2.1 Introduction to AEPI and CFD technology

In Figure 1, two different cleanliness control methods are shown: the solid line represents active mode and the dotted line represents passive mode. Different from the traditional cleanliness control method of AIPE (Figure 1(a)), the AEPI is a combination of active exhaust and passive intake (Figure 1(b)). Outlets are usually connected to a negative pressure device (such as a high-power exhaust fan) and inlets are connected to atmospheric gas (such as clean nitrogen or clean air). Combined with the optimization design of inlets and outlets, the negative pressure pumping (exhaust) actively guides the direction of gas flow through the amplifier cavity. This greatly reduces the probability of turbulence generation and improves the clean purge efficiency. In addition, owing to the active exhaust (pumping), the pressure in the amplifier cavity is slightly lower than normal pressure. Compared with the traditional active intake of high pressure and passive exhaust, the requirement for sealing of the amplifier cavity in the AEPI method is lower. With the presence of negative pressure, the amplifier is better sealed (self-pressing principle).

Figure 1 Cleanliness control methods: (a) AIPE; (b) AEPI.

CFD^[Reference Li, Zhu, Pang, Tao, Jiao and Wu^13, Reference Wang^14] is widely used in the numerical simulation of gas flow. First, it discretizes the hydrodynamic equations followed by the air flow in the calculation domain, transforms the strongly nonlinear partial differential equations into algebraic equations, and then uses certain numerical calculation technology to solve them, so as to obtain detailed information of the flow field distribution in the whole calculation region. Finally, the results can be visualized by computer graphics technology. Generally speaking, CFD usually includes the following steps: establishing mathematical and physical model (pre-processing), numerical algorithm solving, and result visualization (post-processing).

2.2 Numerical calculation model and governing equations

Taking one type of single-aperture slab amplifier (SSA) in the National Laboratory of High Power Laser and Physics (NLHPLP) as an example, half of the amplifier cavity was chosen as the model for checking numerical calculations (Figure 2), because the structures on both sides of neodymium glass (Nd:glass) are the same. The model mainly includes inlets, outlets, and Nd:glass. The air inlets are composed of a row of small holes with a diameter of 10 mm, and the air outlets at the opposite position of the inlets are composed of three elongated holes with a width of 10 mm. The method of combining Tet/Hybrid element with TGrid is used for mesh generation. To obtain suitable mesh generation, three kinds of mesh numbers are divided according to the size of the grid. The mesh densities of inlets and outlets are optimized locally. The simulation results show that the internal pressure and flow state of the amplifier are basically the same under the three grid sizes, so to save calculation time, the final mesh number of 1,357,773 was chosen.

Figure 2 (a) The configuration of SSA; (b) the calculation model; (c) the mesh generation.

Similarly, to simplify the calculation model, we assumed that: (1) the thermal characteristics of the fluid are stable; (2) the flow state of the fluid is determined by the Reynolds number at the suction port speed; (3) the aerosol particles in the amplifier cavity are small enough and light enough to flow with clean gas. Therefore, the flow field can represent the distribution of aerosol particles.

The basic equations of fluid analysis and turbulent flow were used as the control equations^[Reference Wang^14]. These basic equations include the equations for the conservation of mass, momentum, and energy. The turbulent flow model is the RNG k-ε model. The default values of the software were used for the relevant parameters.

As it is difficult to guarantee the same speed of the air outlets during the experimental verification process, we separately added the air chamber under and connecting with outlets on the basis of the original model to strictly approximate and facilitate real experimental verification. The newly improved experimental model is shown in Figure 3. The related parameter settings and process are similar to the previous model.

Figure 3 (a) The modified model; (b) the mesh generation of the modified model.

2.3 Numerical model verification

To verify the correctness of the numerical calculation model, the modified SSA cavity structure shown in Figure 4 is adopted. It is strictly the same as the modified theoretical model in Figure 3, including inlets, outlets, and neodymium glass cavity. To allow the flow field in the cavity to be seen, PVC was used instead of neodymium glass. The inner wall of the cavity is covered with black paper. The experiment was carried out in a dark room.

Figure 4 The modified SSA cavity.

