Exploring Tradeo ﬀ s in Merged Pipeline Infrastructure for Carbon Dioxide Integration Networks

: Carbon integration aims to identify appropriate CO 2 capture, allocation, and utilization options, given a number of emission sources and sinks. Numerous CO 2 -using processes capture and convert emitted CO 2 streams into more useful forms. The transportation of captured CO 2 , which poses a major design challenge, especially across short distances. This paper investigates new CO 2 transportation design aspects by introducing pipeline merging techniques into carbon integration network design. For this, several tradeo ﬀ s, mainly between compression and pipeline costs, for merged pipeline infrastructure scenarios have been studied. A modiﬁed model is introduced and applied in this work. It is found that savings on pipeline costs are greatly a ﬀ ected by compression / pumping levels. A case study using two di ﬀ erent pipe merging techniques was applied and tested. Backward branching was reported to yield more cost savings in the resulting carbon network infrastructure. Moreover, both the source and sink pressures were found to greatly impact the overall cost of the carbon integration network attained via merged infrastructure. It was found that compression costs consistently decreased with increasing source pressure, unlike the pumping and pipeline costs.


Introduction
Increased climate change concerns have resulted in various efforts that aim towards mitigating CO 2 emission footprints. This has created pressure on the industrial sector to reduce emissions, especially since stationary industrial sources account for the majority of global emissions. Multiple methods to reduce CO 2 emission have been proposed, which include carbon capture utilization and storage (CCUS), fuel reduction, or fuel switching, including the use of renewable energy. Since carbon dioxide (CO 2 ) is a primary constituent of greenhouse gases (GHG) emissions, converting CO 2 into valuable products has been the main subject of many recent studies. Various utilization routes exist due to the versatile nature of CO 2 ; it becomes quite a challenge to identify the most viable option to consider [1]. Given that industrial emission sources can both be from energy use or as a product from processing activity, deployment of carbon capture sequestration and utilization infrastructure reduction schemes can be effective [2]. Many works have been published on carbon capture, Leung et al. [3] conducted a review that examines CO 2 capture and storage decisions that can meet a prescribed emission reduction target. Absorption processes were reported as the most utilized option, due to their relatively low cost and high efficiency. Shahbazi and Nasab [4] also investigated various carbon capture and storage (CCS) technologies that were reported to induce a noticeable decrease in the greenhouse gas emissions. Many CCS techniques were reported to be highly effective in serving to decarbonize the energy sector, particularly in countries that highly depend on fossil fuels for electricity production [4]. Having an efficient transportation scheme as part of the (CCUS) infrastructure is vital impact of the presence of impurities on pipeline performance using binary mixtures. Each binary fluid was studied at the maximum allowable concentration, and deviations from pure CO 2 at the same conditions were determined. These deviations were graded to rank the impurities in order of the degree of impact on each parameter. Liu and Gallagher [18] investigated cost-efficient solutions for the transportation of CO 2 in China. While growing efforts in China are underway to understand CO 2 capture and storage, comparatively less attention has been paid to CO 2 transportation issues, as no publicly available China-specific cost models for CO 2 pipeline transportation are available. Hence, a first-order estimate of China's cost of onshore CO 2 pipeline transportation was provided by Liu and Gallagher [18]. For this, an engineering-economic model based on China-specific data and codes was provided in their study. This included a sensitivity analysis in order to examine the effect of pipeline length and soil temperature on pipeline diameter onto the cost of such systems. Mack and Endemann [19] focused on investigating the legal and policy issues surrounding sequestration infrastructure, mainly CO 2 pipelines that will carry CO 2 from where it is removed from fuel or waste gas streams to sequestration sites. Ultimately, Mack and Endemann [19] recommended developing a federally regulated CO 2 pipeline program to foster the implementation of efficient carbon sequestration technologies. Guo [20] combined CO 2 supply of carbon capture and storage (CCS) with CO 2 injection rates to increase the net economic profit using a systematic optimization approach. The work utilized a special type of network structure and simultaneously addressed the optimal location of the potential hubs [20]. The proposed framework was applied to four different instances of the CCS and EOR network design combined with hub location selections.
Previous work focused on matching CO 2 sources with geological sequestration or utilization such as EOR. It explored CO 2 transportation over large distances in CCS/CCUS networks. This leaves a clear gap in CCUS pipeline network design in industrial clusters. Therefore, this study focuses on simultaneously assessing the added benefits of allowing merged pipeline infrastructures for the transportation and allocation of captured CO 2 streams into CO 2 using sinks, within industrial clusters. This approach provides the first assessment of simultaneous carbon integration and pipeline network optimization in close-range clusters. It is enabled by the systematic carbon integration technique that treats, compresses, transports, and utilizes CO 2 into value-added products. This capability gives a comprehensive evaluation of CCUS implementation costs. Pipeline merging methodology has been previously introduced for the design of interplant water networks [21]. The incorporation of such aspects into CO 2 network design has not been attempted before and was inspired by Alnouri et al. [21] since the pipeline merging concept itself is independent of the type of fluid flow involved. The various merging techniques that have been previously introduced simply describe how the pipe segments can be assembled to form a merged pipe, and the different techniques through which a merged pipe may connect sources to sinks. Moreover, since there has been very little research effort that is aimed towards addressing and improving CO 2 transportation across short distances, within industrial clusters, this paper helps demonstrate the sensitivity of carbon integration networks towards various elements that are inevitably important factors that enable cost-effective carbon dioxide transportation, via utilizing the novel pipeline merging techniques for assembling CO 2 networks. The next section outlines the new CCUS pipeline merging method, followed by case study results and discussion.

