Helical gear units H3 motoréducteur flender H3-VV-22B

Flender/Flender Gear Units/Helical gear units H3

cts 8 at 0.5 kg/. This condition allows drying to continue by ambient energy without solar assistance. To decrease losses it is advisable to maintain collector outlet air temperature at below 6 . 3. Additional conditions In the absence of drying, the collector absorber temperature can rise to 1 ; then the fan, F2, is applied for cooling. 4. Support of the material layer temperature measurement procedure 7 CONTROL OF FRUIT STORAGE In developing control strategy for agricultural processes difculties always occur as result of biological elements. The model-based controllers should take into account allthose mechanisms that are rather difcult to describe by mathematical equations (. Forthat reason, biological elements of fruit storage systems can be measured but are notincluded in the control of the process. So far, simple environmental feedback controlhas been applied without accurate responses from the real biological mechanisms. That gives somehow an explanation for those efforts for using articial intelligence (AI) in the control of fruit storage as well (2,. scheme of the speaking fruit approach ( isshown in Fig. 1. The basic control is an environmental feedback control used to maintain the air temperature and humidity corresponding to the set points. The set points, however, canbe adjusted by the system operator. The information for the set points is taken from thefruit measurements. In this loop the fruit responses are identied by an articial neural network (ANN) or by genetic algorithm (GA), which are more effective in this case. The nal decisions on optimal operation can be made by the operator of computer. Sothis loop can be practically operated off-line or on-line. The articial neural network technique is good tool to make dynamic model of an unknown system, especially to identify nonlinear behavior. For training of the ANN,model measurements must be performed. The genetic algorithm (GA) technique serves initially for optimization of multi- vari