An object of research is the unit model for rotational milling of a paving.  Experiments with the milling unit are multiple-factor. They are connected with development of the most rational design of the milling module, with check of working capacity and search of optimum operating modes of the unit model in the process of rotational milling of a paving surface.

The mathematical model of a research object is presented in a look of "a black box", which has 5 influencing factors as entrances: х₁ - type of a processed material (ice, asphalt); х₂ - speed of a milling drum rotation (maximum, nominal); х₃ - projection of a driving force to a longitudinal axis of the unit; х₄ - projection of a driving force to a cross axis of the unit; х₅ - projection of a driving force to a vertical axis of the unit. There are 3 variables at the exit of "a black box": у₁ - unit productivity; у₂ - straightforwardness of lateral borders of a processed strip; у₃ - height of roughness of the processed surface.  Unit productivity у1 is chosen as the main output size because it is unambiguous in statistical sense. у1 gives really effective assessment of functioning of system as optimization parameter. Parameter of productivity of the unit meets the requirement of universality and comprehensively characterizes object, it has physical sense and is easily calculated. Volume productivity of the P_V unit has the dimension [m³/hour], which characterize volume of the cut-off material in a unit of time. It can be calculated by multiplication of speed V_п of milling material strip on the square S of the cross sections y₁ ≡ P_V = V_п·S, where the area of cross section S is defined as multiplication of width of a processed material strip on depth of the removed layer. The remained two sizes у₂ and у₃ serve as restrictions in the solution of a problem of search of optimum operating conditions of the milling unit.

Each factor accepts in experience one of the two values called levels. The fixed set of levels of factors defines one of possible conditions of "a black box". At the same time it is a condition of carrying out one of possible experiences. The number of all possible experiences is determined by the expression N = p^k, р - number of levels; k - number of factors. In the case under consideration р=2, k=5, then N=32.

The following demands are made to factors: they have to be operated, operational, unambiguous, independent, compatible, and all their combinations feasible and safe. Accuracy of measurements of factors has to be high.

All possible options of experiences for field tests in full-factorial  experiment (FFE) like 2^k are presented by a matrix of planning of the field experiment, which contains only 16 options from possible 32 options. It is caused by the fact that if the sign «+» of the factor х₅ associates the effort necessary for creation of pressure of a milling drum on a processed material, the sign «-» of the factor х₅ associates the efforts directed on a drum separation from a processed material. Experiment is directed on clarification of operational properties of the unit, therefore, 16 options from 32 options can be rejected at this stage. The position of a milling drum on height in a model sample of the unit is set by means of the special lath established in the lower part of the case of the unit. The maximum depth of a processed layer is 20 mm.

The signs «+» and «-» have no basic value for the factor х₄ - cross force, therefore 8 of 16 options can exclude from experiment, as equivalent owing to symmetry of a design of a drum of rather processed strip. The mills on a drum are located as teeths of a shevronny wheel. Such arrangement of mills provides stabilization of cross provision of a drum concerning a strip of a processed material. - the direction of movement of the unit is connected with the sign «+» for a factor х₃ - longitudinal force, coinciding with the processing direction. Then the sign «-» will characterize the longitudinal movement of the unit which hasn't been connected with process of processing of a strip. It is possible not to consider 4 of 8 options for this reason during carrying out an experiment.

It is executed 20 measurements for output variables for calculation of statistical characteristics (a population mean and an average quadratic deviation - AQD): y₁ - height of roughnesses (or depth of the processed strip) and y₂ - width of the processed strip. The following values of productivity of the unit are established as a result of carrying out the pilot studies: for ice from 3 to 5 m³/hour, for asphalt from 1,5 to 3 m³/hour with size SKO from 0,0001 to 0,0004 m³/hour.