Review of dynamic line rating systems for wind power integration

When a wind power system is connected to a network point there is a limit of power generation based on the characteristics of the network and the loads connected to it. Traditionally, transmission line limits are estimated conservatively assuming unfavourable weather conditions (high ambient temperature, full sun and low wind speed). However, the transmission capacity of an overhead line increases when wind speed is high, due to the cooling caused by wind in the distribution lines. Dynamic line rating (DLR) systems allow monitoring real weather conditions and calculating the real capacity of lines. Thus, when planning wind power integration, if dynamic line limits are considered instead of the conservative and static limits, estimated capacity increases. This article reviews all technologies developed for real-time monitoring during the last thirty years, as well as some case studies around the world, and brings out the benefits and technical limitations of employing dynamic line rating on overhead lines. Further, the use of these DLR systems in wind integration is reviewed.


Introduction
Taking into account the levels of CO2 emissions [1][2][3][4][5][6][7][8][9][10][11][12][13] and in order to fulfil the Kyoto Protocol commitments, the contribution of renewable energy to the future generation will have to increase significantly from current levels. A good example is that the European Commission has set itself an ambitious target of 20% of total energy consumption to be supplied by renewable energy sources in 2020.
One of the challenges to achieving this goal is the need to expand or strengthen the distribution network in order to accommodate the large penetration of wind power. However, commissioning time of distribution network projects is usually longer than the time needed to build a wind farm. Therefore, recently built wind farms might be ready to generate power, but their evacuation lines have insufficient transmission capacity. Accordingly, wind power plants have to limit their generation in this situation. However, the now ever present solutions for the "smart grid" suggest the possibility of using the existing network more efficiently, so that wind power evacuation is not limited due to network congestion [14].
Another aspect to consider is the repowering of wind farms, i.e. the replacement of existing wind turbines by newgeneration wind turbines [15,16]. The replacement of those first generation turbines has several advantages. For example, modern wind turbines often include control systems of reactive power and immunity to voltage dips, which are very important for the electrical system operation. But the increase of wind power penetration can reach a limit due to an insufficient capacity of the network in which the energy is injected.
Wind power is cheap and clean. If wind power is curtailed due to congestions in the grid, the curtailed amount of power has to be produced in other power plants, usually thermal, which are more expensive and have higher impact on the environment. For this reason, investment in the grids is justified. As an example, in the European transmission network ten-years development plan, the development of renewable energy is found to be the major driver for grid development. Projects of pan- Due to the uncertainties related to wind speed measurements, many DLR systems measure directly the overhead conductor temperature. This magnitude is used to calculate the effective wind speed that cools the conductor. The effective wind speed is the wind that for the measured current intensity, ambient temperature and solar radiation results in the measured conductor temperature. This thermal calculation is carried out by the IEEE and CIGRE models mentioned above. The wind direction is assumed to be perpendicular to the conductor. Finally, ampacity is obtained from the calculated wind speed, which is perpendicular to the conductor, and the measured ambient temperature and solar radiation (Fig. 2). Other DLR systems measure overhead tension or sag. In this case, an additional step is necessary, and the conductor temperature is derived from these magnitudes (Fig. 3). The overhead line mechanical model is used to relate the conductor tension or sag with the conductor temperature. When the conductor temperature increases the sag increases and the tension decreases and vice versa. Once the conductor temperature is obtained, the same procedure described in Figure 2 is followed to calculate the ampacity. Following, DLR systems are described according to the magnitude being monitored.

Weather monitoring
Dynamic line rating by weather monitoring is the simplest system. It is also the least invasive system as it does not need to be physically installed in the line. Hence, making use of weather stations installed in substations, weather monitoring is easy to apply.
The main drawback is that conditions, especially wind, can change along the line due to land irregularities, forests, etc. Hence, there is a degree of uncertainty in the results that can be important in some cases. Another drawback is related to cup anemometers, as these can have measurement errors below 1 m/s wind velocities. This wind velocity range is the most unfavourable from the point of view of ampacity. Figure 4 shows the conductor temperature change as a function of wind speed. When the wind speed is low, low wind speed changes result in high conductor temperature changes. Above 3 m/s, the conductor temperature is almost unaffected by wind speed changes. There is a commercial system, called ThermalRate [49], that calculates wind speed instead of measuring it ( Figure 5). Two identical metallic rods are situated in the tower. The rods can have different emissivity and absorptivity values from the conductor. The tubes are situated parallel to the conductor. In one of the rods, current is injected, while in the other, no current is injected. The injected current intensity is measured. Temperature is also measured in both rods by an embedded thermocouple. One rod is heated and the other rod is unheated. The thermal equation is applied in the rod where current is injected in order to get the equivalent wind velocity. Instead of using ambient temperature Ta, the solar temperature Ts measured in the rod where no current has been injected is used. Ampacity is calculated from the solar temperature T s and the calculated equivalent wind velocity [50,51]. Some application cases are shown in [45,52,53].

