Application: Model testing to establish performance of new inlet works and static mixers

Customer: Biwater Treatment Limited

Service: Physical Modelling and CoV Testing

Problem: Miller Biwater Joint Venture was undertaking the design and construction of a new inlet works at Watchgate WTW. The new inlet was to form part of the AMP3 improvements. Water from the various sources was to enter a reception chamber and pass through a blending chamber housing a Statiflo static mixer. The flow then split between four lanes for dosing and flocculation prior to recombining in the outlet chamber. Testing of the design was required prior to construction, to establish if the full-scale equipment would meet the contractual process performance requirements.

Objectives: To construct and test a scale physical model of the inlet works over a range of operating conditions, to assess the hydraulic conditions generated, the flow distribution and the headloss. Modifications to ensure satisfactory performance would be recommended where necessary.

Approach: A hydraulic model was constructed to include the complete inlet works layout between the reception chamber and the filter approach channels. The blend static mixer was modelled accurately whilst the twelve static dosing mixers were simulated using a mesh which gave the same hydraulic loss as the mixers at the maximum channel flow. Tests were carried out to observe the hydraulic performance of the structure and to assess the performance of the mixer equipment.

The static mixer performance was measured using the Coefficient of Variation (CoV). (CoV is a measure of the mixture quality delivered by the installed mixer at a given point downstream). Residence Time Distribution tests were carried out to examine whether the installed static mixers and penstocks would induce any adverse hydraulic effects.

Outcome/Benefits: In general, the flow through the inlet works was seen to be satisfactory although the flow distribution between the four lanes was not equal. However, the mal-distribution was brought within the limit of +10% by the use of standard stop logs. The head loss across the blend mixer was found to be less than the maximum permissable value. Analysis of the data resulted in a CoV of 0.082, showing that the mixer was well within the specified performance criteria.

The model functionally demonstrated and optimised design, provided a comparison of design solution options, and facilitated performance observation and assessment by the project team.

For more information, or to discuss your case, please contact us.