Conjugate Heat Transfer Analysis and Heat Dissipation Design of Nucleic Acid Detector Instrument

Abstract

Temperature affects both the stability of nucleic acid detectors and efficiency of DNA amplification. In this study, temperature and flow inside a nucleic acid detector were simulated and the results were used to design vents for the instrument casing. A test platform was constructed to collect experimental temperature data that were used for simulation validation. The experimental and simulation results showed that the temperature error was less than ±3 K. A total of two heat-dissipation schemes were designed based on the simulation and a new instrument casing was fabricated based on the scheme with the best results. Nucleic acid amplification was performed continuously for 120 min using a prototype with the new casing. The temperatures of the monitoring points were stable and the maximum temperature measured only 307.76 K (34.61 °C). Therefore, waste heat was effectively eliminated, which ensured safety of the electronic components and stability of the nucleic acid detection process.

1. Introduction

Quantitative real-time polymerase chain reaction (RT-PCR) is a vital technique in the diagnostic industry as it facilitates rapid detection, the importance of which was emphasized by the COVID-19 outbreak in December 2019. In particular, the use of PCR technology in combination with nucleic acid detectors has attracted considerable attention from researchers [1,2,3]. These instruments provide the specific temperatures required for each stage of the PCR process: (1) DNA separation at temperatures greater than 363.15 K, (2) primer annealing at 323.15–348.15 K, and (3) optimal extension at 345.15–351.15 K [4,5]. The precise temperature depends on the reagent system. The stages are repeated until the DNA is sufficiently amplified for fluorescence detection.

However, the internal temperature of the instrument increases as the operation time increases, which increases the failure rate of the components; therefore, heat dissipation is an important consideration in the design of these devices [6,7,8]. Furthermore, efficient heat dissipation increases the cooling rate during temperature cycling, which reduces the time required for DNA amplification and improves the sample detection rate during extensive batch testing. Therefore, it is necessary to research heat dissipation to improve the reliability and stability of PCR devices.

Many studies on PCR instruments have focused on nucleic acid amplification and fluorescence detection. Research on the amplification stage has typically focused on temperature homogeneity [9], temperature control [10], and the rate of temperature changes [11]. By contrast, research on the fluorescence detection stage has typically focused on the fluorescent light pathway, structure of the detection system, and signal-processing algorithms [12,13]. Furthermore, researchers have investigated the heat dissipation and thermal design of PCR instruments. For example, Aziz et al. [14] simulated the 3D heat transfer of a continuous-flow PCR model and used ANSYS CFX to investigate the effect of the fluids on the temperature distribution in the canals. Naghdloo et al. [15] used a numerical simulation of a fully automated PCR system to improve the structure of the operation unit and reduce the operation time. Subsequently, they optimized the time consumption by applying a mixed convective heat transfer mechanism. Parng et al. [16] developed an innovative design in which multilayer microchannels were built in a single tube; thus, droplet generation and PCR could be accomplished in the same tube. Their experiments showed that this design had high heating and cooling rates. Zhou et al. [17] analyzed the pedestal heat transfer of a PCR device and proposed a mathematical model to describe the pedestal temperature field, which was simplified into a simple 2D heat transfer model. Thus, they analyzed the effects of various parameters on the temperature field of the pedestal during uneven heating. Finally, Raza et al. [18] designed a well-shaped sample block with a volume of 50 µL and demonstrated that the fluid sample exhibited a higher temperature response rate and greater uniformity. This pedestal can be manufactured without any special process, paving the way for low-cost PCR.

However, there has been little research on the design of nucleic acid detectors concerning heat dissipation. Therefore, this study aimed to design an instrument ventilation scheme that could provide stable and safe ambient temperatures. This was achieved by using the finite-volume method to simulate the temperature distribution and flow, validating the simulation using experimental data, and using the simulation to design a new instrument casing. Finally, the experimental results prove that the designed heat-dissipation scheme demonstrates high heat-dissipation capability.

2. Geometry and Experimental Method

2.1. Geometry

The structure of the portable automatic nucleic acid detector used in this study is shown in Figure 1. The overall length L, width W, and height H were 0.360, 0.326, and 0.44 m, respectively, and the casing was 0.005 m thick. Numerous techniques can be used to design the vents on the device; however, a closed model was selected for this study, and the vents were designed according to the simulation results.

