Assessment of plum rain’s impact on power system emissions in Yangtze-Huaihe River basin of China

House–time boundary of plum rain

To successfully measure the impression diploma of plum rain on SI and PV era, we first clearly outline the house–time boundary of plum rain. Based on the nationwide customary (GB/T 33671-2017) of plum rain monitoring indices carried out on 12/01/2017, the plum rain-affected area covers the Yangtze–Huaihe River basin, together with Jianghuai District, the Yangtze River Center, and Decrease Reaches and Jiangnan District, with an affected area of almost 500,000 km^2 36. On this foundation, we additional lengthen the affected area by 200 km, acquiring a surrounding distinction area (R2) with an analogous measurement. The typical SI worth in R2 is chosen as a criterion for the area not affected by plum rain. Underneath regular circumstances, the length of the plum rain interval extends from mid-June to mid-July of every yr. Right here, we choose 35-day information from the twenty sixth to the twenty ninth week of the yr, and the corresponding interval lasts from June 18 to July 22 (it’s assumed that the primary day of the yr is January 1, which can also be the primary day of the primary week). This length time can characterize the method of the gradual strengthening and fading of the impression diploma of plum rain.

SI information and CF calculation

A dataset containing 41-year (1980–2020) hourly RSISF information with a spatial decision of 1/2° latitude by 2/3° longitude, retrieved from MERRA-2³⁷, is employed to measure the impression diploma of plum rain on the SI in R1. The 41-year SI worth is averaged to 1 yr with 8760 values at every spatial decision. Contemplating that week-level information can higher reveal the impression of plum rain than can day-level (excessive volatility) and month-level (inadequate accuracy) information, we additional common the 8760 values into 52 week-level values. On this foundation, the typical SI within the area (R1 or R2) or province/municipality might be calculated by additional averaging the SI in any respect spatial resolutions throughout the area or province/municipality. The impression diploma of plum rain on the SI at a given location in R1 might be calculated because the ratio of the distinction between ({{I}}_{{{{{{rm{SI}}}}}}}^{{{{{{rm{MEAN}}}}}}}(w)) and ({{I}}_{{{{{{rm{SI}}}}}}}(c,w)) to ({{I}}_{{{{{{rm{SI}}}}}}}^{{{{{{rm{MEAN}}}}}}}(w)), the place ({{I}}_{{{{{{rm{SI}}}}}}}^{{{{{{rm{MEAN}}}}}}}(w)) is the imply SI throughout the wth week in R2, and ({{I}}_{{{{{{rm{SI}}}}}}}(c,w)) is the SI at a given location c in R1. ({{M}}_{{{{{{rm{R1}}}}}}}) denotes the set of spatial resolutions in R1, and ({Phi }^{{{{{{rm{W}}}}}}}) is the set of weeks throughout the plum rain interval.

The open-source World Photo voltaic Vitality Estimator (GSEE) mannequin³⁸ obtainable on the www.renewables.ninja net platform is adopted to estimate PV CFs, that are utilized within the CUC optimization mannequin. With using meteorological information acquired from MERRA-2, the mannequin considers the inputs of each the instantaneous irradiance and temperature and may regulate the photo voltaic aircraft tilt angle at a given location to yield higher outputs. Contemplating that the plum rain-affected areas happen close to latitude 30 °N, we set the optimum tilt angle of the fixed-tilt system to 26.6 levels, following the examine by Chen et al.³². Within the provinces in each R1 and R2, we take into account the areas in every province with a spatial decision of 1/2° latitude by 2/3° longitude and calculate the imply unit PV output in R1 and R2, respectively (Supplementary Fig. 5). The calculated hourly imply unit PV outputs are thought-about for future projections of the PV potential. Moreover, an in depth description of the GSEE mannequin might be present in Supplementary Observe 1.

