A model that contains comprehensive and useful information is a valuable asset for owners and users of a facility. It can provide information such as the detail of individual assets, risks associated with a space and data to assess a facility’s business case.
But, obtaining a model with comprehensive data can be nearly impossible to obtain. Problems arise for two separate reasons. Firstly, facility owners need to precisely identify the information that is required, right down to the data fields required for each type of asset. This is why all new build projects should have a comprehensive EIR with data requirements (including asset information requirements). A comprehensive EIR, that details your data deliverables, will reduce the risk of you not receiving the information that you require.
Secondly, design authoring applications (such as Revit and ArchiCAD) are optimized for developing models to produce drawings. It’s an understandable focus of the applications, the applications are used by design consultants for the purpose of developing designs to enable buildings to be constructed. And drawings make up the majority of the design information.
This means that adding data is difficult and resource intensive. And when something is difficult and resource intensive it quickly becomes expensive. You end up having to employ individuals or organisations with the technical competency and the expertise to do the task.
I’m fully aware of these problems. I have worked on projects where BIM was implemented too late in the design process and design consultants provided high fee quotations for additional BIM services such as adding asset data to models. The high fee quotations highlighted the cost of adding model data.
Additionally, if the information that you require adding to a model is only useful to you, then the cost and benefits cannot be shared. For example, a reference number for an instant hot water boiler will be useful during construction procurement and during the life of the asset, so the cost of providing this information can be shared. But, the serial number of the same instant hot water boiler, will only be useful during the life of the asset.
This means that numerous organisations, building new facilities, are missing out on obtaining valuable data-rich models. Models which would otherwise prove to be valuable digital assets, enabling estates to be managed in a more efficient way.
Being able to add data to models using an efficient and automated approach, without the need for cumbersome and costly design authoring applications, is the solution to this problem. The data can be collected using simple spreadsheets and added directly to models. Additionally, the data should only supplement the existing model information. It shouldn’t change any of the design information, just add to what has been produced by your design consultants.
Until a year ago, I didn’t think that this would be possible. The problem appeared to be too large and difficult to solve, requiring a deep understanding of model files and data formats.
But, the pieces of the jigsaw began to fit into place. Firstly, the only file format that would enable models to be manipulated outside of proprietary design authoring applications is IFC. Secondly, open source tools IfcOpenShell enabled IFC models to be read and manipulated. And finally, IfcOpenShell can be accessed using the Python programming language which is ideally suited for data-intensive tasks.
This means that it is possible to develop solutions for efficiently adding data to models.
At BIMsense, using the combination of IFC files, IfcOpenShell and Python code, we now have a solution to the model data problem. We have a simple and effective 3 step method of adding data to models.
The first step is to classify objects. Ideally, the classification of objects has been done by your design consultants. If so, this step will just require a verification that classifications have been provided. If not, then we add the classifications to the model.
The reason for adding classifications, such as those provided by Uniclass 2015, is that they provide a means of clearly understanding and grouping together the types of objects within your model. The data required for a maintainable asset such as a fire damper, will not be the same as the data required for a ceiling tile. By grouping assets together, using classifications, enables different data sets to be applied to the different types of assets.
This next step involves adding the required dataset fields for each object type. This could be the information that is required for objects such as fire dampers and ceiling tiles. The data for both types of object would be different, although there would be an overlap both would have an object name and would have manufacturer information.
This step only adds empty data fields. It adds empty fields so that information required for each object can be clearly listed. The data fields for each type of asset are taken from our data template, which will be bespoke for your requirements. We take the data template and add the data fields for each object type.
The completion of this stage involves us providing you with simple spreadsheets. The spreadsheets require completing with the detailed information for each object. This is usually completed with information from your supply chain, your designers or your existing records.
This last step involves us taking the completed spreadsheets with the individual object data and adding the data directly to the data fields from step 2.
The information is verified and the finalised data-rich models are ready to be used for efficient estate management.
We are all aware of the benefits of a data-rich asset model, how the model becomes a valuable asset providing a return on the initial investment throughout the life of a facility. But, the current difficulties in being able to add data results in facility owners missing out on obtaining a valuable digital model asset. At BIMsense we want all facility owners to benefit from BIM which is why we have developed ‘EASY BIM DATA’, 3 steps to developing a data-rich model.
Don’t risk missing out on your data rich model.