Hybrid Communication Network Architectures for Monitoring Large-Scale Wind Turbine

  • Ahmed, Mohamed A. (Dept. of Computer Engineering, Chonbuk National University) ;
  • Kim, Young-Chon (Smart Grid Research Center, Chonbuk National University)
  • Received : 2013.09.13
  • Accepted : 2013.09.23
  • Published : 2013.11.01


Nowadays, a rapid development in wind power technologies is occurring compared with other renewable energies. This advance in technology has facilitated a new generation of wind turbines with larger capacity and higher efficiency. As the height of the turbines and the distance between turbines increases, the monitoring and control of this new generation wind turbines presents new challenges. This paper presents the architectural design, simulation, and evaluation of hybrid communication networks for a large-scale wind turbine (WT). The communication network of WT is designed based on logical node (LN) concepts of the IEC 61400-25 standard. The proposed hybrid network architectures are modeled and evaluated by OPNET. We also investigate network performance using three different technologies: Ethernet-based, WiFi-based, and ZigBee-based. Our network model is validated by analyzing the simulation results. This work contributes to the design of a reliable communication network for monitoring and controlling a wind power farms (WPF).

1. Introduction

Nowadays, most wind power farms are located in remote areas such as in mountains and offshore. While these locations are characterized by richer wind resources, they have a common problem with accessibility. In many cases, the only way to reach the wind turbines is by boat or helicopter, which results in higher maintenance costs. In addition, scheduled maintenance may be postponed due to the dependence on weather conditions. Therefore, a reliable communication network is needed to continuously monitor the WPF and transfer the monitoring data with higher accuracy.

The supervisory control and data acquisition (SCADA) system represents the vital element of WPF, where it enables the control center operators to monitor/control the entire WPF (turbines, substations, and meteorological masts) from a central location. The SCADA system consists of three major components: SCADA master station, remote terminal units, and communication network. The communication network plays an important role in the system as an electric infrastructure, where its failure may cause a serious problem in system operation and control [1]. Conventional communication networks are wired-based architectures, where different wired sensors are installed on different parts (rotor, generator, nacelle, etc.) to collect information. The data collected from the wired sensor nodes, measurement devices, and other sources are processed at the control center, then appropriate commands are executed to manage and control the wind turbines. Wireless technology could be incorporated into the WPF network design as an alternative or backup, where it could offer many advantages compared with wired sensors, such as being easily deployed, having flexible configurations, a reduced installation and operation cost, and communication to remote locations that may not be accessible by conventional wired networks [2].

With the recent advances in wireless technology, smart sensors have emerged as an enabling technology for structure health monitoring (SHM) applications. Smart sensors have the capability of sensing, data acquisition, data processing, and wireless functionality. As wireless technology becomes more popular and matures, several studies are being carried out on using the wireless sensor network for WPF monitoring.

This work considers hybrid communication network architectures for large-scale wind turbines where wireless communication is used inside the wind turbine and an external wired link is used to connect the wind turbine and the control center. Many factors need to be determined in order to design such architectures, such as the types of sensor nodes used, the number of sensor nodes needed inside the wind turbine, the amount of traffic generated from the sensor nodes and how to calculate this. Also, detailed knowledge about measurement devices and applications inside the wind turbine are needed, which may affect the network performance since all types of traffic share the same communication network.

Many authors have addressed the concept of using wireless sensors for WPF monitoring. In [3], the author reviews the technical possibilities of building a wireless sensor network for WTs, including measured parameters of temperature, humidity, wind speed, and heading angle. The authors of [4] described wireless network architecture for monitoring a wind energy generation system. They used a wireless local area network (WLAN) for communication between individual WTs, while worldwide interoperability for microwave access (WiMAX) is configured between the WTs and the control center due to its long range capability and quality of service. Also, the authors in [5] discuss the problems of a wireless sensor network combined with structural health monitoring, including the compatibility between different sensors, the sampling frequencies, the transmission bandwidth, and the energy consumption. Most previous works only considered a few sensor nodes for monitoring the behavior of the wind turbine. However, it is extremely difficult to analyze the behavior of damage with a few sensor nodes [2]. This means that the measurements from a large number of distributed sensor nodes are required for descriptive analysis of the WT.

