Showing posts with label Yilei Jiang. Show all posts
Showing posts with label Yilei Jiang. Show all posts

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.

Wednesday, May 30, 2012

Week 9 - Yilei Jiang

       During the last week, we worked on building our three foot span bridge. we didn't change our design a lot because we thought that is the most serviceable way to build. The bridge we had based on the tension and suspension measured by "Method of Joints", also depended on the predicted price. We then tested the bridge using the parameters we will be using in the next week for the official testing of the bridges. It was able to hold decent amount of weight on the first try (about 25 lbs) so we decided to reduce its cost by removing some of the members connecting the two sides of the bridge. We tested it again but the amount of weight it held dropped significantly. So our plan for next week is to re-design the final bridge that can hold decent amount of weight without failing which we will be test during week 9. The only accomplishment this week was to test the same design with different number of members connecting the two sides of the bridge which gave us an idea on how much the connecting pieces contribute to its weight holding capacity.



      We have completed almost 9 weeks of the bridge module which has brought many new things into my perspective. I learned quite a few things from the bridge design tools such as WPBD, Knex, truss analysis etc. that we utilized throughout the term. One of techniques I learned in designing a bridge when our goal is to have the lowest cost is that you can analyze the compression and tension forces using WPBD and try to reach the ratio of 1. This can be done by using different variations throughout the bridge which include changing the member size, material and length etc. while reaching a ‘functioning’ bridge. However, in real world, the bridge would be undergoing a lot of external forces such as wind turbulence etc. The amount of force applied by the vehicles travelling over the bridge would also vary constantly. So to ensure the safety, the bridge would have to be designed such that it can withstand the maximum amount of force. Every members and gusset plates would have to be analyzed in great detail to ensure that they wouldn’t give up under the normally expected force. In this case, the safety would be the first priority, not the overall cost of the bridge.

Wednesday, May 23, 2012

A3 - Yilei

  • Free Body Diagram & Calculations
     By using the "Method of Joints", the analysis of the truss bridge is showed below. The bridge has a span of 24", a height of 8" and a load of 15 lb at point C.





  • The overall forces can be seen below.



  • The  same results of analysis in The Bridge Designer's version. 



       The online Bridge Designer is a program allows to design a virtual truss, and then put a load on it. It will calculate the tensions and compressions of each members itself. 
       When I did this program, I had to use the same scale of all members and angles to correspond the results of my hand analysis.  I made each grid be 2"so that the length should be 12 grids. And the height should be 4 grids. Then I picked up the middle node of the length to put a load on, which was 15 pound in this case. Consequently, I got the most approximating forces equaling the ones calculated by hand.   



  • The results of analysis of our two foot span bridge in Bridge Designer program.




        Might because of our design, it was hard to computed via the Bridge Designer program. Our two foot span bridge has a bottom truss. So when I completed the adding process and pressed the calculate button, there's an error came out. It said that we had to followed stable structure is M + 3 = 2*N , where M is the number of members and N is the number of nodes. The simulation will not calculate forces unless this condition is met. Thus, we changed our K'Nex design a lot. Besides, we did looked the K'Nex joint test page and it showed the tension would increase if the bridge is symmetrical. So we did the bridge symmetrically.
        That type analysis of the K'Nex truss bridge can find out the strong and weak parts of the bridge by using the average strength of each joint. So we can change our design to reach the proper number. That can make the bridge more stronger.    





week 8 -Yilei Jiang

       I learnt the "Method of Joints" during last lab which is s a way to find unknown forces in a truss structure. This method also would be used in completion of the A-3 assignment.The method of joints consists of satisfying the equilibrium equations for forces acting on each joint. In the lab, we also discussed some ideas of our next three foot span bridge. We agreed that we would use the new method we learnt to calculate our new design to get the most serviceable design.  


       After the amounts of calculating, I think this method of analysis, the "Method of Joints," is not sufficient for a real bridge for some reasons.  First, a real bridge should be tested and analyzed in many more ways than just testing a downward force on a connection joint. Even this method shows the tension and suspension of every members, a real bridge could not be built of this simple databases. Secondly, A real bridge is not only hold its own weight but also deals with dynamic loads and the things like side pushing from the side and even up drafts from underneath. "Method of Joints" does not show that side.  
    
      I am not saying that the "Method of Joints" is completely useless because it at least shows the force when a truss bridge is under a special circumstance, which is  in equilibrium. If a truss is in equilibrium, then each of its joints must be in equilibrium. That's how this method works. This method is one of many methods that are used to analyze bridges. Of course there should have more different calculation for a real bridge. 
      
      There's one thing I'd like to further analyze, which is the accurate breaking point of the K'Nex gussets when they are   experiencing forces. I think that this information would be useful at our next assignment, which is designing a three foot span bridge.
      
