Trans-regional transmission of large-scale hydropower: problems and solutions in receiving power grid

1 Introduction

Due to the increasing concern over the environmental impact of conventional fossil fuels, non-fossil energy has developed rapidly in China in recent years.Hydropower is a clean, renewable and cost-efficient source of energy.More importantly, hydropower can respond quickly to load changes and has huge untapped potential in China, and has therefore earned high development priority in China’s energy plans [1-4].Since the implementation of the westeast power transmission project, the operation of several large hydropower plants was initiated in southwest and central China in the past 20 years [5,6].These include the Three Gorges (22.4 GW), Xiluodu (13.86 GW), Xiangjiaba (6.4 GW), Nuozhadu (5.85 GW), Longtan (4.9 GW), Jinping II (4.8 GW), Xiaowan (4.2 GW), Laxiwa (4.2 GW), Jinping I (3.6 GW), and Pubugou (3.6 GW) power plants, which are the largest hydropower plants in China[3,4], and they account for approximately 20% of the national hydropower installed capacity [7].Limited by the hydropower absorption capacity of local power grids, the produced hydro-energy requires transmission to load centers in eastern and coastal China via ultra-high voltage direct current (UHVDC) or ultra-high voltage alternating current (UHVAC) transmission lines to promote the utilization of renewable energy sources and reduce the hydropower curtailment in southwestern China [8,9].

The Yangtze triangle and the eastern coastal area covered by the East China Grid (ECG) is one of main economic sectors in China.Within this area, there is a large demand for electricity, in addition to a significant shortage of energy resources [10].In recent years, with an increase in the haze and environmental pressure, in addition to an increase in the peak-valley load, the dependence on outer power, especially clean energy, has increased significantly [11].With the rapid development of UHVDC and UHVAC transmission projects nationwide [12-14], there was a rapid increase in the scale of the ECG trans-regional power supply, which reached approximately 70 GW in 2018 [15,16].The ECG is gradually developing into a superlarge-scale mixed operation of an external direct current (DC) power supply and multiple power sources in the region.The dispatching system is very different from several similar-scale power grids such as those in south China [17] and central China, in addition to the United States of America PJM power grid [18, 19] and the south-east Brazil power grid [20,21].The dispatching characteristics are very distinct, and the demands and problems are significant and complex.Moreover, they can be summarized as follows.

(1) The proportion of the thermal power installed capacity exceeds 85%, thus occupying an absolute dominant position with less than 10% of hydropower; which exhibits a decreasing annual trend.(2) Eight of the large pumped-storage power plants with a total installed capacity of 7120 MW will exceed 10000 MW within a few years.(3) The load is large and rapidly increasing.In particular, the maximum load reached 266 GW in 2018, ranking first in the country.(4) The peak-valley difference is continuously widening, with a maximum peak-valley difference of more than 64.44 GW in 2018, which indicates an increase of more than 120% when compared with 2005.Moreover, it reached 75.47 GW in 2019, thus increasing the difficulty of peak operations.(5) There was a rapid increase in the scale of the trans-regional power supply.However, the current “non-peak regulation” or “reverse peak regulation” transmission schedules of DC hydropower cannot fully implement the operation flexibility of large hydropower plants in southwestern China, and there is even an increase in the peak-shaving pressure on the ECG.(6) Load characteristics with significant differences between provinces and cities further increase the difficulty of coordination between provinces and networks.Under these complex characteristics and demand conditions, the effective absorption of large-scale outer hydropower is a bottleneck problem that restricts the safe and high-quality operation of the ECG, and it directly relates to the largescale optimal allocation of hydropower resources in western China.

Hence, in this study, the problems caused by transregional transmission of large-scale hydropower and the corresponding countermeasures were analyzed, with the ECG as the focus of the research.