In the case of real system performance, the authors in [6] installed wireless sensors in three different wind turbines. Two wind turbines were 78 m high (2MW Vestas V-80) and one turbine was 40 m high (250kW NEG Micon). The results of studying the dynamic behavior and response under loading conditions showed that there is very good agreement between the wireless collected data and the wired-based system. The authors in [7] present the design of a data acquisition system (DAS) for a 660kW (Vestas V47) wind turbine. Due to the limitations of the DAS cards, they selected the most critical parameters in the WT for analysis. Also, they divided the parameters into two groups: low sampling rates and high sampling rates. Also in [8], the design consideration of the WPF communication network, based on Ethernet technology, is given in more detail, including environmental issues, redundancy, remote connectivity, physical medium requirements, and network management.

To the best of our knowledge, no simulation model exists for the communication network of a wind turbine or wind power farm. Most previous works consider only a few sensor nodes for monitoring the behavior of the wind turbine which is not recommended in [2], while the other works describe the communication network of existing WPF projects [8]. The main objective of this paper is to design an appropriate wireless internal network to transmit the sensing data inside the wind turbine using different technologies such as Ethernet-based, WiFi-based, and ZigBee-based. Also, an external network is designed to be placed between the turbine towers through a wired link which is connected to the internal network. There are five different sub-networks inside the wind turbine tower: analogue measurement, status information, protection and control devices, video surveillance, and internet connection. We explored the network architectures, topologies, and technologies inside the wind turbine. The data types and numbers inside the wind turbine are defined according to the IEC 61400-25 standard. The data transmission rates are computed for different monitoring devices according to the sampling frequency and number of channels. We evaluated and validated the performance of the proposed network model using OPNET.

This work is related to the communication network architecture for WPF monitoring. The paper is organized as follows. Section 2 describes the related work, where focus is given to the IEC 61400-25 standard. In section 3, the proposed wind turbine network model is given in more detail. Section 4 explains the wind turbine traffic model. Section 5 shows the OPNET simulation results. Finally, section 6 concludes the work.


2. Related Work

2.1 Conventional communication network

The conventional WPF communication network is shown in Fig. 1, where individual wind turbines are connected through a wired link to the control center servers. The Ethernet network with (Transmission Control Protocol/Internet Protocol) TCP/IP has proven to be the most suitable communication network from the WTs to the control center and within the WT itself [9].

Fig. 1.Conventional WPF network structure

The WT controller of each turbine is responsible for continuously monitoring its conditions using wired links. All necessary data are collected from different parts (nacelle, rotor, transformer, generator, etc.) using different types of sensors. The WT controller then transfers the data to the control center through the communication network. The control center processes the received data and appropriate decisions or commands are sent back through the communication network in order to control the WT.

The conventional communication network using a traditional sensors connection requires expensive cabling for both data transfer and the electric power supply. Also, the wired cables (data or power) consider one source for sensor faults and machine downtime. Therefore, using wireless rather than wired sensors could avoid the high costs of network wiring and allow for the connecting of a large number of sensors, which increases the safety and reliability of the system.

2.2 IEC 61400-25 standard

The IEC 61400-25 standard is an adaptation of the IEC 61850 standard series, developed for monitoring and control of wind power plants. It enables components from different vendors to communicate easily with other components, at any location and at any time. The standard only defines how to model the turbine information, information exchange, and mapping to specific communication protocols. It does not provide a definition of how and where to implement the communication interface [10].

The standard defines the logical device (LD), which is decomposed into logical nodes (LNs). According to the standard, the WT is viewed as an LD with different parts (rotor, generator, converter, transformer, etc.), where each part could be viewed as an LN as shown in Fig. 2. The turbine LNs listed in Table 1 are classified into mandatory (M) and optional (O). The WT model shall include all the mandatory logical nodes. Regardless of the optional LNs, it is highly recommended that all logical nodes are considered. According to the IEC 61400-25 standard, we considered that the WT consists of nine LNs, as shown in Fig. 2: WROT, WTRM, WGEN, WCNV, WNAC, WYAW, WTOW, WTRF and WMET. These LNs are well known functions and are modeled by a virtual model related to a real device. For example, LN WROT relates to the WT rotor while WMET relates to wind power plant meteorological information. Table 2 shows the data classes for common equipment at the nacelle that are classified into analogue, status, and control information.

