A new approach to analyzing the power operating modes of electric networks called the energy flow calculation is considered. By solving this problem one can obtain a balanced model for distribution of electric energy flows and losses on the network graph. Evidence is given to the fact that nonrecoverable imbalances arise when operating modes are modeled with the use of conventional equations describing steady state operating conditions. Correlations describing the energy flow problem must be in the form of energy balance equations in electric network nodes and branches. It is shown that these equations retain their adequacy when any changes take place in the network circuit configuration. The energy flow problem can be solved separately for the flows of active and reactive energy. Aspects in which the energy flow observability conditions differ from those used in the classic state estimation problem are also described. Matters concerned with revealing gross measurement errors and localizing commercial energy losses for particular measurement systems are considered.