Method of Generalized Evaluation in the Selecting Factors and Levels of Variation in the Multifactorial Study of Sowing Machines

Abstract

The results of studying complex objects can be presented as a problem with several outputs, and this will require a separate mathematical model for each result. If we use a generalized optimization criterion (super criterion), for example, in the form of the Harrington desirability function, one model will suffice. The authors have drawn an example of applying this methodology to determine the factors influencing the seeding process performed with an experimental seeder.(Research purpose) To show general recommendations for improving the process of planning and organization of scientific experiments in the field of agricultural engineering.(Materials and methods) The authors were guided by the idea of converting the natural values of particular criteria (or characteristics) into a dimensionless scale of desirability with subsequent assessment of the level of preference of a set of factors and their levels of variation in a multifactorial experiment. The object of research was represented by a sowing unit used for combined seeding of two crops. (Results and discussion) The study has established the following levels of desirability of natural output values (combinations of the variation levels of experimental factors): "very good", "good", "satisfactory", "bad", and "very bad". The rate of influence of these factors on corn seed sowing with an experimental seeder was obtained by applying a priori ranking technique. An acceptable set of values (a combination of optimal and maximum values) from the range of factors variation was revealed, which provides for more qualitative choice of the main factors influencing the optimization parameter. The authors have demonstrated the operation of the patented software to calculate data according to the Box-Behnken plan, which allows to obtain sufficient information about the response surface of the process in the experiment center selected as the expected optimum. (Conclusions) The paper presents the way of preparing for an experiment and subsequent processing of the results of complex studies using the super criterion. The necessity for digitalizing the process of planning and organization of scientific experiments has been emphasized.

Keywords

• generalized optimization criterion
• organization of scientific experiments
• experimental seeder
• harrington desirability function
• processing of complex research results