CHEMICAL ENGINEERING TRANSACTIONS VOL. 63, 2017 A publication of The Italian Association of Chemical Engineering Online at www.aidic.it/cet Guest Editors: Jeng Shiun Lim, Wai Shin Ho, Jiří J. Klemeš Copyright © 2018, AIDIC Servizi S.r.l. ISBN 978-88-95608-51-8; ISSN 2283-9216 Total Site Utility System Structural Design Using P-graph Timothy G. Walmsleya,*, Xuexiu Jiaa, Matthias Philippb, Andreja Nemetc, Peng Yen Liewd, Jiří J. Klemeša, Petar S. Varbanova aSustainable Process Integration Laboratory – SPIL, NETME Centre, Faculty of Mechanical Engineering, Brno University of Technology - VUT Brno, Czech Republic bTechnische Hochschule Ingolstadt, Institute of new Energy Systems, Ingolstadt, Germany cFaculty of Chemistry and Chemical Engineering, University of Maribor, Smetanova ulica 17, 2000 Maribor, Slovenia dMalaysia – Japan International Institute of Technology (MJIIT), Universiti Teknologi Malaysia, Jalan Sultan Yahya Petra, 54100 Kuala Lumpur, Malaysia. walmsley@fme.vutbr.cz This paper explores the macro optimisation decisions of energy sources selection and the structural design of the utility system within the framework of Total Site Heat Integration (TSHI). Most TSHI research on utility systems focuses on optimisation of conventional Combined Heat and Power systems. To build a new Utility Systems Planner (USP) tool, P-graph has been selected as the optimisation tool. A critical element of USP is the inclusion of low-grade heat utilisation technologies within the considered superstructure. The USP outputs include the optimal structure of the utility system including the arrangement and size of each component and estimates for Greenhouse Gas and Water Footprints. The successful application of the USP to a representative industrial case study with district energy integration shows an optimal solution with a natural gas boiler, reciprocating gas engine, condensing economiser, steam turbine, thermocompressor, organic Rankine cycle, cooling tower, and electric chiller with a total cost of 14.893 M€/y. The new tool is a platform for launching further research including site-specific application, multi-period optimisation, and sensitivity analysis. 1. Introduction Total Site Heat Integration (TSHI) is a proven method that can lead to practical solutions with high energy and resource conservation that progress towards sustainable design (Liew et al., 2017). The concept of TSHI is to link processes via a central utility system to improve overall performance (Klemeš et al., 1997) and has been extended to non-industrial energy users as a Locally Integrated Energy Sectors - LIES (Perry et al., 2008). Several aspects of TSHI have been improved over the past few years with a focus on the targeting and design of the process-side of TSHI. To mention a few, Liew et al. (2013) looked at Total Site utility systems planning using a sensitivity analysis to set design loads. Chew et al. (2015) analysed the effect of process changes and modifications on TSHI with the view of maximising integration. Tarighaleslami et al. (2017) discussed how to modify and apply TSHI to sites that require non-isothermal utility such as hot water and chilled water. Aziz et al. (2017) introduced a new systemic framework for low carbon emissions industrial site planning. In TSHI, a central utility system connects the various on-site processes and district demand. In most studies, the design and optimisation of the central utility system concentrate on Combined Heat and Power and Steam Turbine optimisation. Tools such as R-curve analysis were applied to mega industrial sites to express the acute trade-off between heat and power (Matsuda et al., 2009). Boiler, gas turbine and steam turbine hardware models have been formulated to conduct co-generation optimisation and determine the marginal steam price (Sun et al., 2016). An area that needs further attention is the macro decision concerning the optimal energy resource selection and structural design of the utility system (Philipp et al., 2017). The aim of this study is to optimise the structural design of the Total Site (or LIES) utility system using P-graph as the optimisation tool. To achieve the aim, a new Utility Systems Planner (USP) tool is introduced, which can perform the optimisation and simultaneously quantify GHG and Water Footprints. The study investigates several DOI: 10.3303/CET1863006 Please cite this article as: Timothy G. Walmsley, Xuexiu Jia, Matthias Philipp, Andreja Nemet, Peng Yen Liew, Jiří Jaromír Klemeš, Petar S. Varbanov, 2018, Total site utility system structural design using p-graph, Chemical Engineering Transactions, 63, 31-36 DOI:10.3303/CET1863006 31 competing and complementary technologies and their application by integrating these into a P-graph superstructure. A LIES case study is presented to illustrate the method and its application. 