Carbon emission flow: from electricity network to multiple energy systems

1 Introduction

Global climate change, which is primarily attributed to excessive anthropogenic CO2 emissions, is a major threat to the sustainable development of the human society [1].Therefore, reducing carbon emissions has become an inevitable trend.Many countries have set ambitious emission reduction targets to mitigate global climate change.For example, the Chinese government plans to achieve 60%–65% reduction of carbon emissions per unit GDP by 2030 with respect to 2005 [2].Energy systems account for a major part of all anthropogenic carbon emissions.Therefore,the decarbonization of energy systems plays an important role in developing a low-carbon society.An effective method for analyzing and calculating carbon emissions of energy systems is key to achieving emission reduction targets and making low-carbon policies.

The calculation and analysis of carbon emissions have been investigated extensively.Most studies calculate carbon emission using macro statistical methods, where carbon emissions are calculated based on fossil fuel consumption and the emission factor [3-5].In addition, some studies estimate carbon emissions based on the life cycle assessment(LCA) method [6, 7].The LCA method considers carbon emissions over a long-term scale that encompasses the raw material, manufacture, construction, operation &maintenance, and decommission stages; this scale is referred to as the “cradle to grave” approach [8].However, carbon emissions could not be linked with the energy consumption behavior in these studies.Although carbon emissions are directly produced by fossil fuel combustion for energy generation, demands on the consumption side are actually the primary driving force of energy generation.Therefore,carbon emissions should be identified and observed from the energy consumption perspective.

The concept of virtual carbon emission flow (CEF)was proposed in [9, 10]to measure the embodied or virtual carbon emissions within international trades using the inputoutput (IO) method.In energy systems, different forms of energy are injected into the network by generators, flow continuously in the network, and are distributed to each consumer.CEF denotes the flow of embedded carbon emissions from the generation side to the consumption side accompanied by the energy flow.Reference [11]proposed the theory framework of CEF in general energy networks and presented an actual analysis of China’s energy pattern.On this basis, the CEF model for the electricity network was formulated in [12].Using this model, carbon emissions resulted from electricity transmission, distribution and consumption can be traced.Furthermore, the concept of CEF was extended to multiple energy systems (MES) in [13],where CEF models for different energy networks and energy conversion processes were explicitly established.Each ton of CO2 emissions can be accounted in the entire process of energy systems by tracing the CEF.Therefore,the carbon emission responsibility of different stakeholders in the energy system can be attributed.

This paper introduces the basic concept and key indicators of CEF and summarizes CEF models in the electricity network and MES that have been proposed in our previous paper.In addition, realistic case studies are implemented to show the virtual flow of carbon emissions in real-world energy systems.

2 Basic concept and key indicators of CEF

Energy sectors have network-constrained product chains,such as power systems and natural gas systems.Different forms of energy flow along the networks from generation to consumption, which leads to the flow of the embedded carbon emissions.CEF is defined as indiscriminate CO2 emissions associated with the energy flow.It is a basic tool for identifying the relationship between carbon emissions and the network-constrained energy flow.The distribution of CEF can be obtained according to the energy flow as they are tightly coupled.Furthermore, the embodied CO2 emissions in the entire process of supply chain, including production, transmission, distribution, and consumption,can be quantified by tracing CEF.Such estimates would be helpful for network upgrades and reasonably allocating emission reduction targets.

The key indicators of CEF include the carbon intensity and carbon emission flow rates.

(1) Carbon intensity (CI): CI denotes carbon emissions associated with per unit of the corresponding energy flow.The unit of CI is kgCO2/kWh or tCO2/MWh.For different parts of energy networks, three types of carbon intensity are defined.

(a) Generation carbon intensity (GCI) denotes carbon emissions emitted by energy generators for producing per unit of energy.GCI depends on the fuel type and energy conversion efficiency of the generator and is a parameter of the CEF model.For example, the GCI of a coal-fired thermal unit is usually larger than the GCI of a gas-fired thermal unit because coal is more carbon-intensive than natural gas.

