Jump to content
Mechanical Engineering


  • Posts

  • Joined

  • Last visited

  • Days Won


Everything posted by dudleybenton

  1. Charles F. Bowman has considerable experience with this topic and has published many related papers. CRC Press just released this book: https://www.routledge.com/Engineering-of-Power-Plant-and-Industrial-Cooling-Water-Systems/Bowman-Bowman/p/book/9780367775285 With his name in the search, you can find many of the publications online.
  2. There are many methods for numerical integration. Most of these are of historical interest and only marginal practical value, because a very few stand far above the others. These are Gauss and Lobatto Quadrature. Gauss does not require the end points, while Lobatto does. Sometimes this makes a big difference. If you must integrate about a singular end point (like mc²=integral(mvdv/sqrt(1-v²/c²),v=0..c) use Gauss because v never gets to c. Of course, this particular integral has an analytical solution, but you can see how v=c results in a divide by zero. Some things reported in the literature are simply not true; for example, 5x5-point Gauss is nowhere near as accurate as 25-point Gauss, yet takes the same time to compute. You may also find inaccurate claims in the literature regarding the Romberg method, which is a pathetic waste compared to Gauss and one of the least efficient methods. I describe and compare every significant method for numerical integration in the attached paper. All of the source code is available in the attached zip file. Comparison of Methods for Computing Cooling Tower Demand Curves.pdf NumericalIntegration.zip
  3. Heat trace tape is usually wound along a pipe in helical pattern like a spring, distributing heat over the surface. While it's not uniform coverage, it's usually enough to somewhat evenly heat. A single source of heat along one side as pictured may be problematic.
  4. I have worked with piping and heat tracing for 40+ years. I suspect you are asking the wrong question. The piping comes first. You size the pipe (and material) based on the flow requirements and properties of the fluid (Is it simply water? Is it oil? Is it chemically aggressive?) You size the heat tracing afterward, enough to cover the pipe and supply adequate heating (kW) to maintain the required temperature inside the pipe, given the expected outside temperature. You will need more heat tracing in Siberia than Minnesota. It is often necessary to cycle heat tracing on and off when it is cold, but not the coldest time of the year. The heat input (kW/m) is selected to match Q=UA(Tpipe-Tair) Tair and Tpipe are site-specific. Typical values of U can be found in the literature.
  5. That depends on what you mean by "work". Can you blow both cold and warm air through your lips? Yes. Will that enable you to cut your air conditioning bill in July? No. Don't forget that the Joule-Thompson coefficient for air is almost zero. Does an eco cooler work? Yes, but it's very inefficient when you're honest about the real cost and not sweeping details under the proverbial rug.
  6. There are two possible root causes: 1) alignment and 2) sealing. Both result in binding. If all the surfaces are clean and smooth and all the seals are resilient, it's alignment; otherwise, it's the seals, which can arise from several problems, including: degradation of the seal material, debris from outside, or oil quality. Either way, disassembly is required. If the problem is alignment, you will probably have to replace the entire unit. If it's seals, you may be able to clean and replace the seals. Once disassembled, you should see physical evidence of alignment problems. If not, then it's probably the seals. For a system like this, it doesn't take much to malfunction; so the cause may not look significant to actually be significant.
  7. A utility burning landfill gas didn't realize their NOx reduction feed led to production of cyanide! Chemical reactions can be quite complex, which is why you need special software that can handle nonlinear and non-ideal problems swiftly and accurately. This figure shows the impact of air:fuel ratio on the production of cyanide, carbon monoxide, and hydrogen sulfide, three toxic gases they didn't realize they were spewing into the environment at this supposedly "green" facility. Search Amazon books for "Thermochemical Reactions" and click on "Look Inside" to read about the solution hidden for 30 years. The software is free and the ebook will be free on Monday (4/12).
  8. This is vague so that I'm not even sure what your question is. Are you suggesting that someone might be able to deduce the type of turbine given two temperatures? I very much doubt that. I have worked with gas turbines for years, yet your terminology is unfamiliar. I suggest you search the web for an actual turbine from one of the big manufacturers (GE, Siemens, Mitsubishi, Solar) and review the section details.
  9. Then no one should mind if you spray a little water around to keep the electronics cool. Evaporative cooling is probably the simplest, cheapest, most reliable, lightest, least complicated method to use. Compressors, refrigerants, thermo-electric, semi-conductor, and other high-tech options sound good but are usually a nightmare to implement and keep working.
  10. There are several options. What is practical depends on the application. Is this robot supposed to enter a burning building or a failed nuclear system or a wood drying kiln or what? I have had various electronic devices fail in conditions I had to put up with, like 50°C and 100% relative humidity atop a cooling tower. Have you ever been inside a coal plant or gas turbine combined cycle plant when it was running? I have been countless times. I once stepped out of a plant in Muskogee Oklahoma into 36°C bright sunlight and shivered at the temperature drop. My coworkers have had to endure 63°C ambient in Saudi and UAE.
  11. This animated model is rendered using OpenGL. The source code is available free online along with other examples, including a T-Rex. [Google "3D Articulation"]
  12. The knight's tour is when the knight, alone on the chessboard, makes 64 moves, landing on each square once and only once. I have written millions of lines of code but have never sat through a single class on computer programming. I had to pass the course to graduate so I made a proposition. The instructor would assign me a problem. If I could solve it, I would get an A and he would never see me again. The year was 1974 and the problem was the knight's tour. Two days later I turned in the solution in FORTRAN on punch cards. I did not know he had been given the problem in graduate school but had never solved it. If I had, I might not have taken the risk. Years later I translated the algorithm into assembler. The entire executable code is only 1665 bytes. Here it is in 3D.
  13. I wrote this screen saver in 1992. It doesn't use one of the algorithms popular today. Instead, it's recursive like qsort, splitting the domain into smaller and smaller pieces. This process naturally results in triangles that are nearly equilateral. When using FEM, you want to avoid acute angles if possible.
  14. When I was in college 40+ years ago, simply running even an elaborate example using somebody else's software was not considered worthy of an advanced degree. We had to write the software ourselves. I am saddened by how many students run a model using Fluent or COMSOL or FOAM or Ansys or ABAQUS or write a bit of Python and consider the job done. You are short-changing yourself if you stop there. I can build a CFD model in 15 minutes. Don't take the easy path. The view at the end of the steep and rocky path is worth the climb.
  15. This isn't the most detailed or efficient model. It's brute force solution of the transient Navier-Stokes equation using finite differences. The nast2d source code can be found on the web.
  16. Resistance (air, road, rolling, drag, whatever) are all forces in the direction opposite of forward motion (i.e., -F). What you're integrating is Newton's 2nd Law or ΣF=d(mV)/dt where F and V are both vectors. In this case there's only one direction so the vector notation isn't necessary and dm/dt is negligible. RK4 (or Euler's method in an Excel spreadsheet) integrates A=ΣF/m=dV/dt and then V=dX/dt. The engine is providing +Power/V less the Drag so that dV/dt=P/V-D. P is a function of rpm, which is a function of V and gear ratio. D is a function of V.
  17. I suggest you get a co-op job for 3-6 months. You may soon find out what you love and hate. Give me a spot under the stairs like Harry Potter so I can work and avoid managers. You may like management. Everyone is different. We need good leaders, nerds, and hands-on workers to get the job done. In my experience, there are exceedingly few people who can actually run a company effectively. I am not one of them.
  18. I have seen this before on a variety of thread sizes, so this is not a unique problem. I run a die on the one and a tap on the other. You may need to turn the die upside down to finalize the thread, because it will be tapered on the starting side. If the two still don't fit right, it's a tolerance issue in the manufacture of the tools. The next thing to try is getting the tap and die from the same manufacturer. If it still doesn't work, you must file the threads. There are special little files in several shapes for this purpose. They're expensive, but worth the investment.
  19. see attached explanation and spreadsheet... expansion.pdf expansion.xls
  20. Isn't this exactly what a RedBox does with DVDs? Unless you want to make a lot of these dispensers and have a lot of money to spend (hopefully to recover later), you will need to find a mechanism that already exists and is already mass produced so that you can adapt it cheaply. When I needed such a thing, I went looking at what people had until I found something I could use. Vending machines would be a place to start.
  21. This is a typical example of solving an ordinary nonlinear differential equation. The details can be found in Chapter 1 of my book, Living Math, which you can read (click on "Look Inside") without buying the book. The eBook will be free on March 9th. The method to use is 4th Order Runge-Kutta, explained in Appendix B. The software is free at the link listed in the Forward. Further details can be found in my book, Differential Equations, which will be free on March 4th. The car simulation in Chapter 3 includes rpm, torque, and gear ratios. The calculations are included in the free source code package (car_simulation.c) and a spreadsheet (car_simulation.xls), also free online. There's another example calculating projectile drag. I have attached several files. https://www.amazon.com/author/djamesbenton car_simulation.c car_simulation.xls car_performance.xls 0to60simulation.xls
  22. Which hydrocarbon? There are lots. I create P-T diagrams all the time. I have attached an example. You can use my Excel AddIn for refrigerants or some of the other software I provide free of charge on my website. I have written several books on the subject. Azeotropes will be free on March 6 at Amazon https://www.amazon.com/author/djamesbenton
  23. dudleybenton


