种鹅舍环境智能监控系统的研制和试验

Development and experiment of intelligent monitoring system for geese house environment

针对种鹅反季节繁殖生产中硬件设备功能低下、难以实施舍内环境操作的适时精细调控、难以获取记录舍内环境数据进行问题溯源等问题,提出一种专门应用于种鹅反季节繁殖生产舍的环境智能监控系统。该系统通过BP神经网络建立温湿度智能调控模型,取代人工手动操作以满足舍内环境要求。通过GPRS模块无线传输舍内环境参数,并利用其GSM功能通过移动终端远程控制风机、照明、水泵等设备。以EXT、Hibernate和Spring为基本框架技术,构建了轻量级、强壮的多级缓存的J2EE企业级Web应用程序,实现鹅舍环境参数的远程监控,并与现有商用人工控制器进行了现场试验和性能对比。试验结果表明:该智能监控系统长期运行稳定、可靠,能够满足鹅反季节繁殖对光照和温湿度的环境调控要求。与人工粗略控制、上海梵龙的畜禽控制器相比,控制精度分别提高5.49%和2.83%。在夏季风机湿帘负压通风降温时测定的舍内温度相对于设定值的均方根误差分别为0.202、0.494、0.372℃,相对湿度相对于设定值的均方根误差分别为1.745%、3.166%、2.621%,控制效果显著优于人工粗略控制和现有控制器(P<0.05)。在精准的光照调控下,种鹅均能按预期的时间开产,并在高峰期长期维持产蛋率35%~45%,表现出稳定、良好的产蛋性能。

Out-of-season breeding technology for goose was proposed due to their obviously seasonality of production. Precisely time control is the most critical factor in this technology. Simultaneously, temperature and humidity control are also needed in case of heat stress. The application of the new technology brings challenges to the traditional breeding methods. Environment control in goose house was mostly artificial or semi mechanized. Manual operation was of poor real-time performance, and could not achieve the organic connection between the various environmental factors. Light, temperature, relative humidity and other environmental parameters could not be recorded real-time. When diseases or some abnormal reproduction happen, managers could not trace causes from the aspect of environment. This study therefore was conducted to develop an intelligent geese house environment monitoring system for assisting geese out-of-season breeding practices. In the system, BP neural network was utilized to construct an intelligent control model with history temperature and relative humidity data. With this model, the system could make comparison between current temperature and relative humidity inside and outside, so as to determine the number of working fans and the on-off state of cooling pad. GPRS wireless transmission module was used to transmit in-house environmental parameters. With this system, mobile terminal GSM function was also utilized to remote control the geese house equipment such as fans, illuminating lights, pumps and cooling pads. On the basic framework technology of EXT, Hibemate and Spring, a J2EE enterprise web application was built. This made it possible of remote control and real time monitoring. In addition, users could also obtain the environment conditions from mobile client or remote control the corresponding equipment. Field tests were conducted in comparison with artificial control and other commercially developed animal house monitoring systems. The system had been running for more than five months. Results showed that this system was stable and reliable during long-term operation, and could meet the requirements of light, temperature and relative humidity of out-of-season breeding. Compared with artificial control and Shanghai Fanlong controller, the control precision was improved by 5.49% and 2.83%, respectively, labor cost was cut down by 50%. Geese elimination rate was decreased by 1.09% and 0.62%, respectively. With the method of cooling pads system in summer, the RMSE between in house temperature of BP neural network （A3）, artificial experience （A2）, setting static operating point（A4） and the set values were 0.202, 0.494 and 0.372℃, respectively, and relative humidity was 1.745%, 3.166% and 2.621%, respectively. Under the precise control of light, geese in all three houses exhibited normal onset and level of egg laying. They began to lay eggs from May I st, and rose to the peak level in about one month. During hot summer, there was no heat stress of geese in A3, they showed steady laying rate, while laying rate of geese in A2 and A4 showed varying degrees of decline. At the peak of egg production, all three groups maintained their laying rate between 35% and 45%. From May 1st to October 1st, the mean daily egg laying rate of A2, A3 and A4 are 24.77%, 31.17% and 27.94%, respectively, which indicated that the intelligent control system of geese house environments allowed geese to exhibit normal out-of-season production performances.