Conductor temperature monitoring
An alternative to reduce the uncertainty related to the calculation of conductor temperature is direct monitoring. Conductor temperature is monitored by a sensor installed in the conductor that measures its surface temperature. However, the conductor temperature can change along the line due to the change of weather conditions. Sag is a function of average temperature, whereas the monitoring system gives a local temperature value. Besides, a radial temperature gradient can be present in the conductor.
Several commercial systems are available ( Figure 6):  The first commercial system based on conductor temperature monitoring, the Power Donut [54], was developed in the early eighties [55,56] (Figure 6.a.). Besides monitoring conductor temperature, the Power Donut monitors current intensity and the conductor angle of inclination. The latter is related to the conductor sag value. It is a toroid coupled to the conductor. It is self-supplied by the tension induced by the magnetic field related to conductor current. It needs a minimum turn on intensity of 70 A. An internal battery allows one hour of operation when current intensity is below this value. It can measure conductor temperature up to 150 ºC. In order to calculate ampacity, the system needs ambient temperature and solar radiation values. These magnitudes can be obtained from a weather station nearby. In case this is not possible, the system offers the possibility of installing a weather station that measures ambient temperature and solar radiation in the tower that is closest to the temperature sensor. Weather values are transmitted to the Power Donut via radio. Some application cases are shown in [55,57,58,59]. monitoring conductor temperature, it monitors its sag. Sag is obtained by using LIDAR technology that determines the height of the conductor from its position to the ground. Furthermore, it measures conductor tilt and roll with a dual axis accelerometer. It is self-supplied with a minimum turn on intensity of 100 A. It can measure conductor temperature up to 250 ºC.  The emo device is a simple system that only measures conductor temperature [71] ( Figure 6.f.). Another difference is the supply option, because instead of being self-supplied it is battery supplied. The standard version can measure conductor temperature up to 85 ºC, whereas the high temperature version can measure up to 210 ºC. An application case is shown in [72].  A different option is monitoring temperature by a device based on surface acoustic wave (SAW) [73]. This device is called Ritherm [74] (Figure 6.g.). The system comprises a radar that sends and receives high frequency electromagnetic waves and a passive SAW sensor installed on the conductor. The SAW passive sensor is a piezoelectric crystal that reacts moving with the electromagnetic wave sent by the radar. In the crystal, some elements convert the movement into electromagnetic waves which are received by the radar. It is possible to determine the position of the elements. The position is a function of the elongation associated to temperature value. Besides, wave velocity in the crystal depends on temperature. With this system, temperature up to 150 ºC can be obtained with an uncertainty of 0.5 ºC. This sensor is installed on the overhead line using the helical preformed rods [68]. Some application cases are shown in [46,69,75,76,77]. Figure 6. Commercial systems based on conductor temperature monitoring From Table 2 to Table 6 different aspects of commercial systems based on conductor temperature monitoring are compared.

Conductor temperature Current intensity Tilt angle Sag Power Donut
x  Apart from the mentioned commercial conductor temperature monitoring systems, another option for measuring the conductor is the distributed temperature sensing (DTS). It is possible to determine temperature distribution along the conductor by using optical fibers embedded in it. Some application examples are shown in [78,79,80,81,82,83,84,85].
On the other hand, in [86] thermal rating calculation methods based on indirect conductor temperature monitoring are compared. These methods use the measurements of the key parameters that affect the allowable conductor temperature, like conductor temperature, ambient temperature, "ground clearing distance", humidity and wind speed.
Finally, some academic projects which have resulted in temperature monitoring system prototypes have also to be mentioned:  The Georgia Institute of Technology has developed two different prototypes. The objective is to design cheap and self-supplied devices. The developed systems measure conductor temperature and current intensity. The first system, called Power Line SensorNet (PLSN), is designed using commercially available low power devices [87]. The second system is a stick-on sensor [88].  The Isfahan University of Technology and the University of Manitoba have developed a device that measures temperature based on radio frequency cavity resonance [89].

