RIASSUNTO
A software is designed for water quality monitoring in the aquaculture industry. It is mainly for dissolved oxygen soft sensing. Dissolved oxygen is an important dependent factor of water quality and has effect on fish growth. At present, most of the dissolved oxygen sensors are expensive for aquaculture farmers, so they won't use them for real-time detection to perform control and optimal operation. To deal with this problem, a soft-sensor software was designed and developed based on a data-driven model which is proposed by partial least squares (PLS) and neural networks. The software included three parts which were control software, monitoring software and model calculation software. An industrial case study demonstrated the feasibility and efficiency of the proposed soft-sensor software, and it was simple to use, real-time and generic. It was also the foundation for the control and optimization of dissolved oxygen in the aquaculture process.