      For next week, we will working on our new bridge and complete the comp2.

Tuesday, May 15, 2012

Week 7 - Yilei Jiang

       Last week, we finally tested our bridge and it loaded 34-pounds of sand. At first, the bridge wasn't long enough to fit in a two foot long span. Therefore, we decided to add an extra 2.125" long chord on the base. By doing this, the bridge became symmetrical.
       The bridge started to twist when we added weights because of the structure of our bridge design. We added an extra 5' long chord on the top truss to make the bridge more stable. However, the 5" long chord was beyond the width of the bridge, so it didn't fit too well with the gussets. When we added the load, the bridge started to twist lightly. There were two groove gusset plates in the middle which pulled apart, at first. Then two members from the support points could not hold the load any more, so the bridge collapsed. We planned to fix these problems on our next three-foot bridge design.
        For K'NEX numbers, I would like to know the accurate tension of every piece truss members during the bridge loading. That would help me to design a more serviceable bridge by calculating some trigonometry functions with those databases.
       Next week, we will start designing the three-foot long bridge on the basis of previous designs.

- Yilei Jiang

Wednesday, May 9, 2012

Week 6 - Yilei Jiang

         This week, we built our own design K'nex bridge to see which one was the best one and the good components of the designs.  We decided to build a design which without bottom truss. Besides, it was also the design which had the lowest budget. After we done, I found one thing during the testing that the weakest part of this bridge was those grooved gusset plates. They were always pulled apart when we was adding the weight. Especially, the gussets from two piers of the bridge were so easy to separate.  I thought that the reason was  the bridge had one-side truss. So I suggested to build the bridge with top and bottom truss. However, the testing result was same. Our second design only could hold 20 pounds before it was fallen to pieces. Then we started to think about the reason. And I felt that we could shorten the height of the bridge because it can support well. Our original base was the structure like this" |\|\|\/|/|/|". And now if we change 5"long chords to 1.25" long chords, the bridge's structure will like " |\/|><|/\| ", which I think can be more stable. The other idea is that we decide to use more groove-less gussets to avoid using grooved gussets. That can not reduce the cost but make the bridge into pieces easily.
      

        For the next week, After working with Knex for a week reread we entry of last week and state how our views of the similarities and differences have changed. Then we will test our design on a 20' span. And we will fill out the bridge results recording before we leave the class.         

Wednesday, May 2, 2012

Week 5 - Yilei Jiang

      The lecture I experienced from this week was abundant and useful. I knew four major components of a truss bridge. And I familiarized with my team member how the K'nex from shapes. I basically knew the steps of designing a truss bridge in reality from the K'nex, like making an idea; choosing materials; build as a model;  budgeting; testing; editing and etc. There must have a lot of different in detail when I am doing a real project. And for K'nex, I felt the materials could predetermine how the bridges look like. I didn't feel that when I made a bridge via WPBD. Besides, K'nex is a real and touchable toy and lots of problems will come out. For instance, in WPBD, the truck has such a strong climbing ability. It can still pass the bridge when it is down to the bottom of canyon.  However, it will never happen in reality. So when I test the K'nex bridge, it will collapse if it can hold the weight any more.
       I also listened a speech in class. It was a speech by Mr. Jay Bhatt, our engineering research librarian, about the most efficient way to look for engineering-related books from Drexel Libraries. He demonstrated the way that searches different resources in the class. I felt that that made my study much easier than before.
       For next week, we are processing the K'nex. We will pick the best idea from our  team and build it with K'nex.
     
   

A2 - Jiang

     The idea of my bridge's shape is from the results of West Point Bridge Design I tried last week. During several times of experimentations, I found that it was much easier to make a bridge for doing bottom and top truss at a same bridge. I tried to do only bottom truss, like Deck Truss. However, the bridge always collapsed when the truck was passing. And the bridge even cost me more money when I finally made it without top truss because I have to add more web members to make the bridge be stable.  My original design, which has both of bottom and top truss, its strength and tension were evenly distributed by the short chords I added on the top. 


side view
top view
Bill of Materials - K'Nex pieces





Height: 5.875"
Length: 27.028"

     I used colors to differentiate the various chords I added. The yellow chord is 3.375" long; the red chord is 5" long; the black chord is 1.25" long.

    About the gusset plates, I used three types of those plates. Two of 360' gusset groove plates could connect the top and bottom truss in the middle. And when I thrusted a 90' gusset plate into a 180' gusset groove plate, the mixture gusset could connect the rest of members. 


About changes
      At first, I tried to make a bridge with arch. But end up with failing. May be because the arch bridge can not be done with K'Nex pieces.  So I started to make a K truss bridge and I thought that it shortened the lengths of the compression members compared to the other trusses. The thing I changes of K truss was I added two short chord under the middle K truss. ( like the picture shows below) 

 The reason why I did that is because this two black chords could support the top better. 


