Intro to Data-Modelling Course
This project was the final assignment of my data-modelling course. The course started with a review of Bayesian statistics, and then started with linear regression and ended with a brief talk on reinforcement learning. For the final project I really wanted to explore a subject in mechanical engineering and with real implications. The project was completed in Python, using libraries such as numpy, scipy, pytorch, and scikit-learn.
Prognostics-based maintenance, which is a typical pattern of predictive maintenance (PdM) has been developed rapidly in recent years. Prognosis, which is defined as a systematic approach that can continuously track health indicators to predict risks of unacceptable behavior over time, can serve the purpose of assessing the degradation of a facility’s quality based on acquired online condition monitoring data. The existing prognostics models can be divided into two main categories, mechanism-based models and data-driven models. Although the real-life system mechanism is often too stochastic and complex to model, a physics-based model might not be the most practical solution. Artificial intelligence-based algorithms are currently the most commonly found data-driven technique in prognostics research.
The data was generated by the NSF I/UCR Center for Intelligent Maintenance Systems with support from Rexnord Corp. in Milwaukee, WI.
Test Rig Setup
Four bearings were installed on a shaft. The rotation speed was kept constant at 2000 RPM by an AC motor coupled to the shaft via rub belts. A radial load of 6000 lbs is applied onto the shaft and bearing by a spring mechanism. All bearings are force lubricated. Rexnord ZA-2115 double row bearings were installed on the shaft as shown in Figure 1. PCB 353B33 High Sensitivity Quartz ICP accelerometers were installed on the bearing housing (two accelerometers for each bearing [x- and y-axes] for data set 1, one accelerometer for each bearing for data sets 2 and 3). Sensor placement is also shown in Figure 1. All failures occurred after exceeding designed life time of the bearing which is more than 100 million revolutions. Each data set describes a test-to-failure experiment. Each data set consists of individual files that are 1-second vibration signal snapshots recorded at specific intervals. Each file consists of 20,480 points with the sampling rate set at 20 kHz. The file name indicates when the data was collected. Each record (row) in the data file is a data point.
The first step was to load all the files in each data set. To do this, Python gets all the files names in each folder and loops through loading them and performing the processing algorithm provided. This data was then saved and loaded later using pickle. Loading and processing the 9000 plus waveforms took over 16 hours using Welch’s method and over 19 hours using the spectrogram method.
While attempts were made to utilize the continuous wavelet transform and consecutive Fourier transforms to create images for use with CNN, this was taking too much time to process. These also did not seem to be promising due to their temporal dependence and lack of visible differences.
The first step when evaluated the effectiveness of a model was to train the model with a portion of a dataset and test it with the remaining portion. When using Welch’s method, 3 out of the 6 datasets displayed a strong correlation with the respectively trained model. Of the 3 that did not display strong linear relationships, 2 had partially linear segments and 1 was almost completely horizontal. While these results were promising, they do not show the ability to predict independent data.
In order to test each a model’s ability to predict independent data, a model was trained with the entirety of at least 1 dataset. It was then tested against a dataset it had never seen. The actual configurations of these tests can be seen on the table below along with examples of the best and worst fits. These results brought about three possible conclusions. The first is that the models need more data in order to capture different types of wear behaviors. The second is that Welch’s method is not a strong enough processing technique to characterize long term changes due to wear. The last conclusion, on perhaps the most likely, is that there is not much difference between a new bearing and a used bearing with no faults.
Another signal processing technique was also used to train new models. This time, the spectral power density was calculated using a spectrogram, leading to more than 2x as many features to use in the neural network. This did not lead to an improvement in performance as can be seen below (these are only 2 of many experiments run).
The overall conclusion of this project is that the estimation of remaining useful life without advanced signal processing algorithms is difficult. Fault detection is more consistent and reliable. This might be the topic of a PhD project but not a month long course project.
Hey there. I’ve gotten back into reading this summer, so I’d like to post some snippets of the books I’ve read. I won’t call these reviews, because I don’t think I’m in the position to judge anyone’s writing. Either way, books are meant to entertain and teach a variety of people, so one person’s trash may be another person’s favorite book. Let’s get into it.
This book is about persuading people through advertising. My little brother was reading this little black book while we were at Disney’s Animal Kingdom Theme Park. I was intrigued so started reading it while waiting in lines. Written by someone who consulted on the show Mad Men, the anecdotes about the show make the book more entertaining than a traditional book on advertising. The book is split into two sections, the first being a general discussion of principles and theory, the other being specific breakdowns on the different type of ads. I only read the first half and took away some key points.