2. Total Site Utility Systems Planner Tool using P-graph This study developed a new USP built within the P-graph framework. P-graph is a combinatorial optimisation framework aimed at optimising process networks with high combinatorial complexity (Friedler et al., 1996). The P-graph framework forms the basis of the software tool P-graph Studio (2017), which is used in this study. P- graph Studio offers users the freedom to build superstructures based on sequences and series of vertices that represent materials/energy (circular) and operating units (rectangular). The performance of operating units is based on proportional ratios of inputs and outputs (y = mx) while cost functions must follow linear functions (y = mx + c). Material vertices have fixed purchase/sale prices for inputs and outputs. Figure 1a presents the USP P-graph superstructure for the utility systems with resources at the top, demands at the bottom and energy generation and conversion unit operations and utility balance vertices (e.g. with suffix “_Bal”) in between. The model is built to meet inputted demands from a LIES. Resource inputs at the top of Figure 1a are grid Electricity (EL) import, Natural Gas (NG), Biomass, Geothermal heat, Fresh Water, and on- site GHG Emissions liability. The main energy conversion unit operations in the USP are: NG boiler, NG Combined Cycle Gas Turbine (CCGT), NG Gas Engine (GE), biomass boiler, geothermal, steam turbine, Thermocompressor (TVR), Mechanical Vapour Recompression (MVR), Organic Rankine Cycle (ORC), Heat Pump (HPP), Absorption Chiller (ACH), and Electric Chiller (ECH). The intermediate energy levels include: Electricity (EL), Very-High Pressure steam (VHP), High Pressure steam (HP), Medium Pressure steam (MP), Low Pressure steam (LP), High Pressure Hot Water (HPHW), Hot Water (HW), Cooling Water (CW), and Chilled Water (CHW). Dark red lines represent electricity flows, other shades of red and orange lines represent different steam levels including geothermal, blue lines are cold utility, and light blue are water flows. The USP model requires specifying (i) the industrial site and district energy demands, (ii) unit operation performances, (iii) electricity and fuel prices, (iv) discrete capital cost, (v) annual operating and maintenance costs, and (vi) direct and indirect GHG emissions and water use factors. These data may be inputted via the P- graph Studio graphical user interface. The objective function for the USP is Total Annual Cost (TAC) as defined in Eq(1). OMCCUCTAC  (1) Where UC is the utility cost, CC is the annualised capital cost, and OM is the operating and maintenance cost. The new USP applies a smart discrete Capital Cost accounting approach, which is carried out by the grey boxes next to each of the major unit operations. Conventional capital cost estimates usually assume continuous power- law functions which relate a key capacity dimension of a system to its cost. This method, although useful, can be problematic for real-world optimisation of utility systems because the capacity is often a series of discrete standard values, not a continuous function. This means, for example, if a site requires a 9.5 MW-fired boiler, which is not a standard rating, it would need to install a 10 MW-fired boiler. The invested capital is then for a 10 MW-fired boiler, which is the next size up of boiler. As a result, continuous power-law capital cost functions are converted into discrete costs based on standard unit operation capacities. These discrete points fall along a common linear function, as required for efficient implementation in P-graph. The applied linear Total Capital Cost, TCC, is presented in Eq(2).   bSaTCC  (2) Where S is the size dimension of the unit operation and constants a and b are specific to a unit operation. OM cost for the major equipment items is determined using Eq(3).  CFTCCiOM  (3) Where i is the capital cost fraction that is required for OM at full load and CF is the capacity load fraction. 