| Artificial light plant factory is the product of technological reform and upgrading of facility agriculture,which can effectively solve the problems that traditional agriculture is facing now,such as food safety,lack of arable land,insufficient yield and climate impact.However,the current level of automation and intelligence of artificial light plant factories remains in a low level now.So it is necessary to develop its automated operation technology and improve its control strategy to reduce the cost and energy consumption for achieving the goal of increasing production and saving energy.A plant factory control system based on Raspberry Pi has been designed and implemented in this paper with hydroponic Romaine Lettuce as the experimental object,and the light-fertilizer environment control strategy has been optimized through experiments on spectral combination,light period,pulsed light and gradient configuration of nutrient solution concentration.Besides,a research on the growth prediction of plant factory lettuce based on machine learning has been done to provide a basis for the further optimization of environmental control strategies.The main work completed in this paper has been shown as follows(1)The plant factory system and an environmental monitoring system based on Raspberry Pi have both been designed and implemented.A plant factory system has been built as the experimental platform in this paper,and the Raspberry Pi has been selected as the main controller,which equipped with multi-type sensors to realize the comprehensive monitoring of the plant factory environment and the automatic control of multiple devices through PWM drive signals.Meanwhile,the upper computer software system has been developed,which can monitor the changes of environmental parameters in real time,send abnormal data prompt messages,and for users to query historical data or adjust control parameters.The system is simple and convenient,with strong scalability.(2)The optimized configuration of environmental control strategies for the goal of high-yield and energy-saving has been completed through experiments.In this paper,for hydroponic romaine lettuce,multiple sets of control experiments have been done by adjusting the four environmental variables of light period,spectral combination,pulsed light duty cycle and nutrient solution concentration gradient,in order to explore the environmental control strategies that can achieve high-yield and energy-saving goals by taking the four morphological indicators of real leaf amount,leaf length,leaf width and fresh weight and the two energy consuming indicators of LUE and EUE into consideration.The experimental results showed that the comprehensive effect of increasing production and energy saving performed the best when the spectral combination was R:B=6:1,the light period was 16h/d,the pulsed light duty cycle was 100%and the nutrient solution concentration gradient was 400μS/(cm·w).(3)Research on the plant factory lettuce growth prediction model based on machine learning has been done.Based on the three algorithms of RF,SVR and LASSO,this paper predicted the growth results of lettuce according to the six sets of results indicators including real leaf amount,leaf length,leaf width,fresh weight,LUE and EUE and the growth trend of lettuce according to the three groups of trend indicators including real leaf amount,leaf length and leaf width.After the selection of hyperparameter,RF’s performance and accuracy were the best in the prediction of each tag.Therefore,lettuce growth prediction model based on RF has been selected,which has guiding significance for the formulation of environmental control strategies under different production goals. |
PDF Full Text Request