International Journal of Computer Applications (0975 – 8887) Volume 101– No.6, September 2014 43 Experimental Analysis of the Linux RT-patched for Data Acquisition applied to Power Sector Roberto Alexandre Dias Federal Institute of Santa Catarina Mechatronics Tiago Emanoel de Souza Federal Institute of Santa Catarina Mechatronics Valdir Noll Federal Institute of Santa Catarina Mechatronics ABSTRACT The use of SoCs allow a compact and low cost post- processing system for monitoring electrical magnitudes. The main idea of this work is to have a small hardware which is deterministic enough to execute acquisition and sufficiently power to process data at a low cost. In this paper, it is presented two problem domains of electrical systems measurement, one with acquisition of 1 kHz and the other with 5 kHz. It was built an acquisition test bed with a SoC with standard and real-time Linux to compare and evaluate its timing constraints. General Terms Data Acquisition Keywords Real Time, Linux, acquisition system, System on Chip (SoC). 1. INTRODUCTION Data acquisition is important to preventive maintenance and monitoring systems for the utilities industry. Specifically in the energy sector, a demand for most efficient and sustainable power generation systems increases the need for a post processing systems and visualization data. Typical scenarios include acquisition of electrical and mechanical parameters for applications in power sector. The monitoring system for power generators group requires compact, low cost and powerful processing systems with high definition graphical interfaces. The use of System on Chip (SoC) platforms meets this goal. SoCs are not expensive, with high processing power and easy integration with measurement systems. In addition, the use of SoC with Linux Operating System(OS) improves the development curve through availability of open source applications and libraries. In this work a data acquisition system based in a SoC over Linux OS (non real-time and real-time) was evaluated. A test bed was developed to measure the data acquisition time for typical applications in energy sector. 2. SYSTEMS ON CHIP – SOC The progress of microelectronic technology allowed miniaturization and higher performance of electronic components. With these advantages, the amount of features in a single Integrated Circuit (IC) increased. A SoC can be defined as “an IC that integrates all components of a computer or other electronic system into a single chip”[1].The benefits are low cost, low power consumption, ideal for complex systems with space limitations [1]. The block diagram of Allwinner A20 SoC in Figure 1 shows the supported variety of features, processors, I/O interfaces and memories. Today’s best examples of SoC’s use are on smart phones and tablets due to space limitations and low power consumption constraints. Figure 1: Block diagram of Allwinner Tech A20 SoC [9]. Since smart phones and tablets are all about graphic and sound features, it is notable the concern of multimedia processing on SoCs. An example is the A20 that has NEON processor with Single Instruction Multiple Data (SIMD). This processor has parallel processing of 8, 16, 32 and 64 bit and single precision floating point. Which is used for signal processing algorithms such as video encoding/decoding, 2D/3D graphics, gaming, audio, speech and image processing, telephony, and sound synthesis [2]. As this project is about signal acquisition and post-processing, it will target especially the FFTW algorithm, which has optimized implementation for NEON and SIMD technology [3]. 3. THE PROBLEM The intended application for this system is the monitoring of electrical systems such as power generating groups. In addition, the main goal is evaluate if the use of SoC is feasible for solving two problem domains: 1. Acquiring and post processing electrical energy parameters like voltage, current, active power, and reactive power. 2. Acquiring and post processing mechanical parameters in power generators like vibrations in bearings In the first domain the involved signal frequency is around 50 and 60 Hz up to tenth harmonic (500 to 600 maximum frequency). In the second domain, the signal frequency is around 500 Hz up to tenth harmonic (5 KHz maximum frequency).