Experimental and Numerical Analysis of a Silica Gel Packed Bed for Passive Humidity Control in Museum Rooms

Abstract

In this thesis the passive humidity control of indoor air is laid under review. This is primarily interesting for museum rooms, where strict requirements apply to the indoor air humidity. Passive humidity control is promising to contribute to stabilizing the relative humidity levels in a room.

Besides, the use of passive humidity control can potentially reduce energy demand for (de)-humidification, and reduce HVAC dimensions as a consequence. This work has zoomed in on the (de)-humidification of air lead through a packed bed with silica gel beads. Silica gel is very promising to use as a buffer material for humidity control as a result of its high hygroscopic capacity, or slope of sorption isotherm.

The work started with uptake rate measurements on different silica gel samples, referred to as Experiment 1. This has yielded adsorption and desorption isotherm between 20 and 80 %RH. The obtained isotherms provide insight in the sorption capacity of different samples, as well as the presence of sorption hysteresis.

Samples with promising sorption and desorption behaviour were selected to include in an experiment with a packed bed. In this experiment, referred to as Experiment 2, silica gel was contained in a packed bed (column) and exposed to cyclic input: alternating high and low levels of relative humidity, using a step function. Both short (1h/1h) as well as long (8h/16h) cycles are executed in the experiment. In both types of experiments, significant dehumidification and humidification of air in the packed bed was measured.

Next to experiments, a numerical model of the packed bed is developed. The results of this model are compared to experimental data from literature and to the experimental data obtained in Experiment 2. The model proves to be able to simulate the course of both temperature and humidity of outlet air exiting the packed bed in a reliable way. The main discrepancy between model and experimental results is found in the response time of the model: it reacts faster to changing inlet compared to measurements. Furthermore the model is sensitive to the inputted isotherm polynomial, RH(w). This polynomial describes the equilibrium of gaseous water in the air and adsorbed liquid water in the silica gel. This equilibrium applies at the interface of air and silica gel surface.

Experiment 2 has shown silica gel is able to both adsorb and desorb water in the humidity range of 40 to 60 %RH. An important observation in the long runs of Experiment 2 is that more water is adsorbed than desorbed in the Ąrst cycles. Two issues can be mentioned to explain this: the samples can experience ŠprimaryŠ hysteresis: some water is permanently retained in silica gel during the Ąrst adsorption/desorption cycle. This is a plausible explanation for the difference in primary and secondary isotherms. The other potential explanation is sway-in behaviour: it takes some time before the effect of initial conditions is eliminated in a cyclic experiment. In this case, it is possible that some water is stored in the deeper layers of a silica bead. The moisture uptake and release by silica gel converges to an equal value.

The performance in buffering moisture is expressed in the Moisture Buffering Value (MBV) in [g/m3/%RH]. This MBVτ accounts for the time period of the expected humidity fluctuations. The maximum, or theoretical, MBV∞ is derived from the equilibrium isotherm. For a given time period of the expected humidity fluctuations, the part of the hygroscopic capacity of silica gel that is effectively used is expressed as:
ξeffective = (MBVτ / MBV∞) · ξtheoretical (1)

The MBVτ is determind based on experimental results in this project. The numerical PGC model can serve as a tool to simulate MBV (and ξeffective) for different geometries of the silica gel container and different humidity input.

Concluding, silica gel is well able to reduce Ćuctuations in relative humidity of indoor air. It is important that the silica gel is well reachable by indoor air. Uniform utilization of the silica gel will increase the effective part of the (theoretical) hygroscopic capacity, ξeffective in [kg/m3], of the silica gel. It is best to avoid long lengths of silica, since the effectiveness of ŠupperŠ silica gel is reduced.

Relative humidity fluctuations can occur either due to changes in absolute humidity (hygric loads) or due to temperature Ćuctuations in a room. If relative humidity fluctuations are due to changes in absolute humidity, performance of the packed bed could be improved by cooling the silica during adsorption (dehumidification of air) and heating during desorption (humidification of air). Energy demand for temperature control of the silica gel should be limited, to not mitigate the profits in energy demand for passive humidity control.