2.2. Experimental Method

As shown in Figure 1c,d, the DNA amplification module used a thermoelectric cooler (TEC) to increase and decrease the temperature of the biochip; therefore, the TEC was considered to be a heat source. A test platform, shown in Figure 2, was constructed to identify other heat sources. An infrared thermal imager was used to determine the positions of the heat sources (FLUKE Ti401 PRO, resolution: 640 × 480, range: 253.15–923.15 K, accuracy: ±2 K) and a thermocouple was used to measure their temperatures (thermocouple: KAIPUSEN K, range: 253.15–473.15 K, accuracy: ±0.1 K). Imaging was conducted at room temperature (approximately 299.15 K) and the results are shown in Figure 3. The other heat sources (called primary heat sources) in the instrument were located on the printed circuit boards (PCBs). Thus, there were three components acting as heat sources in the instrument: TEC and two PCBs.

The temperature measured at the surface of a heat source was assumed to be the temperature of the heat source. The temperatures on each side of the TEC and the outer surface temperatures of the primary heat sources on the PCBs were recorded with a data acquisition interval of 2 s. The temperatures of the heat sources throughout the DNA amplification process are shown in Figure 4. As shown in Figure 4a, a maximum temperature of 368.15 K was recorded on TEC side1 during the amplification cycling stage. Figure 4b shows the temperatures of the primary heat sources in the PCBs, which increased initially before stabilizing.

3. Simulation Process

The heat-source temperature data were used as source terms for the simulation. To improve the efficiency of the simulation, the duration of the simulation was half that of the entire DNA amplification process, that is, 600 s. The simulation and experimental data were compared at 30 s intervals.

All heat sources transfer heat to solids they are in contact with via conduction and with the air via convection [19,20]. However, Fedorov and Viskanta [21] suggested that radiation may be ignored when the typical temperature is less than 373.15 K. The maximum temperature in this study was 368.15 K (TEC side1); therefore, conduction and convection had a dominant effect, and radiation was ignored.

3.1. Material Parameters

The materials used in the modules and electronic components of the instrument included aluminum, steel, acrylic, and copper. PCBs generally have a “mezzanine” design consisting of copper and FR4. Considering the equivalent modeling method for PCBs reported by Mark [22], the plane thermal conductivity $k_p$ and normal thermal conductivity $k_n$ were used to represent the thermal conductivity of the PCBs. These terms are given by:

$$k_p=\frac{{\mathsf{\delta }}_{cu}\cdot {\displaystyle \sum _{j=1}^N{k}_j}+(N-1)\cdot {\mathsf{\delta }}_F\cdot k_F}{{\mathsf{\delta }}_t}$$

(1)

$$k_n=\frac{{\mathsf{\delta }}_t}{{\mathsf{\delta }}_{cu}\cdot {\displaystyle \sum _{j=1}^N\frac{1}{k_j}}+\frac{(N-1)\cdot {\mathsf{\delta }}_F}{k_F}}$$

(2)

where $k_j$, $k_F$, and N are the thermal conductivities of the copper-clad layer, FR4, and the number of layers of the copper cladding, respectively; ${\mathsf{\delta }}_t$ is the total thickness of the PCB; ${\mathsf{\delta }}_{cu}$ and ${\mathsf{\delta }}_F$ are the thicknesses of the copper classing and the unclad layer, respectively; PCBs typically consist of FR4 and copper; therefore, $k_j$ is between $k_F$ and $k_{cu}$. That is:

$$k_j\cong \frac{A_j}{A_t}\cdot k_{cu}$$

(3)

where $A_j$ is the area of the j-th layer of the copper cladding, $A_t$ is the area of the PCB, $k_{cu}$ is the thermal conductivity of the copper. In an equivalent model, ${\mathsf{\delta }}_{cu}$ and ${\mathsf{\delta }}_F$ are considered to be equal; that is:

$${\mathsf{\delta }}_t=N\cdot {\mathsf{\delta }}_{cu}+(N-1)\cdot {\mathsf{\delta }}_F$$

(4)

The parameters of the PCBs are given in

Table 1. Parameters of the PCBs.

	${\mathsf{\delta }}_{\mathit{p}}$ (m)	${\mathit{A}}_1$ (m2)	${\mathit{A}}_2$ (m2)	${\mathit{A}}_3$ (m2)	${\mathit{A}}_4$ (m2)
PCB1	0.0016	0.012	0.012	0.012	0.01
PCB2	0.0016	0.023	0.02	0.022	0.019

and

Table 2. Parameters of the PCBs.

