Optimal Integration of Offshore Wind Power for a Steadier, Environmentally Friendlier, Supply of Electricity in China

Demand for electricity in China is concentrated to a significant extent in its coastal provinces. Opportunities for production of electricity by on-shore wind facilities are greatest however in the north and west of the country. Using high resolution wind data derived from the GEOS-5 assimilation, this study shows that investments in offshore wind facilities in these spatially separated regions (Bohai-Bay or BHB, Yangtze-River Delta or YRD, Pearl-River Delta or PRD ) could make an important contribution to overall regional demand for electricity in coastal China. An optimization analysis indicates that hour-to-hour variability of outputs from a combined system can be minimized by investing 24% of the power capacity in BHB, 30% in YRD and 47% in PRD. The analysis suggests that about 28% of the overall off-shore wind potential could be deployed as base load power replacing coal-fired system with benefits not only in terms of reductions in CO 2 emissions but also in terms of improvements in regional air quality. The interconnection of off-shore wind resources contemplated here could be facilitated by China’s 12 th -five-year plan to strengthen inter-connections between regional electric-power grids.


Introduction
Production of electricity from wind power has expanded rapidly in China since bidding process for offshore demonstration projects . The successful bidding prices for these projects ranged from 0.62 RMB/kWh to 0.74 RMB/kWh, about 9.1 cents/kWh to 10.9 cents/kWh in 2010 US dollars. To ensure profitability, these projects benefit from a favorable concession policy, which guarantees higher bus-bar prices for electricity produced from offshore wind farms as compared with production from either onshore facilities or from conventional coal-fired power plants. imported from inland provinces. Rich onshore wind power resources tend also to be located to the North and West of China. To harvest this renewable source requires significant expansion of existing transmission grid system on a national scale. We shall argue here that investment in offshore wind resources would significantly reduce the demand for coal supply for the south, while providing at the same time a valuable lowcarbon source of electric power.
The intrinsic variability of the output of power from individual wind farms poses a problem for integration at scale of this source into the existing power system. Production of electricity in coastal provinces of China is currently dominated by sources fueled by coal, with percentages ranging from 55% in Guangxi to as high as 91% in Shandong in 2010 (Ma et al., 2011). Coal-fired systems are relatively inflexible, with limited ramp-up and ramp-down capability to cope with the additional variations introduced by wind power. Direct or indirect electricity storage could help address this potential incompatibility. A number of pumped hydro power stations have already been built in China to cope with the increasing diurnal variability of load. Opportunities for further expansion of pumped storage are limited, however, due largely to constraints imposed by geography (compatible topography and hydrology). We shall argue that coupling outputs of wind farms from different regions of the Chinese coast could significantly offset the challenges associated with integrating this otherwise variable source.
The present analysis reports a statistical investigation of advantages that could be achieved in smoothing the variation of offshore wind power supply in the coastal areas of China through an optimal combination of power from geographically distributed offshore sites. A number of studies have examined the advantages of combining geographically distributed wind farms in the U.S. and Europe Huang et al., 2013;Kempton et al., 2010). Kempton et al. (2010), analyzing five years of wind data from 11 meteorological stations distributed along the U.S. east coast, suggested that the output of wind power from an interconnected system there would be more reliable than that from any individual location. Archer and Jacobson (2007) explored the benefits of connecting wind farms from up to 19 sites located in the U.S. Midwest, where annual average wind speeds at 80 m above ground exceeded 6.9 m/s. They found that an average of 33%, and a maximum of 47%, of yearly average wind power from interconnected farms could provide reliable, base-load electric power. Huang et al. (2013), using five years of hourly assimilated wind data, showed that the high-frequency variability of wind-generated power could be significantly reduced by coupling outputs from five to ten wind farms dispersed uniformly over ten states in the middle of the U.S. Their analysis suggests that more than 95% of the variability of the coupled system from this region was concentrated at time scales longer than a day, allowing operators of the overall system to take advantage of multiple-day weather forecasts in scheduling projected future contributions from wind.
Building on the earlier studies, the present analysis will focus on variations of hourly wind power from 12 offshore sites distributed along the Chinese coastline (see minimal attention to the challenges that would be associated with integration of this source into the existing power system. As will be shown here, these difficulties can be significantly mediated by adopting a more coordinated regional approach involving an integration of wind resources for spatially separated coastal regions subject to important differences in prevailing meteorological conditions. Section 2 summarizes the data and methods adopted for the present analysis. The variation of China's offshore wind resources, results of the optimization analysis, and related implications for the electric power systems of coastal regions in China are discussed in Section 3. Concluding remarks are presented in Section 4.