2 Load characteristics and outer hydropower evolution of ECG

The ECG is the regional power grid with the largest load demand in China.In recent years, the maximum power load of the entire grid repeatedly increased.The maximum load increased from 185 GW in 2012 to 266 GW in 2018, which corresponds to an increase of 44%.The average load in 2018 reached 173 GW, and the daily maximum peakvalley difference was 75.47 GW (see Fig.1).Moreover, with the construction and operation of high-voltage grid structures such as “Fufeng DC” and “Binjin DC”, the inter-regional power transmission channels achieved highspeed growth, with a maximum power receiving capacity of approximately 70 GW.Moreover, there were rapid annual increases in the power receiving scale outside the region increased.By the end of 2018, the maximum outer power reached 38 GW, which accounted for 14.13% of the maximum load, with an average annual growth rate of more than 8% over the last 8 years (see Fig.2).The total annual electricity received outside the region was 193.7 billion kWh, which corresponds to an increase of 17% from theprevious year, and accounts for 12.76% of the total annual electricity consumption in east China.From 2012-2018, the cumulative electricity received was 873.3 billion kWh (see Fig.3), which is approximately greater than the annual power generated by the Three Gorges by a factor of 8.7.It should be noted that the Three Gorges is the largest hydropower plant in the world.The outer electricity is therefore a critical factor with a significant influence on the ECG power generation balance of ECG.

Fig.1 Load characteristics of the ECG

Fig.2 Power transmission capacity of the ECG

Fig.3 Annual power transmission for the ECG

3 Problems related to integration large-scale outer hydropower

3.1 Ensuring the power balance

In 2018, three UHVDC transmission channels, namely Fufen, Binjin, and Jinsu, failed to deliver full power due to casing defects.The maximum power of the Lingshao DC in summer was approximately 6.2 GW, the peak power of Yanhuai in summer was approximately 4.2 GW, and Xitai DC was not able to deliver power in summer due to the shortage of electricity on the opposite side.The Jiquan DC does not satisfy the conditions for power transmission in summer.The east China inter-regional DC power supply has gradually changed from a hydropower-based to hydropower source, which is inclusive of new energy sources.In addition to the influence of the flood and dry seasons; the characteristics of new energy peak regulation in the day, the heating demand in the north, and the load growth on the opposite side of the DC are closely related to the power supply of the Yangtze River.Such factors are becoming increasingly significant in the ECG.

In 2018, the flood season in southwestern China was delayed for over a month when compared with previous years, and the inflow of Fujian was relatively low; thus the ECG spring flood peak regulation problem was not significant.For the first ten days of July, there was a significant increase in the incoming water to the Three Gorges power plant, and the power of the trans-regional hydropower transmission reached 23 GW.Since then, it has increased steadily.However, due to the three factors resulting in UHVDC failure to deliver full power, as mentioned previously, the outer hydropower has not reached the scale that was realized in previous years.Due to the late end of flood season and the limited channel in the early stage, the outer hydropower was supplied until the end of November.At this time, the load of the ECG was significantly reduced, and the peak shaving pressure during the autumn flood season was significant due to the superimposition of new energy absorption.In this case, it is very challenging for the ECG to guarantee the power balance [22].

3.2 Determination of hydropower transmission schedules

Chinese energy resources and electricity demand are inversely distributed.It is necessary to optimize the allocation of power resources nationwide through longdistance and large-scale transmission.With the intensive development and utilization of energy bases in western and northern China and the gradual advancement of UHVDC transmission projects, the transmission scale between large energy bases and energy consuming areas is increasing.There are higher requirements for the existing coordination mode of power transmission.For the receiving power grid, more effective methods are required for the efficient use of large-scale external electricity.The objective is the formulation of a power transmission plan, thereby maximizing the role of the optimal allocation of resources of the power grid.The scale of the outer hydropower of the ECG as a large-scale receiving power grid is increasing.The transmission schedules and power distribution methods, in addition to strategies, among provincial power grids have an increasing influence on the receiving power grid.However, the current “linear transmission mode” or “singlepeak load regulation mode” are dependent on agreement and experience, and it is difficult to realize an efficient operation scheme that is consistent with the load demand of the ECG.Therefore, an improvement of the existing methods of DC hydropower transmission and power distribution among subordinate provincial power grids is required [23].