Table 1.Logical nodes in a wind turbine

Fig. 2.Wind turbine model with logical nodes

Table 2.Types of data and attributes of the nacelle

2.3 Communication technologies

The critical part of a wind turbine condition monitoring system is the communication network, which must be secure, robust, and reliable. Two different communication technologies, i.e. wired and wireless could be used for connection between the wind turbines and the control center. It is critical to select the best communication network protocols inside the WT in order to maintain the real time data transmission in the system and also to meet the requirement of industrial application. Recently, different wireless technologies have emerged with the SHM of wind turbines. IEEE 802.11 (WiFi) and IEEE 802.15.4 (ZigBee) are selected among these technologies for the wind turbine communication network in this work.

A) WiFi is referred to as the communication standard for WLAN and is sometimes referred to as wireless Ethernet. IEEE 802.11a, IEEE 802.11b, and IEEE 802.11g are the most commonly used protocols in today’s environment. IEEE 802.11 defines two types of services: the basic service set (BSS) and the extended service set (ESS). A BSS without an access point (AP) is called an ad hoc mode, while a BSS with an AP is called an infrastructure mode [11].

The selection of WiFi for a wind turbine communication network is based on Ref. [12], where the author uses practice tests of prototype installations in the turbine hub and nacelle to demonstrate that WLAN (IEEE 802.11 b/g) works successfully, despite the harsh environment. Also, the prototype tests showed that the communication was stable and reliable. Therefore, IEEE 802.11 is selected as it provides robust, high speed communication, security, and bandwidth requirements. Current WiFi technologies based on IEEE 802.11a and IEEE 802.11g can provide a high data rate of up to 54Mbps. We selected 802.11g as it operates in the industrial, scientific, and medical (ISM) band of 2.4GHz [13].

B) ZigBee is developed for a low rate wireless personal area network (LR-WPAN) and offers a data rate of up to 250Kbps. It supports different network topologies, include star, ring, and mesh which could provide a coverage range of 10-100m. Zigbee devices could be configured in two different modes: full function device (FFD) and reduced function device (RFD). The RFD device is characterized by lower cost due to its limited resources allocated in view of processing capability and communication range. ZigBee is considered an ideal option for smart grid applications due to its low cost of deployment, robustness, simplicity, low bandwidth requirements, and operation within the unlicensed spectrum [14]. The limitations of ZigBee for practical implementations are its small memory size, low data rate, short range, and low processing capability.

Selection of ZigBee for wind turbine communication network is based on Ref. [15]. The authors implemented a data transition interface for a small-scale wind power generator. After long-term practical running, the system proved to be successful and stable. Table 3 presents the network characteristics of WiFi and ZigBee technologies.

Table 3.Wireless network characteristics


3. Proposed Wind Turbine Network Model

In order to design the wireless communication network inside the turbine nacelle, the dimensions of the WT need to be established in order to determine the coverage area for the wireless network. In this work, we considered the Vestas-V164-7MW, one of the largest company products developed for offshore installation. The production of this type is expected to begin in 2015 [16].

The turbine rotor diameter (RD) is 164 meters (80 meter blades), tip height (TH) is 187 meters at the highest blade tip position, and the nacelle dimensions are 7.5m × 24m × 12m for (height, length, width), as shown in Fig. 3.

Fig. 3.Dimension of Vestas V164- 7MW

The proposed communication network architecture shown in Fig. 4 supports five different sub-networks inside the WT; three sub-networks (analogue measurements, status information, and protection network) are sending periodic data traffic to the control center, while the other two subnetworks (internet connection and video surveillance) are only on demand. Logically, the proposed WT communication network model consists of three different levels; instrumentation and sensors level, monitoring and control level, and the control center level.

