Target Parameter Acquisition Method Of Co-regulation Of CO₂-light In Cucumber Juvenile Based On Photosynthetic Rate Prediction Model

Plant factory is an efficient agricultural system.It realizes continuous annual crop production through automatic control of environmental parameters inside facilities.Among them,light is the energy source of plant photosynthesis,CO₂ is an important raw material for photosynthesis,and the lack of light and CO₂ in facility production directly restricts plant photosynthesis,leading to the accumulation of photosynthates reduced,thereby affecting crop yield and quality.Therefore,the regulation of light environment and CO₂environment has become the key of modern agricultural regulation and control research.Traditional methods of regulating environmental factors mostly adopt fixed thresholds,which do not take into account the effects of dynamic changes of environmental factors on crop growth and development.However,the existing multi-factor coupling control modes mostly adopt single-objective control strategies such as light environment or CO₂,but they are independent control systems and fail to achieve the coordination of each factor in dynamic control.Aiming at the existing problems,this paper starts from the influence of temperature,light and CO₂ concentration on Photosynthetic rate,and uses support vector machine regression?SVR?,discrete curvature theory,moving asymptote algorithm?MMA?and non-linear regression to establish the model of the target parameters acquisition method of co-regulation of CO₂ and light,so as to further promote the research of facility environment regulation model.The main work and conclusions are as follows:?1?Study on Modeling Method of Photosynthetic RateIn this paper,a method of establishing photosynthetic rate model based on support vector machine regression algorithm is proposed.In order to improve the accuracy and generalization of the prediction model of photosynthetic rate,the prediction model of photosynthetic rate based on support vector machine regression was constructed,and the generalization ability of the prediction model was evaluated by k-folding cross-validation,so as to select the model.Compared with many regression methods,the same training set and test set are used to verify the accuracy of each algorithm model.In cucumber seedling stage,the average absolute error,determination coefficient and mean square error of the predicted and measured photosynthetic rate based on support vector machine regression algorithm are 0.7518?mol·m^-2·s^-1,0.9895 and 1.1221?mol·m^-2·s^-1,respectively,which are superior to the other three regression algorithms.The results show that the prediction model of photosynthetic rate constructed by support vector machine regression algorithm has better prediction ability.?2?Acquisition method of target region for co-regulation of CO₂ and light based on Photosynthetic modelIn this paper,the target areas of CO₂ and light regulation based on Photosynthetic model were studied.Firstly,the photosynthetic rate response surface of CO₂ and light effects at different temperatures was obtained based on the photosynthetic rate prediction model;secondly,the curvature maximum point?control threshold point?of CO₂ and light response curve at different temperatures was obtained by U-chord curvature-hill climbing method according to the characteristics of photosynthetic response surface.Finally,all curvature maximum points are mapped to the CO₂-illumination plane and polynomial regression is used to obtain the co-regulation region of CO₂-illumination.According to the target region of carbon dioxide-light regulation established in this paper,the lowest photosynthetic rate in the region is the regional boundary point.Taking the temperature of20?as an example,compared with the light saturation point,the photosynthetic rate is3.6%lower and the light intensity is 27.5%lower on average under different carbon dioxide concentration.Compared with the traditional supplementation point of CO₂(1000?mol·mol^-1)under different light intensity,the average photosynthetic rate and carbon dioxide concentration were 8.91%and 31.14%lower,13.53%and 42.62%lower than the traditional supplementation point of CO₂(1200?mol·mol^-1).The results showed that the regulation parameters?CO₂ concentration and illumination intensity?in the target domain could not only ensure that crops improve photosynthetic capacity,but also effectively reduce the cost of regulation.It provides a basis for intelligent agriculture in intelligent dynamic regulation technology.?3?Study on efficient co-regulation of CO₂-light intensity based on Photosynthetic constraintsIn this paper,based on the obtained CO₂-illumination control target area,using the principle of European nearest distance and using the moving asymptote algorithm to obtain the specific parameters of the current environment control target.In practical applications,real-time sensor detection and raspberry pie equipment are needed to optimize online.Firstly,real-time environmental parameters are detected by sensors,and the target area of CO₂ concentration-illumination intensity regulation is obtained according to the current ambient atmospheric temperature.To determine whether the measured environmental parameters?CO₂ concentration,illumination intensity?are in the target area or not,the Euclidean distance optimization mathematical model between the target area and the real-time environment is constructed.Carbon dioxide concentration and light intensity,which are the lowest control cost at present,are obtained by moving asymptote algorithm as the control target parameters.