3. Locally Integrated Energy Sector case study The new USP is applied to meet the energy demands of a LIES, which includes a representative industrial chemical processing site with a Site Pinch between the MP (deficit) and LP (excess) steam mains and district heating and cooling demands. The required energy flows of each utility are provided in Table 1. In P-graph, the performance values for unit operations are defined by proportional ratios of inputs and outputs. The performance values of primary energy generation and conversion operations are given in Table 2, and in Table 3 for the turbine. Unit operation performance ratios are based on standard thermodynamics and unit operation models, using JSteam Excel (2017) software by Inverse Problem for modelling boilers, Organic 32 Rankine Cycle, turbine and CCGTs; General Electric Power Generation (2017) for GEs; Kadant Johnson (2011) for TVR; Piller GmbH (2016) for MVR; and Gebhardt et al. (2002) for HPPs ACH, and Chillers. One of the limitations of P-graph is that these values are fixed and independent of the unit operations production rate and capacity. Once the model is solved, performance values based on the required capacity should be checked to confirm accurate representation of actual performance. It is important to note that energy content of steam and hot/cold water flows only account for the utility’s useful energy. For example, the useful energy of HP steam is the energy available above the liquid saturation point at the given pressure. As steam expands in a turbine, its useful energy flow may increase from input to output accompanied by a downgrading of quality. Table 4 presents the industrial energy and water prices and emission factors that have been applied in the model. Industrial energy prices for electricity, natural gas and biomass was extracted from Eurostat (2017) based on the EU-28 average. The location of the site relative to the biomass affects its price, which ranges from about 8 to 12 €/GJ with the midpoint selected for this study. Table 1: Utility data and net demand for the industrial site and the district heating and cooling Utility EL VHP HP MP LP HPHW HW CW CHW P (bar abs) 100 40 15 4 4 Tsupply / Treturn (°C) 500 250 198 144 144 / 60 60 / 40 30 / 40 1 / 5 Qdemand (MW) 5 0 10 35 -3 0 3* 40 2* Table 2: Performance ratios of major unit operations NG Boiler NG CCGT NG GE Biomass Boiler Geo- thermal TVR MVR ORC HPP ACH ECH Fuel -1 -1 -1 -1 -1 EL 0.300 0.410 0.170 -0.286 -0.060 -0.333 VHP 0.621 0.570 HP 0.122 -1 MP 0.075 0.351 1.915 1.066 LP 0.088 0.178 0.106 -0.872 -1 -1.111 HPHW 0.149 -1 Flue 0.029 0.150 0.050 HW 0.218 1 CW -0.830 0.714 -2.111 -1.333 CHW 1 1 BFW (t/MW) 1.136 0.543 0.301 1.043 0.829 Note: Positive values represent generation while negative values represent consumption Table 3: Turbine stage performance ratios VHP-HP Stage HP-MP Stage MP-LP Stage LP-HW Stage LP-CW Stage LP-CHW Stage Steam in -1 -1 -1 -1 -1 -1 Power 0.104 0.081 0.093 0.143 0.208 0.277 Steam out 1.059 1.060 1.030 1.062 0.997 0.997 Table 4: Industrial energy, GHG emissions and water prices and water footprint and emissions factors Electricity Natural Gas Biomass Geothermal Fresh water GHG emissions Consumption 81.4 €/MWh 6.0 €/GJ 10.0 €/GJ 2.7 €/GJ 0.2 €/t 5.0 €/t Export 56.0 €/MWh Direct EF 55.2 kg/GJ 1.3 kg/GJ Indirect EF 279.5 kg/MWh 5.5 kg/GJ 0.8 kg/GJ 0.2 kg/GJ Indirect WF 24.35 t/MWh 0.11 t/GJ 70.00 t/GJ 0.05 t/GJ The applied linear TCC that is applied using discrete capacity step sizes for each unit operation are presented in Table 5. The cost data sources of are: IEA (2015) for CCGT; Pihl et al. (2010) for biomass boiler; Gebhardt et al. (2002) for HHP, ACH, and ECH; General Electric Power Generation (2017) for GE; US DOE (2016) for steam turbines; Piller GmbH (2016) for MVRs; Ghirardo et al. (2011) for ORC. This study estimates an i value of 4 %, which comprises 2 % for operations and 2 % for maintenance (Smith and Hawkins, 2004). In determining 33 annual Total Cost, total capital investments are annualised using a discount rate of 7 % with an equipment lifetime of 30 y. (a) (b) Figure 1: Total Site Utility Systems Planner Tool – (a) initial superstructure and (b) cost optimal solution. 