1 Introduction

Cogeneration systems enable the simultaneous production of electricity and useful heat, maximizing the overall energy efficiency of the system. Various prime movers are used in cogeneration. However, in Korea, gas engines are price-competitive in a wide range [1]. Megawatt-class gas engines are highly efficient in converting fuel into electricity and heat [2]. They are suitable for large-scale applications in which both electricity and a significant amount of thermal energy are required. These systems are commonly used in industries with high energy demands, large commercial complexes including hospitals, or district heating and cooling systems [3, 4]. Unlike residential spaces, these energy consumers require heat in the form of steam [5].

In gas engines, heat is recovered from the jacket coolant and exhaust gas as shown in Fig. 1a. Heat from the jacket coolant is recovered by a shell-and-tube or plate heat exchanger. The heat is recovered via hot water and used for heating and cooling [6]. Because the working fluid on both the heat supply and recovery sides is water, a plate heat exchanger is advantageous in terms of volume. The effects of pressure and flow rate differences between the jacket coolant and heat-recovery water have been investigated in previous studies [7].

Fig. 1
figure 1

1-MW gas engine cogeneration system with simultaneous steam and hot water generation. a System configuration. b Energy balance

Full size image

Flue gas can produce steam because its temperature is above the boiling point of water [8]. Figure 1b shows the energy balance of the 1-MW gas engine cogeneration system considered in this study. The water tube boilers shown in Fig. 2a are frequently used in heat-recovery boilers that generate steam from flue gas [9]. In a water tube boiler, a steam-saturated water mixture is fed into a drum. Steam is extracted from the drum, and saturated water is recycled. Natural convection is often used in the water circulation of water tube boilers. Unlike direct-combustion boilers, heat-recovery boilers may not have a sufficient density difference, which results in poor water circulation [10].

Fig. 2
figure 2

Heat recovery boiler. a Conventional water tube heat recovery boiler. b Mini boiler constituting a modular heat-recovery boiler

Full size image

To solve the water circulation problem, Ahn et al. [11, 12] proposed a modular boiler using the mini-boiler concept shown in Fig. 2b. As shown in Fig. 2, a modular boiler is composed of modules with a smaller number of tubes in the flow direction than that in a typical water tube boiler. Therefore, as it is difficult to apply the design equation to a conventional tube bank, the effect of uneven tube spacing on heat transfer was analyzed in a prior study [13].

The shape of the inlet of the heat-recovery boiler is different from that of the exhaust port of the engine, resulting in a local nonuniform temperature distribution in the connecting duct. Moreover, the inlet temperature uniformity of the heat-recovery boiler and the water circulation were found to greatly affect the performance of the heat-recovery boiler for small cogeneration [14]. To address this issue, a connection duct shape including guide vanes was designed using computational fluid dynamics (CFD) [15].

In the present study, a model engine that produced combustion gas such as gas engine exhaust gas was fabricated. A mini-boiler-type heat-recovery boiler (Fig. 2b) was connected downstream. A connecting duct derived from previous research [15] was installed between the model engine and the heat-recovery boiler. By using this test rig, the performance of the modular heat-recovery boiler and guide vane designed using CFD was evaluated.

2 Experimental apparatus

2.1 Model gas engine and connecting duct

The experimental rig comprised three main components: a model engine that simulated exhaust gas conditions for a gas engine, a connecting duct featuring a guide vane, and a model heat-recovery boiler. The section responsible for generating combustion gas, mirroring the exhaust gas conditions of the gas engine, was configured in the shape of a coaxial furnace as illustrated in Fig. 3. Natural gas was utilized as the fuel, akin to the case of an actual gas engine, with a flow rate of 12 Nm3/h, equivalent to 100 kW; it was combusted through the burner. To meet the gas engine outlet temperature requirement of 430 °C and a flow rate of 1030 kg/h, amounting to 1/5 of the exhaust gas flow rate, additional air was introduced in two stages from the outer shell.

Fig. 3
figure 3

Gas engine exhaust gas simulator and connection duct to the heat recovery boiler. a Gas engine exhaust gas simulator. b Inlet guide vane to be installed in the connection duct. c Assembly drawing and temperature measurement points

Full size image

The outlet of the gas engine exhaust gas simulation device was fashioned in a cone shape to match the diameter of the actual gas engine [16]. A connecting duct, previously designed in a CFD study [15], was positioned between the gas engine exhaust gas simulation device and the 1/5 scale model heat-recovery boiler as illustrated in Fig. 3c. In the upstream section of the connection duct (the shaded area in Fig. 3c), a diamond-shaped inlet guide vane was installed (Fig. 3b) to ensure a uniform temperature distribution in both the height and span directions over the short duct length. K-type thermocouples were strategically placed at three locations in the flow direction within the connection duct (Fig. 3c) to validate the outcomes of previous research conducted by CFD. The tolerance was ± 1.5 °C when the temperature was 400 °C.

2.2 Model heat-recovery boiler

The model heat-recovery boiler was assembled by linking three mini-boiler modules in series as illustrated in Fig. 2b (see Fig. 4a). A steam separation drum was positioned atop each module, with water supplied in accordance with the water level of the drum, facilitating the measurement of the evaporation rate in each module.