(b) Branch carbon intensity (BCI) denotes the carbon emission flow along a branch that is associated with per-unit of energy flow.The BCI describes the relationship between the energy flow of a branch and embodied carbon emissions.

(c) Node carbon intensity (NCI) denotes the average carbon emissions associated with per unit of energy flow injected into a node.The carbon emissions of different branches flowing into a node are mixed and aggregated.The NCI of the node is equal to the average BCI of all inflow branches, using the branch energy flows as the weight.

(2) Carbon emission flow rate (CEFR): CEFR denotes carbon emissions associated with the energy flow during a unit time period.The unit of CEFR is kgCO2/h or tCO2/h.

The CEFR of a component in energy systems is equal to the product of CI and the corresponding energy flow:

where Ri denotes the CEFR of component i.ρi and Ei are the CI and energy flow, respectively, of component i.

The calculation method of CEF is not unique because it represents the virtual flow of carbon emissions.The proportional sharing principle (PSP) is adopted to calculate CEF.Its basic assumption is that electricity outflowing from a node shares the energy injection of this node at an equal proportion [14].The following two principles are considered to establish the CEF model [13]:

(1) Energy merging principle: the NCI of a node is regarded as the weighted average of all BCIs of the inflow branches:

where and represent the NCI of node n and the BCI of inflow branch i, respectively. denotes the set of inflow branches of node n.

(2) Energy dispatch principle: according to the PSP, we can deduce that the BCI of each outflowing branch is equal to the NCI of its injected node.

where denotes the injected node of branch j.

3 CEF in electricity network

3.1 Characteristics of CEF in electricity network

Different types of electricity generators show distinct emission characteristics in power systems.However,carbon emissions emitted by different generators are mixed.Therefore, the contribution of different generators to electricity consumption should be distinguished.In an electricity network, CEF is defined as the virtual network flow of carbon emissions accompanied by the power flow.CEF can be regarded as attaching power flow in each transmission line with a carbon emission label.

Fig.1 Illustration of CEF in an electricity network.Red and green dotted arrows indicate carbon-intensive and carbon-free flows

Fig.1 presents an illustrative example of CEF in an electricity network, where the solid arrow denotes power flow and the dotted arrow denotes CEF.The dotted arrows are shown in different colors to represent the level of carbon intensity of the branch CEF.The thermal unit is more carbon-intensive as it consumes fossil fuels, whereas the wind farm is carbon-free.Therefore, the BCI of l1 is larger than that of l2.When the power flows of branch l1 and l2 are mixed at node n, the carbon-intensive electricity produced by the thermal unit is “diluted” by carbon-free wind power.Therefore, the NCI of node n is less carbon-intensive than the BCI of l1.In addition, the BCI of both l3 and l4 are equal to the NCI of node n based on the PSP.

3.2 Calculation method for CEF model in electricity network

CEF is based on the power flow.The power flow distribution and GCI of all generators form the boundary conditions for CEF calculation.NCI is crucial to the quantification of CEF and should be calculated first, after which BCI and CEFR can be calculated.A matrix-based calculation method for the CEF model in an electricity network is proposed in [12].A series of matrices and vectors are defined and calculated, including the branch power outflow distribution matrix, nodal active power flux matrix, and vector of the NCI.

Considering that the NCI of a node depends only on the carbon intensities and power flow of its adjacent nodes and branches/generators, a recursive calculation method for the CEF model is proposed in [15].The NCI of all nodes can be sequentially obtained during the recursive calculation process.

The matrix-based calculation method is more direct but may involve heavy computation burden when applied to large-scale power systems due to matrix inversion.This problem can be avoided in the recursive calculation method.However, the recursion times would rapidly increase when applied to the radical distribution network.Combining the advantages of the matrix-based and recursive calculation methods, a coordinated calculation procedure for a realistic electricity network is proposed in [15].First, calculate the CEF results of a transmission network using the recursive method.According to the CEF of the root nodes obtained in the first step, the CEF results of the distribution network are then calculated using the matrix-based method.