34 Table 5: Capital cost function coefficients and size basis NG Boiler NG CCGT NG GE Biomass Boiler Geo- thermal TVR MVR ORC HPP ACH ECH a (€/MW) 150,000 441,000 266,146 378,560 75,000 7,500 68,400 1,414,727 131,817 17,117 184,617 b (€) - - 107,850 3,400,662 - 263,250 499,416 46,766 133,017 46,766 Basis Fuel Fuel Power Fuel Fuel LP in LP in Power HW CHW CHW Unit step 2 MW 2 MW 0.5 MW 2 MW 2 MW 0.2 MW 0.2 MW 0.2 MW 0.2 MW 0.2 MW 0.2 MW The USP tool was applied to minimise the Total Annualised Cost of the considered case study. The TAC includes contributions from fuel use, capital investment for major unit operation components, fresh water consumption, and on-site GHG emissions (from burning fuels). For the given performance and price data, Figure 1b presents the optimal design solution of the utility system given the performance and cost data inputs. For the optimal solution, the Total Cost of the complete utility system is 14.90 M €/y. As overall fuel inputs, Natural Gas (NG) contributes 42.2 MW while geothermal supplies 35.0 MW, which was its set upper limit. NG fuel fires a regular boiler (86.5 % of NG) and a Gas Engine (13.5 % of NG). A steam turbine expanding VHP steam from the boiler generates 3.1 MWe, while the Gas Engine generates 2.5 MWe and Organic Rankine Cycle another 0.8 MWe. Of this, 5.0 MWe is supplied to the industrial site, 1.1 MWe is consumed within the utility system by electric chiller (0.6 MWe) and cooling tower (0.5 MWe), while the remainder is exported to the electricity grid (0.4 MWe). A novel part of the solution is the Thermal Vapour Recompression unit, which upgrades 6.1 MW of LP steam to MP steam using HP steam as the motive force, and the condensing economiser, which extracts 1.3 MW of HW from the two flue gas stacks from the boiler and Gas Engine. Given the Industrial Site Pinch sits between the MP and LP steam levels, it seems unusual to recover HW in a condensing economiser in a temperature range that has excess heat from the industrial site. However, district energy demands, combined with the energy demand of the BFW deaerator and application of Thermal Vapour Recompression and Organic Rankine Cycle operations, creates a lower LIES Pinch compared to the process demands of the Industrial Site. The Thermal Vapour Recompression process forms an open cycle heat pump where excess LP steam is upgraded to the MP steam level and supplies the BFW preheater and deaerator. District heating is another heat demand that requires HW. These additional, non-process energy demands, and utility operations modify the optimal utility system solution to warrant the installation of a condensing economiser to generate HW below the Industrial Site Pinch. This novel result reinforces the need to simultaneously select and optimise a utility system that includes non-conventional Combined Heat and Power operations, e.g. Organic Rankine Cycle and Thermal Vapour Recompression. To quantify the environmental performance of the utility system, GHG and Water Footprints have been estimated by the USP following the methods of Wiedmann et al. (2006) for GHG Footprint and Mekonnen et al. (2015) for Water Footprint. Only direct on-site GHG emissions and fresh water consumption are considered as direct costs to the site. The remainder of the GHG and Water Footprints are indirectly accounted through the energy and capital markets, where applicable. For this case study, the optimal solution has a GHG Footprint of 85,540 t/y, of which 90.8 % is emitted on the Industrial Site, and a Water Footprint of 908,700 t/y, of which 77.3 % is on- site fresh water consumption. Future work will investigate a range of prices for GHG and Water Footprints, including both direct and indirect emissions, to understand the impact of possible increases of emissions prices on the optimal solution. 4. Conclusions This study successfully developed a new Utility Systems Planner (USP) tool using P-graph to support Total Site Heat Integration. A representative industrial site case study with district heating and cooling is optimised to identify the best design of utility system. In this case, 42.2 MW of natural gas and 35.0 MW of geothermal heat are needed to satisfy total demands of 5.0 MW of electricity, 48.0 MW of heating, and 45.0 MW of cooling. The optimal solution contained a boiler, reciprocating gas engine, condensing economiser, steam turbine, thermal vapour recompression, organic Rankine cycle, cooling tower, and electric chiller. The new tool provides a platform for launching further research including site-specific application, multi-period optimisation, and sensitivity analysis of energy prices and unity operation performance. Acknowledgments This research has been supported by the project “Sustainable Process Integration Laboratory – SPIL”, project