Wastewater utilities face the ongoing challenge of having to plan

for operation of their existing system and for potential future

changes due to a changing climate and changing population.

Most utilities have made a significant investment over the past

20 years in the development of hydrologic and hydraulic models

of their wastewater networks to be better prepared to plan for

their systems. These models provide utilities with a means of understanding their existing systems’ hydraulic performance and how that performance might change in the future.


Typically, planning for wastewater networks is based on best engineering judgement. An engineer typically uses a hydraulic model to test alternative capital infrastructure options or operation changes that a utility can implement to overcome existing system deficiencies. In general, they test as many alternatives as they can based on the time and budget available to them, with the primary goal initially being to meet the desired hydraulic design criteria for the network. Once a number of hydraulically sufficient alternatives are identified, their costs and benefits are assessed and presented to key decision makers.


The primary drawback of this approach is that only a small subset of possible options are being enumerated. Furthermore, engineers are limited in their ability to consider multiple objectives at once. An alternative to this methodology is to use data-driven machine learning techniques to help planners screen effective infrastructure and operational outcomes from the vast number of possible choices. The data-driven strategic insights obtained through the use of these Decision Support Systems (DSS’s) can completely redefine how a utility conducts their network planning, and allows them to make robust and adaptive decisions in the face of uncertain economic and climate conditions.


A DSS typically contains four key components:

  1. Graphical User Interface: Allows the user to set up the optimisation problem, visualise their wastewater network and review outputs of the DSS;

  2. Meta-Heuristic Optimisation Algorithm(s): These algorithms are designed to efficiently search through complex solution spaces. There are many alternative types and combinations of algorithms that can be utilised;

  3. High Performance Computing Stack: Often cloud based, this provides the scalable computing power required to deliver results in workable timeframes;

  4. Formulation Module: This defines the problem trying to be solved by the planner or engineer. The formulation module includes objectives, decisions, design criteria, hydraulic model, costs and other relevant asset data.


The first three components of the DSS are not particular to wastewater networks and can be used to aid in the decision making process in other industries including water supply, water distribution, gas distribution, etc.


For a wastewater network, the key is to define the makeup of the formulation module. Firstly, a user must define the objectives for their wastewater network. It is possible to define a single objective statement or to formulate multiple objectives to be balanced against one another. For example, a single objective statement could be “to minimise combined sewer overflow in a sewershed for the lowest possible capital cost”. Alternatively, a multi-objective problem statement could be “to balance capital cost, operation cost and conduit surcharge for a sewershed”.

The most common objectives for wastewater networks consider:

  • Capital Cost (e.g. pumps, storages, SUDS/LID practices, etc.);

  • Operation Cost (e.g. energy, maintenance, treatment, etc.);

  • Risk (e.g. metric based on probability and consequence of failure);

  • Flooding (e.g. surface, basement, etc.);

  • Overflow (e.g. combined, sanitary, downstream discharge);

  • Water Quality (e.g. dissolved oxygen, total suspended solids, total nitrogen, total phosphorus).


The second component of a formulation is the focused on identifying the decisions to be made by the DSS. Effectively, these decisions represent a set of potential options to help the utility meet the objectives defined previously. The DSS will evaluate alternative combinations of these decisions.


Typical decisions include:

  • Conduit Shape: to select the size, shape and alignment of proposed new conduits. Can also be used to consider pipe rehabilitation or abandonment;

  • Conduit Offset: to modify the inlet offset of a conduit and define the vertical profile of a conduit;

  • Storage: identify location and size of new above, or below-ground offline storage;

  • Pump Curve: to size potential new pumps or remove them from consideration. The decision may also be applied to existing pumps where an upgrade or decommissioning is being considered;

  • Direct Inflow Reduction: to select cost effective target levels of inflow reduction for subcatchments by way of applying a percentage reduction in inflow at a junction. The Direct Inflow Reduction decision is applied to Junctions when wet weather inflow is simulated using time series boundary conditions;

  • Weir Crest Height: to select weir crest heights. The decision may be applied to determine the crest height of potential new weirs or to modify the height of existing weirs;

  • Control Rules: to select optimal set levels for controls in the system. This could include alternative set levels for pump operation or gate controls for in-line storage;

  • SUDS/LID: to select type (e.g. biofiltration, permeable pavement, cisterns, green roofs, etc.) and size of alternative green infrastructure techniques.


