For many building owners, maintenance costs are a necessary evil. However, the growth of data-driven maintenance (DDM) means that buildings of the future are becoming a reality today, as landlords utilise Operational Technology (OT) to improve productivity, while generating value and optimising the customer experience.
Traditionally, responsible organisations would devise a scheduled maintenance program, either in-house or contracted out, to proactively ensure all equipment is inspected and serviced regularly. Conversely, companies that are struggling financially might not be able to commit to a regular maintenance schedule, instead being forced to follow a more reactive breakdown maintenance route.
Scheduled maintenance ensures key assets are examined and overhauled regularly and avoids unnecessary faults and downtime from occurring. However, these programs can be time-intensive, expensive and might result in machines being serviced more often than necessary, all while never fully avoiding the potential for unexpected breakdowns. By comparison, breakdown maintenance is cheap in the short term, but can be expensive and disruptive in the longer term as the effects of faults often go unmitigated before significant damage and interruption occurs.
DDM offers a third choice, reviewing the historical data of your equipment to provide a more productive and responsive solution than any traditional option. By capturing and analysing machine data and combining this with conventional data collected on-site by technicians, DDM can provide a more complete picture of the health of an asset than ever previously possible. Moreover, this knowledge can be leveraged to streamline resources and facilitate remote maintenance, maximising both the efficient running of your machines and the extension of their working lives.
DDM has been made possible with the emergence of the Internet of Things (IoT), the growth of OT dataflow through machine learning, and the increasing ability to retrofit smart sensors into otherwise ‘dumb’ equipment. There are now several hundred sensors available that can measure and record all pertinent operating parameters, such as temperature, pressure, humidity, vibration and fluid levels.
While these developments facilitate the collation of a greater volume of performance information than imaginable just a few short years ago, many companies that have invested in such smart technology do not use the data to its full potential. They dutifully collect and store the generated records, but at most employ it to retroactively run reports on past operating behaviour. Relatively few businesses appreciate that this wealth of data can be utilised in close-to real time to enhance their current operational procedures.
A typical existing maintenance contract entails planned maintenance with fixed schedules for inspection and servicing. While on site, the technician will see how that asset is performing at the exact moment he is looking at it—a little like taking a snapshot in time—but effectively knows nothing about how it was operating the previous day, week or month, often under significantly different conditions. He knows even less about how the same asset will fair a day, a week or a month into the future.
Effectively, the role of the technician can be augmented by something far more efficient through the utilisation of OT data, or through IoT sensors that have been retrofitted as part of a DDM upgrade. Where the technician might see how a machine runs for a 30-minute period twice a year, sensors will diligently observe and record all aspects of a machine’s performance every second of every day.
The collation of OT data is crucial to the efficacy of DDM, but having to analyse data is another story. The sensor data needs to be processed with algorithms and passed through advanced analytics software to produce meaningful knowledge that can generate actions.
Grosvenor implements DDM solution via its Grosvenor Actionable Insights infrastructure—InsightsAI™. This continually processes unified OT data to identify deviations outside acceptable operational boundaries, indicative of a fault requiring further investigation. To supplement this, automated remote maintenance checks can be conducted by commanding equipment to run under predefined conditions to measure and recording performance. For example, remote checking of refrigerant levels before the summer season will identify any physical maintenance to be added to the next technician’s visit.
This effectively permanently embeds ‘virtual technicians’ and ‘virtual engineers’ into a building, thereby eliminating requirements for many labour-intensive non-value generating maintenance practices, especially fault finding. This arrangement highlights issues that are then fed through to the Grosvenor Service Teams, delivering the information required to maintain a building’s technical systems in peak operating condition.
Grosvenor monitors its customers’ assets remotely at its National Operations Centre (NOC) and when data anomalies occur, one of three actions will be generated. Firstly, it might be possible to resolve the issue remotely, if the fix involves something as simple as adjusting a schedule or a set point for instance. Interestingly, 10-20 per cent of all faults are rectified without a site visit, thereby minimising machine wear and unplanned downtime, and rationalising technicians’ travel and time on site.
The second action that can be generated from the insights is that a service task might be initialised through linked maintenance or CMMS software. Based on the urgency of the attention required, the software schedules a technician—either a Grosvenor employee or any external subcontractor—to resolve the issue, often enabling several less-serious jobs to be collated for completion during a single visit.
The third potential action that the software can instigate is a requirement for the facility manager to begin the process of acquiring quotes for the work to be completed.
The most recent implementation of Grosvenor InsightAI has involved the delivery of DDM across eight sites of a client that manages a portfolio of offices, shopping centres and industrial facilities. A total of 550 issues have been identified across the portfolio, including fault detection and optimisation opportunities. The data gathered from this implementation has proven that DDM is cost-effective and can help reduce maintenance expenses. One example was the malfunction of an air-handling unit pressure sensor that led to excessive fan operating speeds, thereby using more energy than necessary. Once this fault was identified, the actionable insight was allocated to the next maintenance visit and was efficiently resolved.
Of the 550 total issues identified, 73 per cent have already been fixed and 11 per cent were able to be resolved remotely within 10 days at no additional cost to the customer and without incurring the need for any maintenance hours on site. To date, the client is enjoying a saving of 6,615 kWh/day of energy and 51 kL/day of water.
These improvements have also had a marked enhancement in the NABERS rating of the buildings concerned, with an average increase to date of 0.2 Stars. Not only does this indicate an optimised utilisation of electricity, gas and water, but also signifies a marked elevation in tenant experience through better indoor air quality control.
For one of the client’s buildings, the cost—split over three years—of setting up the analytics was $8,330, and the annual operating cost of the maintenance program is $18,000. However, these outlays can be immediately offset against numerous annual savings: water usage of $4,647; energy consumption of $23,100; reduction in on-site technician time through remote rectification of $9,750; and a reduction in the maintenance contract of $12,000. Overall, the simple payback achieved for this location has been an impressive 6.3 months.
DDM has numerous benefits and few downsides. That said, transitioning from a conventional maintenance service contract to a DDM plan takes time and investment before you start seeing cost savings, requiring a significant change in attitude in order to embrace the potential that DDM offers. Added to this, not every factor can be measured, and for some assets, the cost of installing new sensors can outweigh the benefits.
However, one of the most immediate attractions of DDM is that it is scalable. For any company unsure that this approach is for them—but with concerns over critical equipment—the solution is simply to trial DDM on those key areas. It is a solution that works effectively for OT and IT systems alike, including HVAC, fire, electrical, plumbing, BMS, elevators, computer hardware and software, and even Internet services.
With the insight and the peace of mind that this induces, most businesses will feel the immediate imperative to expand their DDM experience to less-critical peripheral assets. As many problems can be identified before a technician arrives on site—and are often resolved without the need for a site visit at all—DDM will evolve to replace non-productive maintenance and reduce faultfinding hours. This means that overall contract time can be reduced, freeing more of the budget for analytics.
In real terms, as the example above demonstrates, there is huge scope to mitigate utility expenses while extending asset life, and in many instances saving both time and money on unnecessary maintenance. Machine breakdowns become less likely, and resulting downtime greatly moderated. Furthermore, as client data grows, the more this information can be pooled for the mutual benefit of all. For instance, chiller records from one customer will be used to improve service requirements for all Grosvenor clients operating similar equipment.
Data-driven maintenance demonstrates that knowledge is power: the power for a business to leverage the best performance from its assets—and the power to maintain a competitive advantage in an increasingly challenging market.