Materials and Methods
In industrial clusters, many sources and possible sinks of CO 2 exist. CO 2 sources are CO 2-emitting streams within an industrial process or a plant with a given CO 2 purity, pressure, temperature, and known location. CO 2 sinks are CO 2 utilization industrial processes, which can convert or sequester CO 2 at a given purity, pressure, temperature, and have a given location [14]. A single pipe allocation is often used to establish connectivity between any sources to any sink within a network. Figure 1 shows the network superstructure and illustrates a typical connection of CO 2 exchange from source to sink. The connection involves treatment, where CO 2 separation takes place, which is located at the source, compression through a compressor, and/or pumping to deliver to the sink. However, the notion of pipeline merging involves the utilization of common pipe infrastructure to transport material from source-to-sink locations. This is done via common segments that can be assembled together in shared regions, in order to eliminate the use of single pipeline connectivity that establish one on one source-to-sink allocations. Single connections and pipeline merging designs are shown in Figure 2. Pipeline merging greatly eliminates the unnecessary use of parallel pipelines that transport similar materials under similar conditions to and from common locations. It can be observed from the literature that both compression and pumping activities are vital for conditioning the CO 2 into an acceptable form that is safe to transport. Usually, critical or supercritical conditions may be favored for the transportation of CO 2 over large distances. Thus, compressed, treated pure CO 2 is included in this work. source, compression through a compressor, and/or pumping to deliver to the sink. However, the notion of pipeline merging involves the utilization of common pipe infrastructure to transport material from source-to-sink locations. This is done via common segments that can be assembled together in shared regions, in order to eliminate the use of single pipeline connectivity that establish one on one source-to-sink allocations. Single connections and pipeline merging designs are shown in Figure 2. Pipeline merging greatly eliminates the unnecessary use of parallel pipelines that transport similar materials under similar conditions to and from common locations. It can be observed from the literature that both compression and pumping activities are vital for conditioning the CO2 into an acceptable form that is safe to transport. Usually, critical or supercritical conditions may be favored for the transportation of CO2 over large distances. Thus, compressed, treated pure CO2 is included in this work.  It should be noted that compression usually consumes most of the required energy input when compared to pumping requirements. Moreover, booster pumps are only needed when it is desired to go beyond the critical conditions for CO2. Hence, most of the cost expenditures for a standard pipeline that is designed to transport CO2 transport would entail the operating expenditures associated with any compressors and pumps in the system. In order to assess the long-term economic feasibility of effectively running CO2 pipeline networks, the operating costs must be correlated to the energy consumption of both types of pressure changing equipment, and ideally, should be kept to a minimum whenever possible. Given the number of connections possible of similar CO2 qualities (pressure and composition), introducing pipeline merging into the picture would certainly allow additional cost savings to be attained in CO2 networks. Pipeline merging can reduce the overall capital expenditures on the pipelines, in addition to other associated costs. Thus, in addition to CO2 allocation, pipeline installment is optimized as outlined in section 2.1.