The third key component is the hydraulic design criteria that the user wants the wastewater network to achieve. For each combination of options/decisions considered, the DSS identifies how close the network is to satisfying the defined criteria and achieving the desired level of service for a given scenario.

Design criteria may include:

  • Flooding: occurs when the water surface at a node exceeds the maximum defined depth. Flooding relates to the volume lost from the system when this occurs;

  • Surcharge: is the distance between the top of a conduit and the hydraulic grade line once a conduit is flowing full. This allows the user to specify a maximum allowable surcharge for a conduit;

  • Flow Depth Proportion: is often defined as the dimensionless ratio d/D which is termed the proportional depth of flow where d is the depth of flow in a conduit and D is the maximum possible depth of the conduit;

  • Flow: allows the user to specify a maximum allowable flow for a link (conduit, pump, orifice or weir) and penalises flows that exceed this value;

  • Discharge: allows the user to specify a maximum discharge at outfalls and penalise discharges that exceed this value;

  • Maximum Velocity: allows the user to monitor maximum velocities observed in conduits and penalise instances where the maximum velocity is outside a specified range;

  • Freeboard: is the distance between the hydraulic grade line and the ground surface. This criteria allows the user to specify a minimum allowable freeboard at a junction (manhole) or a storage and penalises freeboards observed less than this value;

  • Maximum Depth: allows the user to monitor peak recorded depth at a junction (manhole) or storage and constrain it to be below a specified value;

  • Continuity Error: allows the user to monitor and penalise mass continuity errors if they exceed a specified level. Continuity errors are assessed per node (junction, storage or divider nodes) allowing the user to identify problematic areas of the model. In many cases, continuity error penalties are not included in the objective function but instead are used to monitor model performance and stability.


In addition to defining the objectives, decisions and design criteria the user needs to provide other critical asset and cost data. The most important of these is a hydraulic model of the wastewater network that defines the connectivity of the network and hydraulic properties of its functional elements (e.g. pumps, conduits, storages, weirs, orifices, etc.). Many different hydraulic models are available for wastewater networks and any of these can technically be incorporated into a DSS. Typically used models include InfoWorks ICM, EPA SWMM, InfoSWMM, XP SWMM and Mike Urban.


The DSS also requires the user provide planning level cost data. For example, a table of unit costs for alternative conduit diameters for a given construction depth or a storage cost per cubic meter of storage provided. Finally, other data, not included in the hydraulic model but that is considered critical to the decision making process can be imported into the DSS. A common example is asset condition data in the form of a risk score.


Based on the formulation, the DSS will evaluate alternative combinations of the options defined by the user. For each alternative combination the DSS will run the hydraulic model provided and evaluate whether it met the defined design criteria and sum the costs associated with decisions made. Using the meta-heuristic optimisation algorithms the DSS will efficiently search through the solution space evaluating thousands of possible combinations of options until the near-optimal plans are identified. These plans can be investigated further by the user and used a basis for making a final decision about the infrastructure to be implemented.





Using a DSS for the planning of wastewater networks can yield a variety of benefits. Based on industry feedback using a DSS has been shown to reduce capital and operational costs by up to 40% when compared to a more traditional trial and error (partial enumeration approach). Besides reductions in cost, the DSS approach offers other benefits:


  • Protecting public health through the reduction of combined sewer overflows (CSOs) and sanitary sewer overflows (SSOs);

  • Improving system reliability and sustainability by optimising for specific emergency scenarios;

  • Improving hydraulic performance in terms of meeting design criteria;

  • Gaining confidence in plans developed using a comprehensive, unbiased, defensible approach;

  • Enabling planners to identify the critical decision choices through more powerful modeling;

  • Reoptimising solutions under alternative scenarios and performing unlimited sensitivity analyses;

  • Finding solutions that balance capital vs. operating costs or gray vs. green infrastructure;

  • Staging to identify near-term projects that provide maximum value;

  • Utilising multi-objective optimisation to develop trade-off curves for triple-bottom-line (TBL) objectives.




The use of DSS software typically requires:

  1. A calibrated hydraulic model;

  2. Hydraulic performance criteria;

  3. Operational or infrastructure improvement options;

  4. Alternative option costing data;

  5. Defined objectives.