**1. Introduction**

In the framework of the "20–20–20" targets [1], measures adopted by the European Union to mitigate climate change include the improvement of conventional prime movers' conversion efficiency. As indicated in the European directive 2012/27/EU [2], the promotion of cogeneration, as well as waste heat recovery (WHR) strategies, could significantly contribute to achieving this objective.

Internal combustion engines typically convert just the 30–45% of the fuel primary energy into mechanical or electric power, while the remaining part is rejected as thermal energy with exhaust gases, cooling water, oil, etc. This heat, which would be wasted otherwise, can be used for combined heat and power (CHP) applications, thus improving the overall first law conversion efficiency, but also

for additional electric production, via a bottomer thermodynamic cycle. Heat released by internal combustion engine (ICE) exhaust gases is usually characterized by (i) limited thermal power size (varying typically in the range from kW to tens of MW as order of magnitude), (ii) by temperature often quite lower than 500 ◦C, and (iii) both strongly variable with the engine load. These conditions are not always compatible with the adoption of conventional steam Rankine cycle [3] since problems related to the use of water as working fluid can arise as (i) difficulty in superheating the fluid in order to avoid condensate formation during the expansion process and the risk of erosion of the turbine blades [3]; (ii) working with excessive vapor operating pressure that impose to install complex multi-stage and expensive turbines [3]. A solution is identified in changing the operating fluid with an organic compound in an organic Rankine cycle (ORC). Organic fluids are characterized by higher molecular mass and lower critical temperature compared to water that allows exploiting efficiently, also, low-grade heat sources. Even though the conversion efficiency of ORC is typically limited, compared to steam cycle, this technology features a number of advantages: lower O&M and personnel costs; high molecular weight; low enthalpy drop in the expander, and, as a consequence, higher mass flow rates, if compared with water. Moreover, ORCs are characterized by a wider regulation range, and performances are not particularly penalized at part-load conditions [4].

WHR from stationary ICE prime movers is quite a consolidated option, and several investigations have been proposed on this topic (see for example [5–14]). The studies point out that ORC performance is significantly affected by the choice of the working fluid, the cycle architecture, and the setting of the operating parameters. Investigated fluids are hydrocarbons, refrigerants, and zeotropic mixtures, while the other most important decision variables are identified in the evaporating pressure and the superheating degree [5–7]. In particular, Yang et al. analyzed an ORC system in [5] to recover waste heat from the diesel engine exhaust, where the zeotropic mixture R416A was used as the working fluid for the ORC. Considering various operating conditions of the diesel engine, this study investigated the effects of the degree of superheating on the running performance of the ORC waste heat recovery. In the thermodynamic study by Han et al. [6], a regenerative ORC system was established to recover the waste flue gas of 160 ◦C, focusing on thermodynamic and economic performance while simultaneously considering the effect of superheating and working fluid selection. The optimization of the evaporation temperature was carried out by analyzing the variation of net power output and specific investment cost. Energy and exergy analysis of three ORC–WHR configurations that use a coupling thermal oil circuit was performed by Valencia et al. in [7], where a simple ORC, an ORC with a recuperator, and an ORC with double-pressure configuration were considered; cyclohexane, toluene, and acetone were simulated as ORC working fluids. In addition, the effect of evaporating pressure on the net power output, thermal efficiency increase, specific fuel consumption, overall energy conversion efficiency, and exergy destruction was also investigated.

These studies demonstrated also the importance of taking into account the part-load operation of the ICEs when evaluating the performance of the ORC bottoming solution. For example, in [8] and [9], the performance of combined systems with ORC and ICE was analyzed at steady-state under different typical working conditions, by means of a numerical simulation model of the system. These papers analyzed, in addition, the effect of adjustable parameters on the system performance, giving effective control directions under various conditions. They proved that to get a better system performance under different working conditions, the system should be operated with a slight degree of superheating.

Valorization of waste heat through ORC creates interesting opportunities, especially to combined heat and power generation, in the residential sector and in energy-intensive industries. In this context, a series of hybrid power plants comprising a CHP prime mover integrated with ORC have been investigated [10–13], with the aim of maximizing the share of electricity production of CHP plants rather than a higher heat production rate in a cost-effective manner. The analyzed heat sources are usually cooling water from steam cycles or exhaust gas from the industrial process. An example of a hybrid CHP plant was proposed by Yi et al. in [10], where an ORC and a hydraulic turbine were introduced to reduce exergy losses, which occurred in a thermal storage process for district heating (DH) purpose. Results showed a payback period equal to just 3.5 years. The study proposed by Arabkoohsara and Namib [11] analyzed a waste-driven CHP plant. In addition to the electricity generated by the main steam cycle, the heat withdrawn from the condenser was employed for additional heat and electricity production by employing a small-scale ORC unit and a heat exchanger connected to the local district heating system. This study evaluated an electric efficiency improvement of 20% and a favorable period of return on investment. Marty et al. [12] analyzed the optimization of parallel distribution between electricity and heat production for a geothermal plant. The geothermal fluid was split into two streams, one used for an ORC system, and the other for the DH. An optimization tool was implemented to obtain the sizing of the ORC and the best distribution between electricity and heat production. In the project by Ramireza et al. [13], a large-scale ORC pilot plant along with a waste heat recovery unit in a steel mill was designed. Waste heat was recovered from the fumes of the furnace to produce saturated steam, which was then delivered to a DH network during the cold season and to the ORC for electricity generation during the rest of the year. Results on the performance of the plant during the first months of operation showed promising values. The study by Grljuši´c et al. [14] showed instead an application involving a diesel engine for maritime application. A standard operating propulsion engine at maximum efficiency was assumed to supply the auxiliary power. The aim of this research was to investigate the possibility of using the CHP plant over a ship to meet all heating and electricity requirements during navigation.

