However, as the fault rate continues to increase, the event detection performance of the OTDS and DFDW algorithms decreases significantly, which undeniably indicates that the DSFTED algorithm has better detection performance.
7 shows that when the node fault rate reaches 15%, the fault identification rate is 85% for OTDS and 95% for DFDW, whereas the proposed DSFTED algorithm reaches approximately 97%, which is a significant increase.
8 shows that when the node fault rate is below 15%, the false alarm rates calculated using the DFDW algorithm and the algorithm proposed this paper are significantly lower than for OTDS.
9 shows that as the node fault rate increases, the miss-alarm rate from OTDS tends to increase.
It can be noticed that the event detection rate calculated with the proposed algorithm is far better than with the OTDS and DFDW algorithms.
12 presents the event boundary detection rate (EBDR) versus detection range using the OTDS, DFDW, and DSFTED algorithms when the node fault rate reaches 20%.
It has been recognized that energy savings can be obtained by dispersing computation within the network in a distributed form, such as the above discussed algorithms OTDS , DFDW  and the proposed algorithm DSFTED as well.
During the time window T, the total amount of data exchanging of the entire network generated by OTDS can be calculated as nxNxF/T, where F represents the sampling times within the window T, namely f = F/T indicates the data exchanging frequency.