In this week’s lab session, we used the shock dynamometer to collect data for a “mystery oil” with an unknown weight. Using the same procedure of sweeping intervals of 15RPM and rebound/compression settings of 5 clicks as the previous two weeks, we collected data from the dyno's LVDT and pressure transducers.
Outside of data collection, we had two main goals for this week. The first was to finalize our viscosity model, enabling us to predict the weight of the mystery oil by comparing this week’s data to the previous data for 5W and 20W oil. The basis of our viscosity model comes from the linear relationship between viscous force and a fluid’s velocity gradient, and a plot comparing the force and velocity for each of the three oils can be seen above. Our current guess is a viscosity between 7-10W. As of this writing, we are on track to finalize our model by next week’s lab session, although additional filtering needs to be applied to our velocity data in order to more confidently predict the weight of the mystery oil.
Our second goal for this week, motivated by the February 16th project proposal deadline, was to decide on a final project. After talking through some ideas amongst ourselves and with the teaching staff during Tuesday’s lab session, our group was split between two potential projects: a semi-active suspension system that would mount onto the shock’s rebound pin, thereby allowing us to alter the shock’s stiffness in real-time, and a heated enclosure that would create an isothermal environment around the shock and minimize the effects of thermal drift on the viscosity of the oil.
After taking a night to consider our options, we decided to focus our efforts on controlling the oil temperature within the shock. Although both projects intrigued us, we ultimately chose the temperature control project mainly because of the considerably higher time commitment involved with semi-active suspension. In addition, temperature control has the potential to benefit future ME 107 students by improving the data acquisition of the entire system.
Looking ahead, our main goal for next week, other than determining the viscosities of the three oils and guessing the weight of the mystery oil during our lab session, is to begin work on our final project. Our first steps will include researching necessary microcontrollers and heating elements, creating a model to predict viscous heat generation and convective heat loss within the shock, and starting to consider the advantages and disadvantages of different data acquisition methods in LabVIEW.
Thanks for reading!
-The Hot Shocks
My main goal this week was to fit the load vs. velocity plots to determine which of the parameters (rebound setting, compression setting, RPM) would be best suited for our model to predict the viscosity of the mystery oil. I fit a linear regression to plots with data taken at varying RPMs, and with multiple settings for compression and rebound. In determining which data sets were the “best,” I looked for the plots that were the most linear (R square values close to 1), since the the load should be a linear function of velocity. Second, after completing some research, I found that the 20W should be about 1.2x as viscous as the 5W oil at 50C, so I looked to see if the slope of the 20W oil were 1.2x that of the 5W oil, since the slope should be proportional to the viscosity. With these conditions, it seems that 90 RPM with 0 clicks on compression and 20 clicks on the rebound setting was best.
Our group has gotten along really well so far this semester and we are excited for what’s to come! The freedom to choose our own project for the rest of the semester has definitely helped to engage me and my teammates in this class.
My work this week was mainly focused on data analysis and plot creation with the goal of accurately predicting the weight of our mystery oil. Although our model will not be perfect because of the noise in the LVDT measurements, I believe our filtering methods will allow for suitable linear fits to be made between force and velocity. I was very happy with the Hot Shocks’ communication and teamwork this week; we were able to efficiently assign tasks and brainstorm project ideas, which bodes well for the weeks to come.
The bulk of my work this week focused on data collection and analysis of the underlying assumptions of our predictive model. Because we were able to collect data on both position and force, we believe our best predictive model will be based on the viscous force equation. Unfortunately, sensor noise in the shock dynamometer and the noise from derivating the position readout to velocity hinder the accuracy of our model. I also put a lot of thought into our semester’s experiment and was eager to work on a semi-active shock absorber controller, but we agreed as a team that it might be too ambitious of a project. I’m a little disappointed, but still excited to work on the controller algorithm for our oil temperature regulator. Our team gets along really well inside and outside the lab, and I’m excited to work with them this semester. I feel like I should have contributed more this week, but I appreciate that my team understands I was busy and I plan to devote a lot of time to the control algorithm and microprocessor in the coming weeks.
This week I focused on creating plots that compare the outcome of testing with different oil weights. Keeping all other variables constant, I was able to analyze the effect of altering the oil weight on force and pressure. While the different oil weights did not have a huge impact on the force measurements, the differences between the oils was much more apparent in the oil reservoir pressure plots. All of the plots place the mystery oil in between 5W and 20W. For a majority of the plots, the mystery oil has an almost identical force plot to the 5W leading me to believe that the mystery oil is somewhere between 5W and 10W. The Hot Shocks appears to have really good chemistry thus far and I am excited to start our temperature control project!
I spent the majority of my time working on our data acquisition this week. My focus, in particular, was the filtering of our data. The raw data collected provides only a position measurement. This is insufficient for the characterization of fluid dynamics models as velocity and the velocity gradient are key. Generating a velocity vector by simply taking the derivative of our position measurement resulted in a clear trend, but suffered from significant amounts of noise. This is to be expected as differentiating an already noisy signal will amplify the noise. To deal with the noise, several basic filters were implemented: blurring, exponential moving average, kalman, and basic model based. My future work will include tuning the individual filter parameters, quantitatively analyzing the effectiveness of each style of filter, and deciding on a filter or combination of filters that best suits the job. There is much to be explored and much to learn with regard to our data acquisition. The value of effective, validated statistical analysis cannot be overstated. Go Hot Shocks!
This week I worked on refining our MATLAB dataParser function. During the first week of lab, I wrote this function to pull data from Excel files generated by the shock dyno. This week I added lines to calculate error bars for our velocity. I also looked into using averaging as a filtering method. When we calculate the viscosity of the mystery oil by using a linear regression on a force vs velocity plot, error bars will be useful to visually determine how confident our viscosity guess is. I am happy with the work our group has done so far and excited to start our project! I have high hopes that our attempt to create an isothermal shock body will help future shock dyno groups.