    There's a friendly guy at church, who stands at the door and hands out bulletins. You'd never guess he was a Blackbird pilot. When President Reagan was frustrated with the mess in El Salvador and wanted to send a "message", my friend took the 71 up… flying the whole length of the country at low altitude, throttle wide open with afterburner. Not only did it break some windows, it ripped more than a few drawers right off the clothesline. Another friend from here refueled the 71 several times to pull this off. More than one round of drinks has been wagered on there being a non-stop, round-trip flight from Knoxville to Honolulu.
  24. That's kilo-newtons per square meter. Nm is Newton-meters and is a unit of torque or work or energy, so the two are not equivalent and there is no "conversion" between the two.
  25. I have worked with the equipment you describe and have also looked inside these machines many times over the past 40 years. Your description is not accurate. There are no de-superheating nozzles anywhere near the final stage of this or any other LP steam turbine. The steam would only be superheated at the inlet. Besides, you would only desuperheat steam at the HRSG, not at the turbine, where you want it to be superheated. I have attached an actual expansion line from a similar 3-pressure system, also using an Alstom STG. I have many detailed steam cycle models for this type of system. The steam at the exit will be moist (about 3% to 7%) at the exit of such a turbine, as shown where the red line crosses and extends below the blue one in the figure.
  • Create New...