ENERGIMETER MENGGUNAKAN METODE KALMAN FILTER

AJI PRIATMOKO, : Eka Firmansyah, S.T, M.Eng, Ph.D, Dr. Eng. Adha Imam Cahyadi, S.T., M.Eng.

2017 | Tesis | S2 Teknik Elektro

Permasalahan harmonik telah muncul sangat lama dan merupakan salah satu penyebab terjadinya pengukuran yang tidak akurat. Selain itu adanya white noise pada pembacaan sensor membuat semakin ketidak akuratan dalam pembacaan sinyal, umur komponen yang semakin pendek hingga kerusakan pada alat. Harmonik ditimbulkan karena adanya beban yang tidak linier. Kebanyakan peralatan elektronik sekarang ini menghasilkan harmonik karena peralatan tersebut menggunakan penyearah sebagai salah satu bagian yang menghasilkan harmonik pada arus. beberapa contoh alat tersebut adalah UPS, SMPS, PC dan lain-lain. Penelitian ini menjelaskan metode prediksi harmonik pada listrik. Algoritma Fast Fourier Transform (FFT) sering digunakan namun, memiliki 3 kekurangan yaitu aliasing, leakage dan picket fence effect. Oleh karena itu, dibutuhkan algoritma yang bisa mengatasi 3 hal tersebut. Algoritma Kalman Filter bisa digunakan sebagai solusi untuk memprediksi harmonik. Kalman filter merupakan solusi rekursif tentang masalah filter yang dapat menghilangkan noise dari suatu sinyal yang mengandung informasi dan mengambil informasi tersebut untuk diproses lebih lanjut. Proses dari Kalman Filter terbagi menjadi 2 persamaan yaitu persamaan state dan persamaan output. Perhitungan dari Kalman filter dilakukan dengan meminimalkan rerata kuadrat error. Dari hasil simulasi menunjukkan bahwa dengan menentukan nilai Q dan R yang tepat maka dapat menghasilkan error yang kecil dengan nilai 0:042%. Hasil tersebut memenuhi standar IEC 62053 yaitu Pada standard tersebut dijelaskan batasan saat mengukur arus sebesar 5% dari rating arusnya, persentase errornya harus kurang dari 1% dan beberapa aturan ketat lainnya. Pada penelitian Kf mendapatkan nilai error= 0:042% yang lebih kecil dibandingkan FFT yaitu 14.11%. Dari simulasi dapat disimpulkan bahwa Kalman filter dapat mendeteksi fundamental frekuensi dengan baik dengan error yang kecil. Kalman filter dapat bekerja dengan baik dalam prediksi isyarat fundamental dengan nilai Q=R yang dinamis sesuai dengan galat yang diperoleh.

Harmonic problems have appeared very long time and is one of the causes of inaccurate measurements. In addition, the white noise on the sensor readings to make more inaccuracies in the reading the signal, component life is getting shorter until the damage to the appliance. Harmonics generated because of the burden that is not linear. Most of today's electronic equipment produces harmonics because they use a rectifier as one of part that generates harmonic current. some examples of these tools are UPS, SMPS, PC and others. This study describes the harmonic prediction method on electricity. Algorithm Fast Fourier Transform (FFT) is often used however, have 3 disadvantage that aliasing, leakage and fence effect striker. While the Kalman Filter algorithm used to predict harmonics. Kalman filter is a recursive solution of the problem filter that can remove noise from a signal containing the information and take the information to be processed further. The process of the Kalman filter equations are divided into two state equation and output equation. Calculation of the Kalman filter by minimizing the mean squared error. The simulation results showed that determines the values of Q and R with right value, it can produce a small error with a value of $ 0.042 \% $. These results meet the Harmonic problems have appeared very long time and is one of the causes of inaccurate measurements. In addition, the white noise on the sensor readings to make more inaccuracies in the reading the signal, component life is getting shorter until the damage to the appliance. Harmonics generated because of the burden that is not linear. Most of today�s electronic equipment produces harmonics because they use a rectifier as one of part that generates harmonic current. some examples of these tools are UPS, SMPS, PC and others. This study describes the harmonic prediction method on electricity. Algorithm Fast Fourier Transform (FFT) is often used however, have 3 disadvantage that aliasing, leakage and fence effect striker. While the Kalman Filter algorithm used to predict harmonics. Kalman filter is a recursive solution of the problem filter that can remove noise from a signal containing the information and take the information to be processed further. The process of the Kalman filter equations are divided into two state equation and output equation. Calculation of the Kalman filter by minimizing the mean squared error. The simulation results showed that determines the values of Q and R with right value, it can produce a small error with a value of 0:042%. These results meet the standards IEC 62053 are described In the standard when measuring current limit of 5% of the current rating, the percentage of error must be less than 1% and some other strict rules. In the study Kf get value error = 0042% is smaller than FFT is 14:11%. From the simulations, it can be concluded that the Kalman filter can detect the fundamental frequency of the well with a small error. Kalman filter can work well in the prediction of fundamental cues with a value of Q=R dynamic in accordance with the error obtained.

Kata Kunci : Kalman-Filter, Harmonic.