	${\mathsf{\delta }}_{\mathit{c}\mathit{u}}$ (m)	${\mathsf{\delta }}_{\mathit{F}}$ (m)	${\mathit{A}}_{\mathit{P}}$ (m2)	${\mathit{k}}_{\mathit{c}\mathit{u}}$ (W·m−1·K−1)	${\mathit{k}}_{\mathit{F}}$ (W·m−1·K−1)	N
PCB1	0.000035	0.0000487	0.013	388	0.35	4
PCB2	0.000035	0.0000487	0.024	388	0.35	4

. Using these parameters, the thermal conductivities of PCB1 and PCB2 were determined to be $k_{p,pcb1}$ = 31.83 W·m⁻¹·K⁻¹; $k_{n,pcb1}$ = 0.38 W·m⁻¹·K⁻¹; $k_{p,pcb2}$ = 30.83 W·m⁻¹·K⁻¹; $k_{n,pcb2}$ = 0.38 W·m⁻¹·K⁻¹. The physical parameters of other materials are listed in

Table 3. Material parameters.

Material	Density (kg·m−3)	Specific Heat (J·kg−1·K−1)	Thermal Conductivity (W·m−1·K−1)
6061 aluminum alloy	2750	896	155
Stainless steel	8055	480	13.8
Packaging plastic	2090	900	0.67
FR4	1900	1369	0.3
Pom	1430	1.465	0.23
Si	2329	700	130
Ceramics	1345	710	150
Acrylic	1190	1464	0.18
Red copper	8960	385	401
Wood	700	2310	0.173

. The contact pin between the component and base plate was not considered in the modeling. The component materials used corresponded to the packaging materials listed in Table 3.

3.2. Computational Domain and Boundary Conditions

ANSYS determines the fluid temperature and velocity fields by solving the governing equations, which include the mass, momentum, and energy conservation equations [23,24]. For the simulation, the corresponding boundary conditions must be defined to resolve the corresponding equations [25,26]. The fluid and solid domains are illustrated in Figure 5. To reduce the complexity of the analysis, it was assumed that:

• The flow remains incompressible, ideal, and single phase.
• The initial (ambient) temperature remained constant.
• The heat lost from the enclosure to the surroundings was negligible.
• The connections between parts were considered seamless.

Conjugate heat transfer control equations based on the above assumptions are solved by using the Reynold-average [27,28,29].

Continuity equation:

$$\frac{\partial {\mathsf{\rho }}_f}{\partial t}+div({\mathsf{\rho }}_{f}u)=0$$

(5)

where ${\mathsf{\rho }}_f$ is the density of the fluid, $u$ is the velocity vector. When the fluid is incompressible, $\frac{\partial {\mathsf{\rho }}_f}{\partial t}=0$.

Momentum in x-direction:

$$\begin{array}{l}\frac{\partial ({\mathsf{\rho }}_{f}u)}{\partial t}+div({\mathsf{\rho }}_{f}uu)=div({\mathsf{\mu }}_{f}gradu)-\frac{\partial p}{\partial x}+S_u\\ +\left[-\frac{\partial ({\mathsf{\rho }}_f\overline{u^{\prime 2}})}{\partial x}-\frac{\partial ({\mathsf{\rho }}_f\overline{u^{\prime }{v}^{\prime }})}{\partial y}-\frac{\partial ({\mathsf{\rho }}_f\overline{u^{\prime }{w}^{\prime }})}{\partial z}\right]\end{array}$$

(6)

Similarly, the momentum equation in y- and z-directions can be written. Where $u$, $v$ and $w$ are velocity components in the x-, y- and z-directions, respectively. ${\mathsf{\mu }}_f$ is the dynamic viscosity of the fluid, $p$ is the fluid pressure, $S_u$ is source term, $\overline{u_i^{\prime }{u}_j^{\prime }}$ is the specific Reynolds stress term, and i and j are Cartesian coordinates.

Energy equation for fluid domain is given by:

$$\begin{array}{l}\frac{\partial ({\mathsf{\rho }}_{f}T)}{\partial t}+div({\mathsf{\rho }}_{f}uT)=div(\frac{k_f}{C_{p,f}}gradT)+S_f\\ +\left[-\frac{\partial ({\mathsf{\rho }}_f\overline{u^{\prime }{T}^{\prime }})}{\partial x}-\frac{\partial ({\mathsf{\rho }}_f\overline{v^{\prime }{T}^{\prime }})}{\partial y}-\frac{\partial ({\mathsf{\rho }}_f\overline{w^{\prime }{T}^{\prime }})}{\partial z}\right]\end{array}$$

(7)

Energy equation for solid domain is given by:

$$\frac{\partial ({\mathsf{\rho }}_s{C}_{p,s}T)}{\partial t}=div(k_s\nabla T)+Q$$

(8)

where $k_f$ and $k_s$ are the thermal conductivities of the fluid and solid, respectively, $T$ is temperature, ${\mathsf{\rho }}_s$ is the density of the solid, $C_{p,f}$ and $C_{p,s}$ are the specific heats of the fluid and solid, respectively, $S_f$ and $Q$ are source terms.

