Industrial and commercial power system automation system based on a hybrid network architecture.

C. S. Gehrke3, F. Salvadori2, M. de Campos1, A. C. Oliveira3, D. R. Hller3 and P. S. Sausen1 1 UNIJU Regional University of Northwestern Rio Grande do Sul State 2 UFPB Federal University of Paraiba 3 UFCG Federal University of Campina Grande camilagehrke@gmail.com; salvadori.fabiano@gmail.com
Abstract Monitoring and control facilities in traditional industrial and commercial power systems are usually separated from protection functions. Basically, this is because, they were designed by different suppliers (e.g. different databases and human-machine interface - HMI). The development in microelectronics, digital electronics, wireless communication, and highly integrated electronics in addition to the increased need for more effective control of power electrical systems, turned the development of supervisory control tools the main object of study for many researchers. This paper proposes a digital system for condition monitoring, diagnosis and supervisory control of industrial and commercial power systems. The system is based on hybrid network architecture, consisting of a wired Infrastructure and a Wireless Sensor Network (WSN). The system is based on two hardware topologies responsible for the signal acquisition, processing, and transmission: Remote Data Acquisition Units (RDAUs) and Intelligent Sensors Modules (ISMs). The basic characteristics of the presented integrated system are: (a) easy and low cost implementation, (b) easy to set up by user, (c) easy implementation of redundant routines (security), (d) easy expandability, (e) portability/versatility, (f) extended network lifetime, and (g) open system.

Providing Central Control; Managing Energy Costs; Maintaining Voltage Level; Maintaining Power Factor; Controlling Generation; Load Shedding. Integrated systems let us avoid severe economic losses resulting from unexpected failures, and improving system reliability and maintainability [3]. There are several hardware and software solutions for implementing integrated systems for the most varied scenarios [4]. Integrated systems consist of a number of devices connected to a computer through a Local Area Network (LAN), usually consists of a shieldes twisted pair of wires [2]. These alternatives are based on proprietary solutions for dedicated systems based on wired communication using cables and various types of sensors. The installation and maintenance of such systems is usually much more expensive than the cost of the sensors themselves. In addition to that, systems are not easily extensible, mainly due to the wired communication itself and the lack of flexibility of the proprietary software. There is an increasing interest in applying technology to protect and control supervisory control systems [5]. Today's Intelligent Electronic Devices (IEDs) and robust communications processors contain large amounts of valuable data that have been available for years but largely overlooked. Initial integration efforts by most vendors focused solely on providing data access and control of supervisory control and data acquisition [6]. This paper presents an integrated industrial and commercial power systems automation based on a hybrid architecture network. This systems is subdivided in three subsystems (see Fig. 1): i) Data Acquisition Subsystem; ii) Communication Subsystem; and, iii) Supervisory Controller Subsystem. The data acquisition subsystem is composed of Remote Data Acquisition Units (RDAUs) and Intelligent Sensor Modules (ISMs). Data communication is based on a hybrid network communication architecture: (1) Wireless - Radio Frequency (RF); and, (2) Wired Ethernet based. The supervisory controller subsystem presents an open source implementation for the Human-Machine Interface (HMI). Our supervision system provides flexibility, fault tolerance, high sensing fidelity, low cost, rapidly response, and

I.

INTRODUCTION

Industrial power users are rapidly becoming aware of electronic monitoring and control systems capable of delivering tangible benefits and significant return on investment. Benefits can generally be classified in terms of energy cost savings, better equipment utilization, and increased system reliability. Electric utilities objective include a high degree of security and adequacy of bulk power supply systems, together with effective and economic operation and maintenance. Electric utilities applications have ranged from Supervisory Control and Data Acquisition (SCADA) systems primarily concerned with remote operations to distribution automation, which focuses on operation efficiency. Today, systems designed specifically for industrial customers of electric utilities provide similar functions but are tailored to meet the specific requirements of an industrial power system [1]. In accord with [2], integrated system may be configured to perform one or any combination of the tasks listed bellow: System monitoring; System protection;

interoperability, making the system an ideal platform for power usage evaluation and condition monitoring altogether, and allowing the construction of high level intelligent power management system in industrial and commercial power systems.

