Document Type

Student Research Paper


Summer 2021

Academic Department

Computer Science

Faculty Advisor(s)

Dr. Peilong Li


This a collaborative project between Elizabethtown College and CPNet, LLC that is looking to help apply predictive modeling with CPNet’s domain knowledge to one of CPNet’s clients' IIoT manufacturing problems. CPNet has provided us with datasets taken from one of their clients in the hope that we can build a model that will be able to predict when a part within the machines they are looking at will fail and subsequently shut the machine down. We are trying to take their data and turn it into information that the company can take preemptive action on and save them downtime during operation. In this project, we designed a health index that we used to create y-labels that we did not have in our dataset. We did this so that we could solve the remaining useful life problem of our project, which is where we attempt to determine how long the machine has to run before a failure or maintenance is needed based off of the health index using machine learning. We then went on to build several machine learning models to solve our problem. First, we used traditional machine learning models such as Polynomial Regression, Tree-based Regression, Ridge, Regression and XGBoost. Later we turned to Neural Networks and built a Multilayer Perceptron, a Convolutional Neural Network and a Recurrent Neural Network. From our experiments traditional machine learning models outperformed the neural networks. This was to be expected since the dataset wasn’t that large, but it was worth testing regardless. Also, from our experiments it is apparent that a part of our health index the CPK (CPK analysis will be explained in greater detail later in the paper.) will need to be worked on in the future to provide better y-labels. All in all, this project did show that this problem is worth exploring in the future since the solution could be quite valuable in smart manufacturing.


Scholarship, Creative Arts, and Research Project (SCARP)



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.