4 CEF in multiple energy systems

4.1 Challenges of extending CEF from electricity network to MES

MES provide a pathway to integrate different forms of energy, including electricity, natural gas, and heating,which have the potential to reduce total CO2 emissions and represent a promising prospect.

Compared with the CEF in an electricity network,carbon emissions in MES are associated with both primary energy and secondary energy.The integration of different energy systems leads to the coupling of embedded carbon emissions.Carbon emissions in MES are coupled among two main components: energy transmission and energy conversion.In the former process, carbon emissions embedded in each form of energy are delivered with the energy network.Different energy networks have significant differences in operation characteristics, resulting in distinctions among the calculation models of CEF.In the latter process, the energy conversion results in the flow of embedded carbon emissions among energy systems of different forms.Moreover, the energy conversion between different forms also couples the flow of carbon emissions among different networks.Therefore, the CEF model should be reinvestigated for MES.

4.2 CEF model in ME S

Carbon emissions in MES include both actual carbon emissions that can be physically observed and virtual carbon emissions.Fig.2 presents CEF in MES.Carbon emissions are directly produced in the process of converting primary energy to secondary energy and flows along energy networks.The CEF in the gas network denotes the flow of actual carbon emissions because methane contains carbon, whereas the CEF in the electricity network or heating network is a type of virtual flow.During the energy conversion process, carbon emissions are transferred from one energy system to another.Ultimately, carbon emissions are accumulated at the consumer end.Therefore, the embedded carbon emissions of different energy sectors in the entire process can be calculated by tracing the CEF.

The CEF model in MES is explicitly formulated in [14],including the CEF models of energy networks and energy conversion processes.

Fig.2 CEF in MES.Solid arrows denote actual carbon emissions and open arrows denote virtual carbon emissions

Different CEF models are established for different energy networks linked to their disparate operation characteristics [13].The CEF model of a gas network is the simplest.If carbon emissions incurred by transmission losses are allocated to the network, the CEF is equal to the product of gas flow and the emission factor of natural gas.In the heating network, carbon emissions flow circularly along with heated water, as shown in Fig.3.Both the heating supply and return networks need to be considered,which makes the CEF model of the heating network more complicated than those of the other two networks.

Fig.3 General structure of a heating network

The energy conversion process involves energy conversion and storage devices, such as the combined heat and power (CHP) and electrical boiler (EB), which are generally modeled as an energy hub (EH) [16].The flow of carbon emissions during this process needs to be identified.Various energy converters can be classified into two categories: single input-single output converters and single input-multiple output converters.The CEF models of these two types of converters are first studied.On this basis, the carbon emission coupling matrix is established, which can describe the relationship between the CI of input ports and the CI of output ports for an EH.

According to the CEF models of the energy networks and EH, the CI and CEFR of all energy sectors can be calculated based on the energy flow.

5 Illustrative case studies of CEF

5.1 Energy interconnection system of Southeast Asia

With the conception of a global energy interconnection [17],power systems will be interconnected across countries and continents to promote the utilization of renewable energy.Transnational power flow will increase to 720 million kW by 2050 [18].

Fig.4 presents the transnational electricity trade in Southeast Asia in 2050 according to the planning scheme by the Global Energy Interconnection Development and Cooperation Organization (GEIDCO).

Fig.4 Electricity trade and CEF for the Southeast Asia energy interconnection system in 2050

The renewable energy in Cambodia, Burma, and Laos will be delivered to other countries via the interconnection transmission lines.The electricity trade results in the flow of carbon emissions among countries.Each country can be modeled as a node that is interconnected by transnational transmission lines.The GCI of the thermal units in these countries are set to 0.85 tCO2/MWh.In addition, the average carbon intensity of the Chinese electricity sector is assumed to be 0.6 tCO2/MWh.According to the energy merging principle, the NCI of each node is equal to the weighted average of the GCI of the inflow transmission lines/generators.Considering Burma for example, the electricity generated by thermal units and clean energy generators (nuclear, hydro, wind and PV) are 354 billion kWh and 3035 billion kWh, respectively.In addition, the electricity imported from China to Burma is 30 billion kWh.Therefore, the NCI of Burma is as follows:

