The $A2.4 billion project is the largest of its type in Australia and the third largest advanced water treatment project in the world. The Project comprises a network of pipelines, storage tanks/reservoirs and pumping stations that transports purified recycled water to power stations, industry and agriculture as well as to supplement drinking supplies in the region’s key reservoir, Wivenhoe Dam, via three new advanced water treatment plants from the southern and western areas of Brisbane.

An overall total of about 200 km of large diameter transfer pipelines are designed to transfer up to 232 megalitres per day (ML/d). The majority of the network is to be constructed from rubber ring-jointed, mild steel concrete lined (MSCL) pipe. Current demand for MSCL pipe is high, leading to long delivery times and high procurement costs. Demand was also compounded by the Queensland Governments declaration of the WCRWP as a ‘fast track’ project.

Mechanical design engineer Michael Rodrieguez said “The design team from GHD had to implement strategies for meeting the fast track project schedule, one of which included improving the delivery time for MSCL pipe. We foresaw the procurement of pipe as a major risk and implemented some specific cost and time saving strategies in the design. One of those strategies involved a computational surge analysis for the pipeline to determine the severity of water hammer and explore options to mitigate surge pressures.”

Conventional design of MSCL pipelines for water involves the specification of a maximum allowable operating pressure (MAOP), which is derived from the Design Pressure plus an allowance for surge. An allowance of 25 per cent is usually specified. This often leads to a conservative wall thickness specification and a pressure rating that is unnecessarily high. Through the use of computational surge analysis, designers can instead optimise pipe wall thickness by the utilisation of specific surge mitigating equipment. This could include the strategic placement of surge vessels, pressure relief valves or check valves, as concluded from the analysis.

Michael explains: “We conducted the study using a surge analysis software program known as Hytran. A model was constructed using RL’s and Chainages along the pipeline. The model was then populated with data from the design such as flows, pressures, pipe material and stiffness. A number of boundary conditions were then loaded into the model depending on the scenario of interest. The two main scenarios that were predetermined to produce the largest surge pressures in the pipeline were pump trip (due to power failure) and pump start (or DOL start). These scenarios were applied to a number of operating cases in the pipeline model and studied carefully to evaluate the system behaviour. For each operating case, the results were tabulated and graphed to illustrate pressure versus time and pressure versus chainage.”

Although the mathematics behind such calculations is complex, the graphical results instantly depict the system behaviour. Depicted below is a plot that compares pressure surge profiles for mitigated and unmitigated scenarios (Figure1), and also a pressure-time response plot without surge mitigation (Figure 2).

The results of the computational surge analysis indicated that a MSCL pipe with 16 mm steel wall thickness would be required to contain the level of surge pressures produced during a power failure event. The conventional WT sizing procedure verified this result, indicating a minimum required steel thickness of 16 mm. Pressure versus time plots also indicated a significant level of cyclic surge pressures, providing concern about long-term fatigue failure in the steel pipe – the design life of the pipeline is 75 years.

Using the computational surge analysis software, different methods of surge mitigation could be modelled and analysed to determine the most effective solution for surge pressure reduction.

Michael continues: “Various mitigation techniques were modelled, including the installation of surge vessels and pressure relief valves however the most successful option, as in most cases, turned out to be the simplest. An inline non-return valve positioned approximately midway along the pipeline section proved to mitigate the surge sufficiently to allow the pipe steel wall thickness to be reduced to 10 mm. The non-return valve acted as a partition in the pipeline, separating the upstream and downstream sections of water. In doing so the magnitude of the surge pressure was considerably reduced. Cyclic pressure loading on the pipe was also reduced, minimising the risk of fatigue failure in the pipe.”

Cost savings realised in the pipe steel WT reduction from 16 mm to 10 mm were significant, far outweighing the cost of the non-return valve installation. Manufacturing lead-time for the MSCL pipe was also reduced, enabling project milestones to be met.

Computational surge analysis can be an effective and cheap way of justifying the inclusion of specific surge mitigating equipment necessary for the specification of a realistic pipeline pressure rating. Without it, traditional industry rules-of thumb will generally push pressure ratings up to the next level, providing over-conservative design and increased capital costs. Michael recommends: “Computational surge analysis should be conducted on all non-compressible fluid pipelines greater than 200-300 m long with fluid velocities above 1m/s.”