This article describes the application of a three-layer feed-forward neural network to analyze industrial plant data. To adjust mathematical models (for control or optimization purposes) from plant data, it is necessary to analyze and detect outliers and systematic errors and to remove them. The system studied is the feed preparation of an isoprene production unit and represents a multivariable problem. To detect outliers in a multivariable system is not an easy task. The technique used in this paper is able to identify this kind of error. The methodology employed involves construction of a reliable neural network model to represent the process and its training with a few iterations (a few thousand). Thus, the points at which errors between the experimental and calculated data appear to be scattered far from the majority of the values are probably outliers. In some cases, outlier points can be easily detected, but in others, they are not so obvious. In these cases, they are separated and a cluster with other similar data is built. After analyzing these clusters based on the similarity principle or by hypothesis tests for means, it is then decided whether or not these points can be excluded. At the same time the process is checked for any abnormalities recorded during the specific period. Three year's worth of process data were analyzed and about 30% of the data were excluded.

Foram realizados dois ensaios nos quais compararam-se plantas de girassol, do cultivar Anhandy, livremente visitadas por insetos(testemunhas) e outras impedidas de visitação por dois métodos (cobertura de plantas com gaiolas teladas, de 3m de largura, 4m de comprimento e 2m de altura, e ensacamento de capítulos com filó). O número, o peso e a porcentagem de sementes granadas foram significativamente maiores (P<0,05) nas plantas visitadas (579,28; 41,17g e 82,37%; 456,97; 23,44g 79,70%, nos dois ensaios), em relação às plantas cobertas com gaiolas teladas (81,53; 2,23g e 1,16% no primeiro ensaio e 111,92; 3,16g e 28,49%, no segundo). O número e o peso de sementes não granadas foram significativamente maiores (P<0,05) nas plantas cobertas com gaiolas teladas (416,29 e 6,58g; 375,25 e 5,04g, nos dois ensaios) em relação à testemunha (142,69 e 1,71g; 182,38 e 2,53g). Os capítulos ensacados com filó apresentaram valores de porcentagem de sementes granadas intermediários entre as plantas testemunhas e as cobertas com gaiolas teladas (64,56 no primeiro ensaio e 68,61, no segundo). Pode-se concluir que o cultivar de girassol em estudo apresenta significativo aumento de produção de sementes devido a visitas de insetos polinizadores, principalmente abelhas, e que as gaiolas teladas são mais indicadas como método para impedir o acesso de insetos polinizadores.

Sunflower plants, cv Anhandy, freely visited by insects (control) were compared to plants protected from insects. Two flower isolation methods were used: covering the plants with net cages (width 3m, lenght 4m, height 2m) and bagging the flowers with gauze. The number, the weight and the percentage of seed setting were significantly higher (P<0.05) in the visited plants (579.28; 41.17g and 82.37%; 456.97; 23.44g 79.70%, in two trials) than in plants protected by cages (81.53; 2.23g and 1.16% in the first trial and 111.92; 3.16g e 28.49%, in the second). The number and the weight of no seed settings were significantly higher (P<0.05) in the plants covered with cages (416.29 and 6.58g; 375.25 and 5.04g in two trials) than in the control (142.69 and 1.71g; 182.38 and 2.53g). The percentage of seed setting in the bagged flowers with gauze was intermediate to the control plants and the covered plants. (64.56% in the first trial and 68.61%, in the second). It is concluded that the sunflower, cv Anhandy, presents a significant increase of seed production due to visits of pollinating insects, mainly honey bees. Cage is the most indicated method to protect plants from pollinating insects.