Electrical energy system information

Coal-fired era information, gas-fired era information, PV era information, wind energy era information, different non-fossil power energy era information, and line transmission information for Jiangsu, Anhui, Zhejiang, Jiangxi, Hubei, Shanghai, and Hunan are collected from the China Electrical Energy Statistical Yearbook³⁹, China Energy Information⁴⁰, and energy grid firms. Aside from the PV era capability within the totally different provinces or municipalities, on the finish of December 2020, the whole wind era capability in Jiangsu, Anhui, Zhejiang, Jiangxi, Hubei, Shanghai, and Hunan reached 15.47, 4.12, 1.86, 5.10, 5.02, 0.82, and 6.69 GW, respectively, the whole hydropower era capability in Jiangsu, Anhui, Zhejiang, Jiangxi, Hubei, Shanghai, and Hunan was 2.65, 4.74, 11.71, 6.60, 37.57, 0, 15.81 GW, respectively, the whole nuclear energy era capability in Jiangsu and Zhejiang was 5.49 and 9.11 GW, respectively, and that within the different provinces reached 0 GW. The precise output of hydropower and nuclear energy is calculated primarily based on the annual utilization hours and put in capability of mills. Based on the 2020 provincial statistical yearbooks, the annual interactive electrical energy in Jiangsu, Anhui, Zhejiang, Jiangxi, Hubei, Shanghai, and Hunan reached 120.1, −58.6, 135.52, 15.98, −84.0, 86.96, and 69.25 TWh, respectively.

Based on the China Electrical Energy Statistical Yearbook 2020, the whole thermal era capability in Jiangsu, Anhui, Zhejiang, Jiangxi, Hubei, Shanghai, and Hunan is 100.79, 55.61, 63.58, 24.55, 33.16, 24.50, and 22.69 GW, respectively. The coal-fired era capability in every province might be obtained from the World Coal Plant Tracker⁴¹. Then, the gas-fired era capability might be obtained by subtracting the coal-fired era capability from the thermal era capability. Contemplating that the NG capability largely stays beneath 600 MW, we divide the thermal energy producing models beneath 600 MW into CGs and NGs primarily based on the ratio of the coal-fired era capability to the gas-fired era capability.

As a consequence of information limitations, the hourly provincial electrical masses from 6/18/2020 to 7/22/2020 are collected from provincial energy grid firms or fitted primarily based on typical provincial every day and annual load curves. Within the base case, we take into account that the capability change traits of coal-fired mills, gas-fired mills, PV methods, interchange tie-lines, and electrical energy masses over time within the totally different provinces or municipalities are the identical because the related predicted change traits in China referenced from the China Vitality and Electrical energy Outlook²².

Electrical energy dispatch mannequin

A CUC mannequin⁴² is employed in every province to calculate the CO₂ emissions of CGs and NGs within the totally different instances, with the benefits of using simply obtainable clustered unit information as enter to enhance the computational effectivity³⁴. The target perform in Eq. (2) minimizes the whole price of province-level energy methods throughout the plum rain interval, comprising the era price of thermal mills (TGs, together with CGs and NGs) within the first row, the penalty prices for photo voltaic curtailment in R1 (the second row) and R2 (the third row), and the extra storage price for PV integration and the penalty price for load curtailment within the fourth row.