What I learned from K'Nex is the material I chose can limit my idea of bridge design. But I also also learned how to those prescriptive chords and gusset plates to build different shapes of truss bridges. 

- Yilei Jiang

     

Wednesday, April 25, 2012

Week 4 - Yilei Jiang

    Last week, the major thing we did was building a bridge design via West Point Bridge Design, which is a program to design a virtual bridge. Because of this designed bridge, I knew the basic knowledge of  bridge building components, such as span, material, placement of the travel surface in relation to the structure, and form. The other thing I knew was the cost of a bridge design played an important role. Related to reality, the cost also always decides that a contract is signed or not. 
     
     There's a thing that only can happen in West Point Bridge Design, not in our real world. During the testing, the truck passed every time whatever the bridge I designed looks like. Sometimes, my bridge was down to the bottom of canyon. It's not a serviceable bridge in the real world. 
      
     Next week, we will get the K'Nex and I will put the knowledge I learnt from WPBD on K'Nex. We can use K'Nex as a model to build a low-cost and serviceable bridge. K'Nex is not like WPBD. It is a model made up of solid materials(members), which means it has its own tension and strength capacities. How to hold heavy objects with the K'Nex bridge is the goal during these weeks. 

Wednesday, April 18, 2012

Week3 - Yilei Jiang

This week I experienced how to build a basic truss bridge via a software, WPBD.  I knew two cross sections and three material types. Tube will be much cheaper that Bar if we think about the budget. For CS, HSS, and QTS, my bridge design was more like to use HSS because the bridge passed test as soon as I changed CS to HSS. However, HHS is more expensive than CS. Thus, QTS is a proper choice. QTS is a material that can support more tension than CS but less than HSS. And the price of QTS is between that of CS and HSS. From cross section part, I used Tube because it much cheaper than Bar. However, Bar can hold double or more tension and compression than Tube. So I think that increasing the truss' size is a good idea.

Then the question was coming after. Is there any other way that make a useful bridge and reduce the cost?
As so far, I don't have enough knowledge to support I build a prefect bridge.

So my three question will be?
What is the strongest structure that a bridge has? one side or two side? or other?
Why the bridge always fail when I first designed it, which I just followed the model WPBD gave?
If I want to know different bridge designs, what kind of books should I looked for?

For next week, we will continue the West Point Bridge Design and figure out what are necessary components of a real bridge design.

A1 - Jiang

My goals of designing the bridge are from two sides. One is spreading the tension into top and bottom. Two is using diverse materials and changing the size of them to reduce the building cost. I think that if I use this way to build a bridge, it will be safe and low cost.

Bridge in 2D mode


The Truck in animation mode




The load test results




The original price of the bridge was double as current price due to the cross section I chose, which was Solid bar.   After I using Hollow Tube instead of the bars, the cost was suddenly cut a half, however, the truck didn't pass the test. The compression strength of the tubes was much lower than that of the bars. So I decided to change the size of the truss on the top and the material of all trusses. I compared the cost when I increased the size and changed a stronger material of one same member. The consultation came after testing. Changing a better material is more inexpensive than increasing the size of trusses. For instance, I used quenched & tempered steels instead of carbon steels from the truss on the top. The price increased by $20141. When I kept using carbon steels and I changing the size of the truss, the price was risen by $32204 to pass the test.


Current cost: $292,788.76
I think that I can drop the price by $50000~$60000 with further time.

I learned that where the main compression and tension distribute on a bridge when a truck passes through. I also learned how to analyze the data from the load test results to adjust my bridge accordingly. I found out that the bridge could be failed down without any weight, just itself.



Tuesday, April 10, 2012

Week 2 - Yilei Jiang

      During this week, we set up a team and was getting know each other, like the basic information. This team is made of three different engineering, which are environmental, architectural and mechanical. That might  let the team stronger than other because we have various ideas and advices from diverse aspects. 
      In the team work, conflicts are always existing. For example, one team member has a idea for building   a bridge and another team member has a different idea. Both of them think that they are right. Then we can talk about the possibilities or ask the advisor that which one is better. The big problem of this project is how to design a low-price bridge. So we should discuss the decision before we make it and then give time to "mull" on it. Often I think that time can produce changed minds and got better decisions. 
      I also think that meeting could be happened regularly- ideally twice a week. Once is in the lab and another one can be at weekends. I think that we also can ask a meeting with advisor if it's necessary. We could set up a agenda for meeting.
      We don't have to work together even if it is a teamwork. We can add up the ideas or the work we have done in the end, however, we should first take apart the whole project . For example, the building part, the testing part, the measuring part and etc..  Accomplishing them part by part will make the whole progress a lot of easier.