Sound event detection (SED) assists in the detainment of intruders. In recent decades, several SED methods such as support vector machine (SVM), K-Means clustering, principal component analysis, and convolution neural network (CNN) on urban sound have been developed. Advanced work on SED in a rare sound event is challenging because it has limited exploration, especially for surveillance in a forest environment. This research provides an alternative method that uses informative features of sound event data from a natural forest environment and evaluates the CNN capabilities of the detection performances. A hybrid CNN and random forest (RF) are proposed to utilize a distinctive sound pattern. The feature extraction involves mel log energies. The detection processes include refinement parameters and post-processing threshold determination to reduce false alarms rate. The proposed CNN-RF and custom CNN-RF models have been validated with three types of sound events. The results of the suggested approach have been compared with wellregarded sound event algorithms. The experiment results demonstrate that the CNN-RF assesses the superiority with remarkable improvement in performance, up to a 0.82 F1 score with a minimum false alarms rate at 10%. The performance shows a functional advantage over previous methods.
A6s~ruci-Environmentnl pollutiou can cause the insolator material to become progrcssively coated with dirt and chemicals in tile long run. In the prcseiice of wet atmospheric conditions, the leakage current tlows duc to the development of conducting path across the insulator surface. The level of leakage currcnt depends on the surface wetting and the degree of electrolyte, contamination as well as the environmental factors. An analytical approach bnscd on dimensional analysis technique is applied to develop n niithematical model of leakage current i i i correhtion with the environmental stresses. In order to verify the developed model, an incliiied-plane trnckitig test is conducted on the polymeric insulating materials. The cxperimeiital work is carried out by measuring the magnitude of surface Ieelcage current at diffcrent levels of contaniination. Simulation results of the model have shown good agreemelit to the experimental results and provide useful informatiori on describing the test condition of the tracking test procedures.
A hybrid model incorporating wavelet and feed forward back propagation neural network (WFFB-NN) is presented which is used to detect, identify and characterize the acoustic signals due to surface discharge (SD) activity and hence differentiate abnormal operating conditions from the normal ones. The tests were carried out on cleaned and polluted high voltage glass insulators by using surface tracking and erosion test procedure of IEC 60587. A laboratory experiment was conducted by preparing the prototypes of the discharges. This study suggests a feature extraction and classification algorithm for SD classification, which when combined together reduced the dimensionality of the feature space to a manageable dimension. Wavelet signal processing toolbox is used to recover the surface discharge acoustic signals by eliminating the noisy portion and to reduce the dimension of the feature input vector. The test results show that the proposed approach is efficient and reliable. The error during training process was acceptable and very low which attained 0.0074 in only 14 iterations.
Absfrrrct : The Inclined-Plane Tracking test (IEC 587) is an established standard test for evaluating tracking and erosion resistance of polymeric materials for outdoor insulation. The standard parameter values shall be strictly followed in order for the test result to be accepted. Each selected panmctcr valuc is rclatcd with thc other paramctcr valucs. This paper presents the mathematical approach on development of parameters correlation in the IEC 587 standard test method using diineilsional analysis teclmique. The result from mathematical model is then coinparcd with empirical data from the standard.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.