Fig. 4.Schematic view of proposed WT network model

Level 1: sensor nodes monitoring devices are attached to the wind turbine parts e.g. rotor, nacelle, transformer etc. The instrumentation and sensor nodes collect the operation data from WT parts and other auxiliary systems. At this level, the sensor nodes have the capability of sensing, data acquisition, data processing, and wireless functionality. Sensing data is transmitted to the coordinator in the case of ZigBee-based architecture and to the access point in the case of WiFi-based architecture.

Level 2: The data collection unit (DCU) is responsible for collecting data from the sensor nodes and at the same time, transmitting the control signals and command from the control center level to the actuators. Other types of intelligent electronic devices (IEDs) may be included at this level, such as a merging unit (MU) IED, circuit breaker (CB) IED and protection and control (P&C) IED. A wired link is used to connect the collected data from the turbine nacelle to the protection and control devices at the tower base.

Level 3: the control center is responsible for managing and controlling the entire system. It includes the SCADA station and different system servers. Using a wired link, the hybrid (wired- wireless) internal network inside the turbine is connected with the external wired network (between individual WTs) in the wind power farm.

In the case of the metering server at the control center side, the received data are classified into different categories: electrical measurements (voltage, current, power, power factor, frequency), mechanical measurements (rotor speed, pitch angle, oil level, torque, displacement, vibration, temperature), and meteorological data (wind speed, wind direction, temperature, humidity, pressure).

The following section gives more detail on wind turbine traffic modeling for different sub-networks including the metering network (analogue measurements and status information), video surveillance system, protection and control, and internet connection. The communication network model is configured with hybrid architecture: wireless (WiFi or ZigBee) for level 1 and wired-based for levels 2 and 3.


4. Wind Turbine Traffic Model

4.1 Modeling of metering network

The metering network represents both the analogue measurements and status information. Fig. 5 shows the modeling of the metering network located at the WT nacelle based on LN concepts. Each LN transmits wireless status information and wireless analog measurements directly to the data collection unit (1) located at the nacelle.

Fig. 5.Metering network subnet (Nacelle)

We assumed that LNs are distributed around the DCU (about 24m × 12m), as we are limited by the nacelle dimensions. As different network topologies could be configured for the data collection in level 1, the ZigBee network was configured as a star topology. The WiFi network was configured to work in an infrastructure mode. Table 4 shows the calculated transmitted traffic from different LNs in a standalone WT [2, 17]. Note that we used a different PAN-ID for the ZigBee-based architecture and configured the BSS-ID for WiFi-based architecture to ensure that the sensors data from different logical nodes do not interfere with each other.

Table 4.Calculated Logical nodes data transmission

4.2 Modeling of protection and control network

The authors in [13] showed that a wireless LAN can be used to enhance the distributed protection and automation in a power system such as transformer differential protection and line differential protection as shown in Fig. 6. Applying the same concept for WT protection, we assume that the P&C device is modeled by one wireless CB IED, one wireless MU IED, and one wireless P&C IED [18], [19]. In Fig. 6, the P&C IED is sending updated meter values of 76,800 bytes/sec and CB IED is sending a status change value of about 16 bytes/sec to the second access point (AP2) located at the bottom of the tower. Data is transmitted over a wired-based link (100BaseT) to the protection server located at the control center.

Fig. 6.Wireless LAN network for inter-substation and distribution substation

4.3 Modeling of video surveillance network

The security SCADA system provides a CCTV imagery video to the operator at the control center using the surveillance camera. Also, it enables the operator to observe both inside and outside the turbine nacelle with the naked eye without the need to visit the site [8], as shown in Fig. 7. The thermal imaging camera (IR-camera) is a new technology for monitoring the electrical systems, transformers, gearbox, mechanical brakes, and fire detection in WTs [20]. This technology can be used for both continuous monitoring and for a short time during service and maintenance. The temperature distribution of the gear box using the IR camera is shown in Fig. 7.

