Project plan. Project #14 Estimation of nitrogen oxide (NOx) emissions in marine diesel engines

Transcription

1 Aalto University ELEC-E8004 Project work course Year 2018 Project plan Project #14 Estimation of nitrogen oxide (NOx) emissions in marine diesel engines Date: Student 1 Mikael Helin Student 2 Turkka Helin Student 3 Sauli Sjögren Page 1 of 12

2 Information page Students Sauli Sjögren Turkka Helin Mikael Helin Project manager Turkka Helin Official Instructor Kai Zenger Other advisors Nguyen Khac Hoang Starting date Approval The Instructor has accepted the final version of this document Date: Page 2 of 12

3 1) Background NOx stands for nitrogen oxides that pollute air and accelerate global warming. NOx gases are nitric oxide (NO) and nitrogen dioxide (NO2). These gases contribute to the formation of acid rain, as well as tropospheric ozone. NOx gases are usually produced from the reaction among nitrogen and oxygen during combustion of fuels. Both NOx and CO2 are formed as combustion engines exhaust gasses. NOx gases are considered to be about 300 times more harmful to the environment than CO2 [5]. In addition, NOx gases are formed always while CO2 exhaust gases can be minimized with clean combustion. Recent studies have shown that NOx gases have negative impacts on human health and on the ecosystem. On humans NOx mainly affects respiratory systems causing inflammation, and on long term the exposure can decrease lung function and increase the risk of respiratory conditions [1]. Also on lower atmosphere level NOx emissions increase the formation of fine particles (PM) that are hazardous for humans. NOx emissions also constitute a part of the greenhouse effect causing an impact on climate as NOx is a catalyst in formation of ozone on the troposphere [2]. It has been estimated that diesel engines in ships contribute around % of the worlds NOx emissions that are caused by fossil fueled sources [3]. However, the current trend is that the emissions caused by ships is to be minimized in the upcoming years. Currently there are no replacement for diesel engines in ships, but more strict regulations are formed with respect to the emissions. To meet these regulations new control strategies and less pollutant motors have to be developed. US academic research showed that pollution from the world's 90,000 cargo ships leads to 60,000 deaths a year and costs up to 330 billion dollars($) per year in health costs from lung and heart diseases. And these figures are only from the USA. [1] Current problem in diesel engines is not that they do not run efficiently enough, but that the way they are operated is not minimizing the emissions. This is also due to the fact that the emissions are not easy to measure because of the limitations that current sensors have and because there are no cost-efficient sensors available [4]. Therefore, it would be ideal to have a virtual NOx sensor that can accurately enough estimate the emissions without too much computation. With the assist of a virtual NOx sensor a optimal motor control could be achieved, using the emissions as a constraint for operating the motor. The first implementation of the estimation algorithm will be heavily based on the methods introduced in the article [4]. [1] Kampa, M & Castanas, E. Human health effects of air pollution. Environmental Pollution [2] [3] Lack, D et al. Particulate emissions from commercial shipping: Chemical, physical, and optical properties. Journal of geophysical research, Vol [4] Formentin, S et al. NOx Estimation in Diesel Engines via In-Cylinder Pressure Measurement. IEEE Transactions on control systems technology, vol. 22, no [5] Page 3 of 12

4 2) Expected output The goal is to find an algorithm for evaluating NOx emissions of large marine diesel engines. An approach will be used that utilizes Principal Component Analysis (PCA) and static maps with two parameters: in-cylinder pressure measurement combined with engine speed. After the estimation algorithm works in simulation environment, the algorithm will be tested with a real engine. Expected users for this estimation software would be large maritime diesel engine manufacturers. If the research implementation works, a startup company could be founded. 3) Phases of project The phases of the project are the following: Project planning phase Literature review phase Business aspects phase Algorithm development phase (iterative) Algorithm validation phase (iterative) Final gala preparation phase Final document writing phase As the project is mainly developing a working algorithm for the NOx estimation, the development and validation phases are iterative. Milestones of the project are following: M1 Project plan 29.1 M2 Literature review completed 18.2 M3 Slides for business aspects 2.3 M4 Business aspects presentation 6.3 M5 Business aspects document 9.3 M6 First working algorithm 18.3 M7 Final version of the algorithm 24.4 M8 Final gala poster 8.5 M9 Final gala 15.5 M10 Final report 28.5 Page 4 of 12

