With Denmark set to host the IWA Congress later this year, utility BlueKolding offers a case study from the country on how digital technology has enabled process improvements and multiple benefits. Nikolaj Selmer Mølbye, Ole Mark, Camilla Ravn and Thomas Faarbæk explain.
The early days of real-time control
The first time most of us heard about real-time control was during the late 1980s. Then, the International Conference on Urban Storm Drainage and the Hydroinformatics conferences were buzzing with excitement about new software for personal computers that could simulate the full set of Saint-Venant equations (one-dimensional equations commonly used to model transient open-channel flow and surface runoff) for 1D pipe systems. When real-time control was added, the whole water engineering society cheered as if they had found the Holy Grail. Computer simulations demonstrated a great potential for utilisation of free capacity in sewers, which could reduce overflow and construction costs significantly.
Despite overwhelming results from computer simulations, however, nothing happened in the real world. The reason was simply a lack of the physical means and internet of things (IoT) to make it happen. There were also difficulties convincing operators that the system would work in the case of a power failure or a failure of physical devices. One of the few real-time control systems that was implemented during the 1990s was in Germany, and it had six physical fallback structures in case of power failure or other failures.
Around the millennium, things suddenly changed. Leading water professors joined forces and wrote a paper making a unanimous statement about the way forward (Rauch et al, 2002). The internet, the mobile phone, Wi-Fi and real-time information systems were being used in many places. There was no real operational risk associated with these systems, as they did not directly control anything at all; but the flow of real-time data and forecasts gave the operator the means to interact more intelligently with the sewer systems and the wastewater treatment plants. From this, small, real-time control implementations emerged, later to be followed by a number of real-time control systems working together in a full sewer system.
A digital journey took off in BlueKolding
BlueKolding, a utility company covering the entire municipality of Kolding, located in Jutland, Denmark, has been on a digital real-time control and optimisation journey since 2006. This began with the wastewater treatment plant, which implemented digital real-time technology on a three-year, performance guaranteed, ‘no cure, no pay’ contract. If the new solution didn’t provide savings and a better performance, the consultant would not be paid. The reason for the nature of the contract was that, at this stage in the technology’s development, there was little evidence of substantial monetary gains to be made from real-time control implementations. One challenge in the quantification of gains and benefits from the new system was that the baseline of the operation was constantly shifting, as physical changes in the catchment and operational upgrades were implemented at the wastewater treatment plant during this period.
At the time, the combined sewer network was operated in a rather static way. The pumping stations were used to push the maximum amount of water forward to the wastewater treatment plant, neglecting that water might be sent to a fully loaded wastewater treatment plant, which then bypassed the sewage into an overflow. This ‘pass forward strategy’ was not optimal, as it didn’t use available volume in the sewer system.
In 2011, the first step towards integrated control and optimisation of the sewer system and the wastewater treatment plant was initiated. The objective of the implementation was to reduce combined sewer overflows (CSO), while reducing the needed expansion of basin volume from 6000 m3 to just 3000 m3 to comply with legislation and maintain the level of service. The construction cost of 3000 m3 concrete storage structures to increase capacity would have been in the order of €4.5-6 million. The implementation of real-time control equipment presented huge cost savings.
BlueKolding applied global, real-time control in 16 storage tanks in the combined sewer system. Since then, three additional basins have been included in the control. Over the past couple of years, the setup has been upgraded to a model predictive control (MPC), where controls can anticipate expected capacity of the sewers two hours ahead, enabling proactive monitoring and management.
Real-time input, forecast and control
A real-time inflow forecast was developed for the storage tanks in the sewer system and the inflow to the wastewater treatment plant. The forecasts are based on real-time radar forecasts of rain – i.e. a nowcast with a 10-minute time step and two-hour forecast horizon. The inflow forecasts are based on a lumped conceptual stochastic rainfall-runoff model. The forecast at the inflow to the wastewater treatment plant is updated in real time using a Kalman Filter (an estimation algorithm that can produce estimates of hidden variables and give a prediction of the future conditions of the system) based on flow measurements. The outcome was forecasts of available hydraulic capacity in the storage tanks and inflow to the treatment plant.
A model predictive control (MPC) system was set up. This consisted of:
- A simplified surrogate model containing the storage tanks
- The catchment upstream of the storage tanks
- The sewers between the storage tanks
- The wastewater treatment plant, as the downstream boundary condition
The goal of the MPC in the sewer system was to perform coordinated monitoring of the storage tank volumes to minimise the total combined sewer overflows. The control strategy was based on a dynamic risk assessment (Vezzaro and Grum, 2012 and Vezzaro et al, 2013), where the risk of overflow is calculated for every storage tank, every two minutes, based on:
- Actual storage tank filling and volume
- Forecasted runoff volume to the storage tank from short term rainfall forecast
- The relative cost for an overflow (depending on recipient sensitivity)
- The maximum possible hydraulic load to the downstream storage tank
- The minimum possible runoff to the downstream catchment
- The actual hydraulic capacity of the downstream wastewater treatment plant
The optimal solution – i.e. predicted optimal flows in connections between storage tanks – takes into account that some storage tanks are more ‘expensive’ to use than others and storage tanks are more ‘expensive’ when the available volume decreases. This will locate any unavoidable combined sewer overflows to the ‘least costly’ locations, these being the most environmentally robust ones.