The flow of carbon emissions among countries can be calculated using the CEF model, as presented in Fig.4.The results show that a considerable amount of carbon emissions is transferred from the electricity exporter to the electricity importer.Malaysia becomes the largest carbon emission exporter.The virtual carbon emissions that flow from Malaysia to the Philippines, Singapore, and Brunei are 10620 ktCO2, 8010 ktCO2, and 1068 ktCO2, respectively.In contrast, the Philippines and Vietnam become responsible for an additional carbon emission of 10620 ktCO2 and 42375 ktCO2, respectively.Moreover, carbon intensities associated with power flow of different interconnection lines vary depending on the overall carbon intensity of the importing country.For example, the carbon intensity associated with the outflow power of Cambodia is the lowest because 93% of the electricity in this country comes from renewable energy.

5.2 MES in Northern China

A realistic case study was implemented based on real-world MES in the Jing-Jin-Ji economic circle.The power system in the realistic MES consists of 7 buses that represent seven different regional power systems, which are interconnected through high voltage AC transmission lines of 500 kV.Additionally, most of the natural gas demand in this realistic system is supplied by gas sources in western China via the West-East Natural Gas Transmission Project.The natural gas system is modeled as a 6-node system connected by eight pipelines.

Fig.5 CEF and energy flow in realistic MES

The energy flow and CEF of the electricity and gas networks of this realistic MES at a certain time period are presented in Fig.5.

According to Fig.5(a), Beijing, Tianjin, and Tangshan are carbon import regions, which account for a large proportion of total electricity demands.The remaining four regions are carbon export regions.Approximately 34%of the total carbon emissions incurred by the electricity demand in Beijing come from other regions within the power network.Conversely, 77% (4776 tCO2/h of the total 6237 tCO2/h) of the carbon emissions emitted in Zhangjiakou are attributable to electricity demands in other regions.Carbon emissions flowing among regions account for approximately 38% of the overall carbon emissions(12943 tCO2/h of the 34178 tCO2/h).Moreover, the NCI of node 1 (Zhangjiakou) is the lowest NCI because electricity in this region is generated by carbon-free wind farms.

Fig.5(b) shows that, for a gas network, CEF is proportional to the gas flow as methane contains carbon.Yulin (gas source of Shan-Jing Pipeline) is the largest carbon export region.All embedded carbon emissions in Yulin flow to other regions by transmitting natural gas.In contrast, all embedded carbon emissions in Beijing come from other regions.The gas demands in Beijing,Tianjin, and Qinhuangdao account for 60%, 26%, and 14%,respectively, of the overall carbon emissions.

The results of this realistic case study show that the flow of carbon emissions is ubiquitous in energy systems and accounts for a considerable part of the total carbon emissions.Therefore, identifying the transfer of carbon emissions is significant for allocating the contribution of different sectors to carbon emissions.

6 Conclusions

This paper introduces the concept of CEF and summarizes CEF models in an electricity network and MES.In addition, illustrative case studies are presented to show the embedded flow of carbon emissions in real-world energy systems.The CEF model can effectively quantify carbon emission footprints of energy systems, providing better results than the statistical and LCA methods.

The CEF analysis provides useful and reasonable advice for mitigating global climate change.Emission reduction targets need be set and allocated to countries involved in climate change negotiations.The case study results indicate that a considerable amount of carbon emissions flow among countries.Energy export countries should take additional emission responsibility according to the international energy trade.Without considering CEF, carbon emissions can be attributed only to energy production, and thus, the evaluation results will be unfair to energy export countries.To this end, the CEF model would facilitate rational and equitable agreement in climate change negotiations by reasonably attributing carbon emissions.Additionally, more effective low-carbon measures can be taken by identifying“weak links” in the network using CEF results.

Acknowledgements

This work was supported in part by the Major International Joint Research Project of National Natural Science Foundation of China (No.51620105007), and in part by the Major Smart Grid Joint Project of National Natural Science Foundation of China and State Grid (No.U1766212).