Fig. 2: RDAU block diagram.

Analog inputs Transformers are used for the voltage measurements with a 11:1 relation, and HAIS100 (LEM) for measuring the current. The RMS voltage and current values in the secondary supervision transformer are 75V and 5A respectively. Considering 50% over-voltage and 500% over-current (i.e., short-circuit), the system is able to measure up to 110V and 100A. To achieve better resolution for the nominal current signal, two conditions were developed, one for maximum current of 5A and another for 100A. The transition is implemented via software.
Fig. 1: The proposed system.

II. DATA ACQUISITION SUBSYSTEM Automation processes are extremely associated to Instrumentation and Control. The concept normally used for data acquisition in process control is usually accomplished by placing sensors close to the actual phenomenon [7]. Data gathered by the sensors are then transmitted through a wired communication infra-structure to the processing place. The evolution of sensor technology and communication networks has allowed, to employ, intelligent sensors for improving the processing control. In this case, sensors not only collect data but they also perform some local processing and transmitting their results through: wireless communication (i.e., radio transmission), or for wired communication infra-structure avoiding data redundancy [8]. Concerning to this idea, the system is using two platforms to data acquisition: Remote Data Acquisition Units (RDAUs) and Intelligent Sensor Modules (ISMs). A. Remote Data Acquisition Unit (RDAU) Different RDAUs are responsible for the static monitoring and automation of the industrial and commercial power systems. For this purpose, the RDAU provides six analog inputs and four digital inputs, transmitting the data to a supervisory controller, and if required the platform can operate based on digital outputs. The RDAU system is depicted in Fig 2.

Digital input and outputs Besides the common analog voltage and current inputs, the RDAU has also the capability to analyze four digital inputs: low level, ranging from 0V to 75V, and high level, ranging from 95V to 127V. For the actuation in power systems, the RDAU provides four digital outputs that can be used whenever necessary. TMS320F2812 The TMS320F2812 microprocessor operating at 150MHz clock frequency (i.e., 6.67ns/instruction). The memory architecture is organized as follows: 64kB program memory, 64kB data memory, 18kB RAM, one external memory interface with 1MB, and 128kB Flash ROM. The system also provides an AD converter, PWM, and timers. Software The software was developed based on the C++ language, using the Code Composer Platinum platform (Texas Instruments ). The main routine configures peripherals and interruptions, enabling data acquisition, data processing and data transmission. The data acquisition subroutine is executed in the internal microprocessor AD converter which has two parallel conversion channels with a 12-bit resolution, working with maximum conversion rates of 25 MHz/channel. The sampling is takes places simultaneously for each pair of inputs.

The AD timer is enabled every 130ns, representing 128 samples/cycle and 7680 samples/second for the R, S and T phases of the voltage and current inputs. Aiming at improving the system's performance, some data processing is performed locally to reduce the number of data packet transmissions. The RDAU processing data subroutine performs the computation of the RMS values for the voltage and current phases, active, reactive and apparent three-phase potencies and potency factor, in addition to the computation of Fast Fourier Transformer (FFT). From the RMS values, the operator can analyze the power system behavior (i.e. fault analysis), allowing to cover all rules and international standards. Given that the system has some processing power, it allows running several control algorithms and the implementation of fault detection techniques. All the results obtained from the data processing unit are transmitted through a transmission data subroutine to the Supervisory Controller (SC). This subroutine configures the RS232 communication with two stop bits, no parity, eight data bits, and 115,200 bps data rate, based on the MODBUS protocol. The connection between the RDAU and the SC is performed through serial communication (RS232) or Ethernet. In the next item the possible communication scenarios will be explained. Concerning to the protocol, transmission rate and that data process occurs in the TMS320F2812, the using of the transmission channel is 15%. The Prototype was implemented and can be seen in Fig. 3 and Fig. 4.