Fig. 4
figure 4

Experimental setup. a Drawing. b Photograph

Full size image

Thermocouples were positioned at the heat-recovery boiler inlet and between modules to gauge the enthalpy change of the exhaust gas as it traversed each module. Additionally, thermocouples were placed at four locations along the tube direction between fin-tubes within the module to measure the heat-transfer coefficient for each tube. The comprehensive temperature distribution was determined by recording temperatures at eight evenly spaced locations while moving the thermocouple in the span direction.

The uncertainty of the heat-transfer coefficient was assessed at 20-to-1 odds (95% confidence level) using single-sample experiments [17]. The heat-transfer coefficient (h) was expressed in terms of the Nusselt number, with an uncertainty of 2.1 at its typical value of 20.

3 Results and discussion

3.1 Performance of the inlet guide vane of the connecting duct

An inlet guide vane, shaped according to the findings of the previous research [15], was installed in the connecting duct between the gas engine exhaust gas simulator and the model heat-recovery boiler. The CFD results were validated by measuring the temperature distribution. A comparison of the CFD results with the temperature distribution measured at the location indicated in Fig. 3c (refer to Fig. 5a) revealed that the measured temperature was lower than the CFD-predicted result, which was primarily due to heat dissipation loss to the side wall. As the flow progressed downstream, the disparity between experimental and CFD results became more pronounced.

Fig. 5
figure 5

Effect of inlet guide vane installed on the connection duct. a Temperature distribution within the connecting duct. b temperature distribution inside the heat recovery boiler

Full size image

At measurement points 1 and 2, which are upstream locations, CFD predicted a very uniform temperature distribution; however, the experimental data showed a temperature deviation of approximately 5%. The temperature distribution at the duct outlet location (measurement point 3) exhibited a similar distribution between the experiment and CFD, confirming that a uniform temperature distribution was indeed achieved by the guide vane.

An examination of the temperature distribution obtained from the central plane in the height direction of the heat-recovery boiler shown in Fig. 5b indicates that the upstream area displays a uniform temperature distribution in the span direction. However, as the exhaust gas flows downstream, an asymmetric temperature distribution emerges in the span direction, likely influenced by the downcomer pipe. As shown on the right side of Fig. 2b, the downcomer is installed only on one side for water circulation. The low-temperature area shown in blue in Fig. 5b appears on the side where the downcomer is installed.

3.2 Heat transfer in the heat-recovery boiler

The temperature drop that occurs as the exhaust gas passes through the evaporation tube of the heat-recovery boiler is illustrated in Fig. 6. As depicted in Fig. 4, the heat-recovery boiler comprises three modules with two rows of evaporation tubes. It has six rows in the flow direction, and the numbers on the x-axis in Fig. 6 indicate the row numbers in order from upstream to downstream. To illustrate the effect of the Reynolds number, the results measured at 100% load and 75% partial load are presented in Fig. 6a and b, respectively.

Fig. 6
figure 6

Temperature drop that occurs as the exhaust gas passes through the evaporation tube of the heat recovery boiler. a 100% load. b 75% load

Full size image

In a comparison of the temperature drop that occurred in the flow direction during passage through each fin-tube with the results calculated using the Zukauskas correlation [18], the experimental results and design values showed good agreement overall. Under both design-load and partial-load conditions, the temperature dropped more than the design value in the first and second rows upstream. However, the temperature decrease was less than the design value in the fifth and sixth rows downstream.

The heat-transfer coefficient was computed according to the enthalpy change of the combustion gas as it traversed the evaporation tube; the results are depicted in Fig. 7. In downstream flow, the heat-transfer coefficient is evidently smaller than the value predicted by the Zukauskas correlation. This phenomenon apparently arises because of increased fin density and height, leading to a suppression of flow from the tube surface to the tube direction. Consequently, the actual fin efficiency becomes lower than the value predicted by the Bessel solution [12].

Fig. 7
figure 7

Heat transfer coefficient in the evaporator tubes of a heat recovery boiler. a 100% load. b 75% load

Full size image

To independently examine the two factors, i.e., heat-transfer coefficient and fin efficiency, the evaporation rate derived from the enthalpy change of combustion gas passing through the evaporation pipe was compared with the feed rate as illustrated in Fig. 8. Except for the first module under partial-load conditions, the water supply rate was smaller than the evaporation rate obtained from the enthalpy change, aligning with the outcome of high steam quality. The fin efficiency decreased as the exhaust gas flowed downstream, and in the third module, the water supply rate was less than half of the design value.

Fig. 8
figure 8

Evaporation rate occurring in each module of the heat recovery boiler. a 100% load. b 75% load

Full size image

4 Conclusions

In this study, a 1/5-scale model experiment of a modular heat-recovery boiler for a 1-MW gas engine cogeneration was conducted. The following conclusions were obtained:

  1. (1)

    A uniform temperature distribution was achieved at the boiler inlet by using a diamond-shaped guide vane derived from CFD. However, as the flow progressed downstream, the temperature distribution became asymmetric in the span direction owing to the influence of the downcomer.

  2. (2)

    A heat-recovery boiler composed of a fin-tube-type modular boiler exhibited appropriate heat-transfer performance in each module.

  3. (3)

    When balancing the evaporation rate and water circulation by increasing the density and height of the fin in the downstream direction in a modular heat-recovery boiler, a larger area is required compared to the heat-transfer area predicted by the heat-transfer coefficient obtained by the Zukauskas correlation and the fin efficiency using the Bessel solution.