Making use of geothermal brine in Indonesia: binary demonstration power plant Lahendong/Pangolombian

Abstract

This paper describes a novel concept for integrating binary plants at a geothermal site which has been realized within a demonstration project in Indonesia. By using an intermediate hot water and an intermediate cooling water cycle it is possible to integrate a fully-automated binary plant with flexible operation in an existing, predominantly manually operated geothermal field. The paper gives technical component details and describes specific design considerations of the demonstration power plant. Furthermore operational data and experiences are shared. Current binary plant operation is characterized by many starts and stops because of several technical challenges, e.g. electrical grid conditions, as well as by off-design conditions. The proven maximum power capacity until now is approximately 400 kW. By means of a numerical model it is shown that the design capacity of 500 kW can be reached.

Introduction

Indonesia is known for its tremendous geothermal potential of around 29 GW_e which is dominated by wet steam fields. The currently installed capacity amounts to 2 GW from 15 areas (Darma 2016; Richter 2018a), leaving a significant amount of the geothermal potential untapped. Considering the continuously increasing need for electricity in Indonesia, there are strong efforts to further develop the geothermal power capacity.

The prevailing geothermal plant type in Indonesia is currently the single-flash plant which directly uses the steam phase from the produced steam-liquid-mixture to drive the turbine. Binary plants which transfer the geothermal heat to a separate working fluid are not yet an established technology at Indonesian sites. The first commercial binary units were commissioned at Sarulla field in 2017 (Wolf and Gabbay 2015; Richter 2018b) although due to their adaptability they could be implemented at much more sites and increase the geothermal capacity in Indonesia. Binary power plants can extend the plant capacity at high enthalpy fields but can also be used to exploit low to intermediate temperature geothermal reservoirs and realize small scale geothermal power plants.

In order to successfully demonstrate geothermal binary power plant technology at an Indonesian site and to intensify the know-how transfer in this technology field, a German-Indonesian collaboration project was initiated in 2013 between GFZ Potsdam (Germany), the Agency for the Assessment and Application of Technology in Indonesia (BPPT) and PT Pertamina Geothermal Energy (PGE). The project was based on a bilateral agreement between the German and the Indonesian Ministry of Research (Erbas et al. 2015).

The demonstration plant is located in the Lahendong geothermal area close to the village Pangolombian in the northern part of the island of Sulawesi. The first 20 MW single-flash unit in this area has been commissioned in 2001. Meanwhile six 20 MW units are in operation. The on-site construction phase of the demonstration plant, which is rated with a capacity of 500 kW, started in 2015 (Frick et al. 2015). Technical concept development, component specification, coordination and supervision of the detail planning as well as construction and commissioning were executed by the project consortium under the guidance of GFZ. The operational phase commenced in September 2017. In January 2019 the demonstration plant was handed over to the Indonesian consortium and is now operated and maintained by the Indonesian project partners. Besides commercial operation the plant will be used for demonstration activities and training.

This paper describes the technical concept and summarizes the first operational experiences of the demonstration plant that has been integrated at an already existing site. With this plant concept it was possible to realize a fully-automated binary plant concept with flexible operation in an existing, predominantly manually operated geothermal field.

State of knowledge

Geothermal binary plants use different concepts of site integration. Binary plants complement the direct steam use at a geothermal high enthalpy site by either using the liquid phase from the separator or by utilizing the waste heat from the direct-steam turbine. Geothermal binary plants can also use the complete fluid flow from a geothermal high- or medium-enthalpy field and process the heat contained in the hot water from deep hydrothermal resources.

A comprehensive overview on geothermal binary power plants is given by DiPippo (2015). A description of the technical status-quo of binary plant technology can also be found in Bronicki (2013). Geothermal binary plants are in most cases Organic Rankine Cycle (ORC) plants, which are directly driven by the geothermal heat and which are often realized as single-stage and sub-critical processes. Depending on the plant capacity and the site conditions power plants with multiple heat input, multi-stage cycles and cycles with internal heat recovery can be found (DiPippo 2015; Agahi and Valdimarsson 2015; Heberle et al. 2015). Commonly used working fluids are Isobutane, Isopentane, R245fa, R134a, n-Butane and n-Pentane. Referring to plants with a capacity of 100 kW_e and above, a water-ammonia-mixture (so called Kalina cycle) is currently only used at one site in Germany.

