Connected HVAC systems behaviour analysis: data mining on error codes using DS/ML techniques

• Autor(a) principal: Costa, Luís Miguel Dias dos Santos Pereira da
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/41047
• topic: Data mining; Sequential rules; Association rules; System maintenance; Hvac

With the recent growth in popularity in smart devices, most products have started to offer smart variations. These connected appliances are often accompanied by applications that give the users more control over them. Through them the users are capable of remotely configuring the devices and can also be notified whenever a problem occurs. The smart HVAC solutions offered by Bosch are connected to a backend service that logs and processes the incoming error messages and decides when to notify the customer. With the number of connected devices predicted to increase, it has become appealing to explore the logged error messages. The data was processed and converted to a format that was easier to analyze. After a brief analysis, it was decided that it would be interesting to explore the relationships between error messages. With this goal in mind, a process was set up to analyze the data using Association and Sequential rule mining. The obtained results show that a significant part of the incoming error messages did not correspond to real errors but were instead the result of the control units misinterpreting the lack of response from the appliances sensors.