B. Intelligent Sensor Module (ISM) Unlike the RDAU subsystem that has continuous processing and sending data, the Intelligent Sensor Module (ISM), Fig. 5, is less powerful, transmitting data between longer intervals with less processing time to save battery power. In this case, the ISM module performs data acquisition of metrics which does not vary so often in small periods of times (e.g. temperature).

Fig. 5: ISM block diagram.

Considering that the ISM can be located anywhere, the system can be battery powered or powered by an external source. In case it is battery powered, four AAA 1.2 V rechargeable batteries are employed. The external source has input voltage of 127V, and output of 5.6V. Besides providing the required voltage, it can also be used to recharge the AAA batteries. The Prototype was implemented and can be seen in Fig. 6.

Fig. 6 ISM Prototype.

To provide wireless communication, addressing ISM mobility issues, a transceiver TRF-2.4G was chosen as the RF communication module.
Fig 3 RDAU Prototype.

III. COMMUNICATION SUBSYSTEM Recent developments in communication technologies have enabled cost effective remote control systems which have the capability of monitoring the real time operating conditions and performances of power systems. Each communication technology has its own advantages and disadvantages that must be evaluated to determine the best communication technology for automation system. In order to avoid possible disruptions in power systems due to unexpected failures, a highly reliable, scalable, secure, robust and cost effective hybrid communication network between system under analysis and a remote control center is paramount. This high performance

Fig 4 RDAU Hardware.

hybrid communication network should also guarantee very strict Quality of Service (QoS) requirements to prevent possible power disturbances and outages. This paper presents a hybrid communication network architecture, including wireless (RF), and wired possibilities, to enable minimum cost and very highly reliable communication for several possible operation scenarios of automation applications. A. RS232 The serial interface (i.e., RS232) is employed when the RDAU is nearby the SC. This way, communication between the two modules takes places without the need of any other communication module. The serial communication is based on MODBUS. This protocol has two transmission modes: ASCII and binary. The binary mode (RTU) is the one used in our system. The end of a message is identified after there is no transmission for a minimum period of 1.5 times a data word transmission. The packet structure (see Fig. 7) includes: Network address; Packet type; Packet data; Error detection mechanism (i.e., CRC, to check the message integrity).

The advantages of wireless communication include their commercial acceptance and their application in environments with physical barriers. Some aspects affect the energy consumption of the radio, including type of modulation, data transfer rate and transmission energy. A usual approach in the applications of wireless sensors is that their duty cycle is very short (i.e., around 1%). Nodes can schedule their events and remain in sleep mode while there is no need to be active, saving battery power and extending the network lifetime. Many micro-controllers have an asynchronous oscillator circuit, which remains running while the core and the peripherals are inactive. This capability makes it possible to return from a sleeping state in just a few microseconds. The radio has four operation modes: Transmission (TX), Reception (RX), IDLE and SLEEP. The choice among different transceivers was based on two aspects: the frequency clock and the energy consumption for the different operation modes. The transceivers that operate in 2.4 GHz are easily available and are less prone to noise. Due to their high data rate the energy consumption for a single bit transmission is very low. However the same speed is also required from the microcontroller. This problem is addressed in the nRF2401 transceiver through the application of a data buffer that adapts a low performance micro-controller (e.g., 10 Kbps) with the data rate transmission of the radio (i.e., 1 Mbps) [8]. The main characteristics of the transceivers analyzed in our research are presented in Table 1.

