Hi Yi Liang, I think most of your questions are beyond the scope of this forum and are not with simple answers. So maybe it's more proper to discuss on other professional forums.
1. I never use Matpower for the system as large as 30k bus, but I could predict with high confidence that it cannot be done within one second. I suggest you ask this kind of questions only after you run several given test systems in Matpower. Actually, if you have tried the 2~3k bus systems, the answer is obvious. I ever solved power flow for eastern interconnection by PSS/E. It takes tens of seconds. But the system contains models much more complicated than those in Matpower. 2. No one can give you a simple answer for this question I guess. Because computing time is determined by many factors, including your model in details, algorithm, codes and your computer capability. Several minutes are possible. 3. Again, I suggest you try the given test systems first. 4. It's really a big question. I just want to say Matpower is not for industrial application, especially on real-time platform. I guess you may need even more than one year of study to get the full answer for relevant questions. 5. Again, read the manual first. 6. Did you loss a "price" ? Shiyang Li ----- 原始邮件 ----- 发件人：Yi Liang <[email protected]> 收件人：MATPOWER discussion forum <[email protected]> 主题：Re: A general question about large scale power system 日期：2013年06月04日 13点27分 Dear Dr. Zimmerman, Thanks very much for your reply. I have some more questions. 1, Can I know for 30,000 bus system, how long will it take to run a AC power flow in MATPOWER? Can it be done in less than one second? or in seconds, or in tens of seconds, or in mins? Still for 30,000 bus system, can you also let me know how long will it take by using the fastest power flow solution method, if you know anything about it? 2, As far as I know, OPF is usually slower than PF. If the objective function is the loss of the power at lines, which is a quadratic function, then how long will it take to run an AC OPF on a 30,000 bus system? 3, Compare to using AC model, how long does it take for DC PF and DC OPF in MATPOWER for 30,000 bus system? 4, As far as I know, from the data in PNNL (Pacific North National Lab), the SCADA data is feed into the control center every 4 seconds. If I want to predict normal current flows (consider AC model) and voltages and determine the voltage levels to avoid under or over voltage problems, how fast should it be ? Since the power system can be unstable in seconds, is it correct that we need to finish that in less than 10 seconds? Can MATPOWER be that fast? 5, For the price, do we need to update the electricity price hourly or even every minute? Is it possible for matpower to solve a certain kind of OPF to achieve that goal? 6, And does the industry want to update the electricity as frequently as possible? I really appreciate you reading this email. Thank you for your kindly help. Best Regards, Yi Liang On Mon, Jun 3, 2013 at 2:10 PM, Ray Zimmerman <[email protected]> wrote: 1. Some might consider 30,000 buses to be large. Certainly systems of this size are solved in industry. It depends on the level of the modeling, but I've seen models of the U.S. Eastern Interconnect that are on the order of 60,000 buses. 2. Of course, and the requirement depends on the application. 3. It depends on the system, PF vs. OPF, AC vs DC model, the solver, etc., etc. The range of run-times can vary over several orders of magnitude. -- Ray Zimmerman Senior Research Associate B30 Warren Hall, Cornell University, Ithaca, NY 14853 phone: (607) 255-9645 On Jun 2, 2013, at 7:34 PM, Yi Liang <[email protected]> wrote: > Dear all, > > I am wondering if the performance of current power flow analysis tool is good > enough in terms of speed. > 1, What is the size of a power system that is large in reality? 30,000 bus > -system? > What is the size of the United States power system? > > 2, In the real industrial world, is there any requirements for the speed? > what is it? > > 3, If we use MATPOWER to solve the power flow analysis problems and the OPF > problems, how long will it take? I don't have that kind of test cases in > hand, so I cannot test them. Can anyone theoretically estimate the time > complexity? As far as I know, it depends on the iterations it takes for > newton's method( for example, if we use newton's method) > > > Thank you! > > Best Regards, > Yi