5 4) Work breakdown structure (WBS) Research background information about diesel engine 1 wk Project plan 3 wk Literature review on the topic 2 wk Learn LabView, implement labview NOx measurement block 1 wk Learn GT Suite, engine simulation 1 wk Learn how to apply PCA to a software project 0.5 wk PCA for test data 0.5 wk Receive previous test data: in-cylinder pressure {PCA}, engine speed, NOx-data Business aspects (3wk) Business document 1.5 wk Preparation for presentation 1 wk Other 0.5 wk Design an estimation algorithm 8wk Testing the algorithm (2 wk) Test data 0.5 wk Simulation (GT-suite) 0.5 wk Real engine (AGCO) 1 wk Documentation of the implementation wk Documentation of testing the algorithm wk Iterate and tune algorithm 4 wk Final gala 2 wk Poster 0.5 wk Final report 2 wk Page 5 of 12

6 5) Work packages and Tasks of the project and Schedule 5.1) Work packages WP 1 Project management Turkka WP 2 Labview module Turkka WP 3 Business aspects Sauli WP 4 Algorithm design Mikael WP 5 Algorithm testing Sauli WP 6 Documentation Turkka WP 7 Final gala Sauli 5.2) Tasks T1 Learn LabView T1.1 Implement LabView NOx measurement block 25 h T2 Learn GT Suite T2.1 Engine simulation using GT Suite 20 h T3 PCA for test data 20 h T3.1 Receive previous test data: in-cylinder pressure {PCA}, engine speed, NOx-data T3.2 How to apply PCA to a software project T4 Business plan T4.1 Business document 30 h T4.2 Preparation for presentation 20 h T4.3 Other 10 h T5 Design an estimation algorithm 250 h T6 Testing the algorithm T6.1 Test data 5 h T6.2 Simulation (GT-suite) 15 h T6.3 Real engine (AGCO) 30 h T7 Documentation of the implementation 10 h T8 Documentation of testing the algorithm 10 h T9 Iterate and tune algorithm 120 h T10 Final gala 30 h T10.1 Poster 10 h T11 Final report 60 h Total hours: 665 Page 6 of 12

7 5.3) Detailed schedule Page 7 of 12

8 6) Work resources 6.1) Personal availability during the project Table 1. Number of hours available for the project (excluding lectures and seminars) per week. Student Sauli Student Mikael Student Turkka Week Week Week Week Week Week Week Week Week Week Week Week Week Week Week Week Week Week Week Week Week Total ) Personal goals Turkka: My personal goal is to learn new software skills. On this project I can get familiar at LabView and G-Suite Sauli: I want to learn how to do a proper business model. In addition, I want to learn how to implement good testing and new software skills. Page 8 of 12

9 Mikael: I m having the main responsibility for the estimation algorithm. I want to learn how to lead a software project with good coordination and agile methods. 7) Cost plan and materials Because the project is mainly an algorithm design, there is no need to buy materials. However there are few costs that should be taken into account even though they might not be taken from the project budget. GT-suite license has to be bought, because current license is about to expire. License is bought for the research group. GT-suite is used for simulating diesel engine. License costs 2800 $. Furthermore, The test engine in the Internal Combustion Engine laboratory of Aalto has to be used, so the costs of the laboratory usage should be taken into account. Costs are covered by the cooperation agreement between the two laboratories. On top of above expenses the project has a 2000 budget for unexpected expenses, if needed. Expense Cost GT-suite license 2600 $ Internal Combustion Engine lab usage Covered internally Unexpected expenses ) Other resources The project is done in cooperation with a engine lab that has a big diesel engine. It is well-instrumented with a lot of sensors (in-cylinder pressure, inlet- and outlet temperature, inlet-and outlet pressure, common rail pressure and so on). The engine lab personnel is available to guide us with the LabView program that is used to control and operate the engine. The instructor will provide us a GT-suite simulation tool. A simulation for a diesel engine has been implemented with this tool. This simulation provides us really important data and helps us to understand how diesel engines work. 9) Project management and responsibilities The project manager is responsible for the project as a whole, manages the workloads of project members, follows the project schedule and that tasks are completed on time. Furthermore, the project manager writes the agendas for the project meetings and leads the discussion in the meetings. The supervisor and the instructor are supervising and instructing the project group. They will guide the project group to the end goal. They are responsible to react to possible problems and necessary changes of the project plan, if necessary. Page 9 of 12

