Page 22 - 6686
P. 22

of the process, which is being studied [6].
                Exploitation of oil and gas  fields, artificial  lift of oil and
           gas  to  the  surface,  their  gathering  and  preparation  is
           accompanied  by  changes  of  many  dependent  and  independent
           factors,  therefore  the  study  of  these  processes  should  be
           performed using the rational experiment planning methods.
                Supposing  it  is  necessary  to  explore  the  impact  of  four
           factors  on  some  process  value.  To  obtain  model  information,
           each variable is given with five values and it will be necessary to
                     4
           conduct  5 =  625  experiments.  It  is  not  always  possible  to  do
           such a number of experiments so researchers reduce the number
           of factors that affect the process.
                The  main  feature  of  methods  of  rational  experimental
           planning is that, based on the minimum number of experiments,
           they reveal general patterns within of change of each specified
           factor.  This  is  achieved  when  each  experiment  differs  from
           others by non-recurring combination of selected factors [7].
                According to the results of the experiment  a mathematical
           model of the object of study is needed to be found. Mathematical
           model  refers  to  an  equation  linking  process'  performance  with
           factors affecting on it. This equation is a respond function:
                             у = f(x 1, x 2, …x n).                  (4.1)

                Object  of  the  research  can  have  several  variables,  that
           determine  his  condition,  and  the  less  it  is,  the  shorter  will  be
           the plan of experiment. There are different ways to reduce the
           number  of  variables:  changing  the  purpose  of  the  research,
           separation of an object which is being studied into components,
           establishing links between variables.
                Choosing  a  variable  of  state,  it  is  necessary  to  consider
           following requirements:
                - it must have a numerical value, to be measured;
                - it must clearly describe the object of research;
                                                                         22
   17   18   19   20   21   22   23   24   25   26   27