All these studies make evident that the ORC design for CHP hybrid systems has to be determined in accordance with the specific application. In particular, in the case of a CHP ICE-ORC hybrid system, the optimal ORC sizing, the optimal load allocation, and the resulting off-design performance should be explored more in depth, especially if it is called to operate under variable thermal demand conditions.

#### *The Objective of the Study*

The main objective of this work was the energy and economic assessment of a CHP power plant based on ICEs, feeding a DH network, and coupled with an ORC to optimally recover the ICEs' residual heat. In this work, a specific case study of an existing CHP system was analyzed; however, the same presented approach and similar considerations could be applied to any CHP feeding a district heating network (DHN) as a thermal user.

In detail, as the first step, the optimal operation of the CHP power plant was assessed without ORC: the ICEs load profile was determined, in order to fulfill the DH thermal demand, while minimizing the heat dissipation to the ambient. Then, in order to increase the amount of produced electrical energy and to better exploit the available ICEs residual heat, the opportunity of integrating an ORC as a bottomer was investigated.

The original power plant optimal operation was compared with the new arrangement, where the ICEs were operated at full-load, and the ORC was integrated also with the DH demand, to recover the ICEs' residual heat. The ORC as bottomer of the CHP was an interesting solution also because it could represent a flexibilization tool in the regulation of the whole plant. The ORC allowed operating the ICEs at full-load (i.e., at maximum efficiency point) while not dissipating the surplus of heat production when it is not required by the thermal user. Indeed, the surplus of high enthalpy-heat production was recovered by the ORC and converted into additional electric power, sold to the electric grid.

In detail, three different system architectures were proposed and analyzed. A sensitivity analysis of organic fluid and key cycle parameters was performed in order to identify, for each proposed arrangement, the optimum design of the ORC. A thermodynamic design and off-design model of the ORC system has been developed in the *Thermoflex*™ environment [15]. The additional electrical energy produced during the year, the ORC operating hours, and total investment costs were quantified and compared among the analyzed layouts. Finally, the original optimal operation of the CHP plant was compared with the modified one comprising the ORC, in order to estimate, for each proposed arrangement, the amount of additional generated electric power and the corresponding increase in primary energy consumption.

An energetic index accounting for the global primary energy saving was introduced, in order to evaluate the convenience in producing the additional electric power with the modified CHP plant rather than with the original plant with the help of the electric generating mix. From the economic point of view, considerations on the electric energy sell price to return on investment were also presented.

#### **2. Cogeneration Power Plant Description**

This section presents the description of the CHP plant feeding the DH network and the control logic adapted to regulate the system when working at part-load conditions.

#### *2.1. CHP Plant Arrangement*

The case study system was an existing cogeneration plant intended to satisfy the heat demand of an international hub airport; it comprised three ICEs (model Rolls Royce Bergen engine B35: 40 V20 AG), fed with natural gas, named GR1, GR2, and GR3, as depicted in Figure 1. The CHP single unit prime movers were conceived to provide a nominal electric power equal to 8.5 MW and thermal power of about 4.6 MW (only heat recovery from the exhaust gas is considered). Thus, the nominal electric and thermal efficiency were equal to 47% and 25%, respectively. The DH demand was fulfilled by the ICEs discharged heat and by three additional boilers, producing 8 MW of nominal thermal power output each, also fed with natural gas. The generated electric power was fed into the grid at a connection voltage of 150 kV.

The DH network (DHN) was designed to provide hot water for various types of tertiary building users. It included heat exchangers for heating purposes in winter and heat exchangers working all along the year, providing high-temperature heat to produce food use steam and to feed absorption refrigeration units, but also sanitary water.

For what regards the management strategy, a DHN could be regulated (i) varying the mass flow rate but keeping constant the temperature difference (Δ*T*) on the utilities (this approach is particularly suitable for small size DHNs) or (ii) varying the Δ*T* on the utilities with a constant mass flow rate (usually for large DHNs) [16]. The regulation adopted in this study was the first one: in particular, it was assumed that the delivery temperature was kept equal to 130 ◦C and the return equal to 70 ◦C. This supposition could be reasonable if considering that the DHN inertia was typically such high to presume to be neglectable the return temperature oscillation when relying on an hourly basis (which was the timestep considered in this analysis).

The yearly required thermal production profile is shown in Figure 2. Higher thermal power requests were observed during winter, with a demand peak close to 20 MW; lower heat demand values, around 3 MW on average, were registered during summer months.

**Figure 1.** Schematic of the cogeneration power plant.

**Figure 2.** Heat demand yearly profile.
