Modelamiento matemático y por redes neuronales artificiales del crecimiento de Spirulina sp. en fotobiorreactor con fuente de luz fluorescente e iluminación en estado sólido

Abstract

The Gompertz and logistic mathematical models in the Spirulina sp. growth kinetics were evaluated and were compared with a modeling through Backpropagation Artificial Neural Networks (BP- ANN). Spirulina was cultivated in a (3 L/min) of 500 mL aerated laboratory photobioreactor with 40W fluorescent lighting and 1W lighting Solid State (LED-Light Emitting Diode) obtaining 11.0 klx lighting with both systems. The LED lighting allowed to obtain a (ɑ) 0.90 high biomass value compared with that one obtained with fluorescent lighting of 0.82, as well as a greater growth rate μ=0.63 h-1 preceded by a shorter latency time λ = 0.34 h. The BP-ANN showed a good accuracy with respect to the Gompertz I corrected model for both the Spirulina sp cultivation case with fluorescent lighting and with LED displaying correlation coefficients (R) of the 0.993 and 0.994 order respectively, with regard to the experimental data. Spirulina modeling through the Gompertz I corrected model is advantageous because besides showing R 0.987 and 0.990 values in Spirulina sp. cultures with fluorescent lighting and with LED respectively, it allows to attain the growth parameters kinetics directly.