Based on the control strategy, all the storage tanks (with their local cost functions and the downstream capacity of the wastewater treatment plant) are taken into account, when the combined sewer overflow volumes are minimised.
In the coming months, the final software and hardware connections will be set up and the integrated real-time control equalising the inflow to the wastewater treatment plant will be in operation. This will be followed by the system’s performance being monitored and optimised.
How flow forecast can improve treatment of a storm water event
The software (Hubgrade Performance) optimises the hydraulic capacity of the wastewater treatment plant during stormwater conditions. Reactive mode is based on real-time measurements of the inlet flow, whereas the proactive mode based on flow forecast will increase the hydraulic capacity in advance of the arrival of the stormwater and so avoid or reduce non-compliant discharges.
BlueKolding is taking another step into the intelligent use of data by adding the consumption and the cost of electricity to the optimisation schemes of the sewerage system and the wastewater treatment plant. The idea behind this new ‘SmartGrid’ system is to withhold the wastewater up to 24 hours so that treatment mainly takes place when the electricity tariff is low. As tariffs on the free electricity market fluctuate substantially throughout the day, purchasing prudently brings savings to energy costs. Per Holm, chief executive officer of BlueKolding, says: “We work continuously to energy-streamline our processes. Setting up an advanced SmartGrid system with close interaction between the treatment of wastewater and electricity market tariffs is a huge stepÊforward for us.”
Further, the utility will sell balancing services to the electricity grid with a short response time via up or down adjustments of the energy consumption and production. The need for alternative balancing options increases as more fluctuating renewable energy is produced. This will optimise the purchase of energy and the selling of the energy they produce, without compromising the daily operation of the sewer system and the wastewater treatment plant.
A very important part of the energy optimisation scheme is the reduction of carbon dioxide emissions and the impact on the environment. In Denmark, wind power covers 50% of electricity consumption, so when the wind blows, energy prices go down. This means that the wastewater treatment plant can run on cheap power and its operation is more environmentally friendly.
Over many years, BlueKolding has prioritised and applied software solutions for capacity extension and operations optimisation for the entire sewerage system, from the sewer network into the wastewater treatment plant. The aim was to maximise the use of the facilities and ensure an optimum and compliant operation under all conditions – i.e. for dry and wet weather. At the same time, BlueKolding has fulfilled its aim to provide staff with the most efficient tools to ensure a good overview of all its systems.
Before implementation of the integrated real-time control system, the wastewater treatment plant and network operators didn’t establish any connection between the two systems. Today, they are working as one optimised unit, even adding energy consumption and production to the handling of the wastewater systems.
Rauch, W, Bertrand-Krajewski, J-L, Krebs, P, Mark, O, Schilling, W, Schutze, M and Vanrolleghem, P A (2002) Mathematical Modelling of Integrated Urban Drainage Systems, Water Science and Technology 45(3): pp 81-94.
BlueGrid (2017) development project 2017-2020 funded by EUDP, the Danish Energy Agency, which is part of the Ministry of Energy, Utilities and Climate.
Nielsen, N H, Ravn, C, Mølbye, N (2010) Implementation and design of a RTC strategy in the sewage system in Kolding, Denmark, NOVATECH 2010, Lyon, France, 2010.
A utility company covering the municipality of Kolding in Jutland, Denmark, BlueKolding is inspired by the concept of Blue Economy. It works to find new ways of exploiting the resources in wastewater and improving the processes for cleaning it. The municipality of Kolding manages the treatment of 15 million m3 of wastewater for the city of Kolding and its surrounding area every year, with 12 million m3 of wastewater treated at the Agtrup central wastewater treatment plant.
BlueKolding: key dates
2006: Installation of real-time control (an early version of Hubgrade) at main wastewater treatment plant
2011-13: Expanded to sewer network and three additional wastewater treatment plants
2014-16: Sewer expansion and deployment of spot pricing
2017-21: Sewer Model Predictive Control, demand/response control and weather forecasting implemented
The benefits of digital real-time control
- Compliant effluent quality
- Total Nitrogen in the effluent reduced by 27%
- Energy consumption reduced by 23%
- Chemical consumption reduced by 46%
- Hydraulic capacity of the entire wastewater system increased by 80%
- Combined sewer overflows (CSOs) reduced by 83%
- Very high level of data security
- Provision of balancing to the demand-response electricity market