Tension monitoring
Tension is monitored by a load cell that is installed in series with the insulator strings. The load cell is located between the tower and the insulator string, so that it is electrically insulated from the conductor. There is a direct relation between tension and sag values, and for this reason, tension monitoring is a good indicator of the line condition when sag is the magnitude to be controlled. Sag is calculated based on tension value, span length and the weight per length unit of the conductor. Whereas the conductor temperature measurement is a local measurement, tension represents the average condition of all the spans between two tension towers.
There is only one commercial tension monitoring system called CAT-1 [28,102,103] (Figure 7). This system is calibrated in order to establish the relation between tension and conductor temperature. Besides, it has a special system which enables to measure weather values in an indirect way. With respect to calibration, it is based on measuring pairs of tension-temperature values. In the one hand, a reference for tension and conductor temperature is established. In the other hand, the value of the ruling span is obtained. Once the calibration has been carried out, temperature can be derived from tension measurement. The special system that measures weather values in an indirect way is known as Net Radiation Sensor. It is an aluminium tube with the same emissivity and absorptive values as the conductor. It is installed in the same tower where the load cell is installed. The tube is situated parallel to the conductor. A temperature sensor measures the temperature of the aluminium tube. This temperature represents the temperature that the conductor has with no current intensity. The Net Radiation Sensor is also used for the aforementioned calibration. The conductor temperature value needed for the calibration is not measured directly but it is estimated from the temperature measured in the Net Radiation Sensor. In order to establish a correspondence between both temperatures, current intensity has to be very low. Some application cases are shown in [22,28,50,78,104,105,106,107]. In addition, a university project that has resulted in a tension monitoring system prototype has to be mentioned. The University of the Basque Country UPV/EHU has developed the Tension and Ampacity Monitoring (TAM) system [108,109]. It is based on the monitoring of conductor tension, ambient temperature, solar radiation and current intensity. The system takes into account the creep deformation experienced by the conductors during their lifetime and calibrates the tension-temperature reference and the maximum allowable temperature in order to obtain the ampacity.

Sag monitoring
Although there are a few proposed systems for sag real time monitoring, currently there is only one commercial system, the Sagometer (Figure 8). This system is based on image processing [116]. A target is connected to the conductor in the middle of the span, and a video camera situated in the tower monitors the movement of the target. The system captures and processes images and calculates the sag. Additionally, a current intensity and weather measuring system can be added in order to obtain the ampacity.

Vibration monitoring
One of the last proposals (Ampacimon), developed in Belgium, determines sag value by conductor vibration analysis [118,119,120]. It is based on monitors placed directly on the line (Figure 9). These monitors, equipped with acceleration sensors and electronics, measure conductor mechanical movements with a very high sensitivity. Ampacimon processes these measurements and calculates the conductor sag. Some application cases are shown in [81, 121,122]. Ampacimon (Ampacimon) . It is a noncontact, real time sensor system for the monitoring of HV overhead transmission lines ( Figure 10). The system uses calibrated AC-magnetic field sensors, located roughly under the phase conductors, to accurately and reliably measure the 3 phase AC magnetic fields. Once installed in the transmission row and calibrated, the system reports 3 phase currents, conductor clearance, and maximum conductor temperature [124].
RT-TLMS (Promethean Devices) Figure 10. Commercial system based on electromagnetic field monitoring Table 9 shows different studies developed by companies and institutions during the last years in different countries. The aim of these works is to estimate line ampacity from conductor sag.  Table 9. Studies of the ampacity from sag conductor