Fig. 7: MODBUS RTU framing. TABLE I A COMPARATIVE BOARD OF ENERGY CONSUMPTION FOR SOME COMMERCIAL TRANSCEIVERS

The communication follows a master-slave approach, where only one device (the master) can query the other devices (slaves). The slaves reply back sending the data required by the master. The master can address one slave at a time or all the network devices through diffusion messages (i.e., broadcasting). B. Radio Frequency (RF) In a second scenario where the monitored data is distant and there is not wired communication infrastructure, data gathered from the monitoring devices can be transmitted via a wireless interface, which is easy to implement and a way cheaper compared to standard wire communication (e.g., via fiber optic) [4]. The TRF-2.4G transceiver developed by LAIPAC uses an nRF2401 component (NORDIC Semiconductors). It uses GFSK modulation with a data transmission rate of up to 1 Mbps. The transceiver unit is composed of an antenna, a frequency synthesizer, an amplifier, a crystal oscillator, and a modulator/demodulator. The transmission power can be configured ranging from -20 dBm to 0 dBm, reaching distances up to 250 m (without obstacles).

Model

NordicVLSI/nRF2401 Conexant/CX72303 Conexant/RF109 Ericsson/PBA31301 Ericsson/PBA31305

Tx rate (kbps) 1000 1000 1200 1000 1000

Freq. (MHz) 2400 2400 2400 2400 2400

Sleep (uA) 1 1 5 65 70

Idle (mA) 0.5 0.02 25 21 35

Rx (mA) 15 24 89 40 50

Tx (mA) 6.5 11 31 32 60

C. Ethernet (TCP/IP) Third scenarios, where the monitored data are distant from the SC and a wired infrastructure have already been available. The objective of this module implementation is to make possible the use of wired networks already existing or then in places where the rank of antennas is made it difficult [6]. In this scenario the DE311 from B&B Electronics module is being used. This module provides a data communication solution for connecting Windows and Unix/Linux hosts to asynchronous serial devices over a TCP/IP based Ethernet. As the input data is from asynchronous serial, and RDAU provides RS232, the communication between DE311 to RDAU is directly. The protocol is based on sockets interface. Such technology, improved by the University of California, in Berkeley, from 80s

on, made possible a UNIX implementation of the sockets package to TCP/IP protocols. Nowadays, it is the most usable method to access a TCP/IP network and Internet Originally the sockets were developed as the BSD (Berkeley Software Distribution) UNIX Operational System intrinsic parts. As a consequence, they use lots of concepts found in other Kernel routines. Particularly, sockets are integrated to the I/O routines. This integrated system main advantage lies in its flexibility: can be written a unique applicative that transfers data to an arbitrary localization. A socket is, primarily, a transparent data connection between two computers linked on a net. It is identified by the computers network address as well as by a door localized in each computer. Socket represents a TCP/IP network connection point. When two computers want to maintain a conversation, each one of them uses one socket. In this procedure a computer named server opens a socket, here specifically the Supervisory Controller, and makes a connection check. The RDAU, sends a signal to the server socket aiming to start a connection. In order to establish a connection just the destiny address and the port number are necessary. In the TCP/IP specific door numbers are reserved to specific protocols, for instance, 25 to SMTP (Simple Mail Transmission Protocol) and 80 to HTTP (Hyper Text Transfer Protocol). Sockets present two main operation modes: the connection based mode and the without connection mode. The based connection modes work as a telephone; they have to establish a connection while interrupting the call. Everything that flows between these two events arrives in the same order as it was transmitted. On the other hand, the mode without connection does not guarantee the deliver, and the mail different items can arrive in a different order from that they were transmitted.

Part 1 Monitoring Allow the user to show the graphical representation of all the variable of the platform, with real time values of voltage and current as well as powers values, Fig 8.

Fig. 8: HMI Main Interface.

Part 2 - Alarms Include a sequential listing of alarms and changes of state. This list allows to the operator to take note of any disturbance of the power system, as sag, swell, overcurrent and overload. The list of alarms can be seen in Fig. 8, however, the meters issue alerts in real time, shown in Fig. 9.

Fig. 9: Alarm warning of a meter.

