The convergence in gauss-siedel is very slow when compared to other load flows (like newton-raphson, FDC, NDC). so, In order to increase the convergence speed , acceleration factor is used. too low value of alpha(acceleration factor) will delay convergence. but very high value may skip desired solution/some times divergence. alpha should be chosen based on trail and error method. Generally alpha will be in range of 1.6 to 1.8. However it changes from one power system study to another. if system considered is well connected better to use Newton-Raphson Method.
Now a days no one uses gauss-siedel load flow. FDC is being used in online studies. if accuracy matters ,use of newton raphson method is mandatory.
For Gauss-seidel load flow method, an acceleration factor normally used to lead fast convergence. After doing repeated load flow it has been observed that normally it work good for acceleration factor of 1.6. However, this is good only for standard IEEE test cases. For practical system this factor sometimes do not lead fast convergence.
It is better to do repeated run by varying this factor and select the best one. The followoing link has considered such acceleration factor.
"Effect of Acceleration Factor in Gauss Seidel Method On Load Flow Studies" by Dr. H. V. Ramakrishnan and Hafiz Kashif Khaleel.
In that article, as Gauss Seidel Method is an iterative method and the acceleration factor Alpha is used for rapid convergence, the effect of varying this factor on number of
iterations and the error magnitude was simulated in MATLAB environment that mentioned also in its Appendix.
The real question might be why are you working with a power flow that uses the Gauss-Seidel method at all? In the 1950s and 1960s the method was used when computers had extremely limited memory. It has long since been proven that Newton Raphson methods are far more robust and efficient. I would be asking the question am I using the right program or algorithm.
The convergence in gauss-siedel is very slow when compared to other load flows (like newton-raphson, FDC, NDC). so, In order to increase the convergence speed , acceleration factor is used. too low value of alpha(acceleration factor) will delay convergence. but very high value may skip desired solution/some times divergence. alpha should be chosen based on trail and error method. Generally alpha will be in range of 1.6 to 1.8. However it changes from one power system study to another. if system considered is well connected better to use Newton-Raphson Method.
Now a days no one uses gauss-siedel load flow. FDC is being used in online studies. if accuracy matters ,use of newton raphson method is mandatory.
acceleration factors are used to speed up the convergence of Guess-sidel method. the value of alpha is between 1.4 - 1.6 and depends on the particular network being studied.
Acceleration Factor in the Gauss Siedel Method of load flow study is required for fast calculation and to achieve convergence in a very less amount of time. The best value of Acceleration Factor ranges between 1.6 to 1.8 .
An acceleration factor is a value that can be used to speed up the convergence and reduce the number of required alteration in a Gauss Seidel method of power flow analysis .Very high or very low values may cause the system solution to diverge and slows down convergence.A suitable value for each system can be obtained through trial and error load flow analysis. A generally recommended value of the acceleration factor is 1.6.