Extracted from: Hensen, J.L.M. 1991. "On the Thermal Interaction of Building Structure and Heating and Ventilating System," Doctoral Dissertation, Eindhoven University of Technology.
Introduction
| What ? This problem is an investigation
of the influence of acceleration heating on the behavior of a mechanical
room thermostat.
Why? The objective of the present case study is to see whether the real observations can be repeated - by computer simulation - for a more general case, and to investigate whether decrease of thermostat acceleration heating might be a potential energy conservation strategy. |
 |
Reality Observations...
This case study was inspired by experimental findings from pilot
measurements which were carried out in a relatively small flat with a wet
central heating system that was controlled by a mechanical room thermostat.
Technical considerations led to some modifications of the heating system during the measurements, one of which was disabling the thermostat's acceleration heating (which is used to raise the temperature of the sensor more rapidly towards the switch-off temperature in order to decrease the room air temperature differential).
The acceleration heating is very important with respect to the boiler switch frequency. In this specific case and given the prevailing environmental conditions, the burner cycle time (burner-on till burner-on) was about 90 times longer when the acceleration heating was disabled. The total burner-on time - for an equal period of time - was approximately 50% shorter, suggesting a strong decrease in fuel consumption. It should be noted however, that in this specific case both the number of cycles per hour (at average heating season conditions ~ 30) and the boiler stand-by heat losses may be regarded as well above average. A longer cycle time also has consequences with respect to the fluctuation of the air temperature. Without the acceleration element the fluctuation of the mean room air temperature during one cycle, is much larger.
For the purpose of this case study the problem was divided into building model and plant model. The modeling focused on heating system performance and relative energy consumption. The model was created to represent the dynamic behavior of the heating system during winter conditions. Consequently, the components of the heating system were modelled in detail, while the building description detail was reduced to the minimum acceptable level. ESP-r was used for the modelling and simulation.
The modelled room is facing due south, and is located at ground level of a reference house for energy related research as described in (NOVEM 1990). It represents a typical Dutch, garden-oriented, terraced house.
The room was modeled with a single thermal zone.
| Construction |
Thermal resistance |
|
walls and roof |
2.5 |
|
ground floor |
1.3 |
For the present study, the air temperatures of the spaces adjoining the living room are kept at constant values. See Figure 7.12 (below) for the building and plant model configuration and bounding spaces.
The model uses natural ventilation with infiltration only. The infiltration rate (1.0 air changes per hour) was also kept constant.
A reference profile, as suggested by Van der Laan et al. (1988), is used for casual gains.
The heating system uses forced circulation. The heat source is a gas boiler, with a radiator as the heat emitter. The radiator valve is manually operated. The pump delivers a fixed water flow. The design temperatures of the system are 90/70 C.
For the model the heating system was divided into following elements.
The heating system is controlled centrally by the on-off control of the boiler, which is based on the mechanical room thermostat with acceleration heating.
Fig. 7.12 Schematic representation of a building and plant configuration comprising a living room serviced by (part of) a wet central heating system
The degree of heat input is the primary parameter to be considered in the following results data. To illustrate the influence of the degree of acceleration heating, Figure 7.13 shows simulation results from a two hour period, based on a Dutch climatic reference year for energy research (Bruggen 1978).
Fig 7.13 Influence of acceleration heating on fluctuation of mean living room air temperature and on temperature as sensed by the room thermostat during a two hour simulation period
The simulations were performed for two values of thermostat heat input: 0.05 and 0.10 W. For the given conditions, this gives either approximately 1 or 2 cycles per hour, resulting in air temperature differentials of approximately 1 and 2 K respectively. Figure 7.13 also indicates the set point differential. It may be seen that in the 0.05 W input case, the sensed temperature still rises even after the burner is switched off. This is due to the fact that at those points in time, the room air temperature is actually higher than the thermostat set point. Note that there are two transient factors which play a role in the time lag and damping of the sensed temperature when compared to the room air temperature:
As also demonstrated by Figure 7.13, the resulting average room air temperature is further affected by the degree of acceleration heating; i.e., the average air temperature increases with decreasing acceleration heating.
There is yet another factor which influences the resulting average room air temperature: the thermal load of the system (which affects the length of time the thermostat is switched on).
Figure 7.14 Influence of thermal load on sustained deviation between room air temperature and set point (21.5 C) of the mechanical room thermostat. Note the enlarged top y-axis scaling.
This is clearly demonstrated by Figure 7.14 which shows the room air temperature (and its deviation from the thermostat set point) in relation to the ambient temperature, which is obviously a measure of the thermal load imposed on the heating system. When compared to average climatic conditions for The Netherlands, the data for January 13 represents an extremely cold day, while the data for January 15 represents a fairly average (ambient temperature) winters day. For the simulations presented in Figure 7.14, a heat input to the thermostat sensing element of 0.20 W was assumed, which resulted in burner cycle frequencies of approximately 3, 4, and 5 per hour for the periods around 1:00 and 12:00 on January 13, and around 12:00 on January 15 respectively. This illustrates that the cycle frequency decreases when the thermal load increases. The corresponding relative burner-on periods were approximately 73%, 62%, and 32% respectively.
| To investigate whether the overall gas consumption is also affected, several simulations were performed for various degrees of accelerated heating for the period between January 12 - 15. From the results presented above, it may be clear that when the simulations start from a constant thermostat set point, this would lead to different average room air temperatures. obviously, the results would then be incomparable. Therefore, some of the thermostat set points were chosen (by trial and error) such that the resulting average room air temperature (for January 15) would be equal. The most important simulation results - with respect to the investigated problem - are collected in Table 7.6. When the cycle frequencies and the corresponding air temperature differentials are compared with the thermal comfort criteria, all cases presented in Table 7.6 fall within the comfort limits for transient conditions. Only the cases with the smallest degree of heating acceleration seem to be critical during the extremely cold day; i.e., air temperature differential (= peak-to-peak amplitude) is approximately 3.3 K. |
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
It should be noted however, that the comfort criteria should be applied to the operative temperature, the fluctuation of which is much smaller than the air temperature fluctuation. This is evidenced by Figure 7.15 which shows simulation results for the "0.01 W and O = 21.5 C" case during January 15. From this figure it may also be concluded that in order to create thermally comfortable conditions, the thermostat set point would have to be higher than 21.5 C, because the operative temperature falls below the thermal comfort zone (if we assume normal indoor clothing and nearly sedentary activity).
|
When comparing the gas consumption results for the cases with equal average air temperature, Table 7.6 indeed evidences that it is possible to conserve energy - while maintaining thermally comfortable conditions - by decreasing the burner cycle frequency. Lowering the cycle frequency from 4.5 to 2.0 h-1, results in a gas consumption reduction of only 1% when the whole period is taken into account, but in a 7% reduction when just the "average heating season day" (i.e., January 15) is taken into account. This suggests that the optimal strategy is to apply the "cycle frequency control" strategy selectively; i.e., weather dependent. Obviously, these results require further investigation with respect to what is the optimal strategy (i.e., development of a rules-based system for intelligent controllers), and for which type of systems is it applicable. In the present context, this case study should be regarded only as a demonstration of an application of computer simulation modelling.