Mathematical Model
Pipeline merging modeling equations were obtained from the previous work [21]. The model involve the application of different merging techniques and are independent of the nature of the fluid being transported. The rest of the model has then been modified to accommodate the transportation of CO2 in the context of interplant networks.
Hence, the objective function that was previously utilized by Alnouri et al. [21] for merged water networks, has been replaced by Equation (1) below, which aims to minimize the total cost of the carbon network as follows: where , represents the total compression costs for the network, , represents the total pumping costs for the network, and represents the total pipeline costs for the network. Each of the cost items above has been computed using Equations (2)-(9), which have been adopted from Al-Mohannadi and Linke [14] and are summarized in below.
The overall cost of compression and pumping are given below The annualized capital cost of compressor and pump are given by equations (4) and (5). It should be noted that compression usually consumes most of the required energy input when compared to pumping requirements. Moreover, booster pumps are only needed when it is desired to go beyond the critical conditions for CO 2 . Hence, most of the cost expenditures for a standard pipeline that is designed to transport CO 2 transport would entail the operating expenditures associated with any compressors and pumps in the system. In order to assess the long-term economic feasibility of effectively running CO 2 pipeline networks, the operating costs must be correlated to the energy consumption of both types of pressure changing equipment, and ideally, should be kept to a minimum whenever possible. Given the number of connections possible of similar CO 2 qualities (pressure and composition), introducing pipeline merging into the picture would certainly allow additional cost savings to be attained in CO 2 networks. Pipeline merging can reduce the overall capital expenditures on the pipelines, in addition to other associated costs. Thus, in addition to CO 2 allocation, pipeline installment is optimized as outlined in Section 2.1.

Mathematical Model
Pipeline merging modeling equations were obtained from the previous work [21]. The model involve the application of different merging techniques and are independent of the nature of the fluid being transported. The rest of the model has then been modified to accommodate the transportation of CO 2 in the context of interplant networks.
Hence, the objective function that was previously utilized by Alnouri et al. [21] for merged water networks, has been replaced by Equation (1) below, which aims to minimize the total cost of the carbon network as follows: Min. C comp, TOTAL + C pump, TOTAL + C pipe (1) where C comp, TOTAL represents the total compression costs for the network, C pump, TOTAL represents the total pumping costs for the network, and C pipe represents the total pipeline costs for the network. Each of the cost items above has been computed using Equations (2)-(9), which have been adopted from Al-Mohannadi and Linke [14] and are summarized in below. The overall cost of compression and pumping are given below The annualized capital cost of compressor and pump are given by Equations (4) and (5).
Sustainability 2020, 12, 2678 6 of 14 C pump,CAPEX $ y = 1.11 × 10 6 P pump (F) 1000 + 0.07 × 10 6 × CRF The operating costs of the compressor and the pump are shown in Equations (6) and (7) C comp, OPEX $ y The cost of piping and pipe segments diameter are shown Equations (8) and (9) respectively, In the equations above, ∆P is the pressure difference in pipe segment, ∆P pipe is the pressure drop parameter associated with pipe segment, and Elec is the electricity price in $/kWh. D is the diameter of pipe, ν is the outlet velocity of source s to sink k, m is the molecular mass of carbon dioxide, and CRF is the capital recover factor. F is the CO 2 volumetric flowrate in pipe, T is the temperature of carbon dioxide source, and L is the length of pipe segment. C Pipe is the cost parameter of the pipe segment, P comp is the power parameter for the compressor, C comp,CAPEX is the capital cost of compression, C comp, OPEX is the operating cost of compression, C pump,CAPEX is the capital cost of pumping, and C pump, OPEX is the operating cost of pumping. The rest of the formulation that has been adopted from Alnouri et al. [21], namely equations (10)-(51), describe how the various pipeline merging techniques can be applied. The formulation have been kept the same and can be found in their article [21]. This non-linear problem was implemented using "What's Best 10.0" LINDO Global Solver for Microsoft Excel via a laptop with Intel Core i5 Duo processor, 8 GB RAM, and a 64-bit operating system.