Data and Methods
Wind fields adopted for this analysis were derived for 2009 from the Goddard where V is the wind speed, z is the measurement height, and  Provinces and provincial-level municipalities with coastlines are in pink.
As illustrated in Figure 2, a total of 12 offshore sites were selected for purposes of this study, with sites 1 to 4 in the Bohai Bay region (hereafter referred to as BHB),  This capacity was allowed to vary, depending on the total capacity evaluated for the combined system.    Table 1.   (Figure 4b). Similar clustering is observed for BHB and YRD in fall (Sept. to Nov.) (Figure 4d).

Optimization Analysis
The results in Figures 4a-4d suggest that advantages could be realized by linking these regions. For any given hour, a low contribution of wind-generated power from sites in one coastal zone will be compensated often by a higher output from sites in another zone. If the 12 offshore sites in the three economic zones were all interconnected, the overall power output would be much less variable than the output from any individual site, and less variable than that for any individual zone. This raises the question as to how these zones could be optimally combined to reduce the overall variation of power output.
An optimization model was applied to determine the relative contributions of wind power from the BHB, YRD and PRD zones that would result in the lowest possible standard deviation of the hour-to-hour variation of power output from the combined system. The following objective function is designed to minimize the hour-to-hour variation of electricity produced by the interconnected sites, thus to ease the integration of wind power into the relatively inflexible, coal-dominated, existing electric power system:

Implications to the Electric Power System
The histograms in Figure 6   selected sites from each of the three zones (s3, s7 and s11) together with results for the interconnected composite. The combined system clearly reduces the range of high frequency variability: the frequency distribution is narrowed significantly. Over the course of the year, the largest hour-to-hour increase in CF predicted for any one of the three selected sites was 0.85 (at s11): the largest decrease was 0.44 (at s7). If the capacity of wind farms installed at these individual sites amounted to 100 MW, this would imply that power outputs could increase 85 MW, or decrease by 44 MW from one hour to the next. In the case of the coupled system with capacity of 100 MW in total, hour-to-hour changes would be significantly reduced, ranging from as little as -7 MW to +15 MW.
What this means is that the need for quick-ramping capacity to compensate for variability in the supply from offshore wind resources would be significantly reduced with the combined system as compared to the situation that would apply in the absence of interconnection. The frequency distributions of hourly CFs (as opposed to hour-to-hour changes in CF) over the course of 2009 for the same three offshore sites and for the optimally combined system are illustrated in Figure 7. The distribution of CF values for the combined system is significantly more concentrated (much less variable) than that for any individual offshore site. The variability of power outputs from a single site is evidenced by the wider distribution of CF values, varying from 0 to 1, with large probabilities at the two extremes. In contrast, the frequency spectrum of CFs for the optimally integrated system exhibits a Rayleigh distribution, which a peak at 0.25 combined with a broad tail extending to higher values.   corresponding to this condition is equal to 0.092 as indicated in Figure 8. The firm capacity of the optimally combined system corresponds therefore to about 9.2% of the total installed capacity. The power output consistent with this firm capacity requirement accounts thus for 27.9% of the total power generated from the combined system over the entire year, implying that power of this magnitude from the combined system could serve as base load reducing accordingly the demand for power from coal-fired plants.

Concluding Remarks
An optimization model was adopted to explore the complimentary qualities of variability. An optimal combination of wind power from these coastal zones (24% from BHB, 30% from YRD, and 47% from PRD) was found to minimize hour-to-hour variation of the overall power output. The decrease in hour-to-hour variations from wind power reduces the requirement for quick-ramping capacity in the power system as needed to compensate for variability introduced by offshore wind. The results of the optimization have important implications for planning of future developments for offshore wind resources in China. Should China elect to invest in 60 GW of offshore wind facilities by 2030, the present analysis suggests that 14 GW of this investment should be allocated within BHB, 18 GW within YRD, and 28 within PRD. This combination would ensure maximum reliability of the power supply from an interconnected offshore wind system.
The frequency distribution of the hourly CFs of the optimally combined system is concentrated in a CF range of 0.15 to 0.50. Intermittency is effectively eliminated in this case. The optimally combined system can supply 9.2% of its total capacity as firm capacity, which guarantees reliable power output with an offline record of no more than 7.6%, equivalent to the conditions realized by existing for coal-fired power plants in China. Over the course of a year, about 27.9% of the power generated from the optimally combined system could be deployed as base load, replacing potentially in this case the demand for power production by coal-fired systems.
To realize the advantage of the steadier electricity supply from offshore wind identified in this study, it will be necessary to invest in a significant expansion of the existing transmission grid system, to interconnect the wind facilities contemplated for the three regions identified here. Kempton et al., (2010) proposed an offshore transmission cable to link potential wind farms along the U.S. eastern coast in order to mitigate the variations of overall anticipated power output. An "Atlantic Independent System Operator" was specifically suggested by Kempton et al., (2010) to manage and regulate the market for offshore wind power off the east coast of the US. China, however, may be expected to follow a different path in developing its off-shore resources. The likely strategy in this case may be expected to involve enhancing interconnectivity of the land-