凡纳滨对虾养殖水质预警研究

Water Quality Breeding Prawn Early Warning

研究对虾养殖水质准确预警问题。对虾的活动性极低,对水质的动态变化影响的周期很长,使得对虾养殖水质因子的突变性很小,使得对虾养殖中的水质因子存在非线性、不确定、难以精确评价的难点。传统的水质监测方法多是基于水质瞬间突变,根据水质因子的瞬间异常变化作出报警。如果水质因子的突变性较小,将很难作出及时准确的监测,不能做出及时预警。提出了一种T-S模糊神经网络的水质监测建模方法。该方法结合了模糊系统的推理能力和网络的自学习能力,通过对水质标准模糊系统的结构辨识和参数辨识,建立各种水质因子和水质等级之间的非线性对应规律。实验证明该方法很好的解决了对虾养殖中的水质监测问题。

Research accurate warning of prawn breeding water quality. The dynamic change of water quality is complex. This paper put forward a T-S fuzzy neural network water quality monitoring model. This method combines the fuzzy system reasoning ability and network self-learning ability to identify the strueture and parameters of water quality system. All kinds of corresponding laws between water quality factors and ranks were establishing. The experi- ment results show that the algorithm improves the accuracy of identification.