No amount of advertising or showmanship can fix a bad product or service. Advertising is about nudging people in a direction they already want to go in. Persuading people to get past their inhibitions or doubts is the name of the game. For this reason, be wary of anyone who says, “we need to change the way people think.” To get the best result, you must get to know your customer. Ask a lot of questions. One of the greatest insights of the book is to ask your customer what they want to be celebrating at the end of the project. Then you know your goal. In the modern age, the world and plenty of commerce exist online. The online world is about attention, so advertising there is different.
For those who want to know, the second half is about the four types of ads: intros, calls to action, differentiation, and loyalty builders.
Anyone who has been online has seen the ads for getting rich. Courses teaching us how to make passive income through real estate, blogging, YouTube, and drop shipping are abundant. Entrepreneurship is a buzzword for the ages. However, for those who are not so bold or enjoy working a traditional job, there is still a way to get rich. This is a simple finance book for the masses, and things may be more complicated than the book lays it out to be. I’ll point out those caveats, as I understand them, as well.
The method is slow and doesn’t result in riches until you choose to retire (which may be early if you follow the advice). The first principle is to pay yourself first. This means use your 401k to the max. It takes your money pretax and saves it for you, allowing it to grow. What the book doesn’t mention is how to choose between a 401k and Roth IRA, which is funded by post-tax dollars. When you do retire, you pay taxes on your 401k withdrawals, but not your Roth withdrawals. So, depending on your income at the time and how much you intend to withdraw during retirement, the choice may be different. The second principle is to make contributions automatic. If you do, it’s like the money was never there in the first place and that reduces the perceived burden. After your 401k, you should develop a rainy-day fund of at least six months and then start investing the rest of your savings in low to medium risk accounts. Other tips include using a sizable down payment on a home, paying half of your mortgage payment every two weeks instead of the whole thing every month (which results in an extra month paid every year), and making those payments automatic as well.
This book was my first step into personal finance, a journey which I now find exciting and confusing. I also enjoy minimalism as a lifestyle concept, which fits well with this style of wealth accumulation. With most things, a balance must be struck between saving for the future and enjoying the finer things of the present.
Astronaut’s Guide to Life
I used to want to be an astronaut and would still jump at the chance. For anyone with similar feelings, this may be the best motivational book out there.
Astronauts are incredible people, but not superheroes. They are constant learners. They prepare all there lives to go into space, which, even if they are in the astronaut core, may never happen for various reasons. But preparedness is what makes them incredible. It keeps them calm under pressure and allows them to make the quick decisions necessary for success. For them, preparing and learning is the end. It can never be a means to an end because the end is so rare it is easily discouraging. They learn procedures for when things go wrong because that’s what results in the calmness under fire. Anyone who doesn’t anticipate and prepare for the bumps will fail. Even if things are going wrong, things you can’t control, you can always control your attitude. For them, attitude is not only their emotions but also the direction of a vehicle. So, when things are out of your control, be assured your preparation will see you through and maintain your course.
To beat a dead horse: Always be learning and preparing for the result you want, but you should love the process.
I’ll be honest, these next two were audio-books I listened to, not read. But the following books are science-based, citing many studies, and having them read to me improved my experience.
The world is integrated into the online world. Everyone has a smartphone and media controls much of our perception. How do we remain unchained? How do we break the chains? Many of the points below are tied to research and examples. I know I said these weren’t reviews, but I highly recommend this one.
The book is about behavioral addictions as they relate to technology. These types of addictions are very real, although the effects are very different, and they are easier to hide in a modern world. A key point is that no one is born prone to addictions, we are all equally at risk. These addictions are all about pleasure, escaping pain, and satisfying unconscious needs. They are often tied to our environments. Companies know the formula for addiction and use it against us, but after reading this book so will you. Goal setting and the metrification of everything can be dangerous. Sometimes achieving a goal isn’t enough, there must be more, which can lead to unhealthy actions. If numbers are involved, comparisons with peers can lead to addiction. Phones and email are notorious for ruining our sleep, productivity, and ability to empathize/socialize with the real world. A key to avoiding addictions is to learn how to regulate and deal with emotions away from the screen. Today’s parents taught children how to walk and talk, but not how to use technology in a healthy way. The next generation of parents will be the first to have grown up with technology and will have to change their parenting to adjust for the tech factor.
There are a few different schools of thought for how the next generation should grow up with technology. Today’s biggest tech leaders tend to limit their children’s interaction with screens. Others think kids should be immersed in the digital world they are bound to grow up in. I could not do this book justice with just my takeaways. Don’t judge this book by my post, try it for yourself.
This one is a landmark book on introverts. The possible origins, their strengths, and how the world needs them. And yes, I’m an introvert. This was another audio-book and another science heavy book.
