From an environmental protection standpoint, efforts to reduce energy consumption have intensified worldwide. In large commercial buildings, factories, and datacenters given the scale of their energy use, energy management techniques have been actively implemented. Initially, energy consumption reduction efforts were centered on enhancing the efficiency and energy-saving features of individual facilities and managing them in isolation. However, there is now a demand for these efforts to be balanced with the dynamic needs of various operational circumstances. Furthermore, many countries and regions have initiated green certification programs. To address these complexities and further reduce the power consumption of buildings, factories, and data centers, it is essential to achieve not only total optimization of the entire site but also scheduling of individual facility operations. In this case, it is desirable that the operation of individual equipment can be scheduled to reconcile user circumstances, economic considerations, and production activities. This report aims to demonstrate an operational scenario generation utilizing deep learning technique. A transformer architecture is employed as long-term forecast method. The methodology, which generates individual facility operation scenarios from user needs based on the long-term forecast results is described. In addtion, the operation scenario generation techniques will be demonstrated with a simple data set.