When realizing binary power plants different technical decisions have to be made. Important topics are working fluid selection, process design as well as turbine and heat exchanger specification (e.g. Lakew and Bolland 2010; Gao et al. 2012; Maraver et al. 2014; Toffolo et al. 2014; Walraven et al. 2014). Many authors therefore agree that a thermodynamic analysis is not sufficient and that a techno-economic evaluation based on different criteria is necessary. Different approaches for techno-economic design point evaluation are presented and discussed by Hettiarachchi et al. (2007), Quloilin (2011), Shengjun et al. (2011), Cataldo et al. (2014), Li et al. (2014) and Astolfi et al. (2014).

A specific design constraint for designing geothermal binary power plants is the limitation for the cooling of the geothermal fluid in order to avoid scaling (Franco 2011; Bronicki 2013; Wendt and Mines 2010) or in order to realize hot fluid reinjection close to the production wells (Noorollahi and Itoi 2011). Another important aspect is the consideration of changing or variable operating conditions. Sanyal et al. (2005), Gabbrielli (2012) and Budisulistyo et al. (2017) discuss possible design strategies to deal with the degradation of the geothermal resource over the lifetime of the plant. Astolfi et al. (2011), Manente et al. (2011) and Usman et al. (2017) address the effect of variable ambient conditions when using direct air-cooled condensers (ACC). Astolfi et al. (2019) investigate in this context the implementation of an intermediate cooling water cycle in order to use both, a compact water-cooled condenser and a flexibly operable dry cooler.

This paper gives details on the technical components and describes specific design considerations of a demonstration power plant that is equipped with both, intermediate cooling water and intermediate hot water cycle. Furthermore, this paper presents a numerical model that has been developed based on the technical component specifications and real plant data in order to predict the power output of the demonstration plant for varying hot water and ambient conditions.

Technical concept and design data

The demonstration plant has been integrated at the Lahendong geothermal field close to the village Pangolombian where geothermal brine with a temperature of about 170 °C corresponding to a separator pressure of 7.9 bar_a was available. The cooling of the brine should be limited to 140 °C in order to have the possibility for hot brine reinjection close to the production well. The brine composition was reported with a SiO₂-concentration between 100 and 500 mg/l, a pH-value between 5 and 9 and a TDS-content (total dissolved solids) between 150 and 540 mg/l. More details for Lahendong geothermal area can be obtained from Nugroho and Andi (2007), Koestono et al. (2010), Brehme et al. (2014), Brehme et al. (2016a, b). Based on the available project budget the electrical capacity was rated with 500 kW so that the necessary brine flow rate was estimated with 30 to 35 kg/s. Cooling water was not available.

To be able to handle a broad range of brine compositions and operate the binary plant with variable capacity without changing the brine supply as well as meeting project specific constraints (e.g. pretesting of the ORC-prototype in Germany), it was decided to integrate the power conversion cycle by using intermediary closed water cycles for heat supply and heat removal. Aiming for high reliability, a subcritical, single-stage ORC with internal heat recovery was chosen as the conversion cycle. N-Pentane, a well-known working fluid suitable for the heat source temperature at the demonstration site, was selected as the working fluid. The process diagram is shown in Fig. 1. Solid lines indicate normal operation. Dotted lines indicate operation during start-up or shut-down of the demonstration plant.

Fig. 1

Flow diagram of the demonstration power plant

During normal operation, the heat of the brine is transferred to the hot water cycle using the primary heat exchanger. The hot water is then used to heat and evaporate the working fluid inside the ORC-unit. The hot water is continuously circulated by using a centrifugal pump. In order to keep the pressure in the hot water cycle in a defined operating range, an expansion vessel with nitrogen cushion is used. In the ORC-unit the working fluid vapor drives a turbogenerator. After the turbine, the superheated working fluid vapor flows through a recuperator register before getting in touch with the water-cooled tubes. The heat removal in the cooling water cycle is realized with a dry cooler consisting of 6 units.