There’s nothing wrong with being an introvert. It’s different from shyness or social anxiety. Introverts make up anywhere from 30% to 50% of the population. In fact, there may be a biological basis for introversion that has to do with the sensitivity of the brain to new stimuli. Introverts seem to be shy because they need time alone to recharge their brains. Society has a lot to do with why introverts are perceived poorly. It started with the classic salesmen of the early 1900s and that influenced society to see extroverts as the ideal personality type for success. However, introverts have many strengths that extroverts don’t have and should not be marginalized. Introverts are compassionate, empathetic, calm, and thoughtful. They look before they leap, and many introverts saw housing bubble burst coming. Extroverts may have caused it. Introverts can have social skills and be good public speakers, but just need time to recharge. Their alone time is also beneficial for many skill based and creative pursuits that require practice alone.
The book finishes off with a look into the possibility of changing our personality type. While we may be able to act different, the book shows evidence that our brains never really change.
And that’s that, stay tuned for more books. Suggestions welcomed!
Analyzing Test Data
After the dust settled on some of our initial testing, the team thought to compare the performance of the MATLAB model I had developed to the test data taken in the field. We would compare the path following for the same set of way points to see how well we did. The way points and controller use a threshold of 5 meters, represented by the circles below. Here is a screen shot of the output, and you can see the model worked fairly well.
Testing in Miami
Members of the team flew to Miami in order to test the system portability and meet our stakeholder, Dr. John McManus. The system was checked at the airport and battery cells were carried in our backpacks. While we did get stopped by security for a second check, the trip went smoothly. The system survived the trip and we were able to deploy at a fossilized coral reef at Crandon Park Beach. Here we also met our stakeholder and got his feedback, which was all positive. Unfortunately, our communications stopped before we moved on to autonomous testing, so all we have from that trip is remote control tests. However, the trip did prove something useful. Our stakeholder confirmed the conditions of that day were very characteristic of actual conditions the system would see and the system performed well. Check out my YouTube channel (link in portfolio) for videos!
The Final Result
At the engineering school wide design fair, our project received an honorable mention. The projects were all amazing and we were happy with the result, and happy to have completed a successful senior project!
My team and I started this project as part of our senior design requirement that would span the two semesters of our senior year. The initial idea was given to us by a professor at the University of Miami, who I contacted in my search for project ideas.
The premise was simple: we had two semesters, or about 30 weeks, to create a research tool to help marine ecologists collect data in coastal environments. With the input of our stakeholder, and another scientist at the USGS, we developed the following project goals: way-point guided autonomy, portability (usable and transportable by 1 person), and a suite of sensors including depth, photography, and temperature. After careful considerations of our design space, time, and budget, we decided to build a autonomous surface vehicle (ASV) with a catamaran configuration.
Need Expanded (Bonus Reading)
Climate change and global warming have been talked about for years. There is no clearer demonstration of the destructive nature of these trends than the death of the coral reef.
Collectively, coral reefs have been shown to be worth about 172 billion dollars with their contributions to tourism, medicine, fishing, and storm surge protection. This valuable resource has been disappearing at an alarming rate, with one study showing that the Great Barrier Reef has been reduced to half of its area in the last 30 years. The cause of this disappearance is rising global temperatures and ocean acidification. Researchers studying this on a global scale receive very little funding and often gather data alone. The system we build would replace the labor and time intensive data gathering process. Our project aims to bring awareness to this ecosystem through photos and data. We also hope the data can help conservation efforts and future policy-making.
For a moving story on the problem at hand, I recommend the documentary "Chasing Coral" on Netflix.
The team spent much of the first semester proving need, developing requirements, and exploring the design space. With only a month to work on our solution, we chose to come up with a cheap testing platform for future work on electronics and sensing. In the spring semester, after many redesigns and revisiting goals, we finally built and tested our final prototype. We validated all of our systems separately, except depth readings, and did autonomous way-point testing. The final system also fits in a suitcase and has battery cells that agree with FAA regulations.
My background in mechanical engineering and robotics allowed me to take on multiple roles. One of my largest roles was the design of our center electronics enclosure and telescoping bridge in SolidWorks. I also ran static FEA analysis, drag analysis, static stability studies, waterline calculations, made product animations and helped assemble the electronics. As the boat assembly began, I utilized MATLAB for creating a dynamic simulation and autonomous controllers that would eventually control the boat.
Flash forward to Mechanical Engineering Senior Design Day, we received the Francis G. Tatnall Prize "for an outstanding project showing ingenuity, proficiency and usefulness." The team was very happy with the final product, but we are far from done. We will advance to the engineering school-wide competition and continue working on the project. We hope to have full validation and testing with our stakeholder in the coming weeks.