GEOLOGIC AND GEOCHEMICAL ASPECTS OF GROUND SOURCE HEAT PUMP WELL DESIGN, BOSTON, MA

Geologic and geochemical conditions can significantly affect the design and construction of ground source heat pump wells. Principal challenges include: difficulty drilling vertical wells; borehole stability issues; and groundwater quality issues. The work at this site entailed: drilling pilot test wells, down-hole geophysical logging, groundwater sampling and chemical analysis, and a ten-day thermal loading test. Based on project heating/cooling needs, spatial and other constraints for urban sites, and observed geologic and geochemical conditions, standing column wells are commonly used.

Three test wells were drilled in the Cambridge Argillite. The first two deviated considerably from vertical, and were not considered acceptable for thermal load testing or full scale operation.

The third test well, drilled using a different method, was advanced to approximately 1,200 feet and was sufficiently vertical to be used in subsequent testing. Groundwater temperature ranged from 55^o F near the ground surface, to 62^o F at 1,150 feet. Geophysical logging did not indicate significant fractures, consistent with the observed 5 to 10 gpm well yield.

Geophysical logging and drill cuttings indicated thin sandstone and diabase intervals throughout the borehole. Drilling behavior indicated softer rock at 1,160 to 1,180 feet, which resulted in an unstable collapsing borehole. The softer rock was identified as altered volcanic ash based on observation of the cuttings and a review of the USGS description of the Cambridge Argillite.

Groundwater samples from 300, 600, 900 and 1,100 feet indicated concentrations of chloride, sulfate, sodium, calcium and magnesium increased significantly from 900 and 1,100 feet, consistent with geophysical data. Iron concentrations are high enough to cause significant precipitation problems if the water is aerated.

The well’s thermal properties, thermal conductivity (k_eff) and thermal diffusivity (α_eff), were estimated using the data from the thermal load test and the Line-Source Approximation method. The estimated values were somewhat lower than expected based on values in the literature.