The schematic of the entire waste heat recovery system is illustrated in Figure 1. An automated temperature control unit, incorporating a combustion controller and a thermocouple reader, was installed to maintain the system temperature throughout the process. The volume and flow rate of gases were also controlled with automated devices. A boiler fueled by liquefied petroleum gas (LPG), a hollow steel cylinder with a diameter of 80 cm and a height of 180 cm, was used to simulate the source of industrial waste heat (fiery center at 1,350 °C). A burning control valve was installed at the inlet of the gaseous fuel at the top of the boiler. After natural gas combustion, the hot flue gas flowed through a side pipe to the preheating unit at the bottom of the series of waste heat recovery units. The design of the three heat recovery units is explained in detail in the following subsections.

The preheating unit was a hollow cylinder with a diameter of 50 cm and a height of 30 cm. It was connected to a boiler on the side and an MSR unit on top. The waste heat generated by the boiler was circulated in the coils within the preheating unit before being sent to the MSR unit. An elastic copper tube, 2 m long and with an inside diameter of 5 mm, was used to bring the methanol solution (ambient temperature) into the preheating unit, where the high-temperature air surrounding the tube vaporized the liquid, and then sent it up into the MSR unit. To monitor temperatures, a thermocouple was installed at the center of the preheating unit, and another at the center of the duct connecting the unit to the boiler.

As an endothermic reaction [Equation 1], MSR can use the waste heat from the boiler to produce hydrogen. Accordingly, the MSR unit, consisting of a number of stainless steel tubes, two thermocouples, a pump (Nikuni pump), and a gas analyzer (MRU Vario Plus Industrial), was built in the form of a shell-and-tube heat exchanger [Figure 2] to convert methanol solution into hydrogen through recovered heat. A cylindrical shell with a 0.5 m diameter was designed to contain up to 50 kg of Cu-Zn particles to catalyze an MSR reaction on an industrial scale. A methanol solution with a steam-to-carbon (S/C) ratio of 2.5 was pumped into the reaction space at a volume of 10~100 mL∙min^-1 to react with the aid of the Cu-Zn catalyst. Nineteen (19) stainless steel tubes were vertically installed in the shell to allow flue gas from the boiler to pass through the reaction space, thereby triggering the MSR reaction via heat transfer. To remove the gaseous products of the MSR, pressurized N₂ was directed horizontally through the reaction space from one side to the other at a flow rate of 3,000 mL·min^-1. A flow meter was installed at the outlet to measure the flow rate of the outgoing gas. Its composition (H₂, CO, CO₂, and CH₄) was then analyzed using a gas analyzer (MGAprime H₂, MRU Messgeräte für Rauchgase und Umweltschutz GmbH, Germany). Two K-type thermocouples were inserted into the catalyst bed, one at the middle section and one near the bottom of the reaction zone, to monitor the temperature distribution. The reactor temperature was controlled at 200 or 250 °C, with the thermocouple at the center of the catalyst bed serving as the control point. In the constructed packed bed, axial and radial temperature gradients are inevitable; the mid-bed location was therefore selected as a reference to limit variation in the effective reaction zone and maintain stable operating conditions.

The third heat recovery unit consisted of thermoelectric generators (TEGs), which are considered durable and environmentally friendly devices^[27]. It is especially suitable for recovering low- and medium-temperature waste heat because the TEG’s ability to generate electricity is based on the temperature difference across the generators rather than a high-temperature environment. The TEG unit comprised a square array of thermoelectric modules (TEMs) made of low-resistance metals, so that electrons in the TEMs could more easily be induced to flow from an area with relatively high temperature to an area with relatively low temperature, thereby generating electric current. Figure 3 depicts the apparatus employed in the present study, in which two TEGs (TG1-1018) were attached to the exterior of a rectangular exhaust (691 mm × 236 mm × 122 mm) through which hot air (100 and 300 °C) from the boiler flowed, thereby creating a hot side. On the inside of the flue gas channel, aluminum plate fins (567 mm × 3 mm × 122 mm each) were installed to facilitate heat transfer. The temperature difference was maintained by circulating cool water through aluminum tanks (600 mm × 220 mm × 58 mm each) mounted on the cold side of the TEGs. Lastly, the power generated by the TEGs was measured using an automated DC electronic load system (IT8511A+), which adjusted the load impedance to maximize power output.

