Abstract
In this study, the error of a conventional simple pump model that approximates the resistance of a piping network using a linear or quadratic equation of flow rate was evaluated. For this purpose, a new water circuit model that can easily construct a parallel flow network of free depth without convergence calculation was developed. As shown in Fig. 7, due to a non-uniform distribution of load, it is possible that an error of several tenths in energy prediction should occur in the conventional simple pump model. The magnitude of this error could be higher than the magnitude of the error that is influenced by various parameters, such as the pressure and flow rates of the pump and the pump efficiency, which many previous traditional programs have attempted to precisely represent. Therefore, the resistance curve must be set appropriately by capturing the load distribution characteristics of the HVAC (Heating, Ventilation, and Air Conditioning) system. By considering the heat load characteristics that many conventional heat load calculation programs can express, such as the difference in orientation or the difference between the perimeter and the interior zone, the annual operating point of the pump is predicted, as shown in Fig. 8. In the conventional simple pump model, if the resistance of the water pipe network is represented by a quadratic equation of flow rate, the energy prediction obtained is too small; if the resistance is expressed by a linear equation, the energy prediction obtained is excessive. Compared with the calculation result of the detailed water circuit model developed in this study, the annual error rate of conventional simple pump model was 30~60%, which is not small. For the pump model considered in this study, many operating points were lined up between the two assumed resistance curves represented by linear and quadratic equations. If the resistance of the water pipe network is proportional to the 1.5th power of the flow rate, the error is less than 10%. The chilled and heating water network to be calculated in this study is not complex. It must be confirmed whether characteristic of being proportional to the 1.5th power of the flow rate can also be applied to a complex large-scale building, where there is an extensive merging and branching of pipes. As a result of integrating the model for heat generation due to tenant activities, and the variations in air conditioning operation times, into the aforementioned heat load calculation model, the annual operation point of the pump changed, as shown in Fig. 10. Even when the load on a building as a whole was low, the required minimum differential pressure was high because the load on a specific tenant was large. Thus, the resistance curve shown in Fig. 11 was obtained. Compared to the results of the simulation where the uncertainty of tenants was not considered (Fig. 8), the resistance, according to solid line in Fig. 11, tended to increase at low loads. In actual tenant office buildings, there are uncertainties that cannot be expressed in the model; therefore, the tendency to increase in case of low load is expected to further increase.
