摘要
在凡纳滨对虾集约化淡水养殖中,温度、pH值、非离子氨质量浓度是3项重要的水质因子,三者之间存在着非线性耦合关系。基于这一耦合关系,提出水质信息融合Agent结构和二级信息融合算法。在融合算法的第一级,采用正向推理机制,基于规则库推断水质因子是否超标。在融合算法的第二级,根据水质因子的耦合关系模型,采用数据库搜索和固定步长搜索策略寻找合适的温度和pH值,通过在线调节温度和pH值的方法,同时将温度、pH值、非离子氨质量浓度调整至合理范围,以满足养殖要求。试验结果表明,信息融合算法能够有效地调节养殖水质,减少换水次数。
In the intensive freshwater aquaculture of Litopenaeus Vannamei, temperature, pH value, and non-ionic am-monium mass concentration are the three important factors to evaluate water quality.Based on the existing nonlinear coupling relationship among the three factors, the inner structure of water information fusion agent and a novel two-lay-er-information-fusion-algorithm were proposed.In the first layer of the fusion algorithm, a forward reasoning machine could infer whether or not the water quality factors are out of the standard range based on a rule base.In the sencond layer of the fusion algorithm, database searching and fixed step strategies were adopted to find the proper values of tem-perature and pH value based on the coupling relationship model of water quality factors.Then, the fusion algorithm ad-justed the values of temperature, pH value, and non-ionic ammonium mass concentration to meet the aquaculture re-quirements.Experimental results showed that the proposed information fusion algorithm could adjust the aquaculture wa-ter quality effectively, and reduce the frequency of water exchange.
出处
《山东大学学报（工学版）》
CAS
北大核心
2014年第4期39-45,共7页
Journal of Shandong University（Engineering Science）
基金
国家高科技研究发展计划(863计划)资助项目(2007AA10Z238)
关键词
水产养殖
信息融合
多AGENT
智能调控
正向推理
aquaculture
information fusion
multi-Agent
intelligent control
forward reasoning