Using intermediary water cycles, the net power output is decreased due to the additional heat resistance and the additional power consumption of the pumps. The net power loss caused by the hot water cycle has been estimated with about 13% and the loss caused by the cooling water cycle with about 11%, both compared to a plant without intermediary cycles. However, the intermediary water cycles were realized in this project due to practical advantages which are described below.

Using a hot water and a cooling water cycle it was possible to transport and install a completely preassembled and pretested ORC-unit. Using the hot water cycle, it was further possible to operate, shut-down and start-up the binary plant without changing the existing operational regime of the brine supply. The used prototype ORC-unit can only be started at low supply temperatures since large temperature differences between hot water and the working fluid can lead to steam hammer in the evaporator. Therefore, an additional dry cooler is necessary to cool down the hot water before restart and the control valves are used to realize a defined temperature ramp. The hot water by-pass pump is implemented to realize modest temperature differences around the primary heat exchanger during restart and to reduce the risk of steam hammer on the brine side.

Another advantage of the hot water cycle is that different priorities can be applied for the heat exchanger design. Regarding the primary heat exchanger, which transfers the heat from the brine to the hot water, the accessibility of the tubes for cleaning procedures and corrosion resistant materials has a high priority. The evaporator design focuses instead on the heat transfer. An advantage of implementing the cooling water cycle is that design and operation of the dry cooler becomes easier due to the well-known single phase heat-transfer of water. Furthermore, less working fluid is needed to fill the ORC-cycle and the risk of n-pentane leakage is decreased. The practical advantages of an intermediary cooling water cycle are also addressed by Astolfi et al. (2019).

The 3D-layout is shown in Fig. 2. Table 1 lists the main equipment and its technical specification.

Fig. 2

3D-Plant layout showing primary heat exchanger (a), ORC-unit (b) and dry cooler (c)

Table 1 Component specification of the demonstration plant

The main process design data are listed in Table 2. A T–$\dot{Q}$̇-diagram is shown in Fig. 3. For the geothermal fluid supply, a design mass flow ṁ_geo of 32 kg/s was considered. Supply and return temperature T_geo,i and T_geo,o of the geothermal liquid were assigned with 172.5 °C and 142.5 °C, respectively. The heat capacity rate of the hot water was adapted to the heat capacity rate of the geothermal fluid in order to minimize the exergy destruction in the primary heat exchanger. Due to the fact that the heat transfer area increases significantly with lower pinch points, the minimum temperature difference between geothermal fluid and hot water was set with 7 K.

Table 2 Process design data and design electrical data of the demonstration plant

Fig. 3

T-$\dot{Q}$̇-diagram of the design point

Using a relatively small temperature spread referring to the heat supply, the process design of the ORC-unit had to be adapted. From Fig. 4 it can be seen that the evaporation temperature delivering the maximum power output would lead to a hot water outlet temperature lower than 135 °C. Hence the outlet temperature and not the power output was decisive for the selection of the evaporation temperature.

Fig. 4

Electrical power (left) and hot water outlet temperature (right) depending on the evaporation temperature

The condensation temperature of the ORC-unit was selected based on the maximum net power output considering the power consumption of the cooling water pump and the dry cooler fans. A lower condensation temperature leads to a higher gross power output but also higher fan power consumption. Figure 5 (left) shows the influence of the initial temperature difference ΔT_ITD, which is the difference between cooling water inlet temperature to the dry cooler T_CW,o and air inlet temperature T_a,i, on the gross power and the fan power consumption. In Fig. 5 (right) the influence of the dry cooler size on the net power output and the optimum initial temperature difference are displayed. Based on the evaluation of different quotations with different dry cooler sizes, the dry cooler cross-flow section was defined to be 120 m².