3.3 How to respond to serious peak-shaving demands

In 2018, the average peak-valley difference of power consumption in the ECG was 51.43 GW, which corresponds to an increase of 9.2% (4.32 GW) and a significant increase in peak-valley difference.The average peak-valley difference of power consumption was 26.07%, which corresponds to an increase of 1.2% over the same period.An increasing annual trend was observed.The typical load curve of the ECG is presented in Fig.4, which indicates that the peak-shaving demand of the ECG is very high.The load in the low-waist load period is lower than that in peak period.Owing to the insufficiency of the hydropower peak regulation in southwest China in the flood season, in addition to the non-peak regulation and anti-peak regulation characteristics of new energy sources inside and outside the region, the peak regulation pressure of the ECG is significant, and the peak regulation demand is shifted from the “single-period” in the morning low to “double-period” in the low waist load.

Fig.4 Typical load curve of ECG

On the other hand, most of the outer hydropower is allocated to several provincial power grids using the same fixed proportions in each period.The deviation between the power profile and the load trend of each grid may be greater.In some cases, the opposite trend of the two may occur, which results in the passive absorption of significant lowvalley power by the receiving power grid.This increases the pressure of load regulation at low-load hours [24].

4 Solutions for integration of large-scale outer hydropower

4.1 Coordination of outer hydropower and local power sources

The ECG contains several types of power sources, which include conventional hydropower plants, pumped storage plants, thermal power plants, nuclear power plants, and large-scale external DC tie lines.Fig.5 presents the energy structure of the ECG at the end of 2018.The operation characteristics, constraints and control requirements are different for various power sources such as hydropower and thermal power, conventional hydropower, and pumped storage.Moreover, hydropower and thermal power, in addition to regional power supply and external power, have complementary regulation characteristics.By optimization and coordination for the enhancement of the peak regulation capacity of power grids, the problem of non-peak regulation or inverse peak regulation can be changed.The limitations to peak regulation in the ECG, especially the low-load peak regulation pressure, in addition to the utilization rate of inter-regional power channels and the level of hydropower absorption outside the region, can be overcome.

Fig.5 Energy structure for the East China Grid, 2018

In practice, the original problem can be decomposed into several sub-problems, such as pumped-storage, hydropower, thermal power, and DC power, among other power sources; according to the type of power supply.An appropriate optimization norm such as the minimization of the mean square deviation of residual loads in multiple power grids can be used to determine the power generation schedules of various types of power sources.This objective can be formulated as Eq.(1).Moreover, the power distribution method under the coordination mode of network economy was adopted to further distribute the output process of the power plants to each receiving unit.The distribution methodshould consider an important constraint to guarantee the power balance.Considering a hydropower plant as an example, this constraint can be expressed as Eq.(2).Finally, the coordination strategy can be introduced to realize the optimal coordination among different types of power supplies.

where denotes the original load of the power grid g in period t; denotes the residual load of the power grid g in period t; Xg is the average value of all during the time horizon; and are the power generation of the conventional hydropower plant a, thermal plant b, and pumped-storage plant e allocated to the power grid g in period t, respectively; and A, B, and E are the total number of conventional hydropower plants, thermal plants, pumpedstorage plants, respectively.

where denotes the power generation of hydropower plant a; is the unit generation u of hydropower plant a; and Ua is the unit number of hydropower plant a.

A novel strategy for the joint operation of large-scale UHVDC hydropower and conventional hydropower plants in the ECG was proposed [25].The coordination strategy can be summarized in four main steps:

Step 1: Decompose the joint operation problem into a UHVDC hydropower subproblem and conventional hydropower subproblem.

Step 2: The UHVDC hydropower subproblem is mainly related to the allocation of transmission power among multiple receiving-end provincial power grids.With respect to this subproblem, the generation schedules of the other power plants are considered fixed.A search algorithm is then developed to optimize the UHVDC hydropower allocation.This algorithm involves two steps, as presented below.

