Tuesday, June 5, 2012

A4 - Group 8: Xue, Hayes, Jiang


Background


The Bridge Design Project primarily used West Point Bridge Design Software and K’nex to show engineering students the basic concepts about how to build a bridge and how the design affects its performance. West Point Bridge Design helps students model, test, and optimize a steel highway bridge based on realistic specification, constraints, and performance criteria. Below is the link to a blog from a member of group 8 indicated their feeling of West Point Bridge Design after using only a few time.


During the process of learning how to interpret the data given by West Point Bridge Design, some ideas of design had already been getting rid of because they could not pass the truck test, see below.  
Figure 1: A successful design in West Point Bridge Design



If a bridge could not pass a simulated test how could that design and structure ever be feasible enough to make a prototype. The designs which passed the test, needed to improve because the high cost and inability of the K’nex pieces to be that exact size or lay at that angle. Using the data gather from West Point Bridge Design we could now start building plausible physical bridges out of K’nex.

            K’nex behave more closely to a real bridge than compared with West Point Bridge Design. It has to be built piece by piece from nothing. There’s no sample bridge support be given like one West Point Bridge Design. Thus, building a K’nex bridge a feasible idea and blueprint is needed, a blueprint that could be found on a West Point Bridge Design model that passed the truck test. West Point Bridge Design tests showed that a bridge with top and bottom trusses is more stable than that only with one side of truss, due to the more options the weight has to distribute. Thus our initial, 24 inch K’nex bridge was born.


 
Figure 2: First design of K’nex bridge, 24” span


The goal of this project is designing the most serviceable K’nex bridge by truss analyzing. The ideal bridge would be loaded maximum loading with the lowest cost after several forensic and static analyses.

Design Constraints


            The bridge span for the final design is a thirty-six inch minimum. The width of the final design must be greater than three and a half inch. The bridge must be a feasible prototype design to a real bridge, meaning that a scale car must be able to fit through the bridge, the constraint for this was that a three inch by two inch tube be able to fit continuously through the span of the bridge.

Design Process


             The goal of our bridge is to design a most severable bridge with the lowest cost. Our 24” inch bridge was mostly build using the 1.125” inch chord. In the final bridge, the bridge size is increase, so we use 3.375” long chord instead of the 1.25”one. By testing how different angles work in the bridge. By the calculations we got the middle section is the part which undergoes most of the tension, so we change the middle part to smaller truss, so it can separate the tension of the loads. At the fist design we made the end members very strong as they must negate the force and are under as much force as the middle ones, so we get rid of some chord to short the cost. The original design also had numerous non-fixed member connecting the two halves. This was untimely the reason the first bridge failed, the non-fixed connection allowed the bridge to twist and lean and cause the first bridge to fall over on itself due to too much leaning

The double truss structure was an idea taking from the individual bridge design portion of the project. In the individual project one of designs was to have a single truss structure bridge, the other was a double structure bridge. The data gather by West Point Bridge designer showed that the double structure had a more stable shape and only increase the cost by a minimal amount, thus giving us our basic shape. Our final shape and size of the bridge is decided by testing physical model and data obtained thought experimenting on the bridge designer website. Using the data we gathered in latter we were able get the data we need and improve the bridge without test our real bridge and having to go through rigorous trial and error. This also helped avoid wearing down the pieces and accidentally damage and weaken them avoiding possible unwanted failures but to the condition of the pieces. Of course theory can only go so far so physical test must be done as well. We did test our bridge use reams of paper (500 pages of standard computer printer paper) and book. Upon weighting the books on a scale we found our bridge could hold a load a bit above 30 pounds. It is a good number base on the low cost of the bridge.       

Final Design


            The final design was a top and bottom truss bridge, which used a fair number of pieces, more than the bare minimum but far from excessive to keep the cost to a minimal. The connection between the two halves was by fixed connection on the top, middle, and bottom rows. It also had a few non fixed connections spanning across the halves, this was done to save money and through testing was found that having so few would not compromise the bridge strength and cause it to possible twist or lean to one side when weight was added leading to a premature failure. The idea behind the top and bottom truss is that the more connections there are the more ways the weight’s force can travel, having both a top and bottom mean that the main area the weight will be focused, that being the center and central top have more options to escape to and the weight will be distributed to the top and bottom and could be sent to the end in more options. This design did work much better than just having a top or bottom truss in testing thus it was chosen as the favorable shape.

Figure 3. Final Design in Final Test

            Below is a table of the parts, number of the parts and the cost, with the total piece count and cost at the bottom of their respective columns. Final piece count was 196 and the final cost was $307,000.



Table 1: Bill of Material

 

Testing Results


            The load at failure of the final design was 17.0 pounds. The failure was around the center area and towards the bottom of the bridge and was the result of two grooved gussets being pulled apart as shown below. A very small and simple failure that resulted in the bridge remaining whole and simply falling through the span rather than violently being ripped apart and being almost completely destroyed in failure. In terms of bridge failures this was a very calm and more favorable break.


Figure 4. Failed Connection

Conclusion of Results


            The final test did not behave as predicted and did not behave like any of the previous tests. In terms of the load the final design only held 17.0 pounds where the final version of the twenty-four inch span held 34.0 pounds. The final design only preformed half as well as the twenty-four inch span model meaning the design did not improve but rather became worst. The predicted load was forty pound, which was on the high side, thought multiple test on the final design prior held an average of around thirty to thirty-five pound, in which forty pound would not be too far off. This final test was most likely the fluke of all the testing and may be due to how the bridge failed.

            In many of the prior test and the twenty-four inch span test failed very close to the ends of the bridge due to all the surrounding members being pulled out of the gussets and the center falling straight down, usually resulting in a clean break leaving the bridge in two or three solid pieces and no single loss member or gusset. In the final test however, the grooved gussets pulled apart from each other around the center of the bridge, leaving it whole but with pieces not fixed together. The image below shows where the bridge normally failed (in blue) and where it failed in the last test (in red).


Figure 5. Usual Failed Connections and Final Test Failed Connections



The bridge only failed in the center area and because of grooved gussets being pulled apart once during all the prior testing. The conclusion as to why it failed like this is either due to a missed defect prior to testing or some small difference in this test that did not occur in the previous tests. Regardless, the final test did not behave as the previous test had shown, but that is just how things work out and so long as we can learn and understand from this failure there is always room for improvement in the future.

Future Improvements


Given the chance to modify our design to another version, the largest change would be to build a bridge with only one truss. The two truss worked but when it counted seemed to fail up to its standard and the extra pieces added a fair cost to the bridge, making a one truss bridge would save money and that saved money could be use to add extra support to critical areas such as the center and ends and would allow for more cross connections between the two halve stabilize the bridge, and we feel confident that a one truss bridge could hold more than the final test did and even save money.

No comments:

Post a Comment