Your browser does not support JavaScript!
http://iet.metastore.ingenta.com
1887

access icon free FPGA implementation of adaptive p-norm filter for non-stationary power signal parameter estimation

A p-norm extreme learning machine (ELM) based on sparsity constraint is presented in this study for tracking of fundamental frequency, harmonic and dc in current power signals which finds application in phasor measurement units for wide area power network in smart grid environment. Real-time power applications typically are furnished with on-board controller and hence have constraints to stock a complex architecture. Moreover, the data from online practices are polluted by noises of diverse statistical features obtained on a sample-by-sample basis. Hence, approaches with improved learning paradigm and close model dealing with noises of varied statistical characteristics are essential. The proposed approach formulates a cost function with recursive p-norm error criterion and sparsity penalty that updates the output weights in succession besides adjusting some coefficients of the output weights to zeros that promotes quicker convergence and higher accuracy results. Exhaustive computer simulations have been carried out with synthetic signals and real-time signals to track the dynamic changes in the power signal amplitude, phase and frequency that demonstrate the accuracy, efficiency and robustness of the proposed p-norm ELM. Additionally, the new ELM network also is validated on a field programmable gate array (FPGA) hardware to prove its practicability towards current developments on phasor measurement units.

http://iet.metastore.ingenta.com/content/journals/10.1049/iet-smt.2018.5626
Loading

Related content

content/journals/10.1049/iet-smt.2018.5626
pub_keyword,iet_inspecKeyword,pub_concept
6
6
Loading
This is a required field
Please enter a valid email address