Topic to

where the decision variable set X includes the hourly power of the ith clustered type of TGs ({P}_{i}^{{{{{{rm{TG}}}}}}}(t)), the hourly PV outputs of the l₁th PV generator ({P}_{{l}_{1}}^{{{{{{rm{PV}}}}}}}(t)) in R1 and the l₂th PV generator ({P}_{{l}_{2}}^{{{{{{rm{PV}}}}}}}(t)) in R2, the hourly charging/discharging power ({P}^{{{{{{rm{ES}}}}}}+}(t)/{P}^{{{{{{rm{ES}}}}}}-}(t)), the hourly electric load curtailment ({P}^{{{{{{rm{Loss}}}}}}}(t)), the maximum power capacity of energy storage ({C}^{{{{{{rm{ES}}}}}}}), an integer variable indicating the number of operating ith clustered types of TGs ({U}_{i}(t)), and an integer variable indicating the number of ith clustered types of TGs in the startup/shutdown state ({Y}_{i}(t)/{Z}_{i}(t)). Except for the decision variables in the objective function (Eq. (2)), ({rho }_{{{{{{rm{R1}}}}}}}^{{{{{{rm{PV}}}}}}}(t)/{rho }_{{{{{{rm{R2}}}}}}}^{{{{{{rm{PV}}}}}}}(t)) is the unit mean PV output in R1/R2, ({C}_{{l}_{1}}^{{{{{{rm{PV}}}}}}}/{C}_{{l}_{2}}^{{{{{{rm{PV}}}}}}}) is the installed capacity of the l₁th PV generator in R1 or the l₂th PV generator in R2, ({c}_{i}^{{{{{{rm{TG}}}}}}}) is the cost per MWh of the ith clustered type of TG, ({c}_{i}^{{{{{{rm{U}}}}}}})and ({c}_{i}^{{{{{{rm{D}}}}}}}) are the corresponding startup and shutdown costs, respectively. In addition, ({varphi }^{{{{{{rm{PV}}}}}}/{{{{{rm{Loss}}}}}}}) in the objective function (Eq. (2)) is the penalty factor in regard to solar power/load curtailment, and a sufficiently large value of 100 ¥/kWh is set to avoid solar power and load curtailments as much as possible. ({c}_{{{{{{rm{PR}}}}}}}^{{{{{{rm{ES}}}}}}}), which is equal to ({c}^{{{{{{rm{ES}}}}}}}cdot frac{T}{8760}cdot frac{r{(1,+,r)}^{{n}^{{{{{{rm{ES}}}}}}}}}{{(1,+,r)}^{{n}^{{{{{{rm{ES}}}}}}}}-1}), is the unit investment cost of energy storage during the plum rain period. In this equation, ({c}^{{{{{{rm{ES}}}}}}}) and ({n}^{{{{{{rm{ES}}}}}}}) are the unit investment cost and lifetime, respectively, of electric storage, T defines the number of hours during the plum rain period, and r is the discount rate. ({Phi }^{{{{{{rm{TG}}}}}}}) is the set of TGs, and ({Phi }_{{{{{{rm{R1}}}}}}}^{{{{{{rm{PV}}}}}}}) and ({Phi }_{{{{{{rm{R2}}}}}}}^{{{{{{rm{PV}}}}}}}) are the sets of PV generators in R1 and R2, respectively. ({Phi }^{{{{{{rm{T}}}}}}}) is the set of hours during the plum rain period.