Dynamic line rating and wind power integration
The development of dynamic line rating systems started three decades ago. Dynamic line rating systems are interesting because they give an actual value of line rating. However, the use of these systems in the power network is not widespread because in many cases there is little correlation between need (instants with high flow) and availability (instants with high transmission capacity). This situation has changed with the increase of wind power generation. Therefore, there are more and more grids where there is a high correlation between need (instants with high flow due to high wind generation) and availability (instants with high transmission capacity due to the wind cooling). As a result, several utilities are considering dynamic line rating systems as a practical option for maximizing the capacity of their assets. Examples of studies carried out recently in order to assess the viability of DLR systems in grids with high wind power penetration are found in Ireland, United Kingdom, Canada, Spain, or Germany, as an instance.
Different types of studies have been carried out. Some of them have been performed in real installations, where overhead lines have been monitored and obtained measurements have been analysed. Other studies have been carried out with simulations, including Active Network Management schemes, Computational Fluid Dynamics modelling or Monte Carlo simulations. On the other side, some projects have focused on estimating the additional wind generation integrated using DLR systems, while other research works have evaluated how to integrate DLR systems with control centres or with the congestion management.

Real installations
In Northern Ireland, there is a need to reinforce the network due to the increase of wind energy, and since 2008 the Northern Ireland Electricity (NIE) has installed DLR systems in order to maximize the capacity of the grid. One of the monitored lines is the Dungannon -Omagh 110 kV line, which connects the West of North Ireland, with high wind power production, and the East, with high power demand. This project is described in [42,43,65,130,132]. [42,43,65], compare in different graphs actual weather magnitudes (ambient temperature and wind speed) to magnitudes used for the static method, and conclude that DLR can accommodate the additional capacity required on the line when additional wind generation is introduced. Hence, wind curtailment on existing wind farms can be reduced with DLR systems. Besides, [43] presents the scheme of the system architecture for incorporation of DLR into the Distribution Control Centre. [130,132] show figures where a positive correlation between dynamics ratings and wind speed at the wind farm is observed.
Queen's University Belfast uses weather and line data measured on the Dungannon-Omagh 110 kV line and on the Kells-Coleraine 110 kV line in order to complement the Northern Ireland Electricity's studies. [134,135] show in different figures the improvement in ampacity using DLR. Besides, [134,136,137] describe the experimental program where two test rigs have been constructed in order to examine the thermal behavior of the conductor under controlled weather conditions. The authors analyse graphically the effect of wind speed on conductor temperature and the correlation between wind generation and line capacity. All studies show that the higher wind generation corresponds to a higher line capacity.
In UK E.ON Central Networks has applied DLR for load management and protection of a 132 kV line between Skegness and Boston (North East of England). This project is described in [24,59,138,139] and shows the dynamic ampacity as a function of wind speed for different ambient temperatures and the static ampacity. It can be concluded that for most wind speeds and ambient temperatures, ampacity is larger than the one obtained using the static method. So, it is possible to integrate more wind generation into the grid taking into account the cooling effect of wind. The authors estimate an increase in the wind generation from 20% to 50%.
Furthermore, Durham University, ScottishPower Energy Networks Imass, PB Power and AREVA T&D have developed a DLR technology which is applied to a 132 kV conductor, in order to facilitate the connection of over 200 MW of wind generation [44]. [140] displays two figures: first, the conductor rating exposed to different wind speeds and second, real time dynamic ratings compared to the static rating, with the static rating of the line re-tensioned and with the static rating of the new line reinforced. This research work concludes that "the adoption of a DLR system can provide a 67% gain in energy transfer capacity at 62% of the re-tensioning cost".
The TWENTIES project gathered a group of Transmission System Operators from Belgium, Denmark, France, Germany, Spain, and the Netherlands, generator companies, power technology manufacturers, wind turbine manufacturers and some research and development organizations. Two of the objectives of this project were first, to demonstrate how much additional wind generation could be handled thanks to DLR (NETFLEX), and second, to prove that the current transmission network was able to meet demand of renewable energy by extending the system operational limits, maintaining safety criteria (FLEXGRID) [78]. Part of the conclusions drawn from NETFLEX was that "DLR forecaster enables an average increase of 10-15% in the usable transmission capacity of overhead lines". Additionally, FLEXGRID draws the conclusion that "most days the dynamic ratio throughout the day is more than 10% higher than the seasonal rating. On windy days, it could be more than 100% higher over the day" [81, 103, 106,133,134,135,131].
In Sweden, overhead lines threaten to become bottlenecks for the increase of wind power. [100] compares the traditional static rating and DLR, infers that a high capacity of overhead lines is not used during a large percentage of the time. Other similar studies in this line are the one performed in Spain by the University of Cantabria in [62], or still the research work carried out by Marshall Municipal Utilities, Shaw Energy Delivery Services and Xcel Energy in Minnesota (USA) [53]. The latter underlines that the actual line rating is above the static rating 96% of the time.
In the north of Spain, in the on-shore wind farm La Sía, the University of the Basque Country has installed a DLR system to monitor a wind farm evacuation line of 66 kV. Figure 11 shows the effect of the wind speed on the ampacity. The higher wind generation corresponds to a higher line capacity. Figure 11. Relation between wind speed and ampacity.
Wind is the most influential parameter that affects the conductor temperature and the ampacity. There is a clear relation between wind generation and the ampacity of the line. Figure 12 shows two periods of time, one with wind and wind generation (a) and another with low wind and no wind generation (b). It is observed that in the windy period ampacity is 1200 while the period with little wind this value is reduced to 800 A.