Fig. 5

Generator power and fan power depending on the initial temperature difference (left) and net power and optimal initial temperature difference depending on the dry cooler cross-section (right)

Results and discussion

Operating experiences and plant operation

The geothermal binary demonstration plant has been successfully operating since September 2017 and has produced more than 1.6 GWh gross electricity and 1.3 GWh net electricity as of April 2019. Since the commissioning several technical modifications have been realized in order to improve the plant reliability and availability. Firstly adaptations accounting for the high ambient humidity were realized. Secondly modifications of the prototype-turbogenerator, the hot water cycle and the plant control were necessary in order to manage frequent starts and stops of the plant which are caused by the electrical grid (mainly power outages, phase failure). The binary plant has been designed for grid-parallel and not for electrical island mode operation. Based on the electrical grid quality data that where provided at the beginning of the project, it was not expected to have such a poor grid quality at the used grid connection point. Since September 2017 the plant has experienced more than 130 plant stops and starts. Furthermore, the real brine conditions made it necessary to modify some operating procedures. Instead of pure liquid as assumed for the design, a two-phase flow is supplied to the binary plant. Both the well conditions and the operational regime of the separator have changed since the start of the project.

At the moment, the plant is operating with reduced power, which is sufficient to supply the injection pumps nearby. The operating power is thereby manipulated by controlling the hot water inlet temperature to the ORC unit with the automatic valves in the hot water cycle. The main reason for the power reduction is to reduce the wear on the turbogenerator bearing, which are most stressed during start-up and shut-down in the case of failure in the electrical grid.

Figure 6 shows the power production between March 2018 and April 2019. It can be seen that the plant stopped and started many times. Besides some maintenance work, most stops where caused by electrical grid failure. Figure 7 shows a 3-day-period in April 2019. It can be seen that gross power and net power show a daily fluctuation that is related to changing ambient temperatures (see Fig. 8). The cooling water supply temperature can be kept constant up to an ambient temperature of about 27 °C by means of variable speed fan operation. For higher ambient temperatures the fans are operated at full speed and the cooling water supply temperature follows the course of the ambient temperature. The stronger fluctuation of the net power is therefore caused by the changing ambient temperature and the variable fan power consumption.

Fig. 6

Power output of the demonstration plant in 2018

Fig. 7

Plant data for 48 h operation—gross and net power output

Fig. 8

Plant data for 48 h operation—temperatures heat supply and removal

In Fig. 9 the temperatures and pressures of the brine supply are shown in comparison to the saturation pressure of water. It can be seen that the supply pressure is very close to the saturation pressure and that there is only little cooling of the brine, both facts indicating two-phase flow into and liquid flow out of the primary heat exchanger.

Fig. 9

Plant data for 48 h operation—brine supply temperatures and pressures in relation to the saturation pressure of water

Due to the operation at off-design conditions, the maximum proven gross power until now is 400 kW. In order to evaluate the maximum power output of the demonstration plant, a numerical plant model has been developed and is described in “Model-based power prediction” section.

Model-based power prediction

The plant model has been developed with the software environment Engineering Equation Solver (EES)¹ and comprises the hot water cycle, the ORC-unit and the cooling water cycle. For modeling of the heat exchangers, heat transfer and pressure loss correlations from the VDI heat atlas have been used. The turbogenerator is represented as a turbine wheel with an upstream laval-nozzle stage. More detailed information on the model assumptions are presented in Table 3.

Table 3 Description of the modeled plant components

The model uses the hot water and cooling water supply parameters (temperature and volumetric flow) as input values. Furthermore the degree of superheating of the working fluid vapor after the evaporator is an input parameter. Superheating can occur in case the evaporator tube bundle is not fully flooded with liquid working fluid. The liquid level in the evaporator depends on the operating conditions. For example, a lower liquid level results from higher hot water cycle temperatures.

The flow diagram of the numerical model is shown in Fig. 10.

Fig. 10

Flow diagram of the numerical plant model

The isentropic turbine efficiency has been adapted based on operational data. In Table 4 the used plant data and the resulting isentropic efficiency are shown. For each operating point, the isentropic efficiency leading to the least root means square error between modelled and real plant data has been selected. The used instrumentation is shown in Fig. 11. Sensor data are listed in Table 5. Based on this data the isentropic turbine efficiency increases with increasing pressure difference across the turbine and has a best performance point that depends on the condensation pressure (Fig. 12).

