{"title":"Analisis Pengaruh Diameter Kawat terhadap Distribusi Kapasitansi dari Wire Mesh Sensor Tomography menggunakan Convolutional Neural Network","authors":"Mahendra Satria Hadiningrat","doi":"10.21111/fij.v7i3.9442","DOIUrl":null,"url":null,"abstract":"AbstrakWire Mesh Sensor (WMS) adalah sensor berbasis tomografi yang menghasilkan gambar distribusi aliran fluida. Citra distribusi merupakan pola distribusi kapasitansi yang diukur dengan elektroda sensor. Dari hasil simulasi dilakukan analisis terhadap pola sebaran potensial listrik untuk mengetahui karakteristik potensial listrik dari sistem WMS yang dimodelkan. Ditemukan ada perbedaan parameter berupa variasi jenis larutan yang dapat mempengaruhi distribusi potensial listrik. Hal ini disebabkan adanya perbedaan nilai konstanta dielektrik masing-masing jenis larutan. Kinerja sistem WMS dalam mendeteksi anomali dievaluasi dengan menganalisis perubahan distribusi kapasitansi terhadap pengaruh perubahan diameter kawat. Hasil simulasi menunjukkan bahwa jenis fluida pada kondisi tanpa dan dengan anomali dapat dibedakan dengan jelas melalui pola distribusi kapasitansi yang diukur untuk seluruh diameter kawat. Diameter kawat hanya mempengaruhi kualitas gambar distribusi. Penelitian kualitas citra ini berbasis Convolutional Neural Network (CNN) menggunakan arsitektur MobileNet. Teknik khusus utama pada algoritma CNN adalah convolution, di mana filter meluncur di atas input dan menggabungkan nilai input + nilai filter pada peta fitur. Tujuan akhirnya adalah CNN mampu mengenali objek atau gambar baru berdasarkan fitur-fitur yang dideteksi. Perancangan sistem dibagi menjadi beberapa tahapan dimulai dari penginputan data citra, tahap selanjutnya adalah preprocessing, pada penelitian ini menggunakan dua jenis preprocessing yaitu filter CLAHE dan Filter Gaussian.Kata kunci: Diameter Kawat; Distribusi Kapasitansi; Wire Mesh Sensor; Convolutional Neural Network (CNN)\nAbstract[Analysis of the Effect of Wire Diameter on the Capacitance Distribution of Wire Mesh Tomography Sensors using a Convolutional Neural Network] Wire Mesh Sensor (WMS) is a tomography-based sensor that produces a distribution image of a fluid flow. The distribution image is a capacitance distribution pattern measured by sensor electrode. From the simulation results, an analysis is carried out on the distribution pattern of the electric potential for know the electrical potential characteristics of the modeled WMS system. It was found that the difference parameters in the form of variations in the type of solution can affect the distribution of electric potential. This is due to there is a difference in the value of the dielectric constant of each type of solution. WMS system performance in detecting anomalies is evaluated by analyzing changes in the capacitance distribution to the effect of change in wire diameter. The simulation results show that the type of fluid in conditions without and with anomalies can be clearly distinguished through the measured capacitance distribution pattern for the entire wire diameter. Wire diameter only affects the distribution image quality. This image quality research is based on Convolutional Neural Network (CNN) uses Mobile Net architecture. The main special technique to the CNN algorithm is convolution, in which the filter slides over the input and combines the input value + filter value on the feature map. The ultimate goal is that CNN is able to recognize new objects or images based on the detected features. The design of the system is divided into several stages starting from inputting data image, the next stage is preprocessing, in this study using two types of preprocessing, i.e. CLAHE and Gaussian filters.Keywords: Wire Diameter; Capacitance Distribution; Wire Mesh Sensor; Convolutional Neural Network (CNN)","PeriodicalId":33722,"journal":{"name":"Fountain of Informatics Journal","volume":" ","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2023-05-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Fountain of Informatics Journal","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.21111/fij.v7i3.9442","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Analisis Pengaruh Diameter Kawat terhadap Distribusi Kapasitansi dari Wire Mesh Sensor Tomography menggunakan Convolutional Neural Network
AbstrakWire Mesh Sensor (WMS) adalah sensor berbasis tomografi yang menghasilkan gambar distribusi aliran fluida. Citra distribusi merupakan pola distribusi kapasitansi yang diukur dengan elektroda sensor. Dari hasil simulasi dilakukan analisis terhadap pola sebaran potensial listrik untuk mengetahui karakteristik potensial listrik dari sistem WMS yang dimodelkan. Ditemukan ada perbedaan parameter berupa variasi jenis larutan yang dapat mempengaruhi distribusi potensial listrik. Hal ini disebabkan adanya perbedaan nilai konstanta dielektrik masing-masing jenis larutan. Kinerja sistem WMS dalam mendeteksi anomali dievaluasi dengan menganalisis perubahan distribusi kapasitansi terhadap pengaruh perubahan diameter kawat. Hasil simulasi menunjukkan bahwa jenis fluida pada kondisi tanpa dan dengan anomali dapat dibedakan dengan jelas melalui pola distribusi kapasitansi yang diukur untuk seluruh diameter kawat. Diameter kawat hanya mempengaruhi kualitas gambar distribusi. Penelitian kualitas citra ini berbasis Convolutional Neural Network (CNN) menggunakan arsitektur MobileNet. Teknik khusus utama pada algoritma CNN adalah convolution, di mana filter meluncur di atas input dan menggabungkan nilai input + nilai filter pada peta fitur. Tujuan akhirnya adalah CNN mampu mengenali objek atau gambar baru berdasarkan fitur-fitur yang dideteksi. Perancangan sistem dibagi menjadi beberapa tahapan dimulai dari penginputan data citra, tahap selanjutnya adalah preprocessing, pada penelitian ini menggunakan dua jenis preprocessing yaitu filter CLAHE dan Filter Gaussian.Kata kunci: Diameter Kawat; Distribusi Kapasitansi; Wire Mesh Sensor; Convolutional Neural Network (CNN)
Abstract[Analysis of the Effect of Wire Diameter on the Capacitance Distribution of Wire Mesh Tomography Sensors using a Convolutional Neural Network] Wire Mesh Sensor (WMS) is a tomography-based sensor that produces a distribution image of a fluid flow. The distribution image is a capacitance distribution pattern measured by sensor electrode. From the simulation results, an analysis is carried out on the distribution pattern of the electric potential for know the electrical potential characteristics of the modeled WMS system. It was found that the difference parameters in the form of variations in the type of solution can affect the distribution of electric potential. This is due to there is a difference in the value of the dielectric constant of each type of solution. WMS system performance in detecting anomalies is evaluated by analyzing changes in the capacitance distribution to the effect of change in wire diameter. The simulation results show that the type of fluid in conditions without and with anomalies can be clearly distinguished through the measured capacitance distribution pattern for the entire wire diameter. Wire diameter only affects the distribution image quality. This image quality research is based on Convolutional Neural Network (CNN) uses Mobile Net architecture. The main special technique to the CNN algorithm is convolution, in which the filter slides over the input and combines the input value + filter value on the feature map. The ultimate goal is that CNN is able to recognize new objects or images based on the detected features. The design of the system is divided into several stages starting from inputting data image, the next stage is preprocessing, in this study using two types of preprocessing, i.e. CLAHE and Gaussian filters.Keywords: Wire Diameter; Capacitance Distribution; Wire Mesh Sensor; Convolutional Neural Network (CNN)