Elsevier

Applied Thermal Engineering

Volume 148, 5 February 2019, Pages 1278-1291
Applied Thermal Engineering

Research Paper
Experimental analysis of a micro-ORC driven by piston expander for low-grade heat recovery

https://doi.org/10.1016/j.applthermaleng.2018.12.019Get rights and content

Highlights

  • Experimental characterization of novel micro-ORC driven by radial piston expander.

  • Performance evaluation in off-design conditions.

  • Steady-state detection algorithm tested on-line.

  • Feed pump electric consumption accurately measured.

  • Effect of variable electrical load on expander speed and power output analyzed.

Abstract

In this paper, a full experimental characterization of a micro-scale ORC system is presented. The facility under investigation is driven by a piston expander prototype, made of three cylinders arranged radially around the drive shaft. The system is rated for a thermal input around 30 kW, being suitable for residential, tertiary sector or small industry applications. It is conceived for exploiting low temperature heat sources, such as solar collectors, biomass boilers, geothermal energy or waste heat streams. The facility was provided with an electric boiler as heat source, which warms water up to 90 °C, and cold water at ambient temperature as heat sink. A test campaign was performed varying the hot source temperature and the organic fluid feed pump velocity, in order to characterize the system behavior at different off-design working conditions. The electric consumption of the ORC feed pump was measured, in order to quantify the actual impact of the auxiliaries on the overall efficiency. Moreover, the number of electric loads connected to the generator was varied, changing the equivalent phase impedance value, for evaluating the effect on the expander rotating speed and power output.

The experimental analysis demonstrated that small reciprocating expander is suitable for exploiting low enthalpy heat sources, with quite good performances compared to other architectures like scroll and screw expanders, more applied within low temperature sources. The results show that the gross electric power output varied between 250 W and 1150 W, depending on the expander speed and on the number of electric loads activated. The expander total efficiency showed a barely constant trend around 40%. The pump total efficiency varied between 10% and 20%, increasing with the pump rotational speed. The maximum ORC gross and net efficiency were 4.5% and 2.2% respectively, confirming that the auxiliaries impact cannot be considered negligible in such type of systems.

Introduction

New heat conversion technologies are currently achieving substantial interest both on industrial and research side. The Organic Rankine Cycle (ORC) is well suited for these applications mainly due to its ability to recover low-grade heat and, at the same time, the possibility to be implemented as distributed small-size generator for decentralized end-users’ energy production. Combined heat and Power (CHP) applications of ORC can be considered as alternative to traditional systems, in terms of energy saving and environmental conservation. As pointed out in [1], small-scale and micro-scale CHP systems are particularly suitable for applications in commercial buildings (such as hospitals, schools, industrial premises, office building blocks) and domestic buildings of single or multifamily dwelling houses. Small-scale and micro-scale CHP systems can help to meet a number of energy and social policy aims, including the reduction in greenhouse gas emissions, improved energy security, investment saving resulted from the omission of the electricity transmission and distribution network and the potentially reduced energy cost to consumer.

Thus, micro and small-scale ORC technologies are undergoing a rapid development, stating on the market as emerging and promising device to exploit low grade heat sources [1]. Many experimental small-scale ORC realizations are described and analyzed in literature. Landelle et al. [2] summarize in a comprehensive database data of ORC facilities available in the experimental state-of-the-art analysis. The database allows to evaluate ORC overall performances, in order to perform a fair comparison between realizations. The survey results show that, in the power range between 0.1 kW up to 10 kW, the most diffused applications are solar and waste heat recovery, while the coupling with biomass is less common. Moreover, the results confirm that the ORC achievable performances are strictly related to the operating conditions (hot and cold temperature levels), the selection of the most appropriate working fluid, the choice and sizing of the expansion machine.