To evaluate the integrated ITURS under suitable operating conditions, two operating cases were selected according to the thermal stability of the Cu-Zn catalyst. Cu-Zn catalysts used for MSR generally exhibit stable activity between 200 and 300 °C, whereas prolonged exposure to temperatures above 350 °C may deactivate the catalysts and reduce hydrogen production^[28,29]. In the reactor, temperature gradients developed along the catalyst bed during operation, particularly near the lower section exposed to direct flue-gas heating. To avoid local overheating, the reactor temperature was controlled using the thermocouple positioned at the center of the catalyst bed (thermocouple 1), while the lower thermocouple (thermocouple 2) was used to monitor temperature variation near the bottom region [Figure 2]. In Case 1, the MSR temperature was maintained at 200 °C, and in Case 2, it was controlled at 250 °C to explore the effects of temperature on the efficacy of waste heat recovery. Before use, the catalyst was activated by hydrogen reduction at 350 °C for 24 h to promote high methanol conversion and hydrogen productivity. Temperature control was achieved by controlling the LPG flow rate into the boiler to support combustion, using an automated control unit. While temperature data were obtained with thermocouples installed in a host of different locations across the system, the automated temperature controller was programmed to act based on readings from the one installed in the middle of the shell, which was filled up with 50 kg of Cu-Zn catalysts [Figure 2]. In addition, cooling water at a flow rate of 0.2 kg∙s^-1 maintained at 5-10 °C was circulated through the TEG unit to maintain the cold-side temperature of the thermoelectric generators. After passing through the TEG unit, the cooling water was further directed to a heat exchanger connected downstream of the MSR unit. This secondary cooling stage reduced the temperature of the gaseous reforming products, allowing H₂O vapor to condense and separate from the remaining gaseous products as liquid water.

Figure 4 shows the hourly LPG consumption in the two experimental runs used to simulate industrial waste heat. Conducted with the catalyst bed heated to 200 °C, the first run (Case 1) consumes 4.4 kg∙h^-1. In the second run (Case 2) with the catalyst bed heated to 250 °C and all other parameters held constant, the LPG consumption rate increases to 5.0 kg∙h^-1. Accordingly, when the catalyst bed temperature is increased by 50 °C, from 200 to 250 °C, the system requires additional heat input, leading to an approximately 13.6% increase in the LPG consumption rate under stable operating conditions.

Figure 5 shows the temperature profiles in the boiler, preheating unit, and MSR unit. Temperature measurements are recorded once every 3 min during the experimental runs; Figure 5A shows the measurements from Case 1, and Figure 5B shows those from Case 2. In Case 1, the temperature inside the boiler ranges from 650 to 700 °C throughout the process, while the temperature in the preheating unit rises from approximately 400 °C at the outset to 450-500 °C after 20 min of burning. Meanwhile, the temperature in the MSR unit rises to 200 °C. In contrast, the temperature measurements from Case 2 [Figure 5B] show minimal fluctuations in the boiler and preheating unit, while the temperature in the MSR unit rises from 200 to 250 °C over 40 min. After twenty min, the temperatures in both the boiler and the preheating unit remain nearly constant. The small temperature variation observed during MSR operation indicates that the heat supplied by the boiler is sufficient to maintain stable conditions, while the energy required to vaporize the methanol solution is relatively less.

