Control projects Projects
|
|
| Title |
Large scale issues |
Description:
Irrigation networks can be large with hundreds of kilometers of channels and a large number of gates for regulation of water levels and flows. A typical approach to control design is to start with a small portion of the channel and design a controller for this portion, and then extend the design to incorporate more and more gates and water levels. However, there is no guarantee that a control strategy which works well for a small part of the channel will also work well for the total channel networks. In this project we will investigate which control strategies scales well and if special conditions must be satisfied for the controllers to scale well. In particular we will be interested in how disturbances propagate through a network of irrigation channels, and which conditions, both regarding the controllers and the topology of the irrigation network, must be satisfied in order for disturbances to be attenuated rather than amplified as they propagate through a channel network.
Top
|
|
| Title |
Hierarchical controllers |
Description:
The water level and flow setpoints are inputs to the "low level" controllers for an irrigation channel, e.g. a decentralised PI controller or a multivariable loop shaping controller. There is however, quite a bit of flexibility in choosing these setpoints, particularly the water level setpoints, as a level within say a 100mm band will usually be acceptable. The optimal choice of setpoint values will depend on the current flow, the current and future demand for water, and the setpoint tracking capabilities of the low level controllers. In this project we will develop a controller/algorithm for dynamically changing the water level setpoints such that the water losses are reduced and the level of service to farmers is increased. The developed high level controllers for setpoint selection will take into account the current operating conditions, future demand for water and the available supply of water.
Top
|
|
| Title |
Control under communication constraints |
Description:
Irrigation channels are in remote areas, and all data used for feedback control (water levels, gate positions etc) has to be sent over a radio network. The communication constraints imposed by the radio network limits the achievable control performance. In these projects we will look at several issues regarding the communications constraints. Questions we will investigate are: How do the communication constraints affect the control performance, and How can we design controllers which performs well in the face of communication constraints. Possible designs will include multirate controllers, that is controllers where the different water levels are sampled at different rates according to their importance, and controllers using event driven sampling, that is water levels are only transmitted over the radio network when they have changed with say 3 cm from last time they were transmitted.
Top
|
Fault detection and performance monitoring Projects
|
|
| Title |
Fault detection and isolation |
Description:
Due to the sheer size of irrigation networks and the number of gates and sensors involved, faults are bound to occur. The most typical faults are sensor failure (water levels and gate positions) and actuator failure (motors driving the gates), but other failures such as leaks and alterations to the channel geometry (e.g. channel bank collapse), may also occur. Some failures are easy to detect, but others like a sensor slowly drifting off can be difficult to detect. In other cases it can be easy to detect that something is wrong, but not what has gone wrong, e.g. has the gate got stuck or is it the gate position sensor that is faulty? In this project we will develop algorithms for fault detection and localisation of the fault. The fault detection system will monitor the data from the irrigation channel, extract relevant information and make comparison with the expected behaviour of the irrigation channel. If the difference between what is observed and what is measure is large an alarm is raised.
Top
|
|
| Title |
Performance monitoring |
Description:
In order to operate a network of irrigation channels efficiently, it is important to assess whether the control systems perform satisfactorily. There can be several reasons for non-satisfactory performance; the control design could have been unsatisfactory, the model on which the control was based was not good enough, sensor and actuator failures, unrealistic performance requirements. In this project we will develop a system for monitoring the performance of a control system online, and if unsatisfactory performance is detected, take the necessary steps to improve the performance.
Top
|
System identification Projects
|
|
| Title |
System identification of irrigation channels with siphons |
Description:
In order to achieve better prediction and control and to reduce losses of water, models that capture the relevant dynamics of irrigation channels are needed. Previous works have shown that simple models that describe the dynamics of an open water channel well can be obtained using system identification methods, that is methods which builds models from the observed operational data. In many irrigation channels in Victoria there are siphons, i.e. the water is transported underground through a pipe over some distances before it resurfaces as an open water channel again. Siphons puts restrictions on the flow regimes the channel can operate under, and it is important to understand the dynamic implications of a siphon in order to design a good control system. In this project we will develop system identification models for irrigation channels with siphons. We seek models for the water levels at the very downstream end of the channel reach, immediately downstream of the siphon and immediately upstream of the siphon. Particular emphasis will be put on obtaining models useful for control and prediction.
Top
|
|
| Title |
Approximation of the St. Venant equations |
Description:
Traditionally an irrigation channel is modelled using the St. Venant equations. These equations are hyperbolic partial differential equations and difficult to use for prediction and control. In this project we will use classical approximation techniques (finite element methods, Galerkin approximation) and/or model reduction techniques to find a low order lumped model. In addition to models suitable for control and simulation, we also seek models which are well suited for estimation of the flow through a cross section as well as the total volume of a channel. Understanding the limitations of such approximation methods and establishing their domain of validity is of particular interest.
Top
|
![The University of Melbourne [logo]](/template_images/logos/logo_crest_comb.gif){kind=logo}
CUBIN
CSSIP