Case Study Data
An illustrative example of an industrial cluster was used to study the cost trends and their variation using pipeline merging techniques as introduced by Alnouri et al. [21]. The industrial city considered has 6 carbon dioxide sources and 6 carbon dioxide sinks, which are distributed amongst 4 chemical plants operating within geographic proximity. The layout and distances were adopted from [21], in addition to the same two pipeline-merging techniques: (a) forward branching and (b) backward branching. Tables 1 and 2 below provide CO 2 source and CO 2 sink information in terms of volumetric flowrates under different pressures. * CO 2 sources are CO 2 -emitting streams from an industrial process that has a given CO 2 purity, pressure, temperature, and known location. ** CO 2 sinks are CO 2 utilization industrial processes that can convert or sequester CO 2 at a given purity, pressure, temperature, and has a given location.
Ten different source pressures (ranging from 1 bar to 50 bar) have been considered in this study, in addition to 3 different sink pressures (74, 101, and 151 bar). It should be noted that in real situations, it is unlikely to have all source pressures equal. However, the purpose of this study is to investigate the effects of different source pressures on the cost of the network. Hence, to conduct a fair comparison between the different cases, all 6 source pressures were assumed to be equal, and the same applies to all the 6 sink pressures. For instance, the case of 1 bar source pressure and 74 bar sink pressure, all 6 source pressures were considered to be at 1 bar, and each of those sources may supply various sinks together with 74 bar each as a sink pressure. Therefore, in this study, this extra condition has been assumed for pressure, and was applied for the various cases tested. This greatly facilitated the comparison process between the different cases studied and allowed for some substantiated conclusions in this regard. Source and sink volumetric flow data for the various pressures considered are provided in Tables 2 and 3, respectively. The contamination data for all the carbon dioxide streams (both sources and sinks) are provided in Table 3. The thickness required for carbon dioxide pipes is influenced by the pressure that the pipeline can withstand. In general, higher pressures would require thicker pipes. In this work, three different thickness levels were utilized (5, 10, and 20 mm), depending on the pressure level being applied [22]. The thickness was a specified parameter and was not optimized in this work. All sources have been assumed produce carbon dioxide, at no treatment costs, with the presence of some minor impurities. The respective impurity information (in ppm) for sources and the acceptable impurity levels for sinks is provided in Table 3, in which three different contaminants were considered.

Case Study Results
Tables 4 and 5 presents a cost breakdown for piping, compression, as well as pumping costs that are associated with the two different pipeline merging scenarios. There are no pumping costs associated with Case C for both branching and backward branching scenarios. This was due to the sink pressure setting at supercritical conditions, which can be achieved by compression only.