In particular, the selection of the expander depends on several criteria: pressure ratio, thermodynamic performance, reliability, compactness [3]. However, in micro-power applications, economic, technical and operational constrains may often prove to be more important than efficiency. Also noise, vibration and dimensions aspects should be taken into account. Due to all these variables and limitations, the optimal technology and layout of expanders for micro-scale ORC systems are not detected yet. Positive displacement expanders, which are generally preferable to turbomachines for systems of power lower than 150 kW, have not achieved the technical maturity yet, and their architectures mostly derive from displacement compressors from HVAC&R and compressed air industries. When the difference between the temperatures of the heat source and of the cold sink is relatively low, also the expansion ratio available at the expander is limited. This condition is suitable for the scroll architecture, characterized by a low built-in volume ratio [4]. Most studies in literature refer to scroll compressors converted to expander, while in some cases original manufacturer machines or prototypes were adopted; for example, Abadi et al [5], investigated a 1 kW scroll expander developed by the company Air Squared, adopting a zeotropic mixture of R245fa and R134a as working fluid; they achieved a power output close to 1.2 kW with an expansion ratio of 3.3, with a maximum net cycle efficiency close to 6%. Ziviani et al. [6] tested a 5 kW open-drive scroll expander with R245fa with a hot source temperature equal to 85 °C and 110 °C, varying the ORC mass flow rate; they achieved the maximum power of 3.75 kW at 2500 rpm, with an isentropic efficiency equal to 55%. Ayachi et al. [7] performed a study on an ORC system driven by a prototype scroll expander, using R245fa superheated vapor as working fluid; the expander they tested produced a power output variable between 0.2 kW and 3 kW, with an expansion ratio between 2 and 3.8. In [8] Yun et al. implemented a test facility with two scroll expanders (derived from air compressors) running in parallel. They obtained a maximum power output equal to 1.7 kW and 3.4 kW for single and dual-mode respectively, for pressure ratios between 2.6 and 6.5. The maximum isentropic efficiency and cycle efficiency was 62% and 7.5% respectively, regardless of the operating mode. The screw model is often adopted as well, both in the single and double screw architecture, adapting to expansion ratios higher than scroll expander. Ziviani et al. [9] characterized the performance of a single screw expander modified from an air compressor, with two working fluids (R245fa and the mixture SES36). They reported a maximum power output close to 8 kW and a maximum isentropic efficiency close to 60%, while the ORC net efficiency varied between 2% and 9%. Reciprocating piston expander is less applied in kW-scale systems with low heat source temperature (thus lower available expansion ratio) because it is characterized by higher built-in volume ratio. Unlike scroll and screw models, the piston expander requires valves for suction and discharge, involving more complexity in the machine design and operation. Most examples are related to the swash-plate configuration in waste heat recovery from internal combustion engine (ICE) application: for instance, the authors of [10] presented an experimental study on a three piston swash-plate machine delivered by Exoes, reporting a mechanical power between 0.3 and 1 kW with an expansion ratio between 9 and 12.5. Oudkerk et al [11] presented an experimental characterization of an ORC system driven by a swash-plate expander, producing a mechanical output of 0.3–2 kW, with an expander isentropic efficiency variable between 33% and 53%. Dumont et al. [12] performed an experimental comparison on the same test bench between four models of volumetric expander (piston, scroll, screw and roots expander) in the kW-scale. The aim of their study was to facilitate the selection of the expansion machine in small scale facilities, but the performance they achieved were not optimized since the working conditions were not the most suitable for each machine model. Other studies [13], [14] describe the free piston expander-linear generator as solution for ORC waste heat recovery from ICE, achieving a maximum power output close to 20 W and to 100 W respectively. A comprehensive survey of experimental data on micro-scale ORC systems has been recently published by Park et al. [15]; they collected information from more than 200 references to compare working conditions and performances in terms of the common adopted indicators (power output, expander isentropic efficiency, BWR, cycle efficiency, etc.).

One more critical aspect of micro-ORC systems regards the feed pump selection and dimensioning [4]. The pump operation indeed, has a not negligible impact on the net power output and on the global efficiency of the system. The back work ratio (BWR), defined as the ratio between the pump absorbed power and the expander output, can be considerable if compared to the typical value for large scale and high temperature ORCs (between 1% and 10%) and for steam Rankine cycle (0.1–2%); in fact, the BWR increases if the critical temperature of the working fluid is low, and becomes very high when the evaporation temperature is close to the critical temperature. In some cases with very low power output [16], [17], the authors opted for a solution with pump-less ORC system to avoid the penalization due to the pump consumption. The pump total efficiency becomes therefore a more crucial parameter on lowest size applications (gross produced power < 10 kW), for which the pump consumption can account for a large fraction of the expander generated power. On systems of such size, experimental data on feed pump performance are not frequently and in detail presented in literature. In modeling studies, the pump efficiency is taken between 65 and 85%, but these values are based on pumps manufacturer data and often do not consider the off-design operation performance and the electric motor efficiency, which can be lower than expected especially if the motor is oversized [4]. For example, Quoilin et al. [18] achieved a pump total efficiency of 25%, while Reid et al. [19] report a value of 7%, both in kW-scale ORC units. In [20], Landelle et al. collected from literature several ORC pump efficiency data as function of the hydraulic power. They observed that the average efficiency is around 35% for small ORC power plants, but it is not specified if the electric losses are always included into the efficiency calculation.

In the experimental research, the methodology applied for the data analysis is of primary importance, to produce reliable and repeatable results that can then be compared to other cases with similar features presented in literature. The first step of the data processing is the detection of the steady-state operating points achieved during the experimental tests. Several methods have been proposed in literature, mainly applied in process or chemical engineering [21], [22], [23]; Woodland et al. [24] suggested a standard for ORC systems, which consists in considering the variations of the variables trend on a manually identified steady-state time window, through the comparison between simple average values taken at different time. Lecompte [25] applied a steady-state algorithm (derived from the one implemented by Kim et al. on an air conditioner [26]) on the experimental data of a 11 kW ORC system, based on the calculation of the moving standard deviation of the main process variables. Li et al. [27] adopted a similar method on their dynamic analysis of a transcritical CO2 power cycle for heat recovery from a heavy-duty diesel engine; they computed the absolute deviation on a moving time window equal to 5 s and compared it to pre-set thresholds. All the above-mentioned approaches are suitable for the post-processing application. The on-line implementation of a steady-state algorithm presents the advantage of improving the control during the test operations and of providing better sensitivity to the dynamic phenomena. On the other hand, the duration of the tests and the efforts for the data post-processing can be significantly reduced.