The k−ω shear stress transport (SST) turbulence model was adopted. The k−ω SST model includes all the refinements of the k−ω BSL model, and accounts for the transportation of the turbulence shear stress in the definition of turbulent viscosity. The transport equations for the k−ω BSL model are [30,31,32]:

$$\frac{\partial (\mathsf{\rho }k)}{\partial t}+\frac{\partial (\mathsf{\rho }k{u}_i)}{\partial x_i}=\frac{\partial }{\partial x_j}({\mathsf{\Gamma }}_k\frac{\partial k}{\partial x_j})+G_k-Y_k+S_k+G_b$$

(9)

$$\frac{\partial (\mathsf{\rho }\mathsf{\omega })}{\partial t}+\frac{\partial (\mathsf{\rho }\mathsf{\omega }{u}_i)}{\partial x_i}=\frac{\partial }{\partial x_j}({\mathsf{\Gamma }}_{\mathsf{\omega }}\frac{\partial \mathsf{\omega }}{\partial x_j})+G_{\mathsf{\omega }}-Y_{\mathsf{\omega }}+D_{\mathsf{\omega }}+S_{\mathsf{\omega }}+G_{\mathsf{\omega }b}$$

(10)

where $G_k$ represents the production of turbulence kinetic energy, $G_{\mathsf{\omega }}$ represents the generation of $w$, ${\mathsf{\Gamma }}_k$ and ${\mathsf{\Gamma }}_{\mathsf{\omega }}$ represent the effective diffusivity of $k$ and $\mathsf{\omega }$, respectively. $Y_k$ and $Y_{\mathsf{\omega }}$ represent the dissipation of $k$ and $\mathsf{\omega }$ due to turbulence, $D_{\mathsf{\omega }}$ represents the cross-diffusion term, $S_k$ and $S_{\mathsf{\omega }}$ are user-defined source terms, $G_b$ and $G_{\mathsf{\omega }b}$ account for buoyancy terms. A detailed description of k−ω equations can be found in [30,31,32].

In the simulation, a fan provided the sole source of the flow. The fan was considered to be infinitely thin and the discontinuous increase in pressure across it was specified as a function of the velocity of the fluid through the fan [33,34]. As shown in Figure 6, a polynomial was used to define the pressure gradient caused by the fan. That is:

$$\Delta P=71.053-8.6303v-0.4238v^2$$

(11)

where $\Delta P$ is the pressure jump, and $v$ is the magnitude of the local fluid velocity normal to the fan.

Figure 5 shows a model of the PCB plate and the corresponding heat sources, heat sources were used as the temperature source term, the values are shown in Figure 4b. The experimental data obtained from TEC side1 and side2 were included the final temperature under the effect of the entire TEC, side1 and side2 were used as the temperature wall, The values are shown in Figure 4a.

At the fluid–solid interface [35],

$$u=0;T_f=T_s;-k_f\frac{\partial T_f}{\partial n}=-k_s\frac{\partial T_s}{\partial n}$$

(12)

At the outer wall of the housing,

$$\frac{\partial T_s}{\partial n}=0$$

(13)

where $T_f$ and $T_s$ are fluid temperature and solid temperature, respectively.

In this study, the solar-type pressure-based method was used. The simulation settings are listed in

Table 4. Simulation settings.

Object	Type
Fluid and solid contact surfaces	couple wall
Fan boundary	polynomial
TEC side1 and side2	temperature wall
Heat source on the PCB	values
Fluid	air, incompressible
Inlet and outlet	pressure inlet/outlet
Pressure–velocity couple	simple
The discretization of convection–diffusion	second-order upwind

. A total of twenty iterations were performed for each time step. The simulations were considered to converge when the energy residual curve was less than 1 × 10⁻⁶, and other residual curves were less than 1 × 10⁻³.

