Accuracy of Data from Physical Survey versus BIM Extraction

This Survey uses Revit Model Measures to Compare against SPM Surveyed Data of 26 identical apartments.

The Survey

This is comparison of Duncan Terrace Flats  A01 to A26 which are all similar,( some handed so bathroom/kitchens are flipped- see plans below).

Each individual flat was surveyed by one of 5 surveyors who had a group training prior to the survey to ensure consistency of surveying techniques. The time taken to do the survey (of about 1800 units was 4 hours per unit, this survey recorded the dimensions and the condition of the elements in the room and drew the plan of the unit as well as taking photos (usually 6 per room, one for each wall and one for floor and one for ceiling (1))).

((1) I would note there is the  opportunity to speed up photographing by using 360/180 panoramas of rooms instead. See LINK and  LINK. Only one photo required per room).

Only data tat present for Revit modelling for the housing stock of 1800 units is 1 hour /50m2 (2). A 1 bed flat would be about 50m2 (but stock is estates, single units, family units and stand alone houses so no direct comparison so far).

((2) Time supplied by Matt Cantwell for estimating Facilities BIM documentation (the big buildings, not housing)).

For a direct comparison for this exercise, the flats we are comparing are 30m2. So for creating the BIM model and extracting the data of one unit in 1 hour (not sure if this also includes doing the 2D plans as well, although this was pretty automated at the end of the process),  also allowing surveyor to go into the unit and survey condition of the elements at 20 minutes (see comments at end of article). So round to of 1 hour 30 minutes  (and 2D floor plan and data schedule) compared with 4 hours site measuring(hand sketch) and  surveying.

Both processes need the data to be uploaded to the Asset Management System and site records photos stored on the system.

The data from the survey was put into the SPM Asset Information System. Note, survey data rounded to nearest whole number 1m or 1m2.

The Revit models were developed from Plan drawings from the archives of Wellington City Council (WCC). The data was then extracted using RTV Tools  RTV  Reporter Pro tool developed for WCC.

The data extracted from the model was meant to follow the methodology of the SPM survey, in that elements surveyed would also be captured in Revit & exported.

In some ways, the data from the survey was more explicit, if there were splashbacks behind sinks this was called off as a different wall finish with a seperate area, this could not be replicated in Revit and only the total area of the wall could be obtained.

SPM Survey data has been extracted and compared to Revit Extract for the following:

Ceilings  (SPM) versus Room Area (Revit)

Scotia (SPM) Versus  Room Perimeter (Revit)

Wall Finishes (SPM Paint) Versus Wall Area (Revit- Using Caduceus Wall Lining -Door-Window)

Results.

The Excel file with tables can be downloaded HERE (Excel Spreadsheet 41kb)

1. Ceilings & Perimeters seem to be quite accurate from the Survey, although it would be good to compare irregularly shaped rooms to see if this observation holds (could test on Upper levels where living rooms are L Shaped)
2. Wall areas seem far more variable and need some further investigation.

• The bathroom & Kitchen may have Melamine/Seratone or other non-paintable surfaces .
• There is still a big variance in Surveys measures for very simple shaped rooms.

3. There is a big variation on Total Count for each of the items.

• There are  generally 23 Units that have been surveyed, but in the  Data from SPM some only have 14 (see Ceiling -Hall (14) , Scotia -Hall(12)  & Wall-Hall (20)). If all surveyed why are there not 23 entries?

Note:- Accuracy of Revit model drawn from  Plans- To confirm sizes an actual Site measure check would need to be taken in at least one of these units to validate the accuracy of it.

The 2 plans, one handed with bathrooms on other side.

Comments on this exercise.

1. During the process it was found that creating the floor plan for the units for the surveyors prior to going on site sped up the process of the survey, the surveyors only needed to mark changes, and all the rooms had correct numbering process (this was an issue in an earlier pilot as surveyors would number rooms differently between surveyors, causing some confusion)
2. If the surveyor only had to inspect “CONDITION” of elements and not do the measuring of rooms, the surveys would average 20 minutes a unit as opposed to the 4 hours average for the complete survey.
3. The Revit models quantities were consistent, whereas there was variability in the surveyors information, even at the level of basic shapes of rooms with only simple elements (one door & one window) the variance was noticeable, although, when grouped together they averaged out to within a reasonable degree to the Revit quantity (Just hope the extremes are not used in quoting for work).
4. It is quicker to do a Revit drawing from existing plans, especially on sites with multiple units, to get quantities that can be checked on site, versus full site survey.
5. The skill of the surveyors is in judging condition of objects/elements, the tedious process of quantification can be left to Bim model data extraction.
6. Once the models are made, you do not need to remake them. The spinoff is a set of consistent 2D plans (PDF’s for distribution to the team) and also 3D PDF’s that can be used for distribution to the team and for data extraction (see this post).
7. Other spinoffs from the Model plans were : 1/ calculating gross area for insurance for property stock,2/ Cross referencing all data to correct room number with external contractors, 3/structural inspections post earthquake,4/as base model for any future developments (in as is state).
8. The basic model can be updated and enriched with services data (see this post & this one)  adding more information to the FM’s team knowledge base that can evolve over time for existing buildings. For new builds this can be captured from the outset as long as you get design & construction teams to deliver BIM outputs to the client.
9. When doing Quality Assurance checks on surveyors data you need to know that the rooms are the same, since there are only had sketches (not dimensioned plans (easily created if required)) it is hard to know if their is a variance if you don’t know that the units are the same size. That is an added bonus of having the plans.
10. This work was extended from the Housing to the larger buildings of WCC, including Theatres, Swimming pools, office buildings, libraries , art galleries, and Sports buildings. Getting accurate quantities from these would have been very time consuming from a physical surveying point of view. An added level of development was individual numbering of each door and window with a unique number, so that a specific window could be identified with a defect, rather than a possible one of 20 in a specific room, increasing accuracy of communication between the FM team and their contractors.