(1) Merge all the individual load curves into one load curve using Eq.(3).Then, an initial power distribution scheme is obtained using an improved load shedding method that can efficiently shave the peak load, as proposed in [26].

where is the total residual load for the ith UHVDC tie-line; Xg,max, Xg,min are the maximum and minimum residual load value of the gth power grid, respectively; Ri,g is the power allocation ratio, which is specified in the contracts signed between the dispatch center of the regional power grid and the subordinate provincial power grids.

(2) The previously obtained initial solution requires iterative adjustments to meet the power balance constraint (i.e., Eq.(2)) in each time period.First, the multiple objectives (i.e., Eq.(1)) are combined into an equivalent scalar objective using the weighted sum method.With the combined single optimization objective, the power distribution during the highest load periods is increased, and the power distribution during the lowest load periods is decreased in each iteration.The schematic diagram of the generation adjustment is presented in Fig.6.

Fig.6 Schematic diagram for allocating generation among power grids

Step 3: The objective of the conventional hydropower subproblem is to determine the generation schedule and the power allocation scheme of the hydropower plants.First, a multi-step optimization algorithm, as described in [27], is employed to optimize the generation schedule, and the power allocation scheme is then obtained by the abovementioned method.

Step 4: The entire iteration process is terminated if the deviation between the objective function values obtained in two adjacent iterations is within the allowable range.

4.2 Inter-provincial complementary power operations

Based on the complementarity of the load characteristics and the inter-provincial power network (see Fig.7), the potential of the optimum dispatching of pumped and stored high-quality peak shaving resources can be significantly utilized [28].Considering Shanghai and Anhui as examples, the Shanghai-Anhui joint peak shaving mechanism of the Xiangshuijian and Langyashan pumped-storage hydropower plants (PSHPs) can be established.During the peak period of summer, Langyashan and Xiangshuijian are in the operation “one pumping and two pumping” mode at night.The pumping capacity is shared by Shanghai and Anhui, and the early and late peak power generation capacity supports Shanghai and Anhui, respectively.During the spring and autumn flood season, two pumping operations of Langyashan and Xiangshuijian will be developed.At night, the pumping capacity will be shared by both sides.The waist-load pumping capacity will support Anhui, and the peak power generation capacity in the morning and evening will support Shanghai and Anhui, respectively.In the operation mode, the settlement of the exchanged electricity between two provincial power grids is asked to implement in current day.According to the calculation, the optimal operation of the mechanism involves the increase of the backup support capacity of 880 MW at night and 176 MW at the waist load of the Shanghai and Anhui power grids, respectively, to increase the backup support capacity of 800 MW and 1600 MW for the peak of the Shanghai and Anhui power grids, respectively; and to help the absorption of wind power and hydro energy from the south-west hydropower within the power area.

Fig.7 500 kV-level power network for the East China Grid

The abovementioned inter-provincial complementary power operations require a heuristic algorithm to respond to the peak loads of multiple power grids.Considering the direct operation of the PSHP by the ECG as an example, the heuristic algorithm mainly consists of three steps.

Step 1: As mentioned above, the direct operation of the PSHP by the ECG requires the simultaneous response to the load demands of multiple subordinate provincial power grids.The object function is presented below.

For simplicity, multiple load curves are combined into one load curve using Eq.(3), considering the characteristics of the load demand of each provincial power grid and the specified allocation ratio of energy production from the PSHPs to each power grid.

Step 2: After a total load curve for the PSHPs is created, the schedules of the PSHPs can be determined.Considering that the PSHPs should generate during peak periods and pump water during off-peak periods, the power generation is increased during peak periods, and the power consumption during valley periods is increased until all operation constraints are satisfied.For PSHP j, its detailed output adjustment process is shown below.

