Sr-No | Artificial Intelligence Techniques | Advantages | Disadvantages |
1 | Neural Network | It exhibits high level of accuracy for diverse and mixed power quality disturbance classification. | In noisy conditions its efficiency is very limited or in other words it’s less. |
2 | Artificial Neural Network | It offers mathematical suppleness with high accuracy in real time applications. | Its accuracy and convergence depends on its network architecture and also dependent on noises in the signal. |
3 | Support Vector Machine | It has a high learning process with a prospective to grip hefty features; it also provides a stable solution in quadratic optimization. | Its total dependent on training samples if they are minimum, the classification accuracy is poor and vice versa. |
4 | Fuzzy Logic | It has a fascinated accuracy in analyzing complex systems and it’s accurate in modeling. | It cannot accommodate new disturbances as its training set is limited for every case. |
5 | Genetic Algorithm | It classifies accurately power quality disturbances which are generated due to damped sub harmonic signals and dynamic performance of the power system. | It has a very high computational time. |
6 | Expert System | It has the facility that it can be used with/without limited data. | In this, the actual situation and assumptions do not match exactly and the extract of conclusion is very difficult furthermore its execution is very slow, moreover the system is very expensive. |