Fig. 7.Video surveillance subnet (Nacelle)

The multimedia services have a high transmission bandwidth feature which may affect other network traffic. We considered only one surveillance camera with the resolution of capture image of 128*240 pixels and the transition frame rate is 15 frames per second as shown in Table 5. The data rate can be calculated as follows:

Table 5.Video surveillance camera data

The calculated data rate is about 4.147Mbps. In our network model, we did not apply any video compression standard for the raw data and the scenario is configured in OPNET to support wireless the data transmission of a high resolution video.

4.4 Modeling of control center

In our network model, the control center is modeled by 1 Ethernet switch and 4 different servers: 1 web server, 1 video server, 1 P&C server, and 1 SCADA server as shown in Fig. 8.

Fig. 8.Control center subnet

Note that both APs are connected through a wired link to the WT main switch. Also, the web server located in the control center offers a secure internet connection for maintenance purposes, which enables the maintenance engineer to connect to the internet for downloading specific software and upgrading or configuring files. Table 6 provides a summary of different application traffic inside the wind turbine tower.

Table 6.Total traffic of a wind turbine


5. Simulation Results

5.1 Assumption and metrics

The OPNET modeler is used to evaluate the performance of the hybrid network architecture for a standalone WT [21]. In order to simulate a more realistic communication network model, OPNET is configured with real wind turbine dimensions. The network performance is evaluated in terms of the following metrics:

End-to-End delay (ETE or Latency): is the amount of time (in seconds) for data to be delivered from the source to the destination along the communication path.

Server FTP traffic received (bytes/sec): represents the average bytes per second forwarded to the FTP application by the transport layer in the server node.

We built an Ethernet-based model for the WT with the same specification to be used as a reference, in order to compare it with the performance of the hybrid network architecture [22]. In all simulation scenarios (Ethernet-based and hybrid architecture), we configured the wired-link as fast Ethernet (100BaseT) along the network path inside the turbine tower and between the WT and the control center. Table 7 shows the simulation assumptions.

Table 7.Simulation assumptions

Inside the turbine, different sub-networks are located; each sub-network has a different latency according to its applications. For example, the protection information and commands between IEDs will require the network latency to be lower than the SCADA information messages between electrical sensors and the control center. If the communication delay exceeds the required time window, the information no longer serves its purpose, and damage may be incurred [23]. Table 8 shows the IEEE 1646 message classification in the power system with different categories and different delay requirements. We will use the requirements as listed in Table 8 to evaluate our network model. The performance is firstly evaluated for the individual simulation scenarios and the hybrid network architecture is then given.

Table 8.Communication timing requirements for electric substation automation (IEEE 1646 standard)

5.2 Individual network simulation

A) ZigBee-Based Architecture

The ZigBee-based network architecture for the wind turbine model is configured with a star topology consisting of a PAN coordinator and end devices. In order to validate the simulator, the calculated amount of traffic transmitted from the sensor nodes will be compared with the simulation results from OPNET. If we consider LN WROT as an example, the number of sensor nodes and monitoring devices distributed and installed in the turbine rotor is 14, with a total generated traffic of 5,136 bits/sec. A detailed description of the data type and number of sensors for the LN WROT is given in Table 9. The traffic received at the WROT coordinator as shown in Fig. 9 agrees with the calculated traffic, which validating the network simulator for the LN WROT.

In order to configure the ZigBee-based architecture, each LN is represented by a unique identifier from PAN-ID 1 to PAN-ID 9 to prevent interference between PANs. Six PAN coordinators among the nine LNs were able to successfully receive the amount of generated traffic from sensors. But, it was difficult to receive the traffic from the LNs of WGEN, WTRF, and WCNV due to the high speed of the

Table 9.Amount of data transmission of LN (WROT)

sensing data transmitted to the coordinators. Fig. 9 shows the successful received traffic at the PAN coordinators for different LNs. The results agree with the calculated LNs data transmission in Table 4.

