Claim Missing Document
Check
Articles

Found 2 Documents
Search

Performance Comparison of Ground Source Heat Pump (GSHP) against Air Source Heat Pump (ASHP) for Domestic Applications: A case study in Perth, Australia Aprianti, Tine; Tan, Evan; Diu, Chan; Sprivulis, Ben; Ryan, Greg; Srinivasan, Kandadai; Chua, Hui Tong
Teknomekanik Vol. 4 No. 2 (2021): Regular Issue
Publisher : Universitas Negeri Padang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (406.551 KB) | DOI: 10.24036/teknomekanik.v4i2.11272

Abstract

The goal of this research is to evaluate the efficacy and practicality of installing a ground source heat pump (GSHP) with an air source heat pump (ASHP) in a residential setting. The results of this comparison were obtained by monitoring each of the systems in two almost identical residences in Perth, Western Australia. There is an open-loop groundwater system in the previous home. Each residence has one family whose air conditioning system has its own set of operational needs. The cooling capacity supplied and coefficient of performance (COP) as a function of seasonal changes in ambient temperatures were determined using data gathered from both residences over a two-year period. For both heating and cooling, the GSHP system COP was found to be greater than that of the ASHP system. Furthermore, these two performance measures were independent of ambient circumstances in the former, but they exhibit a noticeable ambient dependent trend in the latter. The GSHP had an average COP of 4.0 for heating, regardless of external conditions, while the ASHP COP ranged from 2.0 to 3.0. Conversely, the GSHP had an average COP of 3.2 for cooling, whereas the ASHP COP ranged from 1.4 to 2.9. It is found that GSHPs result in significant operational cost reductions and a significant decrease in greenhouse gas emissions when a borewell is part of a home.
Analisis Kombinasi Speech Enhancement Dan Keyword Spotting Lightweight Untuk Robust Speech Command Recognition Pada Smart Wheelchair Tan, Evan; Utaminingrum, Fitri
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 10 No 4 (2026): April 2026
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Sistem pengenalan perintah suara berbasis keyword spotting (KWS) merupakan komponen penting dalam pengembangan kursi roda pintar, namun performanya rentan menurun pada lingkungan bising dan dibatasi oleh keterbatasan sumber daya komputasi pada perangkat edge. Penelitian ini bertujuan untuk mengevaluasi kombinasi model speech enhancement (SE) dan model KWS ringan serta menganalisis trade-off antara performa klasifikasi dan efisiensi komputasi. Evaluasi dilakukan menggunakan tiga arsitektur KWS ringan, yaitu BC-ResNet-1, MatchboxNet-3×2×64, dan TC-ResNet8, pada empat skenario pemrosesan sinyal, yaitu tanpa SE, RNNoise, DeepFilterNet, dan GTCRN. Performa sistem dievaluasi menggunakan metrik confusion matrix, sedangkan efisiensi komputasi dianalisis berdasarkan latensi inferensi, jumlah parameter, dan real-time factor. Selain itu, digunakan analisis statistik Generalized Estimating Equations untuk menguji signifikansi pengaruh model SE, model KWS, dan interaksi keduanya. Hasil eksperimen menunjukkan bahwa kombinasi GTCRN dan BC-ResNet menghasilkan performa klasifikasi tertinggi dengan akurasi 98,56%, sementara TC-ResNet8 menunjukkan efisiensi komputasi terbaik dengan latensi inferensi sekitar 1,79 ms. Hasil ini menunjukkan bahwa pemilihan metode SE dan arsitektur KWS perlu dilakukan secara kontekstual.