The speakers were set at both areas B. In the SYSNOISE, the distribution of the temperature and the gas flow in the pipe could not be considered.
At first, the influence of those was investigated through experiments. Figure 4 shows the comparison of the acoustic frequency response functions, identified by the M-Sequence method, between the point B and point Q in figure 3. In figure 4, the solid line shows the acoustic frequency response function at room temperature without gas flow. The broken line shows the acoustic frequency response function, in which frequencies were calibrated by the difference in exhaust temperature, on the engine operating condition. Both show good symmetry. Therefore, we could assume that:
1) The velocity of the exhaust gas flow was about 30 to 40m/s. It was one-tenth less than the sound velocity. In this case, the gas flow was negligible.
2) The sound velocity of the forward wave differed from that of the backward wave slightly by the gas flow. As a result, the differences between the peak and dip were decreased.
3) The temperatures at arbitrary points in the pipe were almost the same. Therefore, the BEM could be applied for detailed analysis.
3. Acoustical Optimization
The position of the speaker affects the performance of the noise reduction by the active exhaust noise cancellation system (AENC) significantly. Considering the durability of the speaker, it was better to locate the speaker outside the exhaust pipe5). This was due to the conditions of high temperature and high humidity inside the pipe. On the other hand, with the system where the speaker was positioned outside the exhaust pipe, the noise reduction was different at selected measurement points6). Consequently, the speaker was positioned inside the pipe. In order to protect the speaker, the branch pipe, which is called the secondary speaker pipe, was installed as shown in Fig.3. Next, the speaker was installed at the end of the branch pipe. Inside the exhaust pipe, there are many resonance and anti-resonance modes, caused by wave reflection at the bent parts and enlarged-shrink parts of the cross section. It is well-known that the exhaust noise has the line spectrum. If the frequency of the dip of the secondary noise source corresponds to the frequency of the major order of the exhaust noise, then enormous speaker power would be required. Therefore, in order to reduce the exhaust noise efficiently and minimize the occupied space of the system, it was very important that the acoustic behavior was analyzed in detail and optimized.
3.1 AENC for Main Engine
In order to reduce the exhaust noise of the main engine, such as the variable speed engine, by the AENC, the secondary noise level should be larger than the exhaust noise level under all engine speed conditions. For discussing the arrangement of pipes, the "Acoustic Ratio" is defined as follows:
1) Calculation of Exhaust Noise Lex(f): the sound pressure level at Point P shown in Fig.3 can be calculated by SYSNOISE, when the unit particular velocity (1m/s) is acted on aria A.
2) Calculation of Secondary Noise Lsp(f): the sound pressure level at Point P can be calculated, when the unit particular velocity (1m/s) is acted on aria B shown in Fig.3.
3) Calculation of Acoustic Ratio η(f): According to Eq(1), the acoustic ratio should be calculated.
η(f) ＝ LSP(f)/ Lex(f) ... (1)
By having a flat acoustic ratio for the frequency axis and then enlarging it, the capacity of the speaker can be minimized.
3.1.1 Investigation of Mixing of Sound
The number of loud speakers was considered to be two. For the mixing mode of the acoustic wave, as shown in Fig.5, the acoustic ratios were calculated as shown in Fig.6.
1) Model A: There is peak of acoustic ratio at 80Hz, in other words, the efficiency of the system is increased. The reason for this is that the secondary speaker pipe works as a resonator. On the other hand, the efficiency decreases at other frequency ranges. This means that this model is not adequate for variable speed engines.
2) Model B: The acoustic ratio has a high level, that is of 350Hz. Although there are dips at each 45Hz, the shape of the acoustic ratio frequency is mostly flat. Hence, it is possible to apply to variable speed engines.
