The Data Strategy There are data and information that you'll want for the life cycle of the building, and there are analytic opportunities in long-term data you'll want for comparison and trending, plus opportunities for greater correlation between data, improved data analytics and the possibility of developing or identifying new building data metrics. “There is nothing permanent except change.”

Jim Sinopoli PE, RCDD, LEED AP Managing Principal, Smart Buildings LLC Contributing Editor

Buildings should be managed and operated with accurate data. It takes Artificial Intelligence via machine learning to manage and analyze masses of data. Humans can data mine, but humans must be involved creating naming conventions and acronyms for the equipment, where it is located, what building, what floor, what space, etc., acronyms for many types of mechanical equipment, electrical panels, plumbing, etc.

Intelligent humans create Artificial Intelligence. But what is Artificial Intelligence? Will it change everything? Will it cause world war three (via Elon Musk)? Will it kill all the regular jobs? Will the country leading the way become ruler of the world (via Russian President Vladimir Putin)? “Robots are coming for your Job.”

AI is one of the transformations of the 21st century; it will infiltrate all aspects of our lives. For buildings, AI can operate the buildings, but related controls are likely to be quite different. Automation will take on a whole new vocabulary that will include artificial intelligence, robotic process automation, open edge software, self-managing systems, self-organizing networks, self-optimization systems and self-learning systems.

The idea of AI is that machines can accurately mimic intelligence. Yes, machines and computer systems can perform tasks that would normally require human intelligence.  Underlining AI is massive data sets. The machines can “learn” and “solve problems” and have a “cognitive function” somewhat like humans. Statistical methods are used, and systems can understand human speech, identify visual perception, decision-making, and translation between languages.

Machine learning uses computer programs that can teach themselves to grow and to change when new data is available. The process of machine learning is like that of data mining. Machine learning uses that data to detect patterns in data and to adjust program actions accordingly.  Companies related to healthcare, financial trading, insurance, fraud detection, etc. are already using Artificial Intelligence(AI). AI mimics cognitive functions of humans. The machine “learns,” solves problems, and identifies patterns. 

Engineers need to have experience with AI before they see it as another tool, but the results of AI can be compelling. The FDD analytics created by NIST and other industry experts a few years back has been a very useful tool for facility management, but FDD is not in the league of machine learning, self-managing networks, or robotic process automation.
 
It’s likely that artificial intelligence will soon supplant the traditional building management systems from the major building automation manufacturers. The AI platforms are likely to deliver more reliable building information and building performance showing better-operated buildings; employee productivity, energy consumption, comfort, etc.

Artificial Intelligence can be an incredibly positive tool. One of the most interesting applications is medical diagnoses. A type of AI, called deep learning helps medical experts pinpoint faster, and more accurately problems. Radiologists can examine many images of concern in the patient’s body; MRI and CT scans. In 2016, AI systems in healthcare revenue were $1.06 billion. That was up from $811 million in 2015 and $633 Million in 2014.  International Data Corp predicts the AI market could deliver $47 Billion revenue by 2020 – as well as generate revenue for organizations using AI as part of their processes driving profitability, efficiency, and market.

What do people think of AI?   70% say they understand AI.  33% are uncomfortable with AI.

The weak link is the humans. How does the building owner manage building data? Are both facility management and IT involved? Does facility management have any IT staff? Are there equipment naming conventions and acronyms, for MEP, electrical panel naming conventions, mechanical acronyms, electrical acronyms, plumbing, document revision history, etc.

There are many “data "repositories" in a building, such as asset management, inventory, maintenance, building management systems, independent control systems, facility management systems, business systems, BIM data, as well as construction drawings, and product data, and data in the hands of third-party contractors that install, service, and maintain building equipment.

Much of the data is stored away in varied electronic and paper formats. The typical building has several "silos" of data scattered throughout the organization with no cohesive strategy for data management and little coordination. Also note that it's not only the data that is in silos but also the underlying technology systems for data management, different data management processes, and even the people involved.

There is a good case for bringing all the facility data into a unified database architecture and putting into practice standard methodologies and processes to manage the data. There are several benefits to this approach.

The industry has realized that data, data analytics, and even machine learning are emerging major tools for improving building operations. Data applications, such as energy management, and fault detection, and diagnostics, are probably the best early examples of the effectiveness of managing and analyzing data. The effort for many building owners to acquire and manage facility data, however, appears either ad hoc or narrowly focused on specific aspects of the building, such as energy and HVAC systems.

Many data "repositories" currently used in buildings provide a substantial amount of data. They include building management systems, independent control systems, facility management systems and business systems.  In addition, there is the "umbrella" of Building Information Modeling, which addresses design and construction drawings, equipment and product data, as well as data in the hands of third-party contractors that install, service and maintain building equipment.

The typical building has several "silos" of data scattered throughout the organization with no cohesive strategy for data management and little coordination.

A structured approach can improve the archiving, preservation and retention of data for the long-term. There are data and information that you'll want for the life cycle of the building, and there are analytic opportunities in long-term data you'll want for comparison and trending, plus opportunities for greater correlation between data, improved data analytics and the possibility of developing or identifying new building data metrics.

[an error occurred while processing this directive]If you're involved with new construction and going through the programming and conceptual design of the facility, the project team should establish rules for the data management that will be generated throughout the project with some thought given to the data that will need to be exported into operations and facility management systems. Yes, the focus in new construction is typically the construction schedule and budget, but any acknowledgment and appreciation of long-term operations and rules and standards for data management would be positive.

During commissioning and project closeout, data and information such as commissioning reports, project record documents, contract drawings, project manuals, contract modifications, startup logs, test reports, certifications, the complete as-built BIM and other documents and data are generated. All this information should be permanently retained and accessible. Some documents may be paper, such as certifications, but all documents and data should be submitted electronically and stored. The importance of many of these documents is that if the building or its systems are modified the designers and contractors will want to use the original record document as the baseline.

An immense amount of building data is created during the design, construction, and operation of a facility but we've only managed and analyzed a relatively small amount of the available data. The industry foray into AI, data management, and analytics is just in its infancy. The initial results, however, especially analytic applications, show impressive results and are very promising. We should expect the model to apply to other building systems and additional data to be generated by new building systems.