Table 4 Calibration of isentropic turbine efficiency based on minimum deviation between modelled values (M) and measured data (D) for 17 operating points

Fig. 11

ORC process diagram showing the used instrumentation

Table 5 Instrumentation and sensor data

Fig. 12

Isentropic turbine efficiency estimate derived from real plant data

From Figs. 13 and 14 it can be seen that the model gives a good representation of the plant operation. Despite the gross power, which is underestimated by about 3%, all other values can be reproduced with good accuracy.

Fig. 13

Gross and net power output—comparison of modeled and real plant data for Plant data for 48 h operation

Fig. 14

N-pentane temperatures and volumetric flow—comparison of modeled and real plant data for Plant data for 48 h operation

Using the numerical model the power output of the binary demonstration plant can be calculated for varying hot water and cooling water temperatures. In Fig. 15 the gross and net power output for a saturated vapor cycle is shown. The net power reaches a maximum depending on the cooling water inlet temperature and increases with higher hot water temperatures. In Fig. 16 (left) the power increase with increasing hot water inlet temperature is shown. In Fig. 16 (right) it can be seen that the working fluid mass in the ORC-unit that is necessary to realize a saturated vapor cycle also increases with the hot water temperature. This results from changing working fluid distribution in the n-pentane cycle and therefore also changing n-pentane liquid-level in the evaporator. Superheating occurs if the liquid-level in the evaporator is lower than the tube area.

Fig. 15

Estimated generator power, net power and auxiliary power (left) and net power for different hot water inlet temperatures (right) depending on the cooling water temperature

Fig. 16

Estimated generator power, net power and auxiliary power (left) and working fluid mass (right) depending on the hot water temperature

In real operation, the binary demonstration power plant is operated with a constant working fluid mass. In this case, the degree of superheating increases with higher hot water temperatures since the liquid level in the evaporator decreases. The power estimate with constant working fluid mass is shown in Fig. 17. Due to the superheating, the power output is lower than the estimate for the saturated vapor cycle in Fig. 16.

Fig. 17

Estimated generator power, net power, auxiliary power and superheating depending on the cooling water temperature (left) and the hot water temperature (right)

Conclusions

The presented technical concept was developed to produce power from a small temperature difference in order to avoid scaling and to preserve the possibility of hot reinjection. Another aspect for the technical concept was the integration of a fully automatic binary plant at an already existing geothermal site. Within this paper the feasibility of this technical concept has been shown.

The presented technical concept integrates an Organic Rankine Cycle unit using an intermediary hot water and cooling water cycle. Even though the net power output is decreased by using intermediary cycles, the technical and organizational advantages explained in the paper prevailed. The operational phase has even shown that the demonstration plant could not be operated without an intermediate hot water cycle. Due to changing well (operating) conditions the binary plant is supplied with two-phase flow instead of pure liquid, as previously designed. It can be concluded that intermediate cycles, especially the hot water cycle, are interesting technical solutions that can be applied to other projects. However, cost reduction potentials (e.g. decreasing expansion vessel size or using thermal oil instead of water) should be evaluated.

Besides high ambient humidity, the availability of the electrical grid is the largest technical challenge for the operation of the demonstration plant. The binary plant has been designed for grid-parallel operation and is not capable of electrical island mode operation. As a result of power outages and phase failure, the plant has experienced over 130 plant stops and starts until now. Due to technical modifications an automatic re-start is now possible for various situations. For the future it is planned to change the grid connection point in order to connect to a more stable electrical grid. A modification of the ORC-unit for grid-parallel and island operation would be technically possible but is much more costly.

The plant is currently operated with limited power by controlling the hot water supply temperature to the ORC-unit in order to meet the consumption of the injection pumps nearby. Real plant data for typical operation are shown in “Results and discussion” section. Due to the technical concept it is possible to adapt the plant operation in order to realize low auxiliary power consumption at different operating conditions. Due to the operation at off-design conditions, the maximum proven gross power until now is 400 kW.

Using the presented numerical plant model which incorporates detailed component data and was validated with measured plant values, it can be shown that the rated capacity of 500 kW can be reached when the working fluid mass in the ORC unit is sufficient to realize a saturated vapor cycle. Technical and operational solutions in order to adapt the working fluid mass to different operating conditions will be investigated in the future.