Fig. 9.Received traffic at PAN coordinators

The calculated amount of traffic from LNs WGEN, WTRF, and WCNV represent 590,112bps, 592,480bps, and 589,920bps, respectively. The reason why these generated amounts of traffic from the three LNs did not successfully receive at the coordinators is due to the limitation of the maximum data rate of ZigBee, which is about 250Kbps. The current or voltage sensors participate with about 589,824bps among the total generated traffic. In order to model the remaining three LNs, we did not consider the voltage and current sensors. Fig. 10 shows the successful received traffic at the PAN coordinators for the three LNs WGEN, WTRF, and WCNV. In this case, the calculated amount of traffic of LNs WGEN, WTRF, and WCNV are 288bps, 2,656bps and 96bps, respectively, which concurs with the results obtained from Fig. 10.

Fig. 10.Received traffic at PAN coordinators

The total end-to-end delay of the modified ZigBee-based architecture for the 9 LNs is shown in Fig. 11. The highest ETE delay is about 10ms for LN WTRM and about 4ms for LN WCNV. The remaining WT LNs are less than 4ms. Based on the timing requirements of monitoring and control information shown in Table 8, the proposed network model satisfies the requirement of the power system with a total ETE delay of less than 16ms. In spite of the ZigBee agreement with the timing requirement of the metering network, we will exclude ZigBee architecture from comparison with WiFi-based architecture and Ethernet-based architecture. This exclusion is due to the limitation associated with the three LNs WGEN, WTRF, and WCNV due to the higher generated traffic (about 589,824Kbps) from the voltage and current sensors in each LN.

Fig. 11.ETE delay for WT LNs (ZigBee-based)

Considering the ZigBee-based wireless network model of the wind turbine, we can conclude that it was possible to monitor the LNs with low speed sensing data such as WROT, WTRM, WNAC, WYAW, WTOW, and WMET. Furthermore, it was difficult to monitor LNs with high speed sensing data of WGEN, WTRF, and WCNV. In terms of the delay characteristics for WT LNs, the WTRM was the highest ETE delay value of about 10ms, while other LNs were less than 4ms.

B) WiFi-based Architecture

The WiFi network configuration differs from the ZigBee-based architecture. In the case of ZigBee, one LN is modeled with one PAN coordinator and different end devices, while in WiFi-based architecture the LN is modeled by only one workstation. This workstation is configured to transmit different traffic profiles wirelessly to the AP located in the turbine nacelle. The AP will provide

Table 10.Received traffic from sensor nodes at SCADA server

Fig. 12.Received traffic at SCADA server (WiFi-based)

the connectivity for all turbine LNs. All traffic profiles are configured and assigned to LNs according to the IEC 61400-25 standard. Both the AP and LNs are configured with the same BSS-ID in order to operate in the infrastructure mode.

Fig. 12 shows that the received amount of traffic at the SCADA server for different measurements agrees with the calculations of Table 10. In the case of tower displacement, the number of monitoring devices is two devices, located at LN WNAC (WT nacelle), with each device sending 40 bytes/sec, and the total received data at the server is then 80 bytes/sec. The same results for status information, temperature, frequency, and humidity are as shown in Table 10. Fig. 13 shows the amount of received traffic at the protection and control server from the P&C IED located at the turbine tower. During the 20 minutes of simulation, the received traffic using WiFi was stable and this concurs with the traffic calculation of Table 6.

We configured the video camera to work for 5 minutes. During the simulation time, the video camera was requested for video surveillance with a start time offset of 300 seconds and duration of 300 seconds. In order to configure the video streaming in OPNET, we configure the attribute of the video conference to support the outgoing stream frame size only. The results shown in Fig. 14 demonstrate that the WiFi network is capable of receiving the video surveillance traffic during the configured time.

Based on the results of the WiFi-based architecture, it was possible to validate the network simulator by the

Fig. 13.Received traffic at P&C server (WiFi-based)

Fig. 14.Received traffic at video server (WiFi-based)

received amount of traffic at different servers located at the control center side. Therefore, the WiFi-based architecture will be considered for monitoring all the sub-networks inside the turbine. The following section will consider the timing requirement of the power system in order to evaluate the network performance of the WiFi-based architecture.