The suction port is connected to a high-power fan^[15] (model MF-150P), which can realize a pumping speed of 360–530 m³/h. The suction port is connected to the air chamber, which in turn is connected with outlets. When the exhaust fan starts to work, an active air extraction speed of 20–30 m/s at the outlets can be realized (20 m/s was selected in this experiment). The air inlets connected to the external atmosphere consist of a series of 10 mm diameter holes evenly distributed along the edge. A pressure gauge is placed in the modified SSA cavity to measure the internal pressure. Three smoke cakes^[16] were placed near air inlets, and the smoke they produced accompanied the air into the amplifier cavity. By monitoring the fluid state of smoke in the amplifier cavity, the flow field in the amplifier is characterized, and then the pollutant removal efficiency is reflected^[Reference Ren, Zhu, Liu and Yang^12]. The steady-state flow field in the amplifier cavity was recorded. The correctness of the theoretical simulation is verified by comparing the experiment with the theory.

3.1 Theoretical simulation and experimental results

Using the AEPI method, the intake air speed is 20 m/s. The theoretical and experimental results obtained are shown in Figures 5 and 6. The pressures in the figures are relative to standard atmospheric pressure. It can be seen from Figures 5(a) and 5(b) that under the experimental parameter conditions, the pressure within the entire modified SSA cavity is between –2000 and 2000 Pa (relative to standard atmospheric pressure). In Figure 5(c), the experimental results are 0 Pa in the position of the bottom surface of the cavity and –113 Pa in the position of outlets, which are relatively consistent with the theoretical results.

Figure 5 Pressure in the modified SSA cavity: (a) pressure of the whole cavity; (b) pressures in the positions of Z = 0, 0.1, 0.2, 0.3 and 0.4 m; (c) experimental result.

Figure 6 The flow field in the modified SSA cavity in steady state: (a) contour of velocity; (b) slices of velocities in different positions; (c) vector of velocity; (d) experimental result.

Therefore, based on the above experimental and theoretical results, we believe that the CFD model is correct and the internal pressure of the amplifier cavity obtained by AEPI method is very small. To minimize turbulence, it is necessary to further optimize the SSA cavity structure and air exhaust velocity.

3.2 Design optimization

For the SSA amplifier, to achieve the goal minimal turbulence, some optimization designs of the inlets, outlets, and speeds of air exhaust have been made. We set up a series of circular inlets and rectangular outlets along the other two sides of the triangle, with different air exhaust speeds (20, 40, and 80 m/s). The final results are shown in Figure 7. When the speed is 20 or 40 m/s, two huge eddies are generated along the two acute angles of the triangle, so it is hard to remove pollutants. When the speed reaches 80 m/s, there is only a small vortex, and then the SSA cavity can be quickly cleaned. The pressure in the cavity is very low. Therefore, using the amplifier cavity structure shown in Figure 2 and an exhaust speed of 80 m/s, the cleanliness of the SSA cavity can be controlled effectively.

Figure 7 Vectors of different pump speeds and pressures: (a) 20 m/s; (b) 40 m/s; (c) 80 m/s; (d) 80 m/s.

In addition to SSAs, the method of AEPI is also applicable to multi-segment disk amplifiers (MSAs). Taking one type of MSA as an example, its structure is shown in Figure 8. Similarly, we choose the 1/4 structure. The simplified figure is shown in Figure 8(b). A series of circular inlets and outlets are provided along the three sides of the triangle. The exhaust speeds used are 10, 20, 40, and 80 m/s. The final results are shown in Figures 9 and 10. The results show that under the condition of exhaust speeds of 10 or 20 m/s, two larger vortexes are generated in the direction of the two acute angles close to the triangle. As the exhaust speed increases (40 and 80 m/s), the vortex gradually decreases and disappears. Although Figure 9(d) produces a vortex in the direction close to the air inlet, the vortex is small, and the vortex will instead bring out the pollutants in the sharp corners, so cleanliness control can also be well achieved. In addition, at exhaust speeds of 40 and 80 m/s, the pressures in the amplifier cavity are also small. Therefore, MSA can achieve cleanliness control by adopting the AEPI method and reasonable structure design.

Figure 8 (a) The configuration of MSA; (b) the calculation model.

Figure 9 Vectors at different air speeds in the outlets: (a) 10 m/s; (b) 20 m/s; (c) 40 m/s; (d) 80 m/s.

Figure 10 Pressures at different air speeds in the outlets: (a) 40 m/s; (b) 40 m/s; (c) 80 m/s; (d) 80 m/s.