Artificial Intelligence Techniques for the Control of Refrigerated Potato Stores. Part 2: Heat and Mass Transfer Simulation

The simulation described in this paper is designed to form a component part of an expert system controller for refrigerated potato stores. The simulation is designed to function in two ways, first to provide information to the controller to allow it to make the required management decisions, and second to provide data otherwise unavailable from sensor inputs, either due to physical difficulty of measurement or due to the cost of it. The development of the program has been in C ++ and illustrates some of the benefits that can be gained from using an object orientated (OO) approach. The simulation consists of modelling the heat and mass transfer functions of both the crop and the store, and is based on dividing the crop into arbitrary layers with plug airflow between them. Although the model was developed initially for a bulk store, it was subsequently modified to deal with box stores, and it is its role in modelling the latter that is discussed. This paper describes the methods used for developing the simulation, the program structure and the results of the validation and verification against real store data. The results show a good correlation under two types of test conditions, steady state and pull-down to a lower temperature.