Equation (3) expresses the balance for the power supply and load demand, where ({P}_{k}^{{{{{{rm{WT}}}}}}}(t)), ({P}_{m}^{{{{{{rm{RG}}}}}}}(t)), and ({P}_{n}^{{{{{{rm{TL}}}}}}}(t))are the hourly power levels of the kth wind power plant, the mth renewable generator (comprising hydroelectric and nuclear power units), and the nth interchange tie-line, respectively, and ({P}^{{{{{{rm{Load}}}}}}}(t)) is the hourly total electric load. ({Phi }^{{{{{{rm{WT}}}}}}}), ({Phi }^{{{{{{rm{RG}}}}}}}), and ({Phi }^{{{{{{rm{TL}}}}}}}) denote the sets of wind plants, renewable power generators, and interchange tie-lines, respectively. Constraint (4) expresses the system reserve requirements, where ({P}_{i,,max }^{{{{{{rm{TG}}}}}}}) is the unit mean capacity of the ith clustered type of TG, ({S}^{{{{{{rm{ES}}}}}}}(t-1)) is the ending hour-t-1 state of charge (SOC) of electric storage, and (RS(t)) is the hourly upward reserve margin⁴³. Constraint (5) bounds the lower and upper limits of the ith clustered type of TG, where ({P}_{i,,min }^{{{{{{rm{TG}}}}}}}) is the minimal power for one unit when the ith clustered type of TG is brought online. Constraints (6)–(8) express the relationships between ({U}_{i}(t)), ({Y}_{i}(t)), and ({Z}_{i}(t)), and limit the minimum on/off hours of the ith clustered type of TG, where ({T}_{i}^{{{{{{rm{U}}}}}}}) and ({T}_{i}^{{{{{{rm{D}}}}}}}) are the minimum on and off hours, respectively, of the ith clustered type of TG, and ({N}_{i}) is the total number units in the ith clustered type of TG⁴⁴. Constraints (9)–(10) impose the commitment order that TG 1 is committed first and subsequently TG ({N}_{i}) is committed last, where ({u}_{i}^{g}(t)) is a binary variable indicating the on/off states of TG g in the ith clustered type of TG⁴⁶. Based on similar nameplate capacities, six CG groups are clustered, and four NG groups are clustered (please refer to Supplementary Table 3 for details). Constraints (11) and (12) define the lower and upper limits of PV power generation in R1 and R2, respectively, where ({gamma }^{{{{{{rm{aerosol}}}}}}}) is the percentage increase in the PV potential caused by anthropogenic aerosol emissions. Following the study of Sweerts et al.¹³, we assume that Chinese anthropogenic aerosol emissions in 2060 are consistent with those in 1960, which could yield a 12–13% increase in PV generation by improving the all-sky SI. Although the CO₂ emissions in China will continue to increase and a peak is expected to occur in 2030, the implementation of ultralow emissions standards for coal-fired power plants⁴⁵ and alternate electric power policies such as electric heating policies⁴⁴ have improved the clean use of coal. Therefore, starting in 2020, we consider the PV potential to increase by 3% every 10 years. Constraints (13)–(15) impose power capacity and energy constraints on the charging, discharging, and SOC levels of energy storage, where ({eta }^{{{{{{rm{ES}}}}}}+/-}) is the charging/discharging efficiency of energy storage. Constraint (16) ensures a nonnegative SOC value and that the starting and ending states remain the same. Here, Li-ion storage with a duration up to 4 h at the rated power²⁵ is chosen as a representative option, and the energy capacity cost is set to 1380 ¥/kWh⁴⁷. Constraint (17) defines the bounds of the integer variables.

The CO₂ emissions of TGs are calculated as follows:

where ({e}^{{{{{{rm{C}}}}}}}) and ({e}^{{{{{{rm{N}}}}}}}) are the emission factors of the CGs and NGs, respectively, ({Phi }_{{{{{{rm{C}}}}}}}^{{{{{{rm{TG}}}}}}}) and ({Phi }_{{{{{{rm{N}}}}}}}^{{{{{{rm{TG}}}}}}}) denote the sets of CGs and NGs, respectively, ({P}_{p}^{{{{{{rm{TG}}}}}}}(t)) is the hourly power of the pth clustered type of CG, and ({P}_{q}^{{{{{{rm{TG}}}}}}}(t)) is the hourly power of the qth clustered type of NG. The sum of the two amounts equals the outputs of the TGs at any time. The fuel costs and emission factors of CGs and NGs are summarized in Supplementary Table 4.

By replacing constraint (11) with constraint (19) and re-optimizing the CUC model, the power system operation status and corresponding carbon emissions eliminating any plum rain effects can be obtained.

Where ({{{theta }}}_{{{{{{rm{P}}}}}}}^{{{{{{rm{MEAN}}}}}}}(w)) is the provincial mean impact degree during the wth week and ({Phi }_{w}^{{{{{{rm{T}}}}}}}) is the set of hours during the wth week.

Finally, the ICEs (varDelta E) caused by plum rain can be obtained as the difference between the CO₂ emissions determined via Eq. (18) under the first optimization ({E}^{(1)}) and those determined under the second optimization ({E}^{(2)}).

Comparison of the various technologies

According to the different mitigation principles of carbon emissions, we consider four pathways to address the negative impacts of plum rain: C2N, C2N + DR, C2N + CG with CCUS, and C2N + LD. The CUC models considering C2N, C2N + DR, C2N + CG with CCUS, and C2N + LD are presented in Supplementary Note 2. To compare their performance levels, we further define the LCCM as follows:

where (varDelta C) is the total additional cost when offsetting the ICEs caused by plum rain.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.