A submersible solid distribution transformer (SDT) (Fig. 1) has been developed at CITEQ, a joint venture between ABB Canada and Hydro-Quebec. It enables electric utilities to operate their distribution systems economically while meeting strict environmental standards. The SDT is corrosion-free, silent, withstands severe weather conditions and contains no liquids or gases of any type. It eliminates the risk of soil or groundwater contamination.

Direct Burial Now Possible A growing number of utilities are now considering switching distribution from overhead to underground where the transformers, switches and fuses can either be located in an underground vault or be padmounted. To reduce the cost of underground residential distribution (URD), many attempts have been made to directly bury the transformer. The outcome of these initial installations was disappointing. Two major causes were identified: 1. Corrosion of the oil tank often leads to oil leakage and ultimately to breakdown. The many attempts to use epoxy painting or to replace the metallic tank by a polymeric one have all failed because of wear and moisture penetration. 2. The soil creates an additional resistance to heat dissipation from the transformer to the ambient air, leading to an important derating of the permissible load.

The SDT is an ideal method for preventing or alleviating problems such as these because of its corrosion-free characteristics and the solid insulation that causes no environmental hazards.

The installation procedure for direct burial requires local excavation, cable connections, backfilling and landscaping. In our prototype installation, a 50-kVA SDT is installed inside a thin-walled, cylindrical-shaped plastic tube. This facilitated the installation of several temperature and current sensors to monitor the transformer's performance. The primary and secondary cable connections were made inside the pit prior to backfilling. In future installations, these connections could be installed inside a small, buried junction box in the immediate vicinity of the transformer.

Backfilling is an important step in any directly buried installation. The backfill needs good thermal conductivity that is retained during the equipment's operating life. Many developments have been made in the past 20 years to improve the backfilling around buried equipment. Today, it is quite feasible to prepare and test an appropriate mixture of locally available materials that are cheap and remain efficient in conducting the heat, even when completely dry. This practice is used frequently in underground cable installations.

Our facility has used a mixture of 0-3/4-inch stone screening, sand and water. A small amount of flyash was also added. The mixture was vibrated during the installation to get rid of trapped air. The measured thermal resistivity varied between 50 and 60 cmoC/W at the beginning and during the preliminary testing period, which was carried out at one pu load for several weeks and interrupted by emergency overload tests at 1.4 and 2 pu. To guarantee the thermal properties of the backfill over the years, a small amount of Portland cement could be added. At this stage, we decided not to do so in order to have easy access to the transformer and its connections.

Mini-Vault Installation For utilities that are not yet ready for the direct burial of their distribution transformers, mini-vaults are an interesting alternative to padmounted transformers. They are aesthetically appealing but can be slightly more costly. In comparison with the direct-buried practices, a transformer buried at the same depth inside a closed vault will normally have a more or less equivalent loading performance. If the vault is ventilated-as is the case for large installations-the cooling brought on by ventilation will improve the general performance of the transformers. Unfortunately, providing openings for ventilation will inevitably lead to littering and added pollution.

At Hydro Quebec, there is a need to increase underground distribution systems and to reduce costs. Thus, the mini-vault total solution concept came into existence. Bottomless vaults made of fiberglass with high-density polymer cement covers can be easily installed with a crushed rock foundation, avoiding the need for heavy cement components. The vault can be separated into two compartments with the transformer, along with the primary connections in one compartment and the secondary connections in the other. Both secondary fusing and fault indication can be added.

Figure 2 shows a 50-kVA SDT inside of a mini-vault that can be buried with its covers flush with ground level. This concept using a 100-kVA transformer is what Hydro Quebec has chosen to be used in several locations in its residential distribution network.

Thermal Rating The heat transfer between a transformer and its surroundings is a complex problem. In the case of oil transformers, there are a number of guides and standards (IEC, IEEE) that accurately deal with this matter. In the case of completely encapsulated transformers, like the SDT, there are none. The heat generated in the core and the windings has to dissipate by conduction through the various non-homogeneous materials present inside the transformer, then by convection and radiation from the outside surface. The matter is further complicated when the transformer is installed underground. In this case, the heat also has to go through the layers of soil and backfills before dissipating into the ambient. This is a full three-dimensional problem where time has to be considered because of the large time constant of the surrounding ground. The following partial differential equation is used to analyze the complex heat dissipation inside the transformer and its surroundings: where k and (rc) are the thermal conductivity (W/moC) and volumetric thermal capacity (J/m3oC) of the transformer and its surrounding materials, respectively. Both can be temperature and space dependent. Q is the various heat sources (W/m3) located in the core, and in the high- and low-voltage windings, they are temperature dependent. T and t are the temperature (oC) and time(s), respectively.

