Award-Winning Pump Optimisation Solutions from Deritend

Date Posted: 12/07/16
Our client Severn Trent Water engaged with us to deliver real time pump and turbine network optimisation. The organisation’s Melbourne Area Network near Derby has always suffered from large variations in energy use and numerous historic investigations had not uncovered the reasons or best operating policies.

With the start of the sixth Asset Management Period (AMP6) in 2015, companies such as Severn Trent are taking a longer term view on how costs should be saved. The company has embarked on a significant programme of operational efficiency savings and capital expenditure efficiency savings through targeted investment. A key part of its infrastructure is the Melbourne Area Network.

Our project won the Pump Centre’s Project of the Year 2016.

Working in partnership with Severn Trent Water we applied our patented pump performance monitoring systems across the Melbourne Network pumping stations to assess individual pump and station efficiencies and derive live network models.
Information-driven approach
The project required two distinct pieces of work:

• Provision of real time measurements of pump, station and network performance in terms of efficiency and specific cost.

• Design of a custom Decision Support System (DSS) to enable operators to have real time visualisation of pump efficiency and provide instructions on how to operate the network, meeting demand, at the lowest cost.
We installed our monitoring equipment on all machines at the network’s stations, a total of 26 pumps.

We then measured pump and turbine (hydraulic) efficiency using a thermodynamic method (according to ISO5198:1999).

Utilising the data we gathered, it was possible to implement an automated Decision Support System (DSS) to provide recommendations as to which pumps to operate, what speeds, and when, to achieve the lowest overall operating cost whilst taking account of variable tariffs, storage, demand and asset performance.

To do so, the project pioneered the use of machine-level efficiency and performance measurements for the minimisation of energy costs, forming the foundation for the ‘Decision Support System’ (DSS). The DSS works by assessing multiple combinations / scenarios to achieve a given demand – a system able to translate huge volumes of real time data to accurately optimise an entire sub network.

Due to the sheer number of combinations available at Melbourne, a simple statistical approach would not be practical.

So at the centre of the DSS is some of our most innovative patent-protected technology, comprising Artificial Neural Networks – to measure variable speed drive efficiency – and Genetic Algorithms – to compute lowest cost pump schedules, with accuracy, across a network. It is technology that has been developed to vastly improve best practice in water distribution and address modern challenges in the UK water industry.

The complex algorithms developed as part of the project enabled Severn Trent Water to accurately schedule operation over a set horizon, meeting the demand with very appreciable energy cost reductions – a process simply too difficult for operational staff to do by experience alone.

Savings attributable to the installation of the network optimisation system were categorised on three levels:

• Station Level – Those achieved through optimisation of individual stations in isolation through pump scheduling and speed selection.

• Network Level – Those achieved through optimisation of water routing, tariff and storage.

• Capital Savings – Savings potential identified through additional capital expenditure to reduce operating costs in the longer term.


• From a benchmark demand of £4.5m/annum, an overall saving of 9% (£400,000) was realised through operational changes only with no additional CAPEX beyond this project resulting in a <1 year payback.

• In addition, a further 4.5% (£200,000) of annual savings were identified through targeted CAPEX of £300,000.

• If additional capital expenditure was made and all possible sources of saving were obtained by upgrading the system to ‘best practice’, a saving of 19% could be achieved, equivalent to £855,000/year.

• Taking the problem at its most basic level, the network pumps water from its source to the water treatment works which is it a higher elevation. Conducting a simple hydraulic calculation to calculate the power required to raise water by this height reveals a minimum theoretical energy cost of £1.45m; the current cost is some three times greater highlighting the degree of losses (and so potential) in water distribution systems.

Further roll out:

• The impressive results have changed the way the network operates and it is intended to roll the system out to other networks across the AMP6 period to utilise this new information driven approach as a key feature in targeting resources for OPEX reduction and CAPEX allocation – fulfilling the TOTEX mandate.