Introduction
. Effective management of building engineering systems is a key challenge in modern construction and maintenance industry. The complexity of integrating individual subsystems, such as power supply, heating, gas supply, and ventilation, calls for implementation of new approaches to analyzing and predicting their lifecycles. As the demand for energy efficiency is growing, reliability, and resource conservation, development of intelligent control methods that minimize the risk of failure and optimize maintenance costs is of particular relevance. In spite of the considerable progress attained in automation and digitalization, existing solutions typically fail to provide a comprehensive assessment of the condition of engineering systems at all stages of their lifecycle. The aim of the study is to develop and test an intelligent model for managing the lifecycle of building engineering systems using artificial intelligence methods in order to improve operating efficiency of capital construction projects.
Materials and Methods
. The object of the study are building engineering systems examined within the context of their lifecycle — from design to decommissioning. Lifecycle analysis of complex engineering systems, as well as machine learning and artificial intelligence methods, served as the methodological framework. In order to achieve these aims, a "Lifecycle of Engineering Systems" subsystem was developed as part of the energy resource management system. Time series analysis methods were applied, and a recurrent neural network with Long Short-Term Memory cells was implemented to predict the remaining service life and failure probability. Data on energy consumption and engineering system performance parameters were compiled based on real-world data. Model training and validation were performed by means of standard machine learning procedures, including data normalization and division into training, validation, and test sets.
Research Results
. Throughout the course of the project, an intelligent model was developed and tested for predicting operating costs, failure probabilities, and optimal upgrade times for utility systems. The use of a neural network ensured high forecasting accuracy (the determination coefficient was 0.9873 on the control sample), as confirmed by a comparison of the actual and predicted failure rates. Implementation of the suggested subsystem automated the process of analyzing the technical condition of utility systems, to identify potential failures in a timely manner, and to schedule preventative maintenance. The economic impact of the implementation amounted to over 200 million rubles per year. Discussion and
Conclusion
. The results demonstrate the high effectiveness of the suggested approach to intelligent lifecycle management of engineering systems. The developed model not only improves forecasting accuracy but also helps optimize operating costs, reduce the risk of accidents, and extend the service life of equipment. The theoretical significance of the work lies in the development of a methodology for comprehensive lifecycle analysis of engineering systems by means of artificial intelligence. The practical value lies in the scalability of the suggested solutions for various types of buildings and infrastructure facilities. Prospects for further research include expanding the functionality of a system, integrating additional data sources, and adapting the models for other industries.