10 Work package leaders will be responsible for managing the specific work packages and ensure that all the tasks of the work package will be completed on time. They are especially responsible for the content, the whereas project manager will be responsible for all work packages as a whole. 10) Project Meetings Project meetings will be held weekly by default at least in the start of the project. Later meetings can be held more seldom or frequently if necessary. The next meeting will always be agreed in the previous meeting and possible changes in meetings can be communicated via Telegram. Meetings will be held punctually at the agreed time. A default agenda of the meetings will be following: 1. Review of last meetings minutes 2. News and information 3. Results/Progression 4. Discussion of Project topic 5. Next steps 6. Next meeting The agenda will be presented beforehand for each meeting by the project manager. It can be commented and complemented by group members before the meeting. The agenda will be available in Google Drive. Memos of the meeting will be written by other members of the group and they will be available in Google Drive before the next meeting. Official memos will be compact and include briefly the discussion topics. Unofficial memos will have more information about the discussions, but they will not be transcribed. 11) Communication plan With Instructor Face-to-Face Meetings weekly unless agreed otherwise. Telegram Google Drive Internally Face-to-Face whenever needed Telegram Google Drive Overleaf Trello Weekly Face-to-Face Project Meetings are mostly held with the Instructor unless agreed otherwise. In addition, additional meetings will be held internally with the project group when needed. Internal communication is done via Telegram and communication with the Instructor is done via and Telegram. All materials are stored in Google Drive, except for LaTeX documents which Page 10 of 12

11 are stored in Overleaf. Overleaf is a cloud service for collaborative writing. Trello will be used for managing ToDo of the group. 12) Risks Risk Severity of the risk How is it taken into account Test engine breaks down High If the test engine breaks down, fixing it can take several weeks. In that case the algorithm cannot be tested with real data. Therefore, if this happens, the algorithm development will be based only on literature. Sensor breaks down High Especially if the in-cylinder pressure sensor breaks down, measurements cannot be performed. Replacing the sensor can take time. If this happens, at the worst case the algorithm development will have to rely on literature. Results are not as expected Low If the results are not as expected, the project is still successful, because one non-working method for estimating NOx emissions can be ruled out. 13) Quality plan Project quality is assured by multiple different methods. The project is based on sound and solid mathematical estimation methods that have already been used in similar kind of research before. The supervisor and instructor of this project are both experienced in the field of automation and control. In addition, the estimation method is based on research literature. Furthermore, the estimation algorithm will be tested with real engines for the sake of final quality. 14) Changing this plan If needed, the project plan can be changed by mutual agreement of the project group including the Instructor after discussion. An example of a situation when a change would be needed would be if the test engine breaks down and real engine measurements cannot be performed. Changes are documented to this document and saved as revisions. If a schedule change is needed, the protocol will be the same as for changing this plan. 15) Measures for successful project The main goal is to find a more accurate algorithm for estimating NOx emissions than the one that is currently known. The project would be very successful if such an algorithm were found. However, since whether the approach is good enough for the problem is unknown, the project will be successful even if a non-optimal algorithm were found. This is because non-optimal approaches Page 11 of 12

12 can be ruled out for future use. Hence, the project will be successful if a working algorithm is found, even if it is non-optimal. The algorithm is evaluated with GT-suite tool and compared to existing current algorithms. If the algorithm tests can be performed with a real diesel engine, then validation data can be gathered from it. Currently, real engine may be unavailable. The procedure is an iterative process, because the goal is to find an accurate estimation algorithm. Hence, the evaluation will be performed in the same way as the evaluation of the final outcome: the results will be compared with currently known algorithm and the previous algorithms. Results will be documented. The project group will keep track of tasks and completing them with cloud service Trello. Trello is a ToDo software for collaborative use. It will help the project group handle tasks, subtasks and their deadlines. Reaching the milestones will be documented in a Google Drive spreadsheet describing what has been done and when. Page 12 of 12