The aim of this paper is to provide a full characterization of a kW-scale test facility, conceived for exploiting low-enthalpy heat sources in micro-cogenerative applications for residential, small industry or tertiary sectors. The Authors’ opinion is that the improvement and the deployment of low-temperature heat conversion technologies is one of the key solutions contributing to primary energy savings worldwide and to the reduction of global emissions of pollutants and greenhouse gases.

The novelty of the system presented is the expansion machine, which is the radial piston prototype already introduced in [28]. To the Authors’ knowledge, reciprocating expanders with radial architecture are quite uncommon in the dedicated literature, especially with heat source temperature lower than 100 °C. Hence, the analysis presented in this study intends to help to understand potentialities, criticalities and possible improvements of power systems with such characteristics, providing a full set of experimental results that can be compared to other more conventional systems, in terms of working conditions and performance. With respect to the mentioned previous work of the Authors, the operating conditions investigated were extended, and a deep analysis was conducted on the experimental data, obtaining detailed operation and performance maps of the ORC system, that are presented in this paper. Moreover, the analysis of the expander performance was carried out changing the number of electrical loads connected to the generator, focusing on the effect on the expander rotating speed and power output. On the other hand, accurate experimental data on the circulation gear pump implemented on this test bench are provided: the actual power absorbed by the pump electric motor was measured, instant by instant, allowing to evaluate the pump total efficiency and the real net power output of the ORC system. The steady-state detection algorithm called R-test was tested online and the results compared to the post-processing applicable method. The faster calculation compared to the other mentioned methodologies makes the R-test a helpful tool for the comprehension of the dynamic phenomena and for the system control, if opportunely calibrated. A detailed analysis of the global uncertainty related to the performance evaluation was conducted too; the uncertainty contributions of the single components of the acquisition system have been estimated, considering the accuracy of the different sensors (including the signal cables), and the error introduced by the acquisition devices (responsible for analogical/digital conversion, cold joint compensation, etc.).

Section snippets

Micro-ORC and test bench description

A small-scale test rig was developed at the Laboratory of the University of Bologna for investigating the global performance of the prototypal micro-ORC energy system. The test system, presented in Fig. 1, consists of three loops, namely the hot source, the cooling system and the ORC circuit. In the hot source loop, the thermal input is provided by a 500 L electric water boiler, rated for a thermal power input of 32 kW. The variable flow centrifugal pump P2 circulates water with flow rate

Instrumentations and acquisition system

The test bench has been instrumented with a number of sensors in order to investigate the system behavior under different steady-state operating conditions. The main specifics of the measurement devices are listed in Table 1 with their associated Component Off-the-Shelf (COTS) accuracy values.

To measure extensively the performance of the system and its main units, temperature and pressure sensors have been placed at the inlet and outlet of each component of the ORC circuit (see Fig. 1). T-type

On-line steady state detection methodology

In this study, the on-line implementation of the steady-state detection algorithm R-test [32] was experimented. This approach was selected after a comparison against two other methods found in literature, namely the Moving Standard Deviation (MStD) [26] and the Wavelet transform [23], where the R-test demonstrated to be the most effective between the considered procedures, showing a good time response in all the tested conditions as well as the best match with the mean values calculated with

Measurement uncertainty

The uncertainty evaluation is applied to pressure and temperature measurements. In this section, the contributions to the overall uncertainty such as calibration process, connections, and acquisition devices are taken into account. Calibration procedure, described below, was performed for the overall pressure and temperature measurements chains, in order to reduce and compensate for noise originated from the electric connections.

Measurement instruments are installed at the inlet and outlet of

Test set points and operating conditions

The tests were conducted imposing different set points of the water temperature at the evaporator inlet (hot source temperature, Thot). The hot source temperature was varied in the range 65–85 °C. The effect of the cooling system temperature (Tcold) was also investigated, as the tests were performed in two different ambient conditions (namely in winter and summer time), affecting the condenser inlet temperature, which varied between 18 °C and 27 °C. The hot source and cooling system water flow

Conclusions

This paper presents an experimental study on a micro-ORC energy system suitable for low-temperature heat recovery. The peculiarity of this system is the expansion machine, which is a radial piston prototype, more often adopted with high temperature thermal sources. The other components are a gear-type feed pump with variable speed, two brazed plate heat exchangers as evaporator and recuperator and a shell and tube condenser. The heat source is made by water heated by an electric boiler with

Acknowledgments

The authors would like to thank StarEngine S.r.l. for the technical support during the performed test campaign. Gratitude must also be expressed to Dr. Eng. Valentina Orlandini for her precious contribution to this work. University of Bologna and University of Ferrara founded this project.

References (36)

Cited by (0)

View full text