Discussion
From the results obtained, it is evident that backward branching was able to yield more cost-effective network schemes when compared to forward branching, regardless of the source or sink pressure being considered. The total carbon dioxide network cost was found to be the most expensive for Case A (considering the highest sink pressure), and the least expensive for Case C (considering the lowest sink pressure). The source pressure had a great effect on the overall network cost. It was found that higher source pressures (50 bars) tend to reduce the compression cost requirements. While the highest compression costs were associated with the lowest source pressure (1 bar) scenarios, for forward branching, as well as backward branching. This trend was a result of the higher-pressure difference between sources and sinks that needed to be supplied for all lower source pressure scenarios. Higher source pressures tend to reduce the pipeline cost requirements, while the highest pipeline costs were associated with the lowest source pressure scenarios, for both forward and backward branching. However, in varying sink pressures, the higher sink pressures (Case C) tend to increase the pipeline cost requirements. Whereas, the lowest pipeline costs were always associated with the highest sink pressure scenarios (Case A). This was seen in both forward and backward branching pipeline merging scenarios. It was due to the lower volumetric flowrate (and subsequently lower segment diameters) that needed to be transported and delivered to the various sinks, in case of all high sink pressure scenarios. The relative costs have been compared for the three different entities. The attained trends associated with backward pipeline branching are depicted in Figures 3-5 for compression, pumping, and piping costs, respectively. The trends associated with forward pipeline branching are depicted in Figures 6 and 7 for compression, pumping, and piping costs, respectively.         From the results obtained, both the source and sink pressures have a great effect on the overall cost of the merged network. The compression costs always tend to follow a decreasing trend with increased source pressure, unlike the pumping and pipeline costs. Hence, when more compression and pumping were required, pipe costs tend to generate more savings in the carbon integration design attained via merged infrastructure. On the other hand, compression costs tend to generate From the results obtained, both the source and sink pressures have a great effect on the overall cost of the merged network. The compression costs always tend to follow a decreasing trend with increased source pressure, unlike the pumping and pipeline costs. Hence, when more compression and pumping were required, pipe costs tend to generate more savings in the carbon integration design attained via merged infrastructure. On the other hand, compression costs tend to generate more savings when pipeline costs were reported to be the highest.

Conclusions
This paper discusses the various tradeoffs between compression, pumping, and pipeline costs associated with carbon networks. It investigated the application of the two different merged infrastructure scenarios, forward branching and backward branching, which was previously studied for water networks. A case study that involved the exploration of 30 different scenarios was tested for each merging technique, in order to identify which merging technique leads to more cost savings. It was found that backward branching was able to yield up to 15% more savings compared to forward branching schemes in carbon integration networks. Moreover, for cases where more compression/pumping is required, compression costs tend to increase, while the resulting pipeline costs decrease as a result of the lower volumetric flowrate transported. In contrast, pipeline costs tend to be on the higher side when the network compression requirements reduced, as a result of the higher volumetric flowrate transported. Additionally, compression costs consistently decreased with increasing source pressure, unlike the pumping and pipeline costs in the case of both forward and backward merging infrastructure.

Conflicts of Interest:
The authors declare no conflict of interest.

C comp, Total
Total cost of compression ($/y) C pump, Total Total cost of pumping ($/y) C Pipe Cost of pipe segment ($/y) ∆P Pressure difference in pipe segment (bar) ∆P pipe Pressure drop parameter associated with pipe segment (bar) Elec.
Electricity price in $/kWh D Diameter of pipe (m) ν Outlet velocity of source s to sink k (m/s) m Molecular mass of carbon dioxide g/mol CRF Capital Recover Factor F Mass flowrate in pipe (m 3 /s) T Temperature of carbon dioxide source ( • C) L Length of pipe segment (m) C Pipe Cost parameter of the pipe segment ($/y) P comp Power parameter for the compressor ($/y) C comp,CAPEX Capital cost of compression ($/y) C comp, OPEX Operating cost of compression ($/y) C pump,CAPEX Capital cost of pumping ($/y) C pump, OPEX Operating cost of pumping ($/y)