5.3 Hybrid network simulation

All applications explained in section 4 are configured to run simultaneously on the same network. The video camera is configured to work for 5 minutes as explained in the previous section and the internet connection is configured for 5 minutes (start time offset of 720 seconds and time duration of 300 seconds). The total end-to-end delay of the full wind turbine network architecture is shown in Fig. 15. It consists of two types: Ethernet delay and wireless LAN delay. The Ethernet delay of the wired link is the same for both wired-based and hybrid architecture with a slight difference, while the wireless LAN delay of the wireless links has a significant impact on network performance which related to the data rate. The WiFi-based architecture

Table 11.End-to-end delay

Fig. 15.Total ETE delay for large-scale WT

is configured to support different data rates of 54Mbps and 24Mbps. Fig. 15 can be divided into three parts depending on the running applications: normal state, video surveillance ON, and web enabled.

In a normal state part, it was observed that the avg. ETE delay is about 0.876ms and 0.609ms for a hybrid model with data rates of 24Mbps and 54Mbps, respectively. In the video surveillance enabled part, for the duration of 5 minutes (the time for which video camera was requested for video surveillance), the delay increased to about 7.71ms and 3.82ms as listed in Table 11. This difference in ETE delay between normal state and the video surveillance enabled part is related to the multimedia service, a feature of which is a high transmission bandwidth which affects the network performance by adding more delay in the network. In the web enabled part, the HTTP traffic is light and has no effect on the total ETE delay, which is similar to the normal state.

As we expected, a better performance is achieved with the higher data rate. In this case, during the period where the video camera was on, the maximum ETE delay for the hybrid architecture with the wireless data rate of 54Mbps is about 3.82ms. Considering 4ms as the requirement of protection information in a power system, we can conclude that the WiFi-based architecture with a 54Mbps data rate satisfies the requirement of the power system.

In this work, we proposed the hybrid network architectures for a large-scale wind turbine (WT) based on the logical node (LN) concepts of the IEC 61400-25 standard. The hybrid network consist of two levels; level 1 connects the metering data (analogue measurements and status information) and video surveillance system to the first access point (AP1) located at the nacelle, while the second access point (AP2) located at the bottom of the tower provides a connection between the protection and control (P&C) unit. Both APs collect the different sensing data using IEEE 802.11g technology and offer internet connection to the maintenance engineer in the case of remote equipment configuration. The control center collects and analyzes all data from WT. We validated our network model from the received amount of traffic at the servers. Considering the ZigBee-based architecture, it was possible to monitor the LNs with low speed sensing data, while it was difficult to monitor the LNs with high speed sensing data of WGEN, WTRF, and WCNV. The simulation results showed that the proposed wireless LAN model satisfied the delay requirements of the power system. Variant MAC schemes can be applied for improving network performance. In future work our network architecture will be expanded for large-scale WPF by using WLAN/WiMAX.


Supported by : National Research Foundation of Korea (NRF)


