The future is now, and you can hear it.
It’s more important than ever to think about how we can employ predictive maintenance to keep everything, from our air conditioners to our cars, in working order. Most people only know of two ways to maintain machines — reactive maintenance or preventative maintenance. But these aren’t always the best or most cost-efficient processes.
Think about it in terms of risk management. You can take the risk and be surprised when a failure happens (reactive maintenance), pay an ongoing premium for scheduled preventative maintenance, or find a middle ground and optimize the process through predictive maintenance. With predictive maintenance, a facility looks for clues about the working condition of its mechanical equipment and acts accordingly before a catastrophe happens.
The biggest question then becomes: How do you observe the condition of the machine and then understand what your observations mean?
Listening to the Problem
A number of different technologies are used in predictive maintenance today. My company, Augury, utilizes ultrasonic and vibration analysis to listen to heating, ventilation, and air conditioning systems (HVAC) in order to figure out where problems exist.
Ultrasonic diagnostics can help technicians pick up on problems and pinpoint the sources. This is the first step; the one where you can definitively say, “Something is wrong.” Similarly, with infrared thermal imaging, heat is often a quick indicator that there’s something wrong with a part in a machine.
Vibration analysis can then be used to better understand what the issue is. It will help you figure out exactly what’s wrong and how to fix it.
Oil analysis is used in conjunction with these technologies to gain better insights of the inner workings of a motor. By analyzing its oil’s properties — such as consistency, viscosity, and free metal particles — in a lab, we can point to wear and high-running temperatures.
By combining these methods into a predictive maintenance plan, a facility manager can plan ahead. That way, events where machines suddenly stop working will no longer be an issue. This translates into less downtime.
But not everyone is on board with predictive maintenance.
Jumping Hurdles
The good news about moving toward more predictive diagnostics is that it could help cut maintenance costs, reduce productivity loss, increase revenue, and even reduce energy consumption.
But some companies have not completely embraced the ideas and technologies of predictive maintenance for two main reasons: expense and training.
Historically, predicting a breakdown could be very expensive. Some of the sensing equipment could reach $20,000. And even if you have the equipment, you need someone with expertise to make it work. Vibration and ultrasonic analysis require years of training and certification to use the current systems. Setting up an in-house predictive maintenance program will cost more than $120,000 for the first year.
But these costs are worth it in the long run. While predictive maintenance can be initially expensive, the long-term price tags of preventative and reactive maintenance will dissipate — and the people who are impacted by machine downtimes will be in better shape because of it.
Saving Dollars and Lives
Due to the inherent costs of predictive maintenance, facilities tend to focus on their most critical and most expensive machines while leaving the auxiliary machines to lament their time with reactive or preventive maintenance. But a chain is only as strong as its weakest link, and in the case of HVAC, a failure can cause unexpected costs — or even the loss of lives.
Just last month, it was reported that a hospital in Dallas had air conditioners that weren’t working as they should, causing problems for newborns in the neonatal intensive care unit. Additional cooling sources were brought in to the building to prevent any issues from escalating. HVAC failures in NICUs can be the difference between life and death for newborns who can’t regulate their body temperatures as easily as adults.
With more predictive maintenance, an issue like this would have been foreseen and even prevented.
How can big data help with these types of preventive maintenance? One option is to use machine learning driven HVAC software. This technology has a great track record for boosting performance.
Luckily, the cost of predictive technology continues to drop, so everyone can continue to make strides toward this more effective and efficient model of predictive maintenance.
The only question that remains is whether you’ll be ahead or behind the future of diagnostics.