Simulations
The integration of DLR systems with control centres has also been studied. In [58,141,142] initial results from the deployment of a DLR system on the Onkey Isles (isles where the connection and operation of distributed generation is high) and its integration with the Active Network Management (ANM) scheme are presented. The authors conclude that the addition of DLR to the existing ANM scheme reduces potentially curtailment in the region between 47% and 81%, and so, it allows the connection of additional 4 MW. [44] introduces a DLR system trialed in open-loop control mode. Authors show graphically a representation of a set of field data as ambient temperature, current, wind speed and conductor temperature, and it is observed that the line operates with very wide thermal margins. Therefore, the authors present a strategy which combines the DLR system with a network flow management system, which will be at the disposal of Scottish Power as part of its operation and control system. SP Energy Networks and Durham University use FLUENT 12.1 for studying a real wind farm connection in north Wales [143]. This study shows that "a wind farm of 140MW can be connected to a conductor which could only support 90MW based on its static rating, and if the route is chosen correctly only 1% of the energy yield will be constrained".
There are other studies that consider the use of DLR systems for increasing line capacity based on simulation models. Natural Sciences and Engineering Research Council of Canada and the University of Alberta analyse in [26] the reduction of power generation emissions that can be achieved using DLR systems to integrate more distributed generation on the existing power grid. Different operating cases are simulated and the increase in available capacity using DLR systems is presented. The authors show that, in the worst case, there is a loss of 149.9A in capacity while in the best case a gain of 1956.17 A is achieved.
In Germany, DLR systems have been examined in order to improve the integration of wind power generation into the transmission system and the congestion management [23,144]. [144] modelled a reduced sample network (28 nodes) to reproduce realistic situations in the current German transmission system, while the case study in [23] presents a future scenario (2015) with increased installed capacity of wind power. The authors simulate the congestion management with and without DLR systems, and they conclude that redispatched energy during a whole year is reduced by about 85% with DLR systems.
Other studies where a significant increase of transmission capacity compared to the static rating is indicated are [21] where a multi-period AC Optimal Power Flow (OPF)-based technique is used, [145,146] where Monte Carlo simulation is used, and [25] where a probabilistic approach is applied to generate different scenarios of wind power output and thermal ratings of overhead lines. Furthermore, in [147] a thermal model is used and the cooling effect of wind in overhead power lines required for the evacuation of energy generated by wind farms is analysed.

Conclusions
DLR systems are based on methods that allow increasing line power flow securely and safely. So, there is no need to strengthen the towers. These systems are particularly useful on lines which are connected to wind farms, often with an insufficient transmission capacity. Wind plays an important role in line cooling, and therefore, under most circumstances, DLR limit is higher than static line rating, and DLR based methods are more accurate. So, the application of DLR will effectively increase line ampacity and postpone the necessity to upgrade the existing network to integrate more renewable energy generation.
This paper has introduced a literature review of all technologies developed for real-time monitoring, as well as some case studies around the world. In addition, the benefits and technical limitations of employing dynamic line rating on overhead lines versus traditional static methods have been described. And the use of these DLR systems in wind integration has been reviewed.