a choice and the description of a technique of carrying out of experiments

Mapping out of experiments spend for the purpose of a choice of a minimum quantity of spent experiences under optimum conditions of target parametres [1,51].

For carrying out of experimental researches has been selected central composite rotatabelnyj the plan of full factorial experiment (PFE TSKRP 23) as the given type of the plan of experiment gives the fullest and exact picture about the received information and in regular intervals distributes it on all spacing of factorial space. Thus there is a minimisation of the regular errors connected with inadequacy of representation of results by polynoms of the second order [1,7].

On the received data during carrying out of experimental researches the mathematical model representing a piece of a polynom of the second order is made:

For check of a mathematical model on adequacy it is convenient to use F-criterion (Fisher's criterion)  as this method is the most suitable statistical method for check of adequacy of a model and check is reduced to comparison with tabular values [65,67].

The importance of each factor is carried out by check by criterion Stjudenta [65,92] under the formula:

72

gdekoeffitsient at variation factors in the regress equation;

- Square-law error of factor of regress.

For an exception of influence on function of the response of the regular errors caused by external conditions, provided by a negative mould of mapping out (tab. 3.1) experiences were spent in casual sequence. For this purpose it is necessary experiences randomizirovat in time and for these purposes the table of random numbers  was used.

It is known, that the result of repeated measurement comprises both casual, and instrument (regular) errors. The total error of measurements is defined as follows [1,92]:

Where - size of a casual error of measurements;

- Size of an instrument error.

As the casual error decreases with increase in quantity of measurements, it is expedient to make such quantity of measurements that the casual error could be neglected in comparison with an instrument error.

The number of repeated experiences at experiment carrying out is defined as follows:

Where n - number of the spent measurements;

- Size of the received error of measurements (provided that number of measurements nтаково, that we have neglected factor Stjudenta);

- Size of an instrument error.

Table 3.1

Negative mould of mapping out PFE TSKRP 2