In the facilities management space, the convergence of Artificial Intelligence (AI) and predictive maintenance is urgent and innovative. By melding AI’s analytical prowess with real-time data, this transformative approach averts operational disruptions, slashes costs, and elevates facilities’ performance.
Explaining this much better is Javeria Aijaz, Managing Director of HITEK. She says: “Since the inception of AI, it has found widespread application and endorsement across all sectors. Over time, it has become a subject of extensive discussion and implementation efforts within facilities management. This transition aims to move away from traditional modes of reactive, corrective, and planned maintenance towards the realm of predictive maintenance. AI brings forth a multitude of significant benefits, including cost savings, prolonged equipment lifespan, heightened reliability, streamlined maintenance scheduling, data-driven insights, minimised downtime, safety enhancements, condition-based monitoring, efficient resource allocation, and a competitive edge.”
HITEK, a UAE-based provider of smart facilities management (FM) solutions, is part of Farnek group of companies.
Aijaz adds that in contrast to conventional maintenance methodologies that rely on fixed schedules or respond to failures, AI-driven predictive maintenance operates proactively and is tailored to the specific attributes of each piece of equipment. As a result, it not only delivers cost savings but also enhances resource scheduling and utilisation, thereby amplifying overall facility efficiency. Traditional maintenance approaches can sometimes result in inadequate or excessive equipment upkeep, leading to increased operational expenses and heightened downtime due to unforeseen breakdowns.
“While AI offers a plethora of advantages, its integration must be executed judiciously into existing maintenance strategies. This involves considering pivotal factors such as data quality, model precision, and the expertise of maintenance personnel. It is through this strategic fusion of AI capabilities with the existing framework that its true potential can be harnessed to drive efficiency, effectiveness, and lasting operational gains within facilities management,” says Aijaz.
In response to the growing demand for AI, an application called CAFMTEK was developed by HITEK. This smart, secure, and sustainable maintenance management tool incorporates AI throughout its processes, starting from the assets level. By harnessing predictive maintenance approaches, this tool assists organisations in transitioning their processes from traditional to predictive maintenance.
Sangeetha B, CEO and Founder of Amantra FM, adds: “In a traditional maintenance approach, the core objective is to increase an asset’s useful life, efficiency levels, and availability. That is to say, it is concerned about the asset alone and, as such, is one-dimensional. Conversely, AI-driven predictive maintenance harnesses connected intelligence working in tandem with Machine Learning and other algorithms, analysing not only the asset’s performance but also its relationship with the surrounding environment, including other critical systems and functions in the facility — characteristic of a multi-dimensional approach. The integrated insights help identify anomalies and patterns before a potential breakdown, thus averting a service disruption, associated downtime, and repair costs.”
Paul Bogan, Chief Digital Officer at Serco, lists down some of the benefits of AI-driven predictive maintenance: “Cost savings are a main benefit. AI-driven predictive maintenance can help reduce operational costs by minimising unnecessary maintenance activities and optimising maintenance schedules. Traditional maintenance approaches often involve routine or time-based maintenance, which can lead to over-maintenance and increased costs.
“AI can predict potential equipment failures and issues in advance, allowing maintenance teams to address problems before they become critical. This can extend the lifespan of equipment and assets, reducing the need for frequent replacements. Because AI can predict when equipment is likely to fail and recommend proactive maintenance, another benefit is minimised unplanned downtime. Traditional maintenance approaches might lead to longer periods of downtime due to reactive responses to equipment failures.
“Enhanced operational efficiency is another key benefit: AI algorithms can analyse large amounts of data to identify patterns and trends in equipment behavior. This information can be used to optimise maintenance schedules and improve the overall efficiency of facility operations”
Other benefits include: Data-Driven Decision Making, Reduced Maintenance Disruption, Safety Improvements, Customised Maintenance Plans, Resource Optimisation, and Improved Asset Performance.
Challenges
However, Bogan notes that while predictive maintenance using AI offers many benefits, it’s not without its challenges, which include initial setup costs, data quality and integration, and the need for skilled personnel to interpret and act on the insights provided by AI models.
Aijaz adds: “Human nature tends to resist change initially, as people require time to visualise the benefits before embracing new concepts. In the realm of technology, adapting to change has become a challenge that every organisation must confront. This challenge is evident in the adoption of AI within facilities management, manifesting as hurdles such as data quality and availability, data privacy and security, skill gaps, change management, integration with existing systems, initial investment, model accuracy and interpretability, cultural shifts, vendor selection, and scaling up implementations.
“To address the challenges, organisations can adopt a holistic approach that encompasses technical, organisational, and cultural considerations. Forming multidisciplinary teams, executing pilot projects, selecting experienced vendors, and providing continuous training, monitoring, and optimisation of AI models are measures that contribute to the successful adoption and long-term benefits in facilities management.
“Several challenges and barriers must be addressed before embarking on any AI implementation. As an organisation, we have integrated AI into our overall strategy to maximise efficiency and reap its benefits. We have established working groups comprising FM professionals, system designers, and AI developers. Their collaboration has resulted in the creation of standardised operating procedures for training and operating AI models, as well as end-user training.”
Resistance to change, a lack of capital investment, and continued indifference toward empirical evidence and monitored reports related to AI-driven predictive maintenance are the key barriers, says Sangeetha. “They can be addressed by raising awareness of the potential rewards, the business risks of resistance to new technologies, and, most importantly, the sustainability outcomes achievable through predictive maintenance in buildings. The growing sustainability imperative will make AI a necessary adoption instead of a discretionary one.”
