions that work in 3D bring to the table for designers and surveyors?Today’s software provides designers with a strong toolkit for developing projects in 3D. But, it should be noted that the major civil/survey software manufacturers have had these tools for almost 15 years, and one wonders why designers aren’t already well along the path to building 3D models. I was told by one design firm recently that they weren’t going to promote or train staff in working in 3D until the government began requiring 3D for submissions. So it will likely be up to the surveyors to prepare their data using this same 3D software in a manner similar to what they have always done, except now they aren’t creating points for stakeout but rather preparing entire surfaces. Hopefully the advent of strong 3D software will embarrass designers into working in 3D if not for data prep purposes, at least as an aid in value engineering, QA/QC and presentations for which they are responsible.

How do you handle the poor constructability conditions in the plans we are to use for construction?Deflect as much responsibility back onto the design firm as is applicable. The engineer should be controlling all design intentions, and the contractor or surveyor should document all of the engineer’s comments in this regard.

What about the increased liability associated with providing 3D model data to a contractor?This is decided on a state-by-state basis. Ensure that you are up-to-date with all of the position statements your state has enacted.

What is the impact of additional liability on the surveyor when he/she creates 3D modeling data and provides it to the contractor for construction?This is why we hold professional licenses. As I was once told, that license constitutes “our right to be sued.” Therefore, we shall be liable for our work. So we need to have all of the skills and education possible when we take on this task. Our contracts need to be vetted by a legal professional to cover us and protect us when appropriate.

Does the 3D model constitute intellectual property?In my opinion, it does. As such, it should be contracted to what we deem fair use to be. For example, we may simply supply the data for “use only” and strictly prohibit any changes to that data. This may require the supplier of the data to perform all change orders, which will have ramifications such as response times, turnaround times, etc. Another consideration is whether we will allow a single change under the requirement that the contractor will take liability for the resulting model. All of this should lead us to an attorney for review of our documents.

How accurate does the model have to be?That depends on your agreement with the client. I have personally witnessed accuracy requirements in the 3/8" range. It will often have to match the inspection requirements for the project. That is why I usually refer to these projects being done to “millimeter-level” accuracies. Much of the 3D machine control equipment these days can build very accurate surfaces based on our models and are only off by small amounts (which can be corrected with proper maintenance). Errors may occur due to blade wear, calibration errors and other items, but these are usually quite small and avoidable. The idea here is that only rarely can contour data be used for these high quality models--although rough grading can sometimes use contours effectively.

Is anyone aware of how many firms are currently building models in 3D?This question was raised by an audience member and the collective response was that maybe one firm in the region was actively working in 3D as a normal course of business. When you consider that the tools have been around for more than 15 years, having only one firm in this category shows how slow to change many design firms are. This is why contractors and construction surveyors are taking the bull by the horns and learning to prepare 3D models--they can’t get them from the design firms. And it doesn’t look like they will be able to obtain them for the foreseeable future, either. So this is the perfect opportunity for surveyors to step up and take over this line of work for themselves.

Why don’t engineers work in 3D?There are several reasons for this. Their business models are not set up to prepare construction plans; they are essentially preparing review and approval plans. Contracts, staffing, training, and use of software and other resources are all tuned for this result. Constructability and review of same is a different goal. Until owners begin requiring design firms to provide 3D data, designers will continue to offer the status quo of 2D data. They don’t realize the incentives of preparing this data yet.

What is involved in the data prep that seems to be causing concern for people?The concept that seems to be throwing most firms and staff is what I will attempt to sum up as: “Contours are output not input.” Designers often sketch in contours to reflect the surface conditions and CAD technicians digitize them in, often with elevations. These contours only reflect maybe a 2-foot interval with little to no information as to what is happening between the contours. Many audience members pointed out that contours won’t show how a sideslope actually hits the existing ground; it only shows where the contour actually resides. Contractors want to know everything possible about that connection of proposed to existing. This is easily accomplished by using breaklines combined with spot shots such that the breaklines contain the actual intended slopes, grades and elevations to properly define the surface. This requires a different mentality than that for drawing contours and it requires some skill in using the 3D tools within their respective software applications.

Can you take contours and spot shots provided by the engineer and move them into the field?Yes, but they often do not have the accuracies needed for top-of-the-line machine control equipment. Hence the potential of delivering an insufficient work product. For instance, contours created to an interval of 2' do not provide enough accuracy for a site that will be inspected to accuracies of ranges from 3/8" to 0.2'.

Have the methods for inspections changed as a result of the advent of 3D machine control?Yes, in some cases they have changed. Extra attention to quality may occur during inspection on the first few projects that use machine control. In one case cited, a DOT purchased a total station and used its capabilities to provide an as-built of a roadway since no stringline was used.

Discussion Takeaways

These are real questions from real people voiced during a recent forum on the topic of machine control. The lesson that came from this particular gathering was that surveyors should be the clear providers of this service of preparing 3D models. However, they need to rewrite their contracts, update their skill sets in building models and consider their rates for taking on the liabilities associated with these deliverables.