A gain scheduling approach to improve pressure control in water distribution networks
Introduction
Real Time Control (RTC) of service pressure plays a fundamental role in the context of Water Distribution Networks (WDNs) management, allowing for leakage reduction (Creaco and Walski, 2017, Farley and Trow, 2003), pipe burst abatement (Creaco and Walski, 2017, Lambert et al., 2013, Thornton and Lambert, 2006) and overall infrastructure life extension. The WDN is first subdivided in homogeneous pressure zones (Walski et al., 0000). On this basis, hierarchical control schemes are developed, with high-level optimal controllers defining the pressure setpoints for the different pressure zones, according to an economic cost–benefit evaluation over the whole network, and RTC controllers working as low-level controllers to ensure regulation to the assigned setpoint (Cembrano et al., 2000, Grosso, Maestre, et al., 2014, Grosso, Ocampo-Martínez, et al., 2014, Grosso et al., 2017, Ocampo-Martinez et al., 2012, Ocampo-Martinez et al., 2013, Toro et al., 2011). Recently, the spread of wired water distribution networks, promoted by the Water 4.0 approach, is attracting new interest on the design of low level controllers. In fact, in wired WDNs, sensors and actuators are connected by wire to control units: this allows developing new control approaches working with high sampling rates. Some recent works (Fontana, Giugni, Glielmo, Marini, and Verrilli, 2017, Fontana, Giugni, Glielmo, Marini, and Zollo, 2017, Galuppini et al., 2019) try to investigate this approach and compare it to more traditional control strategies (Campisano et al., 2009, Campisano et al., 2016, Campisano et al., 2011, Creaco et al., 2018, Creaco and Franchini, 2013), to understand how to exploit the new communication possibilities at best. The design of such control algorithms results particularly challenging, due to the complexity of the nonlinear system under control, and the topic of guaranteed closed-loop stability, which was previously disregarded by the hydraulic community, is getting more and more attention (Galuppini et al., 2020, Janus and Ulanicki, 2017, Janus and Ulanicki, 2018). In particular, Fontana, Giugni, Glielmo, Marini, and Verrilli (2017) and Galuppini et al. (2019) propose a model-based frequency domain approach, with regulators based on a linear, local model of the system dynamics around the working point. The issue of stability is faced by providing robustness margins against gain and phase variations. Galuppini et al. (2020) further investigates the problem of stability, and demonstrates that a poor description of the high frequency behaviour of the system, together with the strong gain nonlinearity affecting the Pressure Control Valves (PCVs) exploited as actuators, may result in highly overestimated robustness margins and, eventually, in closed-loop instability as the process moves from its nominal working point. The aim of this paper is to improve the design methodology originally developed in Galuppini et al. (2019), to explicitly account and compensate for the main static nonlinearities characterising the process under control. This is achieved by means of a proper choice of the control variable (inspired by Creaco, 2017), a nonlinearity inversion block, and a gain scheduling approach. The gain scheduling policy is further extended to better manage the regulation error vs cost of control trade-off at different operating points of the system. The correctness of the approach is verified by means of closed-loop simulations, performed on a detailed model of two different WDN topologies and different user demand scenarios. This paper is organised as follows: Section 2 describes the different case studies adopted for simulations while Section 3 reports the details of their numerical modelling; Section 4 discusses the control methodology and Section 5 its application to the case studies. Finally, a detailed discussion of the results, including a comparison with similar algorithms, is given in Section 6, while the conclusions are reported in Section 7.
Section snippets
Case studies
The control algorithms developed in this work are tested over two different WDN topologies, characterised by very different dynamic behaviours. For sake of comparison, reference is made to the same case studies adopted in Galuppini et al., 2019, Galuppini et al., 2020, which can be consulted for further details. For both case studies, the control goal is the regulation to the setpoint of the measured pressure at the critical node, in presence of process disturbances generated by the time
Numerical model
Unsteady flow modelling (Creaco et al., 2017, Streeter et al., 1998) allows a proper analysis of the hydraulic transients due to rapid nodal demand and/or valve setting variations, and is therefore adopted in this work to develop a simulated environment for the WDNs considered as case studies.
For a generic pipe of a WDN, the one-dimensional unsteady flow equations take the form: where and are the pressure head and the flow discharge along
Control algorithm design methodology
This section is devoted to the description of the control algorithm design methodology. In particular, the design consists of three main phases:
- •
Nominal design of the regulator.
- •
Definition of a gain scheduling policy to compensate for process gain nonlinearities.
- •
Definition of an additional gain scheduling policy to balance between cost of control and regulation error at different working points.
Each step of the procedure is described in detail in the following subsections.
Application
This Section is devoted to the application of the methodology to the two case studies introduced in Section 2. To quantify and compare the performances of the different control schemes, two metrics are introduced. All signals are sampled with a 1 s sampling time. Let be the current discrete-time instant. Let be the measured pressure, be the pressure setpoint, be the valve closure and be the variation of the valve closure over a single sampling time. Let
Discussion of results
The control algorithms described in this work was successfully tested in two case studies characterised by different dynamic behaviours. The dynamics of Case Study A, a water distribution system with a simple topology, is in fact dominated by the water hammer effect. On the contrary, in case of Case Study B, the dynamics is mainly determined by its complex topology. In both cases the control algorithm provided satisfactory results all over a wide range on operating conditions, with
Conclusion
This paper proposed a frequency domain control design with a gain scheduling approach to improve real time pressure control in water distribution networks. The algorithm was extensively tested with simulations on a numerical model of the WDN. In particular, two different topologies and different demand profiles were considered, to assess its performances over a wide range of situations. The algorithm delivered satisfactory results, and detailed comparison with a simpler class of frequency
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (36)
- et al.
Optimal control of a water distribution network in a supervisory control system
Control Engineering Practice
(2000) - et al.
Service pressure regulation in water distribution networks
Control Engineering Practice
(2019) - et al.
On the assessment of tree-based and chance-constrained predictive control approaches applied to drinking water networks
IFAC Proceedings Volumes
(2014) - et al.
Chance-constrained model predictive control for drinking water networks
Journal of Process Control
(2014) A pragmatic approach to robust gain scheduling
IFAC Proceedings Volumes
(2012)- et al.
Tuning of predictive controllers for drinking water networked systems
IFAC Proceedings Volumes
(2011) - et al.
On the lpv control design and its applications to some classes of dynamical systems
- et al.
Rtc of valves for leakage reduction in water supply networks
Journal of Water Resources Planning and Management
(2009) - et al.
Field-oriented methodology for real-time pressure control to reduce leakage in water distribution networks
Journal of Water Resources Planning and Management
(2016) - et al.
Calibration of proportional controllers for the rtc of pressures to reduce leakage in water distribution networks
Journal of Water Resources Planning and Management
(2011)