Learning and Control of Energy Consumption in Large Company

Digital control of energy consumption in a large corporation is considered on the base of concept of INDUSTRIE 4.0. Along with the use of big data and machine learning, the human factor takes into account. The problem of data mining is in identification of stochastic energy-conservation capabilities of plants included in a large company by machine learning. The asymmetric awareness of the company governing body and the staff of these plants about energy conservation capabilities is supposed. Using own awareness, the visionary staff of the plants can choose energy consumption in such a way as to achieve own goals. Such activity of staff can lead to non-use of existing capabilities of reducing energy consumption of the company. In addition, assessments of the energy consumption of plants obtained using standard procedures processing big data, such as learning identification procedures, are being distorted. The mechanisms of the digital control of the energy consumption are proposed that regulate the behavior and interaction of the governing body and plants in the face of uncertainty. The results of their operation are digital assessments, standards, and inducements to reduce energy consumption in plants. Sufficient conditions have been found for the synthesis of such mechanisms of the digital control that provide identification of energy consumption capabilities of company plants. These conditions are illustrated by the example of digital control of plants repairing locomotives and wagons during the implementation of the Energy Conservation and Efficiency Program of the large company Russian Railways.