  1. M. Shahraeini, M. H. Javidi, and M. S. Ghazizadeh, "Comparison Between Communication Infrastructures of Centralized and Decentralized Wide Area Measurement Systems," IEEE Transactions on Smart Grid, vol. 2, no. 1, pp. 206-211, Mar. 2011.
  2. S. Wijetunge, U. Gunawardana, and R. Liyanapathirana, "Wireless Sensor Networks for Structure Health Monitoring: Considerations for Communication Protocol Design," 17th ICT Int. Conf. on Telecommunication, pp. 694-699, Apr. 2010.
  3. H. Saffour, "Wind Turbine Wireless Communication Network & Heading Measurements System- Feasibility Study," Congress DRIVES Electric Automation - Systems & Components, Nuremberg, pp. 1-6, Nov. 2010.
  4. Z.H. Khan, D.G. Catalot, and J.M. Thiriet, "Wireless Network Architecture for Diagnosis and Monitoring Applications," MASAUM Journal of Computing, vol. 1, no. 2, Sep. 2009.
  5. P. Wang, Y. Yan, G.Y. Tian, O. Bouzid, and Z. Ding, "Investigation of Wireless Sensor Networks for. Structural Health Monitoring," Journal of Sensors, Hindawi Publishing Corporation, pp. 1-7, May. 2012.
  6. R.A. Swartz, J.P. Lynch, S. Zerbst, B. Sweetman, and R. Rolfes, "Structural Monitoring of Wind Turbines using Wireless Sensor Networks," Journal of Smart Structures and Systems, vol. 6, no. 3, 2010.
  7. G. Swiszcz, A. Cruden, C. Booth, and W. Leithead, "A Data Acquisition Platform for the Development of a Wind Turbine Condition Monitoring System," Int. Conf. on Condition Monitoring and Diagnosis, pp. 1358-1361, Apr. 2008.
  8. M. Goraj, Y. Epassa, R. Midence, and D. Meadows, "Designing and Deploying Ethernet Networks for Offshore Wind Power Applications - A Case Study," 10th IET Int. Conf. on Developments in Power System Protection, pp.1-5, Apr. 2010.
  9. Z. Hameed, S.H. Ahn, and Y.M. Cho, "Practical Aspects of a Condition Monitoring System for a Wind Turbine with Emphasis on its Design, System Architecture, Testing and Installation," Journal of Renewable Energy, vol. 35, no. 5, pp. 879-894, 2010.
  10. IEC 61400 Wind Turbines - Part 25-2: Communications for Monitoring and Control of Wind Power Plants - Information Models, IEC Std., 2006.
  11. B.A. Forouzan, "Data Communications and Networking," Fourth Edition, McGraw-Hill, 2007.
  12. L. Nilsson, "Lifetime Monitoring of Wind Turbines," Master Thesis, Dept. of Automatic Control, Lund University, Nov. 2005.
  13. P. Parikh, M.G. Kanabar, and T.S. Sidhu, "Opportunities and Challenges of Wireless Communication Technologies for Smart Grid Applications," IEEE Power & Energy Society Meeting, pp. 1-7, Jul. 2010.
  14. V.C. Gungor, D. Sahin, T. Kocak, S. Ergut, C. Buccella, C. Cecati, and G.P. Hancke, "Smart Grid Technologies: Communication Technologies and Standards," IEEE Transactions on Industrial Informatics, vol.7, no.4, pp. 529-539, Nov. 2011.
  15. C-L Hsu and W-B Wu, "The Practical Design of Constructing Data Transition Interface with ZigBee WSN and RS-485 Wired Interface - Example with Small-Scaled Wind-power Electricity Generator System," Journal of Software, vol. 3, no. 8, 2008.
  16. Vestas Website,,V164-7 MW.
  17. M.A. Ahmed, W.-H. Yang, and Y.-C. Kim, "Simulation Study of Communication Network for Wind Power farm," Int. Conf. on ICT Convergence, pp. 706-709, Sep. 2011.
  18. R. hunt, J. Cardenas, V. Muthukrishnan, and D. McGinn, "Wind Farm Protection Using an IEC 61850 Process Bus Architecture," Distributech Conference and Exhibition, Mar. 2010.
  19. T.S. Sidhu and Y.J. Yin, "Modelling and Simulation for Performance Evaluation of IEC61850-based Substation Communication Systems," IEEE Trans. Power Delivery, vol. 22, pp.1482-1489, Jul. 2007.
  20. B. Anjar, M. Dalberg, and M. Uppsall, "Feasibility Study of Thermal Condition Monitoring and Condition Based Maintenance in Wind Turbines," Elforsk Rapport 11:19, May. 2011.
  21. OPNET Modeler, OPNET Technologies.
  22. M.A. Ahmed and Y.-C. Kim, "Network Modeling and Simulation of Wind Power Farm with Switched Gigabit Ethernet," 12th Int. Symposium on Communications and Information Technologies, pp.1009-1014, Oct. 2012.
  23. W. Wang, Y. Xu, and M. Khanna, "A Survey on the Communication Architecture in Smart Grid," Computer Networks, vol. 55, pp. 3604-3629, 2011.

Cited by

  1. An Investigation on Thrust Properties under Wind Shear for an On-Shore 2 MW Wind Turbine vol.19, pp.6, 2016,
  2. Hierarchical Communication Network Architectures for Offshore Wind Power Farms vol.7, pp.12, 2014,
  3. Modeling and simulation of communication networks for use in integrating high wind power generation into a power grid vol.7, pp.4, 2015,