3.3. Grid Independence

Mesh independence is an important aspect of the computational fluid dynamics simulation. Although the model used in this study was simplified, it still exhibited a high degree of complexity. The grid independence of the model was verified to ensure the accuracy of the simulation [36]. The average fluid temperature $T_{ave}$ in the last moment of the simulation was used as the verification index for grid independence. The relative change rate $\mathsf{\eta }$ was defined as:

$${\mathsf{\eta }}_i=\left|\frac{T_{ave,i}-T_{ave,i-1}}{T_{ave,i-1}}\right|\times 100\%$$

(14)

where $T_{ave,i}$ is the average value of the simulation of $T_{ave}$ at the current grid scale and $T_{ave,i-1}$ is the average value at the previous grid scale. The calculated value of $T_{ave}$ is shown in Figure 7. When the mesh size was less than 0.008 m, the $T_{ave}$ was approximately unchanged and $\mathsf{\eta }$ was less than 0.5%. This agrees with the grid independence [37]. The mesh size of the last refinement was used to determine the simulation accuracy. As shown in Figure 8, the maximum element size was 0.005 m and the number of grids was 1,271,934.

4. Simulation Results and Discussion

In this section, the temperatures of all the monitoring points (K1 and K2, B1–B4, p1–p9) were measured using the platform shown in Figure 2 (using thermocouples). The measurements were repeated five times and the system was allowed to cool naturally to room temperature (approximately 299.15 K) after each measurement.

4.1. Amplification Module

The simulation results for the amplification module are shown in Figure 9. Figure 9a,c show that the temperature decreased as the distance outward from the heat source increased. In addition, there was an apparent temperature gradient across the fins. The frame and fins conducted and dissipated heat. Figure 9b shows that the temperature was higher in the central region, owing to the position of the amplification module.

The temperatures at points K1 and K2 were measured to verify the accuracy of the amplification model. A comparison of the experimental and simulation data is shown in Figure 10. During the reverse transcription phase, the temperature of TEC side2 was more consistent. However, as the instrument entered the pre-denaturation and cycle phases, the overall simulated temperatures of TEC side2 were slightly higher than the experimental values. In the model, there are gaps between the parts and the air in these gaps increases the thermal resistance. The simplified simulation model ignores the thermal resistance of the connections between the parts; hence, the thermal conductivity of the simulation is higher than that of the actual model. Therefore, when the temperature changes, the simulation value is always greater than the experimental value. Overall, the trends of the simulated and experimental results were consistent.

4.2. PCB

Figure 11 shows the simulated and actual temperature distributions for the two PCBs, which were similar in both cases. There were a large number of relatively centralized heat sources on PCB1 which were clearly shown in the simulation. However, there were some differences between the simulated and experimental results for PCB2. In particular, a minor heat source on the right side of PCB2 was captured by the thermal imager but not considered in the simulation.

To verify the accuracy of the PCB model simulation, points B1–B4 were studied without affecting the operation of the PCBs. A comparison of the experimental and simulation results is shown in Figure 12. During the recording process, the initial value was always higher than room temperature (approximately 299.15 K). This is because there were many components and copper wires in the PCBs which caused Joule heating during operation and affected the PCBs. In addition, the secondary heat source was ignored in the model. Therefore, the simulated temperature at each point was lower than the experimental value. Moreover, it was difficult to maintain a stable room temperature during the tests, which presented another source of error. We predict that the simulated temperature will still be lower than the experimental temperature after 600 s; however, the error will gradually decrease. Overall, the simulation and experimental results show similar increasing trends, which indicates that the model and simulation can be used for reference.

4.3. Overall Device

Figure 13 shows the simulated temperature distribution for the entire solid between 60 and 240 s. To show greater distinction, the upper limit for the displayed temperature was adjusted to 333.15 K (60 °C). The figure clearly shows that heat was conducted through the solid over time and a temperature gradient formed over the entire solid domain. Under the action of the fan, the temperature also increased in nearby objects that were not in contact with the heat source.

As shown in Figure 14, two sections were created to show more detail regarding the temperature inside the casing. The temperatures at z = 0.135 m and x = 0.036 m are shown in Figure 15 and Figure 16, respectively. The temperature field shows that the temperature was highest at the heat source and that the heat spread through the entire space with the aid of the fan. The flow velocities of the two sections and the temperature of the air domain at the last moment of the simulation are shown in Figure 17. The air passed through the fan and over the cooling fins. The flow rate was highest at the fan outlet and gradually decreased as the heat-dissipation structure blocked the air and caused it to diffuse. Figure 17a,b show the cross-sectional flow velocity and the path of heat diffusion. The temperature spread along the wall under the action of the velocity field and finally gathered at the top near the rear-left corner.

