Active Jet Noise Control of Turbofan Engine Based on Explicit Model Predictive Control
Round 1
Reviewer 1 Report
The article deals with the current topic of reducing the noise of aircraft jet engines. Unlike traditional noise reduction technologies, which are mostly based on a passive approach, the authors of the article have developed an active jet noise control system. An integrated model of turbofan engine and jet noise was created to calculate the engine parameters and jet noise in real time. The individual modules of this computational model are described in more detail, including references to the relevant literature.
After a detailed study of the article, I can state that I have no fundamental reservations or comments.
In point 3, on line 120, the authors mention comparison experiments. Here, it would be useful to clarify whether these are only "computer experiments" or whether it was possible to measure some quantities on a specific aircraft jet engine.
Even in terms of form, the text is well structured in accordance with the template. Only minor corrections can be recommended, e.g. in some places (lines 330, 340, 372) figures are quoted "Fig." instead of "Figure".
Numerous abbreviations are used repeatedly in the text, which are explained when used for the first time. For readers, however, it may be appropriate to add a list of abbreviations used at the end of the article.
Author Response
Much thanks for spending your valuable time reviewing our work and your praise of our research has greatly inspired us. We have studied the comments carefully and made some corrections which we hope to meet with approval. Here are the responses. Point 1: In point 3, on line 120, the authors mention comparison experiments. Here, it would be useful to clarify whether these are only "computer experiments" or whether it was possible to measure some quantities on a specific aircraft jet engine. Response 1: Thank you very much for your suggestions. The comparison experiments mentioned in point 3, on line 120 are just numerical simulations, which are carried out on computer. The main goal of these experiments is to verify the effectiveness of the designed controller. The future study will focus on hardware-in-loop simulations. Point 2: Even in terms of form, the text is well structured in accordance with the template. Only minor corrections can be recommended, e.g. in some places (lines 330, 340, 372) figures are quoted "Fig." instead of "Figure". Response 2: We are sorry for our writing mistake.The minor corrections mentioned in Point 2 have been implemented. In the article, all figures have been quoted as “Figure”. The form of this article is also checked in detail. Point 3: Numerous abbreviations are used repeatedly in the text, which are explained when used for the first time. For readers, however, it may be appropriate to add a list of abbreviations used at the end of the article. Response 3: Appendix A which contains the explanation of all abbreviations, has been added to the end of this article. More modifications can be found in the revised manuscript.Author Response File: Author Response.docx
Reviewer 2 Report
This paper proposes an active jet noise controller of turbofan engine based on explicit model predictive control (EMPC). An integrated model of a turbofan engine and jet noise, which can calculate the engine parameters and jet noise in real time, is claimed to be established. The online computational burden of MPC is transferred to offline computation by multi-parametric quadratic programming (MPQP).
The proposed method itself is quite interesting. A model for SPL noise prediction is used, but no clear discussion regarding the motivation of this choice is presented. The full set of equations is not discussed in detail, too. I think that to be important in order to better understand what is going on. Therefore, the authors should discuss advantages and limitations of the chosen models in more detail.
The description in section 3.2 is difficult to follow, too, as the authors try to set up everything very briefly.
The model is later tested in Matlab and it is shown to produce a desired solution, but I would like to have a comparison with experimental data of some sort, as we can't know whether this model will work and how it would perform in reality.
How would a real jet noise control process happen? It would be nice if a situation could be described how one would proceed to control SPL in real-time for real jet engines. How fast would the controller need to act and achieve desired results to be judged successful, as flight conditions change all the time when air density changes, turbulence levels increase, flaps are extended, jets interact with flaps, etc.?
Noise levels are usually irrelevant in cruise flight but important for landing and take-off. Is the situation depicted in the paper realistic?
To summarize, the paper needs more clarity and context to be able to judge whether the proposed method has clear benefits and provides a mpc which is able to really do mpc in real time.
Some other remarks:
- I don't see how Table 3 and Figure 8 fit together.
- Oftentimes a relative change is shown, in SPL, for example (e.g. Fig. 7), but the caption refers to the quantity itself.
- Language needs improvement. Oftentimes a new sentence is not started with a capital letter. Words are missing, etc.
Author Response
Much thanks for spending your valuable time reviewing our work and your praise of our research has greatly inspired us. We have studied the comments carefully and made some corrections which we hope to meet with approval. Here are the responses.
Point 1: The proposed method itself is quite interesting. A model for SPL noise prediction is used, but no clear discussion regarding the motivation of this choice is presented. The full set of equations is not discussed in detail, too. I think that to be important in order to better understand what is going on. Therefore, the authors should discuss advantages and limitations of the chosen models in more detail.
