Automated data audit in BIM: checking information model for errors
Errors may appear during the process of creating BIM-model. The reason for that might be a human factor or mistakes in shop drawings. In this case the advantages of BIM – saving of time and money – might be lost. The way to avoid it is to use data audit.
In this article we will talk about how these errors appear and how BIM-models can be checked in an automated mode.
There are 2 ways to run a data audit – manual and automated.
The automated one can lower the time of data checking from days and weeks to a few minutes. A BIM-manager doesn’t need to look for errors and write down the comments for a designer, designer will see what needs to be corrected himself and a customer will get a high-quality model with the most accurate calculations.
Artem?Knyazev, Product Director at “Tangl” shares a case from his manual?check?practice.
Manual check of information modeling
I would like to share with you my own case, how and why data audit?can speed up the processes by several times. In 2018 I worked as a BIM-Manager in a big development company. It was my pen test, my first project. The?task?was?standard – to?calculate the scope of work. At?that?moment?the?only?possible way for me to collect data from the model was to use a tool classification.?I?needed?to?assign?codes?to elements of the model, and the number of elements might be hundreds of thousands. The total number of elements (including engineering structures) for a small residential complex would be around 300 thousand.
Problem?with?elements?classification
In Revit or Navisworks a planner can classify elements easily: depending on the standard size or other properties. But there are complex elements, which don’t have standard sizes. For example, a part of a multilayer construction, which is not recognized by Navisworks as a multilayered - the wall - includes masonry, insulation, facade. Which code should it be assigned to, if there are many variations of masonry, insulation and facades, and they can be combined in different variations also? It means such elements need to be divided into parts first. I made a note in my technical task to the designer which codes need to be assigned to which properties. The designer classified 150 thousand elements, but he was not an economist or a quantity surveyor and he couldn’t understand the difference between some positions. Taking into account the volume of construction, the price difference might be quite big.
One section took a few days
I started to check the model and understood quite fast that these codes would be needed for the future tasks too, where qualitative data are a must to complete the task successfully. As it turned out in the process of testing, half of the elements didn’t fall into Navisworks search sets, which were added to the technical task. Out of this number, around 40% were reclassified into the wrong property, 30% incorrectly: Cyrillic alphabet was mixed up with Latin, the space was added by mistake etc. And 30% didn’t have a classifier code at all. I started to check the elements manually, but checking 70 thousand elements might take a lot of time – it took 12 hours to go through one section. Of course I could send the project back tp the designer, but I wanted to do the job well. I put down my comments, for each element separately, it took a few more hours. For one section I spent several working days.
?At the same time I received 4 more sections, the situation with each one was same. I had enough time, approximately a month, because it was a pilot project. And even this month was not enough to check all the elements. Usually a BIM-manager doesn’t have such amount of time at all.
Work with changes
The designer made changes and was swearing that he checked all the data. I opened the project and noticed, that half of the ?corrected? information was wrong again. The cycle repeated. And the most unpleasant thing: the data, that was validated at the first time partly failed validation this time. Why? Because the designer made some changes: deleted a wall or moved it, it automatically caused changes in floors etc. Settings were configured manually, that’s why when the element was deleted and the new one was modeled the code wasn’t saved. I started checking and putting my comments again. This experience helped me to understand that data need to be checked in a different way. I wanted to do my job well, but couldn’t succeed. Now our Tangl Control program helps to check all the data automatically.
Tangl is a platform for deep analysis of BIM data.
领英推荐
How Tangl works:
Tangl Control. Tangl allows complex checks to be made under the governmental standard, products standard of the company or EIR – customer information requirements. It’s very helpful at the stages of expert evaluation, preparations for a tender and interaction with customers.
Tangl provides multi-stage data verification according to the technical task: from general to particular. At first elements are selected, then validated, tested for the correctness of filling out.
Automated BIM-audit in Tangl Control has 3 stages:
As a result a designer will get an accurate report with all the details: where the data is wrong or mixed up.
Data audit in BIM is not a regulatory requirement, but it could improve the accuracy of BIM-models and bring work processes to a new level. Data audit is first of all the saving of time, and secondly the facilitation of the future tasks fulfillment. The less errors there are at the beginning, the better result will be reached.
Artem Knyazev, Product Director at “Tangl”
Fields of activity: product and project management, digitization of construction.
Autodesk Certified Professional
Cases: