Experimental Characterization and Response Surface Modelling of Electrostatic Charge Dissipation on PET Insulators Under Controlled Thermo‐Hygrometric Conditions

ABSTRACT

Insulating polymer materials are widely utilized in electrical and electronic applications. During their operation, they have the capacity to accumulate electrical charges on their surface or in their volume, which can cause their premature aging and therefore alter their long‐term performance. In the present work, the flow of ions deposited on the surface of polyethylene terephthalate (PET) is analysed using the surface potential decay (SPD) method. Therefore, an experimental study is carried out, followed by modelling of the results obtained. A negative corona discharge, produced in a needle–grid‐plate electrode system, is used to charge the surface of the samples. In order to obtain reproducible results, all SPD measurements are carried out in a commercial climate chamber, where relative humidity (RH) and temperature ( T ) are rigorously controlled. The effects of initial potential, relative humidity, thickness and polarity are also analysed. The results showed that the SPD is strongly conditioned by these factors. It should be noted that temperature and humidity do not contribute equally to the neutralization of deposited charges. Temperature acts as an accelerator of the process, whereas humidity enhances its effect. In addition, the experimental results demonstrate that the SPD depends on the internal structure of the polymer and confirm the influence as well as the role of the electric field in the decay of the potential. The experimental design methodology was used to quantify the effects of different influential factors (i.e., initial potential, humidity and temperature) and to optimize the efficiency of the characterization of the electrical properties of PET. Under identical thermo‐hygrometric conditions ( T  = 55°C, RH = 80%), the SPD increases by about 20% as the time increases from 5000 to 80,000 s (i.e., a 16‐fold increase in time). Two mathematical models were developed to estimate these parameters and their interactions under short‐ and long‐term conditions, demonstrating excellent statistical performance, with goodness of fit ( R ²  > 97.5%) and goodness of prediction ( Q ² > 87%).