Multi-level Numerical Simulation And Optimal Control Decisions For The Precooling Process Of Apples

Precooling is the first phase of the postharvest cold chain before fresh fruits and vegetables are put in long-term low temperature environment such as refrigerated storage,transportation or display cabinets,which is a critical step toward ensuring the quality and safety of fresh horticultural produce,extending the shelf life and sales zone of the produce,and improving the overall economic benefits of the cold chain.Therefore,the rapid precooling after harvest to remove field heat have become the most essential of all the value-adding services and been demanded by the increasingly more sophisticated consumers.A variety of precooling techniques are available in the agriculture industry?e.g.,room cooling,forced air cooling,hydro-cooling,vacuum cooling,liquid icing,etc.?;of these techniques,forced-air cooling?FAC?is the most prevalent and low-cost precooling technique used to remove field heat.This thesis is based on the forced-air precooling process of Fuji apple,a 3D CFD?computational fluid dynamics?model of a horticultural produce packaging system and the produce contained therein?i.e.,corrugated carton,tray,and apples?was developed,which was used to comprehensively and detailed research the cooling characteristics of produce from microphenomenon?e.g.,the cold air movement,heat and mass transfer in inside a single fruit-packing crate or between two packages?to macroscopic phenomena?e.g.,precooling energy consumption,chilling injury and mass loss of fruits?by combining with field tests.Furthermore,based on a large number of validated simulated data,an optimized BP neural network was developed by using Genetic Algorithm?GA?,which provides a reliable theoretical and experimental basis for further achieving an integral performance evaluation,the optimization of control decision and intelligent control for the FAC process of apple.The main reseach contents of this thesis are stated as follows:?1?Selection of high-precision turbulence model:This study was based on an experimental platform which was similar to a refrigerated vehicle or a small-scale cold storage.The objective of this study was to establish a comprehensively verified 3D CFD model of this experimental platform to simulate the airflow and heat transfer at different two-equation eddy viscosity turbulence models such as standard?-?,RNG?-?,Realizable?-?,standard?-?,SST?-?and RSM,and to predict the temporal-spatial temperature and velocity variations during cooling.After comparing the accuracy of six two-equation turbulence models,the SST?-?model gave more accurate predictions by a comparison of the experimental and simulated results.This research thus provided reliable method for better understanding and selecting high-precision CFD turbulence models to the FAC process of apple,and further ensuring the accuracy of subsequent research.?2?Construction of the energy model:To improve the accuracy and authenticity of the simulation results,the phenomena of respiration,transpiration,condensation,and convection in the produce zone are modeled with a user-defined function?UDF?written in the C programming language.?3?Dynamics simulation study on the level of single fruit:This work established a quasi real three-dimensional model of Fuji apples,by comparing the wall shear stress,drag coefficient,separation angle and recirculation length of an apple with those of an equal-diameter sphere over a large Reynolds number range?10³0000?,and verify whether an apple can be directly replaced by an equal-diameter sphere in the study of CFD models.Moreover,to achieve the optimal CFD modeling for the contact points between apple and apple?or the wall of boxes?and fully consider heat transfer by contact between the produce,this study is also based on two apples contact with each other,and determine the optimal size range of 4 different ways of the contact point modifications?e.g.,reduction,expansion,smoothing,bridge?for different Reynolds number range from the accuracy of the simulation results of the drag coefficient and temperature distribution along a line from the center of an apple to the center of another apple.The research results show that an apple can be replaced by an equal-diameter sphere in academic research,and the maximum relative deviation is about 15%for all the physical parameters in the whole range of Reynolds number;the accuracy of bridge methods is best in 4 different ways of the contact point modifications.?4?Dynamics simulation study on the level of single box:This study presents a market survey that studies samples of typical apple cartons used in China.Furthermore,by combining experiment and CFD modeling,a novel integral approach is proposed to evaluate cooling rate and uniformity,energy efficiency,and fruit quality?including safety?as a result of FAC for different ventilated-packaging designs.The process uses CFD to simulate the three-dimensional spatio-temporal distributions of airflow and product temperatures during precooling.In addition,experiments on chilling injury and mass loss are also reported.The results show that the optimum fresh-fruit packaging design depends on the product size and the location of the product and tray inside the packaging.For all existing package designs,the optimal air-inflow velocity is found to lie in the range 0.4–1 m/s(or 3-5 L s^-1 kg^-1).Furthermore,to obtain the optimal packaging design for a specific product,we need to consider not only how vent area,shape,number,and location affect cooling efficiency,but also the product size?e.g.,diameter or volume?and the location of product and tray inside the package.?5?Dynamics simulation study on the level of palletized fruit:This study established a three-dimensional model of palletized packages,which was used to evaluate the cooling rate,uniformity,moisture loss,and energy consumption during precooling with different pressure differentials.An exponential relationship exists between fan power?P_w?and seven-eighths cooling time?SECT?(P_w=a*SECT^-b)as well as between total fan energy consumption?E_c?and air-inflow rate?E_c=hx^k?.Furthermore,the mass loss of fruit is directly related to the cooling time of the FAC rather than the changes of differential pressure.?6?Particle swarm optimization?PSO?for theproblems of single objective optimization:For the stardard PSO and constriction factor PSO,an adaptive inertial weighting method is designed to balance exploration and exploitation in generating offspring according to the diversity among individual particles?i.e.,the population distribution charact eristics of the PSO?.Furthermore,an exemplar is the minimum from Lagrange interpolation with three points,this process ensures the learning direction is always flying to a theoretical point which is better than the global optimal particle?i.e.,gbest?.Finally,when the gbest has not been further improved,only one dimension of gbest is randomly choosed to perform a mutation operation through a Gaussian perturbation.In the operating process,the new particle wll be accepted if its fitness is better than the current gbest.Otherwise,the new particle is used to replace to the particle with the worst fitness in the swarm.?7?The Construction of the BP model and system development for the prediction of cooling performance:This work established an optimized BP neural network by using PSO based on the large number of validated simulated data.The air-inflow velocities?or pressure drop?,cooling temperature,initial temperature of fruits,cooling time,and hole ratio of package as input vectors,and the average temperature and temperature uniformity of all fruits as output variables.The BP model can be used to evaluate the real-time cooling performance for different operating conditions during precooling,which is very meaningful to optimize the current operating conditions and thus improve the cooling performance of fruit.The thesis provides reliably theoretical basis for improving the accuracy of CFD numerical simulation in cold chain logistics application,increasing the cooling uniformity,precooling throughput and the economic performance across the entire cold chain,reducing precooling energy consumption and fruit quality loss and achieving an integral evaluation of the performance of FAC as well as optimizing the precooling control decisions.