By: Parizad Shojaee Nasirabadi, Jesper H. Hattel
INTRODUCTION
Condensation and moisture related problems are the cause of failures in many cases and consequently serious concerns for reliability in electronics industry. Therefore, it is important to control the moisture content and the relative humidity (RH) inside electronics enclosures. Understanding the transport phenomena and the effects of different parameters on the RH management and local climate inside the enclosures and applying this knowledge during the design phase is crucial for reducing the chance of failure and controlling maintenance costs. For decades thermal management has been extensively studied; however, the RH of the operating environment is commonly disregarded during the design of electronics enclosures [1]–[3].
Numerous parameters affect the local climate inside electronic enclosures, such as material properties, dimensions of the enclosure and the electronics, components and their configurations and finally environmental conditions. Ambient condition changes significantly affect the local climate inside electronic enclosures.
In applications such as military, industrial, commercial or consumer electronics, certain equipment may contain devices highly sensitive to environmental conditions [1].
In order to precisely predict the local climate inside electronics enclosures, one has to perform coupled momentum, heat and mass transfer analysis on the system composed of several components of various sizes (such as PCBs, heat sinks …). The method of choice for making such predictions is mostly CFD (computational fluid dynamics). Application of CFD analysis has the potential to provide accurate solutions and allow the user to assess them in different cases [3]–[5].
The main objective of this work is study the effect of ambient conditions on local temperature and RH inside the enclosure so that to prevent the 100 (%) RH and the consequent condensation. Exposure to high RH leads to condensation of water on the electronics. The concentration of water molecules rises as the RH increases. The thickness of the molecular layers of water eventually permits ionic conduction which can lead to changes in electrical resistance and even short circuits. This phenomenon accelerates the rate of corrosion [2].
In this study, the effect of ambient conditions in three cities on a summer day (July 1st, 2016) is studied on local RH and temperature on a PCB (printed circuit board) in a cold-water pump enclosure. Furthermore, the results from the transient simulations can be used to improve the humidity management inside the enclosure. Figure 1a shows a 2D schematic view of the pump. The enclosure consists of two chambers, connected with a small tube.
The chamber on the left side (the control box) contains the electronics and the chamber on the right houses the copper coils, coil keeper and the shaft. The 3D geometry of the pump is shown in Figure 1b. The dimensions of the pump are listed in Table 1.
NUMERICAL SIMULATION METHOD
Temperature gradients in the system generate buoyant forces that lead to an air flow. In this work, to estimate the velocity profile caused by these volumetric forces, the energy equation is fully coupled with the momentum and continuity equations. Figure 2 demonstrates the way that these equations are coupled. The continuity equation or equation for the overall mass balance is:
The equation for the momentum transport is:
To consider the buoyant flow, the volume force is considered as given below:
The energy equation reads:
Several material properties are included in these equations, namely density, viscosity, thermal conductivity and specific heat capacity. Both temperature and RH influence these thermophysical properties of air. Thus, it is crucial to consider the changes during the calculations [6]. The effect might be significant or insignificant depending on the temperature and RH range of the study. Tsilingiris [7] evaluated the thermophysical properties of moist air as a function of mixture temperature (from 0-100°C) and RH, ranging from dry air to saturation conditions. In another study, Melling et al. [8] provided simple analytical correlations for humid air. The correlations were derived from theory and numerical curve fitting in the temperature range of 100-200°C. Zhang et al. [9] also estimated the thermophysical properties of humid air using the same concepts for the binary mixture of air and water vapour as the two other studies. In the current study, the thermodynamic correlations from the Tsilingiris’ work [7] are used.
The commercial software package COMSOL Multiphysics TM version 5.1 was used for all the CFD simulations. Due to the complicated structure of the modeled system, an unstructured mesh composed of tetrahedral elements is applied on the computational domain in all the simulations. The adaptive mesh refinement method was used to improve the mesh quality.
The average temperature on the PCB was monitored to examine the mesh independency. The final mesh consists of 1,055,444 elements with an average quality of 0.79. The element sizes in x, y and z direction are on order of micrometers. With such a fine mesh, each simulation took about five hours on 20 nodes of scientific Linux 6.4 cluster; where each node was configured with an Intel Xeon Processor X5550.
The solver split each problem into one or more linear systems of equations by approximating the given problem with a linearized problem. In this work, the parallel direct solver (PARDISO) was utilized as the linear system solver. This memory efficient solver works on general sparse linear systems of the form Ax = b and use LU factorization on the matrix A to compute the solution x.
RESULT AND DISCUSSION
The temperature and RH on the PCB are calculated while the pump enclosure is exposed to the ambient conditions in Copenhagen, Atlanta and Singapore for a day (July 1st, 2016). Figure 3 shows the ambient conditions for these three cities, using data from [10]. The working cycle of the pump is shown in Figure 4. For all the three cases, the initial temperature, gauge pressure and velocity are 25°C, 0 Pa and 0 m/s) respectively. The cold wall temperature is always the same as the cold water (the water entering the pump), which is 5°C. The initial RH follows the conditions of each city. The RH of the trapped air inside the enclosure follows the ambient changes transiently, due to the relatively large drain hole and the internal air flow.
Singapore.