IV. SUPERVISORY CONTROLLER SUBSYSTEM (SC) The Supervisory Controller Subsystem was developed in order to emphasize data presentation versatility. It runs on standard PC architecture using Linux or Windows Operational System, once the interface was developed using JAVA language, and IBM-DB2 Express-C and MySQL to database. The application was developed to support on-line data presentation through reading of the received data files, including as well as a Database system, which stores the data. Since the main function of the SC subsystem is receive and store the data derived from the RDAUs, it can process and present them to operators through: i) SCADA (proprietors systems); or, ii) Human Machine Interface (HMI) specially designed for this purpose. A. Human Machine Interface (HMI) The HMI, Fig. 8, used in the SC as mentioned before, was developed in JAVA language and is divided in four parts described follow:

Part 3 Database The measurement of voltage and current are stored in a hard disk (HD) permitting after analyzes anytime that is necessary. Part 4 Historic graphics and reports This function was developed for the user can observe the evolution of the current, voltage, and power values in a determined time (see Fig. 10 and Fig.11).

Fig 10.Spreadsheet Report.

A. RDAU Maximum practicable data acquisition The RDAU improved the sampling rate, because it employs a DSP (TMS320F2812). In this case, the RDAU could achieve 128-samples/cycle, translating to 7680-samples/s. These results allow analyzing up to the 64th harmonic. Considering that the RDAU is used for the signals acquisition with fundamental frequency of a 60Hz, and as performs an A / D conversion for each 260ns, comes to a sampling frequency of 7.68kHz, resulting in a maximum bandwidth, or the Nyquist frequency of 3.84kHz. These values fully attend the technical standards for the monitoring of electrical power substations. B. ISM maximum achievable data acquisition The Intelligent Sensor Modules are used only for the acquisition of signals with large constant times, and therefore do not need a high processing capacity (e.g. temperature). The maximum capacity of acquisition of ISM was 12-samples/cycle. C. RDAU/ISM - Transmission Rate Both the RDAU and the ISM using the same transceiver for communication (TRF 2.4G). Radio Frequency (RF) communication is well suited for environments with physical barriers. However, it is prone to interference. D. RDAU/ISM - obstacles susceptibility The ISM/RDAU systems were evaluated for two different scenarios (with and without obstacles). Preliminary results for the first scenario show that the sensor node is able to transmit without any packet losses up to a distance of 80m, regardless of the antenna orientation. At a distance of 90m, and with the antennas directed at each other, we notice around 10%packet losses. Distances larger than 100m showed an unacceptable number of packet losses. As for the second scenario, we noticed that obstacles such as a 30cm concrete wall allows transmitting to distances only up to 60m without any significant packet losses.
Fig. 12. Parameters.

Fig 11. Historic Graphics.

Part 5 - Parameters To a better versatility the parameter of sag and swell can be configured by the user. In this form the nominal values and alarm enabling will run under the usual norms. The configurable parameters are: nominal voltage, maximum swell, minimum sag and overcurrent. (See Fig. 12)

V. EXPERIMENTAL RESULTS To evaluate our systems performance, we have defined a set of experiments for analyzing: Remote Data Acquisition Unit (RDAU) maximum practicable data acquisition and obstacles susceptibility; Intelligent Sensor Module (ISM) - maximum practicable data acquisition, obstacles susceptibility and maximum lifetime batteries; Communication (Ethernet (TCP/IP) and Wi-fi) transmission rate Human-Machine Interface (HMI) - easy of understanding information and time response;

E. RDAU/ISM - maximum lifetime batteries Considering that a WSN is energy constrained, it is paramount to reduce the consumption of energy due to modulation, filtering, and demodulation. The energy consumed by the radio depends on the type of modulation, data transfer rate, and transmission energy.

