model for DQC control that can be agreed upon at a (inter)national level in due
course. Given that the customers are relatively the same in the country, this implies
that the water utilities can develop an intercompany data model for water accounting,
with the possibility to extend it in the future as new variables are incorporated in their
data warehouses.
Data Collection Redundancy
As presented in this chapter, the analysis of the data diet of large data collections
may help utilities identify which sensors are more reliable and how much data
storage is required. It is not yet clearly known by utilities which variables are
correlated among different sensors. An effort should be made to develop an analysis
of redundancy with the utilities using their data mining techniques discussed in this
chapter.
Aggregation of Data
Aggregation of data is performed by all companies, for specific purposes. Additional
data analysis which is of interest for all utilities can be introduced to obtain regular
water balance calculations. All utilities have to account the water that is produced
and billed. However, it was identified that the utilities have serious concerns about
the limits of the anomalies in the water balance. Given that non-revenue water
(NRW) is not a serious concern in the Netherlands, the main issue is to be able to
identify when the water supply system presents a deviation from its regular pattern.
As demonstrated in this chapter, for some utilities, the need to establish such
boundaries for water balance is a current issue. Even with advanced tools for
water accounting, there are deviations present in the data of water balance for all
utilities. Therefore, it is recommended to implement advanced techniques such as
model based validation to tackle this issue. For this, a pilot for a DMA configuration,
with an optimal time step and additional info (e.g. pressure, flows) and expert
knowledge (e.g. operators’ knowledge and logbooks of operations and maintenance), can be compared with a simulation model to determine anomalies.
Data Correction Techniques
Most methods applied for incomplete time series have been developed for surface
and groundwater hydrology [39, 94, 95], and some applications have been made for
WDN [61]. This chapter and its supporting research did not address the issue of data
correction techniques. A proposal was made to continue the process of using
hydroinformatics tools in the Dutch drinking water companies. The goal here
would be identifying specific techniques for data correction which may be suitable
for a subset of WDN variables. If for the case of data validation the spectrum of
A Bird’s-Eye View of Data Validation in the Drinking Water Industry of the. . .
103
course. Given that the customers are relatively the same in the country, this implies
that the water utilities can develop an intercompany data model for water accounting,
with the possibility to extend it in the future as new variables are incorporated in their
data warehouses.
Data Collection Redundancy
As presented in this chapter, the analysis of the data diet of large data collections
may help utilities identify which sensors are more reliable and how much data
storage is required. It is not yet clearly known by utilities which variables are
correlated among different sensors. An effort should be made to develop an analysis
of redundancy with the utilities using their data mining techniques discussed in this
chapter.
Aggregation of Data
Aggregation of data is performed by all companies, for specific purposes. Additional
data analysis which is of interest for all utilities can be introduced to obtain regular
water balance calculations. All utilities have to account the water that is produced
and billed. However, it was identified that the utilities have serious concerns about
the limits of the anomalies in the water balance. Given that non-revenue water
(NRW) is not a serious concern in the Netherlands, the main issue is to be able to
identify when the water supply system presents a deviation from its regular pattern.
As demonstrated in this chapter, for some utilities, the need to establish such
boundaries for water balance is a current issue. Even with advanced tools for
water accounting, there are deviations present in the data of water balance for all
utilities. Therefore, it is recommended to implement advanced techniques such as
model based validation to tackle this issue. For this, a pilot for a DMA configuration,
with an optimal time step and additional info (e.g. pressure, flows) and expert
knowledge (e.g. operators’ knowledge and logbooks of operations and maintenance), can be compared with a simulation model to determine anomalies.
Data Correction Techniques
Most methods applied for incomplete time series have been developed for surface
and groundwater hydrology [39, 94, 95], and some applications have been made for
WDN [61]. This chapter and its supporting research did not address the issue of data
correction techniques. A proposal was made to continue the process of using
hydroinformatics tools in the Dutch drinking water companies. The goal here
would be identifying specific techniques for data correction which may be suitable
for a subset of WDN variables. If for the case of data validation the spectrum of
A Bird’s-Eye View of Data Validation in the Drinking Water Industry of the. . .
103