CANDAS Dataset: A Cooling Fan Sound Dataset with Modeled Disturbances and Controlled Experimental Conditions
Henrique Lima, Cristofer Silva, Ricardo Nery, Wilmer Córdoba Camacho, Ricardo Brazileiro, Rodrigo de Paula Monteiro, Andrea Maria Nogueira Cavalcanti Ribeiro, Carmelo José Albanez Bastos-Filho, Mariela Cerrada, Diego Pinheiro

Predictive maintenance cuts economic and safety risks in rotating machinery by leveraging vibration and acoustic data, which machine-learning models convert into intelligent fault detectors. Acoustic signals are especially powerful for early fault detection and cooling fans, with simple rotational dynamics, are convenient proxies for complex rotors. Yet existing fan datasets lack disturbance models and controlled conditions. We present CANDAS, a controlled sound dataset featuring 28 h of recordings from two cooling fans under five modeled disturbance conditions.

• Setup

The experimental setup was carefully designed to collect sounds from cooling fans under different weight configurations and voltage levels. To compose our experimental setup, we selected two models of cooling fans: the Delta AUC0912D, referred to as A2, and the Intel AUC0912D, referred to as A3—both commonly found in electronic devices. In order to obtain a greater variety of sounds, an experimental arrangement was defined in which five weight configurations were applied to each cooling fan, under two voltage levels: 9 Volts and 12 Volts.

• Vibratory disturbance model proposed

A model of vibrational disturbances was employed using neodymium magnets as weights, motivated by recent works on the creation of cooling fan datasets with the introduction of weights on the cooling fan blades. In this model, a magnet is fixed to each blade of the cooling fan using superglue, forming a layer of magnets. This first layer makes it possible for new magnets to be stacked on top of the existing ones, creating new configurations of weight distribution

The black dots represent first-layer magnets attached to the cooling fan blades, while the blue dots indicate additional magnets stacked on top. It is possible to design both balanced configurations (i.e., Config 1) and severely unbalanced ones(i.e., Config 5).