VI. CONCLUSIONS This paper presents a system for automated industrial and commercial power systems based on the concept of hybrid architecture involving the use of network with infrastructure network and WSN acting together or separately. Several scenarios could be analyzed considering the infrastructure network or WSN connection. The tests objective to verify the operation of the three subsystems that make up the system for substation automation. For data acquisition subsystem the tests were conducted aiming to verify the signals bandwidth acquired by RDAU. This analysis was performed from the Nyquist Theorem and Fourier

The RDAU/ISM systems were evaluated for different scenarios: monitoring a power transformer with 23kV input voltage, and 380/220V output voltage, and a 75kVA nominal power.

Series. Concomitant use of these methods has the ability to verify analysis of harmonics of the unit. The analysis of the RDAU bandwidth presented results consistent with the established standards [9][10]. Whereas the RDAU bandwidth is 3.84kHz. It was found that the RDAU to identify until the 64th harmonic. The ISM tests were performed in this case: the distance that the module is capable of transmitting and the number of packets sent X time. Preliminary results show that the sensor node is able to transmit without any packet losses up to a distance of 80m. At a distance of 90m, and with the antennas directed at each other, we notice around 10% packet losses. Distances larger than 100m showed an unacceptable number of packet losses. With obstacles such as a 30cm concrete wall allows transmitting to distances only up to 60m without any significant packet losses. ACKNOWLEDGMENT The authors gratefully acknowledge the financial support supplied by the Companhia Estadual de Distribuio de Energia Eltrica - CEEE, Rio Grande do Sul State, Brazil.

REFERENCES
R. A. Kennedy and D. D. Rickey, Monitoring and Control of Industrial Power Systems. IEEE Computer Applications in Power, [2] J. A. Bright and Wei-Jen Lee, Integrated Monitoring, Protection, and Control Systems for Industrial and Commercial Power Systems, IEEE Transactions on Industry Application, vol. 36, no. 1, January/February 2000. [3] B. Lu, L. Wu, T. G. Habetler, R. G. Harley, and J. Gutirrez, Wireless sensor networks: a survey. IEEE Transactions on Computer Networks, vol. 38, no. 7, pp. 393422, july 2002. [4] F. Salvadori, P. S. Sausen, L. Hartmann, M de Campos, E. Padoin, G. V. Leandro, Acquisition and Transmission Data Monitoring System Applied to Energy Substation. I Industrial Informathics - IEEE, 2003, Calgary. Proceedings of the I Industrial Informathics, 2003; [5] G. T. Heydt, C. C. Liu, A. G. Phadke, V.Vittal, Solutions for the Crisis in Electric Power Supply IEEE - Computer Applications in Power. Vol. 14, Issue 3, pp.23-30, July, 2001. [6] Dolezilek, D. McDermott, B. Remote Data Monitoring and Data Analysis for Substations: A Case Study in Implementation. IEEE Power Systems Conference: Advanced Metering, Protection, Control, Communication, and Distributed Resources, PS '06, pp. 496-500. Clemson, SC, 2006. [7] I. F. Akyildiz, W.Su, Y Sankarasubramaniam, E. Cayirci, Wireless sensor networks: a survey, Computer Networks, Vol. 38, pp.393-422, July 2002. [8] F. Salvadori, M. de Campos, R. C. de Figueiredo, C. S. Gehrke, C. Rech, P. S. Sausen, M. A. Spohn and A. C. Oliveira, Monitoring and Diagnosis in Industrial Systems Using Wireless Sensor Networks with Dynamic Power Management, IEEE International Symposium on Intelligent Signal Processing, October 3-5, 2007, Alcal de Henares (Madri), Spain, 2007. [9] IEEE Std 1613-2003 - IEEE Standard Environmental and Testing Requirements for Communications Networking Devices in Electric Power Substations. 2003. [10] IEEE Std C37.1-1994 - IEEE Standard Definition, Specification, and Analysis of Systems Used for Supervisory Control, Data Acquisition, and Automatic Control. 1994. [1]