(1) Set the index of iteration i as 1;

(2) Search for k consecutive valley periods with load values that can satisfy Eq.(5).Moreover, t1 is the first index of the valley periods, and k is the minimum required pumping duration, which is assumed to be equal to 4 in this subsection.Hence, the final index of the valley periods can be represented by t4.The pumping power during these periods is then increased at a fixed step size.The residual load value is updated via Eq.(6) accordingly.It should be noted that the pumping power of the PSHP should satisfy constraint (7).

where ΔPPj is a fixed value, which is typically set as the rated power value of a single unit; and PPj,max and PPj,min are the maximum and minimum pumping powers of PSHP j, respectively.

(3) Similarly, the initial generating power can be determined using Eq.(8) and should satisfy constraint (9).

where ΔPGj is a fixed step size for the generating power adjustment; are the indices of the peak periods; and PGj,max and PGj,min are the maximum and minimumgenerating powers of PSHP j, respectively.

(4) Calculate the water storage and water discharge of the PSHP, and then determine if the water storage limits and water discharge limits can be satisfied (i.e., Eq.(10) and (11)).If not, the generating power or pumping power requires modification.

Fig.8 schematic diagram for determining the generation schedules of the PSHPs

where and are the upper and lower bounds of the upper reservoir, respectively;and are the upper and lower bounds of the lower reservoir, respectively; and are the maximum and minimum generating water discharges of the PSHP, respectively; and and are the maximum and minimum pumping water discharges of the PSHP, respectively.

(5) Determine if Eqs.(12) and (13) can be satisfied.If so, the iteration procedure is terminated.Otherwise, set I = I + 1, and repeat Steps 2)-4).The generation schedules of the PSHPs can then be obtained.The schematic diagram for the determination of the generation schedules of the PSHPs is shown in Fig.8.

Step 3: The algorithm described in Section 4.1 is then used to optimize the power allocation among several provincial power grids.

More detailed procedures of the operation strategies can be found in the previous work [29].

4.3 Introduction of a reasonable market mechanism

With respect to the market mechanism and operation optimization, the construction of the auxiliary service market for peak shaving in the ECG will be significantly promoted [30].The peak-shaving ancillary service market can realize the mutual benefit of inter-provincial peakshaving using the market-oriented mechanism, thus establishing the operation mechanism of the “demand declaration-provincial transaction organization-unified clearing-two-level security check” for network-provincial coordination, mobilizing the interest of power generation enterprises in participating in inter-provincial peak-shaving, and alleviating the complex peak-shaving problem in the ECG.A method is required by the peak-shaving assistantservice market for the determination of the proportion and cost of peak-shaving power allocation with respect to the multi-generation capacity of the power generation enterprises with insufficient peak-shaving, which effectively solves the peak-shaving cost allocation problem.

5 Conclusions

Based on this study, the use of a large grid platform to carry out a variety of load allocation methods and strategies of multi-grid power supply will help to improve outer hydropower integration.In addition, the findings of this study can effectively alleviate the complex problems related to peak load regulation, which commonly occur in the load centers in China.For large hydropower projects under construction in China, changing the existing dispatching methods and transmission schedules will promote hydropower integration into the eastern power grids such as the ECG.Moreover, it provides important technique supports for current and future large-scale hydropower transmission in China.

In summary, this paper suggests the following three suggestion strategies.First, the coordination of power sources within and outside the region should be implemented, thus fully utilizing the operation characteristics of different power sources and improving the regulation ability of the system.Second, the complementary operation among provinces should be coordinated to improve the resource allocation ability of the regional power grid through the power mutual assistance of two or more provinces.Third, it is essential and important to design a reasonable market trading mechanism to fully use price to efficiently guide the allocation of power resources.The above strategies are mainly designed and developed from the perspective of receiving power grids.Extensions of the research with sending power grids may provide good insight for finding other effective and efficient methods for hydropower transmission.This problem is more complex, which is recommended as another subject for further work.In addition, to promote the implementation of the suggested strategies, future research can focus on some specific models and methods for the coordinated operations of multiple power sources, complementary operations among provinces, and price compensation strategies.

Acknowledgements

This work was supported by the National Natural Science Foundation of China [No.51579029], and Fundamental Research Funds for the Central Universities (No.DUT19JC43).