As increasingly pumped-storage power stations are built

designing the upper reservoir will encounter more complicated geological issues

and better evaluation on cement grouting projects is demanded. However

the invisibility and complexity of the underground project make the current grouting theory difficult to guide the grouting practice effectively. Artificial intelligence technology has significant advantages in addressing problems characterized by fuzzy constitutive relationships

bringing a new trend in geotechnical engineering field in big data era. But its main inputs are from geological datasets that are usually too small because of difficulties in practical collection limited severely by technical means

time constrains

and costs. Instead

the grouting data

which can be easily acquired in large amounts

is not being fully utilized. This paper presents a new grouting quality evaluation method based on descriptive statistical analysis and spatial statistical analysis on grouting data. It is based on the deep mining of information contained in the 4350 groups of grouting datasets collected from 1104 grout holes and a small amount of geological data from the Wuyue Pumped-Storage Power Station project. The results show that the descriptive statistical characteristics of grout consumption can be used to identify fresh bedrock

general fractured rock mass

densely fractured rock mass

and rock masses that features a higher risk of substandard grouting quality. The grout consumption is correlated with geological conditions to a certain extent

but their impact varies in different grouting sequences due to spatial variations in geological features and different fracture fillings. Neither grouting quality nor grouting efficiency can be evaluated only based on the consumption in the first order grouting. Compared to the limited data from inspection holes

massive grouting data in high density help evaluate grouting quality more comprehensively. Full mining of the information hidden in grouting data is a new approach to full utilization of massive grouting data

which would greatly promote the intelligent development of grouting engineering.