Figure 6A shows the morphological features of the Cu-Zn bimetal catalyst used in this study, which is used for the MSR reaction. The catalyst consisted of cylindrical particles with dimensions of 5.0 mm in height and 5.12 mm in diameter. The catalyst bed has a porosity of 0.614. Figure 6B shows the scanning electron microscope (SEM) image of the catalyst, which exhibits a rugged, porous surface. The elemental composition of the catalyst examined via energy-dispersive X-ray spectroscopy (EDS) is shown in Figure 6C, and the results reveal that it is largely composed of CuO and ZnO, with oxygen accounting for over 60% of the mass, while Cu and Zn account for approximately 35% and 1.0%, respectively. In the MSR reaction, Cu and its oxides (Cu₂O and CuO) serve as the primary active sites for methanol adsorption and dehydrogenation, leading to the formation of key intermediates, such as formate. ZnO functions as a structural and electronic promoter, stabilizing highly dispersed Cu species and tuning Cu surface properties. In addition, ZnO facilitates water activation and suppresses CO formation, thereby promoting the preferred CO₂ pathway^[30,31]. A comparison between the fresh and used catalyst compositions shows no significant change in the amounts of carbon and copper, indicating that coking is minimal after MSRs.

Figure 7 shows the concentration time profiles of CO₂, CO, CH₄, and H₂, the main gas species from MSR, under the operating conditions of Cases 1 (200 °C, blue curve) and 2 (250 °C, red curve). The temperatures are set at the usual operating temperature of the Cu-Zn catalyst for MSR^[32,33]. After the MSR reaction reaches a steady-state, the CO₂ concentration [Figure 7A] ranges from 20.06% to 22.26% at 200 °C (Case 1) and from 20.74 to 25.21% at 250 °C (Case 2). The CO concentration [Figure 7B] is in the range of 0.03%-1.18% in Case 1 and 0.20%-1.14% in Case 2. As reported in previous studies^[31,34], Cu-Zn catalysts typically produce small amounts of CO due to their intrinsic catalytic properties, resulting in low CO concentrations. These results suggest that conditions in the MSR unit favor the MSR or WGSR [Equations 1 and 2] over the competing reaction [Equation 3], producing a high CO₂-to-CO ratio.

Compared to the CO concentration, the CH₄ concentration [Figure 7C] showed minimal fluctuations between the two cases, ranging from 1.34% to 1.67% at 200 °C and from 1.19 to 1.38% at 250 °C. Lastly, the H₂ concentration remains comparable in both cases, ranging from 61.19 to 69.13% at 200 °C and from 62.51 to 71.73% at 250 °C [Figure 7D]. Based on the measured gas concentrations, the H₂ yield and CH₃OH conversion can be calculated, as expressed below:

(4) $$ H_{2} \text { yield }\left(\mathrm{mol} / \mathrm{mol} \space C H_{3} O H\right)=\frac{\dot{n}_{H_{2}}}{\dot{n}_{C H_{3} O H}} \\ $$

(5) $$ C H_{3} O H \text { conversion }(\%)=\left(\frac{\dot{n}_{C O_{2}, o u t}+\dot{n}_{C O, o u t}+\dot{n}_{C H_{4}, o u t}}{\dot{n}_{C H_{3} O H, i n}}\right) \times 100 \\ $$

where $$ \dot{n} \\ $$ represents the gas molar flow rate, and the subscript out designates the outlet. While the theoretical maximum level of H₂ yield is 3 [mol∙(mol CH₃OH)^-1] [Equation 1], obtained H₂ yield ranges from 2.19 to 2.77 [mol∙(mol CH₃OH)^-1] at 200 °C and from 2.25 to 2.76 [mol∙(mol CH₃OH)^-1] at 250 °C with a uniformly high CH₃OH conversion of over 80% [Figure 8A and B] across both cases (81.02%-97.49% at 200 °C and 80.36%-99.65% at 250 °C). These parameters for hydrogen productivity and reaction efficiency do not vary significantly between the two cases, indicating that the Cu-Zn catalyst utilized in the experiments has sufficient activity to promote effective steam reforming at a temperature as “low” as 200 °C. The experimental outcomes further demonstrate that the Cu-Zn catalyst delivers consistently high hydrogen production, with H₂ concentrations exceeding 60% under both operating conditions. Meanwhile, the H₂ yield approaching the theoretical limit confirms the catalyst’s strong hydrogen-generation capability in the MSR operating conditions of this study.