The temperature and consequently the RH of the PCB are affected by the heat capacity of each of the components inside the enclosure. This effect is extensively studied in another work [1]. Thus, to make a comparison between the cities, all the parameters are the same in the depicted cases.
In the first 10 hours that the pump is not working, the temperature of the PCB follows the same trend as the ambient. For Copenhagen and Atlanta, the initial temperature is colder than the 25°C PCB. Thus, the RH on the PCB is lower compared to the ambient value (see Figure 5). For Singapore, the temperature of the PCB is initially lower than the ambient. That leads to a 100% RH on the PCB. After about 10 hours, the pump starts working. During this time, the heat generated by the electronics on the PCB increased the temperature, which significant decreases the RH on the PCB.
Atlanta and Singapore.
It is worth mentioning that, despite the fact the PCB temperature is mainly affected by the heat flux, it still follows the peaks and valleys of the ambient changes. Regardless of the working cycle of the pump, the condensation risk is generally higher in the first 6 hours of the day due to lower ambient temperatures and higher RHs. From 8:00 to 18:00 hours, the temperature increases and consequently the RH decreases. On the other hand, in most of the applications, pumps work some time from 8:00 to 18:00. This causes even lower RH on the PCB.
Therefore, while the electronics are safe during the working hours, they are exposed to high RH while not they are not operating. In general, the RH on the PCB can be reduced by increasing the temperature (maximizing the saturation limit) or by reducing the moisture content of the air in contact with PCB. Considering the fact that there are drain holes in many electronics enclosures, it is quite challenging to control the moisture content. Furthermore, the natural convection of the trapped air makes the moisture transfer even faster. In the case of this pump enclosure, the presence of the drain hole is necessary to eliminate the condensed water on the cold wall. Thus, temperature control of RH, such as with internal heaters, is a more feasible approach.
CONCLUSION
A 3D finite element based CFD model is developed for investigating the RH evolution on a PCB inside a pump enclosure. The pump is exposed to the ambient conditions of Copenhagen, Atlanta and Singapore in a summer day. The following concluding remarks can be made:
The condensation risk is higher in the initial hours of the days. This is due to the low ambient temperatures and the fact that, in many cases, the electronics are not operating during that time of the day.
From 8:00 to 18:00, both ambient conditions and the generated heat by the electronics reduce RH on the PCB. This demonstrates that a heat source in the PCB box can be very effective for reducing the condensation risk on the PCB in the early morning and during the time that the pump is not operating.
It is worth mentioning that the power consumption of the heat source should be optimized according to the ambient conditions, working cycle of the electronics and other components inside the enclosure.
ACKNOWLEDGEMENT
The current research has been conducted as part of the ICCI project from the Danish Council for Independent Research, Technology and Production (FTP) and the IN SPE project from the Danish Innovations Fonden which are highly acknowledged. Moreover, the authors would like to acknowledge the commitment and help of the industrial partners in this project.
REFERENCES
[1] P. Shojaee Nasirabadi and J. H. Hattel, “A 3D numerical study of humidity evolution and condensation risk on a printed circuit board (PCB) exposed to harsh ambient conditions,” Microelectron. Reliab., vol. 83, 2018.
[2] P. Shojaee Nasirabadi, M. Jabbari, and J. H. Hattel, “CFD simulation and statistical analysis of moisture transfer into an electronic enclosure,” Appl. Math. Model., vol. 44, pp. 246–260, 2017.
[3] P. Shojaee Nasirabadi, M. Jabbari, and J. H. Hattel, “Numerical simulation of transient moisture and temperature distribution in polycarbonate and aluminum electronic enclosures,” in International Conference on Thermal, Mechanical and Multi-Physics Simulation and Experiments in Microelectronics and Microsystems, EuroSimE, 2016, pp. 1–6.
[4] P. Shojaee Nasirabadi, M. Jabbari, and J. H. Hattel, “Estimation of water diffusion coefficient into polycarbonate at different temperatures using numerical simulation,” AIP Conf. Proc., vol. 1738, no. 1, 2016.
[5] M. Tencer, “Moisture ingress into nonhermetic enclosures and packages. A quasi-steady state model for diffusion and attenuation of ambient humidity variations,” Electron. Components Technol., pp. 196–209, 1994.
[6] P. Shojaee Nasirabadi and J. H. Hattel, “A Numerical Investigation of the Effect of Ambient Conditions on Natural Convection Cooling of Electronics,” in 23rd InternatIonal Workshop on Thermal Investigations of ICs and Systems, 2017, vol. 1, pp. 5–10.
[7] P. T. Tsilingiris, “Thermophysical and transport properties of humid air at temperature range between 0 and 100 °C,” Energy Convers. Manag., vol. 49, no. 5, pp. 1098–1110, 2008.
[8] H. V. Adrian Melling and Stefan Noppenberger and Martin Still, “Interpolation correlations for fluid properties of humid air in the temperature range of 100 °C to 200 °C,” J. Phys. Chem., vol. 26, no. 4, 1997.
[9] J. Zhang, A. Gupta, and J. Baker, “Natural Convection Heat Transfer Coefficients Effect of Relative Humidity on the Prediction of Natural Convection,” Heat Transf. Eng., vol. 7632, no. August, 2007.
[10] “COMSOL Multiphysics 5.1 User Guide,” 2015.