The temperatures at points p1–p9 (indicated by red crosses in Figure 18) were used to verify the utility of the simulation. The electronic components and screen control panel in the upper part of the instrument are vulnerable to heat; therefore, the points were selected from this region. The simulation and experimental results are presented in Figure 19. The errors in the simulation were mainly caused by the model simplification. In addition, it was difficult to model the wiring in the instrument for the simulation process. The wires exchange heat with the air, which affects the local airflow and temperature. Hence, there were some differences between the experiments and simulations. The temperature error (obtained by subtracting the experimental value from the simulated value) of each point is shown in Figure 20. The errors were within ±3 K and the maximum relative error for p4 was 0.9%. Therefore, the results of the finite-volume simulation can be used to guide the optimal design of the temperature field for nucleic acid detection instruments.

4.4. Heat-Dissipation Scheme

During operation, the fan circulated the air inside the instrument, and over time the ambient internal temperature increased. High ambient temperatures threaten the stability of the electronic components. The analysis in Section 4.3 showed that the heat was concentrated at the top-back section of the instrument; based on this, two heat-dissipation schemes were designed, as shown in Figure 21. In Plan 1, an inlet was positioned at the bottom of the front of the instrument and an outlet was positioned at the top of the back (Figure 21a). In Plan 2, inlets were positioned around the bottom sides of the instrument and an outlet was positioned at the top (Figure 21b). In both cases, the area of the inlet was approximately equal to the area of the outlet.

There are fans for the amplification modules inside the instrument. According to the pressure difference obtained by simulation, combined with the pressure polynomial in Section 3.2, it is estimated that the velocity is between 5.5–6 m·s⁻¹ (flow rate is between 0.014–0.016 m³·s⁻¹). Considering that the volume of the fluid domain inside the entire instrument was 0.038 m³, a silent fan with a maximum flow rate of 0.028 m³·s⁻¹ was preliminarily selected (Delta Electronics, AFB0924HH). Thus, the pressure polynomial was:

$$\Delta P=53.935-12.953v-0.3001v^2$$

(15)

The temperatures at p1–p9 at the last moment of the simulation were used as the evaluation indexes. Figure 22 shows the temperatures at each point under each scheme. At each point, the temperature under Plan 2 was lower than that under Plan 1; therefore, Plan 2 had a superior heat-dissipation effect. Figure 23 shows the temperature and flow velocity of the section and the overall streamline of the Plan-2 model at the last moment of the simulation. The fluid entered the instrument at a high velocity and then slowed as it became blocked by the internal structure of the device. The flow velocity was increased to remove heat under the action of the fan on the amplification module.

The trends in the simulated and experimental temperatures at points p1–p9 under Plan 2 are shown in Figure 24. The relative error of each point did not exceed ±2 K and the maximum relative error of p9 was 0.54%. The acrylic casing of the prototype based on Plan 2 was used to evaluate the heat-dissipation capability. Figure 25a shows the prototype of the casing and instrument body. Nucleic acid amplification was conducted continuously for 120 min and the temperatures of p1–p9 were measured simultaneously, as shown in Figure 25b. The maximum temperature was 307.76 K. During the first amplification process, the temperature increased rapidly and then stabilized in subsequent amplification cycles. After each amplification, the temperature decreased significantly, which indicates that the residual heat was removed from the instrument immediately after each amplification. This shows that the proposed heat-dissipation scheme for nucleic acid detectors is feasible.

5. Conclusions

In this study, a vent scheme was designed to provide a safe working environment for the reliable operation of a nucleic acid detector. A simplified 3D model of the instrument was produced and the finite-volume method was used to simulate the temperature and velocity fields within it. The simulation showed that heat accumulated in certain areas of the instrument. A temperature-measurement platform was constructed to measure the temperature experimentally and the results were used to validate the simulation and its accuracy was confirmed. A comparison between the experimental and simulation results showed that the temperature error did not exceed ± 3 K and that the maximum relative error was 0.9%. Therefore, the finite-volume simulation results can be used to guide the design of the temperature field for nucleic acid detectors. A total of two ventilation schemes were designed based on the simulation results and the best scheme was employed to fabricate a new casing for the instrument. In a long-term nucleic acid amplification experiment, the monitoring points reached a maximum temperature of 307.76 K. Therefore, the vents in the designed casing showed excellent heat-dissipation performance; hence, they can remove waste heat and create a safe operating environment for the instrument.