Applying ANN technology to determine acceptable control parameters for the National Library of Scotland’s collections to inform energy efficiency improvements in the UK heritage sector

Abstract

The National Library of Scotland (NLS) uses purpose-built storage enclosures to protect their heritage collections. These enclosures can moderate micro-environmental temperature and humidity fluctuations inside. This study aims to determine an acceptable macro-environment in storage room to inform energy efficiency improvements based on a relaxing macro-environmental control. There are four objectives: 1) to assess the feasibility of using the enclosure’s buffering capacity and to obtain its hygrothermal properties; 2) to determine an acceptable macro-environment; 3) to achieve real-time micro-environment predictions; and 4) to assess potential energy savings from the relaxed control strategy. Correspondingly, the methodology comprises four parts: 1) using laboratory measures to quantify the buffering capacity of an enclosure and associated hygrothermal properties; 2) using a heat, air, and moisture (HAM) transfer model to simulate the hygrothermal interaction between macro- and micro-environments, and using a trial-and-error method with this model simulation to determine the acceptable macro-environment; 3) training a long short-term memory neural network; and 4) using a transform function to create the energy consumption model. The results show that 1) The enclosure’s buffering capacity is feasible to moderate the short-term micro-environmental temperature and RH fluctuations. 2) The acceptable macro-environment was determined to be 33%~65% RH and 15-25 °C control bands with ±16% RH and 5 °C 24 h fluctuations while there is no any detrimental effect on collections. 3) The trained Long Short-term Memory (LSTM) neural network can is robust for real-time prediction of micro-environment. 4) Implementing the relaxed control strategy presents a promising way to achieve the NLS's targeted annual reduction rate of 7.6% over the next decade. In conclusion, this study confirms that relaxed macro-environmental controls, enabled by the enclosure’s buffering capacity, ensure collection safety while achieving significant energy savings. Additionally, this control strategy advances the NLS’s building management toward smarter, energy-efficient control and offers scalable solutions for other heritage institutions.