According to Sangeetha, despite its well-documented benefits, AI-driven predictive maintenance has not found mainstream adoption, partly due to “shoestring budgets” in FM. The hesitancy can also be linked to the nascency of AI in real estate operations. “Thus far, the manufacturing industry has been the greatest beneficiary of AI-driven predictive maintenance. Soon, with more mass-marketing and the resulting increase in awareness, predictive maintenance could become a norm in FM as part of the IFM offering.”
HITEK has successfully implemented AI-driven predictive maintenance within CMMS solutions, providing FM teams with the tools to make data-driven and informed decisions. This empowers teams to optimise energy usage, lower operational costs, and elevate facilities’ overall efficiency. Notably, other prominent players in the market have also embraced AI-driven predictive maintenance. Examples include Royal Dutch Shell, Siemens Gas Turbines, and Honeywell Building Management System. All these entities share a common goal: to achieve enhanced efficiency, cost reduction in maintenance, and increased energy output.
Aijaz adds: “These instances underscore how AI-driven predictive maintenance has the potential to revolutionise efficiency, minimise downtime, and elevate operational performance across diverse industries and sectors within facilities management. By harnessing insights derived from data, organisations can confidently make informed choices, efficiently allocate resources, and ensure the seamless operation of critical equipment and systems.”
Bogan says that effective predictive maintenance through AI relies on a combination of data and sensor technologies that provide accurate and timely insights into equipment health and performance. The types of data and sensors used can vary depending on the industry and the specific equipment being monitored. Bogan lists some essential types of data and sensor technologies commonly used for predictive maintenance, along with strategies to ensure data quality and reliability:
- Vibration Sensors: Vibration sensors detect abnormal vibrations in equipment, which can indicate misalignments, imbalances, or wear. They are crucial for detecting issues in rotating machinery like motors, pumps, and turbines.
- Temperature Sensors: Temperature sensors monitor the heat generated by equipment. Sudden spikes or fluctuations in temperature can indicate problems like overheating, friction, or cooling system failures.
- Pressure Sensors: Pressure sensors measure fluid pressure in systems such as pipelines, hydraulic systems, and boilers. Anomalies in pressure levels can signal leaks, blockages, or other operational issues.
- Current and Voltage Sensors: Current and voltage sensors are used to monitor the electrical characteristics of equipment. Changes in electrical behavior can reveal issues like motor faults or insulation breakdowns.
- Ultrasonic Sensors: Ultrasonic sensors detect high-frequency sound waves that are not audible to humans. They can identify issues like leaks, friction, and air/gas flow irregularities.
- Oil Analysis Sensors: Oil analysis sensors monitor the condition of lubricating oils in machinery. They can detect contaminants, wear particles, and degradation of the oil itself.
- Infrared (IR) Sensors: IR sensors measure infrared radiation emitted by equipment. They can identify hotspots, worn bearings, and electrical malfunctions.
- Optical and Visual Sensors: Optical and visual sensors capture images and videos of equipment. They can be used for visual inspections and to detect visible defects or anomalies.
- IoT and Connectivity: Internet of Things (IoT) devices and connectivity enable real-time data transmission from sensors to central systems. This connectivity is crucial for timely analysis and decision-making.
Aijaz says: “AI-driven predictive maintenance relies on data-hungry models that necessitate a continuous flow of precise, consistent data for making accurate predictions. This practice helps avert potential issues or challenges. Achieving this data feeding process is feasible through various methods, with manual data entry and the utilisation of available sensor technologies being the most vital approaches.”
As an organisation, HITEK also employs sensor technologies to gather data from assets, yielding positive outcomes. Among the key IoT sensors widely utilised by HITEK are those for current, temperature, voltage, vibration, and the Building Management System (BMS). Leveraging this data, the CAFMTEK platform generates and assigns work orders for assets requiring maintenance. Emphasising both intelligence and security, HITEK places significant importance on data security and employs industry-standard encryption technologies. This ensures that all assets are meticulously monitored and maintained to operate at their utmost efficiency.
Sangeetha says: “The efficacy of AI-driven predictive maintenance is as good, and as bad, as the data feed. The integrity, comprehensiveness, and history of data are vital to unlocking desirable outcomes in facilities. Most commonly, in old buildings, the heterogeneity of the type of devices and sensors, due to their adoption at different stages of technological evolution, undermines the quality and reliability of data. At times, the use of different sensor types could be due to vibrations, infrared, etc. Facility managers can ensure better data quality by making it an enterprise-wide priority, establishing effective data governance guidelines, and investing in training and learning.”
She adds that the impact of AI-driven predictive maintenance on the business bottom line can be immediate, but it becomes more apparent cumulatively in the long run. However, the benefits are contingent on data quality and sensor connectivity with the central control and monitoring system (CCMS). If implemented effectively and methodically, AI can transform the maintenance strategy from reactive or preventive to predictive — a paradigm shift with great implications for energy optimisation and asset performance and lifespan.
Aijaz concludes: “As we all know that AI is the future as its is advancement is increasing day by day, so in order to adapt the continuously changing and enhancing technologies, organisations must have a strategy in place to create multidisciplinary teams that are working and collaborating together to bring the AI to existing organisational set up so that it can be utilised to drive efficiency and cost optimisation.”
In an era where efficiency and uptime are paramount, the marriage of AI and predictive maintenance in FM offers not just a glimpse, but a tangible path to a smarter, more resilient future. By harnessing the potential of AI, organisations can navigate maintenance challenges with foresight, ensuring facilities operate at their best while embracing a new standard of operational excellence.