【正文】 seen from Fig. 4 that entrained air downflows are established in pipes 1, 6 and 14 as expected. However, the entrained airflow in pipe 19 is into the network from the sewer. Initially, as there is only a trickle water flow in pipe 19, the entrained airflow in pipe 19 due to the . discharges already being carried by pipes 1, 6 and 14, is reversed, . up the stack, and contributes to the entrained airflow demand in pipes 1, 6 and 14. The AAV on pipe 12 also contributes but initially this is a small proportion of the required airflow and the AAV flutters in response to local pressure conditions. airflows during appliance discharge.Following the . discharge to stack 4 that establishes a water downflow in pipe 19 from 2s onwards, the reversed airflow initially established diminishes due to the traction applied by the falling water film in that pipe. However, the suction pressures developed in the other three stacks still results in a continuing but reduced reversed airflow in pipe 19. As the water downflow in pipe 19 reaches its maximum value from 3s onwards, the AAV on pipe 12 opens fully and an increased airflow from this source may be identified. The flutter stage is replaced by a fully open period from to s. Fig. 5 illustrates the air pressure profile from the stack base in both stacks 1 and 4 at s into the simulation. The air pressure in stack 4 demonstrates a pressure gradient patible with the reversed airflow mentioned above. The air pressure profile in stack 1 is typical for a stack carrying an annular water downflow and demonstrates the establishment of a positive backpressure due to the water curtain at the base of the stack. pressure profile in stacks 1 and 4 illustrating the pressure gradient driving the reversed airflow in pipe 19. The initial collapsed volume of the PAPA installed on pipe 13 was , with a fully expanded volume of 40l, however due to its small initial volume it may be regarded as collapsed during this phase of the simulation. 7. Surcharge at base of stack 1Fig. 6 indicates a surcharge at the base of stack 1, pipe 1 from to 3s. The entrained airflow in pipe 1 reduces to zero at the stack base and a pressure transient is generated within that stack, Fig. 6. The impact of this transient will also be seen later in a discussion of the trap seal responses for the network. pressure levels within the network during the . discharge phase of the simulation. Note surcharge at base stack 1, pipe 1 at . It will also be seen, Fig. 6, that the predicted pressure at the base of pipes 1, 6 and 14, in the absence of surcharge, conform to that normally expected, namely a small positive back pressure as the entrained air is forced through the water curtain at the base of the stack and into the sewer. In the case of stack 4, pipe 19, the reversed airflow drawn into the stack demonstrates a pressure drop as it traverses the water curtain present at that stack base. The simulation allows the air pressure profiles up stack 1 to be modelled during,and following, the surcharge illustrated in Fig. 6. Fig. 7(a) and (b) illustrate the air pressure profiles in the stack from to s, the increasing and decreasing phases of the transient propagation being presented sequentially. The traces illustrate the propagation of the positive transient up the stack as well as the pressure oscillations derived from the reflection of the transient at the stack termination at the AAV/PAPA junction at the upper end of pipe 11..(a) Sequential air pressure profiles in stack 1 during initial phase of stack base surcharge. (b) Sequential air pressure profiles in stack 1 during final phase of stack base surcharge. 8. Sewer imposed transientsTable 2 illustrates the imposition of a series of sequential sewer transients at the base of each stack. Fig. 8 demonstrates a pattern that indicates the operation of both the PAPA installed on pipe 13 and the selfventing provided by stack interconnection. airflows as a result of sewer imposed pressure transients.As the positive pressure is imposed at the base of pipe 1 at 12s, airflow is driven up stack 1 towards the PAPA connection. However, as the base of the other stacks have not a yet had positive sewer pressure levels imposed, a secondary airflow path is established downwards to the sewer connection in each of stacks 2–4, as shown by the negative airflows in Fig. 8. As the imposed transient abates so the reversed flow reduces and the PAPA discharges air to the network, again demonstrated by the simulation, Fig. 8. This pattern repeats as each of the stacks is subjected to a sewer transient. Fig. 9 illustrates typical air pressure profiles in stacks 1 and 2. The pressure gradient in stack 2 confirms the airflow direction up the stack towards the AAV/PAPA junction. It will be seen that pressure continues to decrease down stack 1 until it recovers, pipes 1 and 3, due to the effect of the continuing waterflow in those pipes.The PAPA installation reacts to the sewer transients by absorbing airflow, Fig. 10. The PAPA will expand until the accumulated air inflow reaches its assumed 40l volume. At that point the PAPA will pressurize and will assist the airflow out of the network via the stacks unaffected by the imposed positive sewer transient. Note that as the sewer transient is applied sequentially from stacks 1–4 this pattern is repeated. The volume of the high level PAPA, together with any others introduced into a more plex network, could be adapted to ensure that no system pressurization occurred. pressure profile in stack 1 and 2 during the sewer imposed transient in stack 2, 15s into the simulation. volume and AAV throughflow during simulation.The effect of sequential transients at each of the stacks is identifiable as the PAPA volume decreases between transients due to the entrained airflow maintained by the residual water flows in each stack. 9. Trap seal oscillation and retentionThe ap