An analytical solution to this equation is practically impossible. Instead, finite element numerical methods are better suited to solve such a complex problem with great accuracy. Today, these methods are widely used in all engineering disciplines. The first step in applying these methods is to subdivide the domain into small elements, obtain the finite element mesh, then assemble the corresponding matrix resulting from the equation (1). The solution gives the temperature distribution everywhere in the domain including the transformer. The hot spot is then easily identified.

Thermal Rating in Free Air The main thermal resistance to heat dissipation in free air is caused by the insulating epoxy. The epoxy used in a conventional dry-type transformer has a relatively low thermal conductivity (0.2-0.3 W/moC). Better epoxies were developed previously with higher conductivity and better mechanical and electrical characteristics. Such an epoxy was used in the SDT with a thermal conductivity exceeding 0.8 W/moC at 20oC.

The other major parameter that affects the heat dissipation is the ratio of the exposed outside surface of the transformer to the amount of heat generated inside. For relatively small and medium size transformers (25, 50 and 100 kVA), this ratio is high enough so that natural cooling is sufficient. Beyond this limit, i.e. for 167 kVA and above, the smaller surface-to-heat ratio makes it difficult for the natural cooling alone to be sufficient.

Therefore, we conceived a new, state-of-the-art system based on the heat-pipe technology. This is a totally passive device that consists of an evaporator and a condenser located inside a small copper tube requiring no outside power supply and posing no risk to the environment.

The hot-spot temperature rise in free air is below 100oC at full load, which is the targeted temperature rise for this product.

Thermal Rating in Direct Buried In the case of a constant load, any heat-generating electrical equipment is bound to be less efficient when buried in the ground compared to a similar installation exposed to free air.

This is because of the additional resistance created by the surrounding soil. To analyze the heat dissipation of the complete direct buried (DB) system, we proceeded first with the generation of the finite element mesh.

Compared with a similar installation in free air, it can be seen that a direct-buried SDT would eventually run hotter than one exposed to free air for constant long-term loading. However, a DB transformer has a much longer time constant, allowing for significant fluctuations in short-term overloading. In fact, the impact of large variations in load on temperature fluctuations inside a DB transformer is much less severe than on one exposed to free air. The damping effect in the case of a DB transformer leads to smaller thermo-mechanical stresses and, consequently, to longer service life.

Thermal Rating in Underground Mini-Vaults A transformer installed inside a closed vault will normally have a more or less equivalent performance to a direct-buried transformer. The thermal rating of such a transformer installed in a closed shallow mini-vault is estimated to be 15% lower than a comparative installation in free air. This is true if the transformer is loaded at 100% for a prolonged period of time. At Hydro-Quebec, the average residential daily load cycle rarely exceeds 75%. In such case, the relative drop in performance of a DB transformer is negligible.

Conclusions The SDT, installed either in a mini-vault or directly buried, is well suited for underground residential distribution. The reasons include corrosion and moisture immunity, adequate loading performance and operational benefits. Moreover, the use of cheap, simple but effective thermal backfill improves the overall loading characteristics, especially in emergency situations where the longer time constant plays a significant damping effect on internal temperature variation. Acknowledgement: The authors wish to thank C. Paradis for his primary role in developing the SDT and Jim Ferrero for his role in preparing this article. Both are members of CITEQ.

M. Chaaban has been a research engineer at IREQ, Hydro-Quebec since 1981. He is also a project leader at CITEQ. He has the BSc from the Technical University of Budapest, the MScA from Ecole Polytecnique de Montreal and the PhD degree from Concordia University, Montreal. He has been working on the modeling of heat transfer phenomena related to energy conservation and electrical equipment and has developed several codes for heat transfer calculation and cable ampacity calculation using the finite element technique.

J. Leduc joined IREQ in 1988. He has the BEng degree from Universite de Sherbrooke, and the MScA from Ecole Polytechnique de Montreal. He has been working with the high-voltage cable group on heat transfer calculations and computer program developments. His research activities are related to numerical methods in heat transfer.

All physical events in nature, such as heat transfer, are governed or can be described by partial differential equations (PDE). An analytical solution to a PDE in 3-D space and time is practically impossible without numerous simplifying assumptions.

A finite element method, on the other hand, is a powerful tool that can numerically solve this type of equation with great precision. The power of today's computers has lead to the increased use of this method.