No. CZ.02.1.01/0.0/0.0/15_003/0000456 funded by EU “CZ Operational Programme Research, Development 35 and Education”, Priority 1: Strengthening capacity for quality research under collaboration agreement with the University of Maribor and the Universiti Teknologi Malaysia, and the Technische Hochschule Ingolstadt. References Aziz E.A., Alwi S.R.W., Lim J.S., Manan Z.A., Klemeš J.J., 2017, An integrated Pinch Analysis framework for low CO2 emissions industrial site planning, Journal of Cleaner Production, 146, 125-138. Chew K.H., Klemeš J.J., Wan Alwi S.R., Manan Z.A., 2015, Process modifications to maximise energy savings in Total Site Heat Integration, Applied Thermal Engineering, 78, 731-739. Eurostat, 2017, Database - Eurostat, Eurostat accessed 24.4.2017. Friedler F., Varga J.B., Fehér E., Fan L.T., 1996, Combinatorically Accelerated Branch-and-Bound method for solving the MIP model of Process Network Synthesis. Chapter. In: Floudas C.A., Pardalos P.M. (Eds), State of the Art in Global Optimization Vol 7, Springer, Boston, USA, ISBN: 978-1-4613-3437-8, 609-626. General Electric Power Generation, 2017, Case studies accessed 12.7.2017. Gebhardt M., Kohl H., Steinrotter T., 2002, Derivation of Cost Functions for Components of Energy Generation, Institute of Energy and Environmental Engineering, Duisburg-Rheinhausen, Germany. Ghirardo F., Santin M., Traverso A., Massardo A., 2011, Heat recovery options for onboard fuel cell systems, International Journal of Hydrogen Energy, 36 (13), 8134-8142. IEA (International Energy Agency), 2015, Projected costs of generating electricity, International Energy Agency, Paris, France. JSteam Excel, 2017, Inverse Problem Limited, Auckland, New Zealand. Kadant Johnson Inc., 2011, Thermocompressors, Kadant accessed 2.2.2017. Klemeš J.J., Dhole V.R., Raissi K., Perry S.J., Puigjaner L., 1997, Targeting and design methodology for reduction of fuel, power and CO2 on total sites, Applied Thermal Engineering, 17 (8-10), 993-1003. Liew P.Y., Theo W.L., Wan Alwi S.R., Lim J.S., Abdul Manan Z., Klemeš J.J., Varbanov P.S., 2017, Total Site Heat Integration planning and design for industrial, urban and renewable systems, Renewable and Sustainable Energy Reviews, 68, 964-985. Liew P.Y., Wan Alwi S.R., Varbanov P.S., Manan Z.A., Klemeš J.J., 2013, Centralised utility system planning for a Total Site Heat Integration network, Computers & Chemical Engineering, 57, 104-11. Matsuda K., Hirochi Y., Tatsumi H., Shire T., 2009, Applying heat integration Total Site based pinch technology to a large industrial area in Japan to further improve performance of highly efficient process plants, Energy, 34 (10), 1687-1692. Mekonnen M.M., Gerbens-Leenes P.W., Hoekstra A.Y., 2015, The consumptive water footprint of electricity and heat: A global assessment, Environmental Science: Water Research & Technology, 1 (3), 285-297. Perry S., Klemeš J.J., Bulatov I., 2008, Integrating waste and renewable energy to reduce the carbon footprint of locally integrated energy sectors, Energy, 33 (10), 1489–1497. P-graph Studio V5, 2017, University of Pannovia, Veszprém, Hungary. Philipp M., Schumm G., Peesel R.-H., Walmsley T.G., Atkins M.J., Schlosser F., Hesselbach J., 2017, Optimal energy supply structures for industrial food processing sites in different countries considering energy transitions, Energy, DOI:10.1016/j.energy.2017.05.062. Pihl E., Heyne S., Thunman H., Johnsson F., 2010, Highly efficient electricity generation from biomass by integration and hybridization with combined cycle gas turbine (CCGT) plants for natural gas, Energy, 35 (10), 4042–4052. Piller GmbH, 2016, MVR Blowers and Squeeze Oil Damper Bearing, Piller Blowers & Compressors GmbH, Moringen, Germany. Smith R., Hawkins B., 2004, Lean Maintenance: Reduce Costs, Improve Quality, and Increase Market Share, Butterworth-Heinemann, Oxford, UK. Sun L., Doyle S., Smith R., 2016, Understanding steam costs for energy conservation projects, Applied Energy, 161, 647-655. Tarighaleslami A.H., Walmsley T.G., Atkins M.J., Walmsley M.R.W., Liew P.Y., Neale J., 2017, A Unified Total Site Heat Integration targeting method for isothermal and non-isothermal utilities, Energy, 119, 10-25. US DOE (US Department of Energy), 2016, Steam Turbines accessed 24.7.2017. Wiedmann T., Minx J., Barrett J., Wackernagel M., 2006, Allocating ecological footprints to final consumption categories with input-output analysis, Ecological Economics, 56 (1), 28-48. 36