Response 1: Thank you very much for pointing out the shortcomings of this article. The semiempirical Model of Jet Noise which uses the calculation method ST2JET, is introduced in this article. However, the description of the calculation method of jet noise is really incomplete. To further illustrate the advantages of ST2JET, the detailed discussion that compares three semiempirical calculation methods is added in the first paragraph in Section 2.1. Besides, a more detailed derivation process of the calculation formula of sound pressure level is also added.
Point 2: The description in section 3.2 is difficult to follow, too, as the authors try to set up everything very briefly.
Response 2: Thank you very much for your suggestions. We are sorry for our mistake. Section 3.2 is rewritten and checked in detail to ensure that it is reasonable. The goal of section 3.2 is to introduce the construction of an explicit MPC model. We derive the MPQP problem based on the linear discrete state-space model and the linear quadratic objective function. For the MPC algorithm, this is the problem that needs to be solved online by optimization algorithms. However, it can be solved offline and be located online for the explicit MPC algorithm. The way to achieve the above goals is introduced in section 3.3.
Point 3: The model is later tested in Matlab and it is shown to produce a desired solution, but I would like to have a comparison with experimental data of some sort, as we can't know whether this model will work and how it would perform in reality.
Response 3: Thank you very much for your suggestions. We don’t mention the verification of the model, especially of the semiempirical model. It is unconvincing. Therefore, we carry out the comparison verification simulation with NASA experimental data on the basis of the designed model software. The corresponding simulation results are added in section 2.1. The calculation accuracy of the model is acceptable. In the following research, we will establish a platform for jet noise measurement based on the scale model of the nozzle of turbofan engines. The goal is to further modify the semiempirical model and improve its performance in reality.
Point 4: How would a real jet noise control process happen? It would be nice if a situation could be described how one would proceed to control SPL in real-time for real jet engines. How fast would the controller need to act and achieve desired results to be judged successful, as flight conditions change all the time when air density changes, turbulence levels increase, flaps are extended, jets interact with flaps, etc.?
Response 4: Thank you very much for your suggestions. The goal of active noise reduction is to reduce the noise level without affecting the engine’s performance as much as possible. At present, a real active jet noise control process is achieved by microjet technology and guide blade technology. The microjet technology reduces jet noise by injecting air or water into the tail jet flow field. The guide blade technology changes the condition of the bypass jet flow field, shields the strong noise generated by the central gas in a specific direction, and reduces the noise radiation in that direction. Both two technologies are based on hardware components. Therefore, compared with control simulations, a real jet noise control process must consider more factors, such as the rate of actuators, the rate of valve opening, and so on. Because the main control cycle is 20ms in the aero-engine field, the controller must finish the control step in 20ms, otherwise, it is unsuccessful. Through the simulation in this paper, the mean consuming time of MPC is more than 20ms at each step which is unacceptable. This also illustrates the complexity of three-variable control. This is why we choose to study the explicit MPC controller. For the controller, control quality is also one of the criteria to judge whether it is successful, such as overshoot, regulation time, and steady error. No matter how the external environment changes, the controller has to be able to keep track of the target commands in time. The change in the external environment can be regarded as external disturbances, which test the controller's robustness.
Point 5: Noise levels are usually irrelevant in cruise flight but important for landing and take-off. Is the situation depicted in the paper realistic?
Response 5: Thank you very much for your suggestion. Actually, noise levels are usually irrelevant in cruise flight but important for landing and take-off. We conduct the simulation at sea level, which is the altitude of take-off. The simulation includes two parts. One is to determine the best prediction horizon by step simulation. The other is to compare the performance of different controllers by continuous step simulation. In general, the simulation in this paper is to prove the good performance of explicit MPC controller based on binary search tree, especially the real-time performance. Of course, the superiority is relative to the other two controllers, which are the MPC controller and explicit MPC controller based on sequence search. As for the specific flight segments, namely landing and take-off, we can use the designed explicit MPC controller to hold the level of thrust and ensure that the sound pressure level does not exceed the limit. However, it needs to combine the take-off data or landing data about turbofan engines. In the following study, we will focus on the practical application of the designed explicit MPC controller, and carry out hardware-in-loop and semi-physical simulation.
Point 6: I don't see how Table 3 and Figure 8 fit together
Response 6: We are sorry for this writing mistake and the maximum of MPC is corrected as 261.33ms.
Point 7: Oftentimes a relative change is shown, in SPL, for example (e.g. Fig. 7), but the caption refers to the quantity itself
Response 7: Sorry for our mistakes and we change the captions in Figure 5 and Figure 7.
Point 8: Language needs improvement. Oftentimes a new sentence is not started with a capital letter. Words are missing, etc.
Response 8: Thank you very much for your suggestion. We examine the whole paper and improve the language.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The paper is now improved enough to warrant publication. Maybe a bit of language fine tuning could be done.