Whether you want to call it a smart wastewater network, or a real-time controlled network, or an actively controlled network, the UK water market is gearing up for some sort of instrumentation and automation changes within the wastewater collection network in the next investment period (2015-20). Some would say it is about time. Mainland Europe has had intelligence in some of its wastewater networks for over ten years now, and some of the oldest real-time control systems in the U.S. date back to the 1980’s. So why has intelligence in the wastewater network not been more abundant?
Drivers For An Intelligent Network
It could be said that in the past there were simply not the drivers for widespread adoption of intelligent networks— be they regulatory or financial. An example in Europe occurred in Barcelona where a real-time control system was put in for the Olympic Games in 1992. The initial solution was a huge interceptor tank at €150 million. This was reduced to €56 million when a real-time control system was considered. The driver was quite clearly financial.
Financial drivers are unlikely to be the main driving factors because they are simply unproven. It will take one of the water companies to make a leap of faith to do things purely on financial grounds.
However, there is at the moment a building of political pressure. Over a year ago all of the water companies received a letter from the then Secretary of the Environment, directing the installation of event detection monitoring on combined sewer overflows (CSOs) to find where the worst offending “polluters” to the water environment were. The environment agency has more recently recognised that the industry is approaching “best available technology” at the treatment works and a more holistic catchment approach is needed for permitting. Studies are starting to show that controlling the network can actually mean that tighter environmental standards can be seen for a less environmental impact (better for the environment agency) and a lesser operational cost (better for the operating company).
Internal Flooding and unsatisfactory CSO’s are just two of the drivers that could be minimised or mitigated by Intelligent Networks
What Is Needed And The Technological Barriers
So, what is actually wanted from an “intelligent network” and do we have the technology to do it?
Of course the usual mantra in this situation is – “there just isn’t the technology to do it,” and what usually happens is that the house of cards comes tumbling down and the issues and projects get forgotten.
In terms of what the industry wants that is, of course, different in some aspects for the different companies (not all of them have bathing waters or shellfish fisheries for example) and also very similar (all the companies have customers that are at risk of having their houses flooded and all have problems with infiltration). So what is needed?
In an ideal world what is needed is a sewage collection system that is aware and/or can predict what volume of water that is going to pass through it, that can convey sewage to the treatment works in the most efficient way possible, and can minimise losses (spills) and detect any unwarranted gains (infiltration and cross connections).
Some would say ‘this is technologically impossible,’ others would say ‘it is too expensive’ and some would say, ‘yes it can be done.’ Let’s look at what would be needed and what is possible.
Rainfall Forecasting And Inputs Into The Sewer Network
The accuracy of weather radar has always been questioned, but the UK Metrological Office has a “Nowcasting System” which does predict rainfall accurately, although it does not have 100 percent coverage of the UK in the correct detail for wastewater network use. The technology that is under development within Europe at the moment is smaller polarmeteric X band radar which gives a much higher resolution than the more traditional C Band radar that is currently used by the 15 Met Office radars that are used in the UK.
What of course must also be used is the network of rain gauges that cover the UK. A lot of these rain gauges are owned by other companies and stakeholders other than the water companies, but the data is collected.
Technologically it can be said that there are enough systems in place to feed enough data to a model to predict “natural” inputs into the system
C Band Radar versus X Band Polarimetric radar gives a different level of scale of data allowing a different level of radar data resolution for the Water Industry (photos courtesy of TU Delft)
Human Inputs Into The System
Of course rainfall inputs are not the only inputs into the system. Finding out what contribution is from the customer is also an important factor. Studies have been done in the UK by the Water Research Council that have looked at the consumption by different groups of customers split by such things as location, wealth, and ethnicity, and accurately found the ratios between the amount of water consumed and the amount of wastewater disposed of. Used together with smart metering this could be used to assume with a high degree of accuracy the amount of wastewater that an area should be creating. Thus an input into the system of the baseline flow through the system can be assumed.
How Much Water Is In The Network?
Measuring the amount of water in the network is fraught with difficulties. The main ones are the power requirement of the instrument and the communication back to a centralised point. This problem is exacerbated if the data is needed in what is called “real time.”
The main parameters that would need to be measured for at least a basic level of network are flow and level. Instruments do exist and have been in use for many years. In the UK, one supplier company uses radio networks and GSM communication is also available amongst others.
The monitoring of the network is the main cost and can be done with current technology as evidenced by networks such as those in South Tyrol in Italy but there is a balance between the resolution of the data and the cost of installation and maintenance
In-sewer monitoring of flow and level has been undertaken for many years (photos courtesy of Flowline and Radio Data Networks)
Modelling The Network
All the natural inputs from rainfall, the human inputs from the customer, and the flux of the wastewater passing through the network feed the various models that would be needed to run an intelligent network. This takes the form of two separate models. The first model includes these natural and human inputs and where this has been done in Europe it has typically been done with an artificial neural network model. This then feeds forward to the second model which is the networks models that the water industry has been using for years to design and operate the various sewer systems of the world.
Modelling of rainfall in the Aquasafe project for Lisbon (courtesy of Bentley)
So Are Intelligent Networks Possible?
The big question that still has to be answered is whether or not intelligent wastewater networks are possible? The obvious answer is— of course they are as evidenced by the 20 or more networks in some degree of complexity that already exist. The next question to ask is whether or not they are financially viable. This depends upon the political and economic drivers that exist.
In the UK the situation is certainly right for intelligent wastewater networks. But what shape, form, and complexity they will take and whether the full potential of the technology is available is the biggest question of them all.
Image credit: "Algorithmic Contaminations," derekGavey © 2011, used under an Attribution-ShareAlike 2.0 Generic license: https://creativecommons.org/licenses/by-sa/2.0/