To find the cause of malfunction of HVAC(Heating Ventilation air conditioning) system using machine learning technique

HVAC systems are often non-linear, poorly behaved, noisy, and very unpredictable under real-world circumstances. These complex systems require expert knowledge to maintain, and optimizing their performance can depend on factors that are unrelated to the systems themselves. Factors may include the environment, the systems use profile, the cost of energy and interactions with other systems. Unfortunately, it is difficult to optimize traditional hand-developed, rule-based approaches or personalize them for a given environment because of these non-linear complexities and variations. As an alternative, systems can take data driven approach in the form of measurements over time from a variety of sensors to intelligently learn the effective strategy for operating the building’s HVAC systems.

Machine Learning is increasing in both power and scope every year and it has huge impact on automating industry. In particular, if we talk about large HVAC industry, it has real promises. It can change the way we approach things like: Fault Detection, Fault Prediction, Energy Optimization, System Modelling and System Prediction. HVAC system requires a lot of care and maintenance and is very difficult for skilled technicians to know the best way to control such equipment. Assuming each machine malfunction follows a normal distribution, a technician who sees a new HVAC malfunction every single day could take years to learn the diagnostic skills needed to analyze each function alone. This doesn’t even account for the various combinations and permutations of problems that can exist on a piece of equipment. It’s nearly impossible for a single person to know everything about these machines. This is where machine learning can help by predicting the most accurate cause of malfunction of the system which will lead to performance optimization. In our case, the HVAC systems from where data got collected have more than traditional 200 rules. We have applied various Machine learning algorithms like SVC, RandomForest, Gradient boosting and Artificial Neural Network etc. in order to fulfill our objective and chosen the one which gave better result over others.

Our method is of prime industrial interest because technicians would no longer need to guess what operating parameters should be used after lengthy trial and error. Instead, intelligent controls could tell the technician what the problem with the equipment is and, more importantly, what to do about it. So, these changes will reduce stress for clients and technicians alike while improving our ability to avert disaster.