Articles
CONTROL OF THE COMPOSITION OF THE NUTRIENT SOLUTION IN AN AUTOMATED NFT SYSTEM: A SIMULATION STUDY
Article number
304_32
Pages
281 – 290
Language
Abstract
A simulation model is developed to follow the change in concentration of macronutrients in an automated NFT system equipped with ion-selective electrodes.
The main emphasis is on the effect of random measurement errors.
It is assumed that the cumulative uptake of each macronutrient is given by a logistic model.
The measured concentrations are simulated by assuming that they are normally distributed around the true concentration.
Any other possible systematic errors are assumed to be absent.
From the simulated measured concentrations the cumulative uptake is computed and compared with the assumed logistic cumulative uptake.
The goodness of the estimated uptake is expressed by a correlation coefficient.
The automated system configuration described by Heinen et al. (1991) is used in this study.
The two aims are to keep the concentration of the nutrient solution constant, which is of particular interest to growers, and to obtain reliable cumulative uptake patterns from the measured concentrations, which is of particular interest to researchers.
These two aims are characterized by the average deviation of the mean concentration during the growth period from the set point concentrations (Cdif), and by the minimum time interval for which the correlation coefficient between assumed and computed cumulative uptake exceeds 0.9 (t0.9), respectively.
The deviation from set point concentration Cdif decreases as the measuring error decreases, or when the system volume or set point concentrations are increased.
The correlation time t0.9 decreases as the measuring error decreases, or when the system volume or set point concentrations are decreased.
It is concluded that growers and researchers require different systems.
The main emphasis is on the effect of random measurement errors.
It is assumed that the cumulative uptake of each macronutrient is given by a logistic model.
The measured concentrations are simulated by assuming that they are normally distributed around the true concentration.
Any other possible systematic errors are assumed to be absent.
From the simulated measured concentrations the cumulative uptake is computed and compared with the assumed logistic cumulative uptake.
The goodness of the estimated uptake is expressed by a correlation coefficient.
The automated system configuration described by Heinen et al. (1991) is used in this study.
The two aims are to keep the concentration of the nutrient solution constant, which is of particular interest to growers, and to obtain reliable cumulative uptake patterns from the measured concentrations, which is of particular interest to researchers.
These two aims are characterized by the average deviation of the mean concentration during the growth period from the set point concentrations (Cdif), and by the minimum time interval for which the correlation coefficient between assumed and computed cumulative uptake exceeds 0.9 (t0.9), respectively.
The deviation from set point concentration Cdif decreases as the measuring error decreases, or when the system volume or set point concentrations are increased.
The correlation time t0.9 decreases as the measuring error decreases, or when the system volume or set point concentrations are decreased.
It is concluded that growers and researchers require different systems.
Authors
M. Heinen
Keywords
measurement error, logistic uptake
Online Articles (42)