Periodic Reporting for period 2 - PURE-WATER (Improved Estimation Algorithms for Water Purification and Desalination Systems)
Berichtszeitraum: 2022-07-01 bis 2023-06-30
Sustainable access to drinking water and providing usable water supply for adequate sanitation and also for irrigation based agriculture forms one of the major challenges for the global society in the 21st century. Accordingly, finding respective solutions became part of the Sustainable Development Goals (SDG 6) formulated by the United Nations in 2015. Politics shows broad interest in view of the growing population in many arid and semi-arid regions on the one side, and the loss of fresh water resources related to global warming on the other side. The clear demand for technological approaches focusing on efficient water usage and the exploitation of alternative sources led to the emphasis of this consortium activities on waste water purification as well as seawater and brackish water desalination processes. As a crucial part of a functional water resource management system, besides water distribution itself, the information processing and monitoring of the respective water filtration and refinement procedures are subject to high requirements for accuracy, real-time standards and reliability. From a system engineering point of view, major issues with respect to the corresponding complex underlying physical principles are to gain an appropriate mathematical description of the dynamic behavior combined with an adequate parameterization and knowledge about the internal state conditions of the distributed processes via intelligent sensor data evaluation in spite of external perturbations.
There remain several problems in the supervision and dynamical analysis of modern water purification and desalination processes. Irrespective of the significant social task behind saving secure and accessible water supplies, the following current technical challenges are emphasized by the consortium:
- Several advanced control and estimation concepts for water processes are reported in the literature which do not get applied to actual plants despite satisfactory simulation results. Accordingly, realistically implementable observer schemes (soft sensors) have to be developed with respect to reliable validation standards.
- Often, specific approximations of the underlying complex dynamical processes are made which concentrate on very specific conditions and fail to capture all significant influences. For applications with high safety requirements such as water treatment which directly or indirectly relates to nutrition, a reliable system description is indispensable. Thus, more accurate procedures should be enabled. Compositions of membrane, thermal and biochemical processes, as they occur in water purification, may then be treated in a unified way.
- In membrane based filtration facilities, fouling of membranes is a major challenge (especially RO). This may fully change the plant behavior and lead to complete plugging. Treating the phenomenon as a perturbation, a novel fault isolation concept for membranes shall be developed, combining the competences of the consortium. This includes to classify the membrane condition for identifying fouling early and to adapt the process control or to initiate adjustment actions.
- The overall efficiency of a water resources system is crucial with respect to distribution and energy consumption. A proper information processing unit integrated into the management system builds the basis of an optimal control scheme which may reduce the partially unacceptable high distribution losses and helps increase the applicability through better energy recovery, avoiding uneconomical operating costs.
There remain several problems in the supervision and dynamical analysis of modern water purification and desalination processes. Irrespective of the significant social task behind saving secure and accessible water supplies, the following current technical challenges are emphasized by the consortium:
- Several advanced control and estimation concepts for water processes are reported in the literature which do not get applied to actual plants despite satisfactory simulation results. Accordingly, realistically implementable observer schemes (soft sensors) have to be developed with respect to reliable validation standards.
- Often, specific approximations of the underlying complex dynamical processes are made which concentrate on very specific conditions and fail to capture all significant influences. For applications with high safety requirements such as water treatment which directly or indirectly relates to nutrition, a reliable system description is indispensable. Thus, more accurate procedures should be enabled. Compositions of membrane, thermal and biochemical processes, as they occur in water purification, may then be treated in a unified way.
- In membrane based filtration facilities, fouling of membranes is a major challenge (especially RO). This may fully change the plant behavior and lead to complete plugging. Treating the phenomenon as a perturbation, a novel fault isolation concept for membranes shall be developed, combining the competences of the consortium. This includes to classify the membrane condition for identifying fouling early and to adapt the process control or to initiate adjustment actions.
- The overall efficiency of a water resources system is crucial with respect to distribution and energy consumption. A proper information processing unit integrated into the management system builds the basis of an optimal control scheme which may reduce the partially unacceptable high distribution losses and helps increase the applicability through better energy recovery, avoiding uneconomical operating costs.
Major outcomes of the works performed within the PURE-WATER project are summarized as follows:
(i) Non-asymptotic observer for distributed systems - One of the most important outcomes of this project topic is a Modulating Function (MF) based state estimation framework for systems governed by Partial Differential Equations. Its algebraic nature allows for the reconstruction of the distributed state in fixed time while resulting in an efficient realization scheme. Due to the context of water filtration plants, the class of Advection-Diffusion-Reaction equations is considered as the prime target of application. Extensions towards nonlinear terms from biochemical reactions, coupled equations from combining different water processing steps as well as higher spatial dimensions aiming for more complex setups have been pursued.
(ii) Fault diagnosis scheme - As a central result from combining different estimation steps, a comprehensive supervision and diagnosis architecture was designed using the MF Method. The classical parity space approach has been extended by algebraic estimation steps. Furthermore, application oriented fault detection and isolation was realized for leak localization in water transport systems as well as for detecting membrane fouling inside desalination plant components.
(iii) Smart sensors using exact state observers - The members of the consortium put an emphasis on the concept of distributed and intelligent data processing. To this end, a sensor fusion scheme was developed using real-time capable MF based algorithms that can efficiently be implemented as FIR filters. In addition to that, these state observers guarantee fast convergence while leaving degrees of freedom for adaptation and robustness tuning.
(iv) Validation and testing - A major goal of the cooperation between the application oriented participants within the PURE-WATER project is the development of test benches related to water processes for assessing the performance of the information processing algorithms. First, a simulation environment was built using identification techniques from measurement data that enables digital testing before applying the estimation concepts to real plants. Then, experimental setups were installed for reverse osmosis desalination, water transportation systems and waste water treatment using biochemical reactors.
(v) Networking and training activities - Extensive training of seconded staff members via seminars and personal tutorials led to a strong experience sharing in the fields of non-asymptotic estimators, nonlinear systems theory, experimental setups, sensor design, distributed and efficient data processing, fault diagnosis methods, adaptive control, distributed parameter modeling of water filtration processes, advanced simulation techniques as well as real-time implementation on hardware devices. Planned dissemination activities such as the PURE-WATER workshop in 2021 and a special session at the renowned 2023 IFAC World Congress were pillars of direct contact between the participants and the larger scientific community. Furthermore, members of the consortium demonstrated and used the project results in running bilateral activities.
(i) Non-asymptotic observer for distributed systems - One of the most important outcomes of this project topic is a Modulating Function (MF) based state estimation framework for systems governed by Partial Differential Equations. Its algebraic nature allows for the reconstruction of the distributed state in fixed time while resulting in an efficient realization scheme. Due to the context of water filtration plants, the class of Advection-Diffusion-Reaction equations is considered as the prime target of application. Extensions towards nonlinear terms from biochemical reactions, coupled equations from combining different water processing steps as well as higher spatial dimensions aiming for more complex setups have been pursued.
(ii) Fault diagnosis scheme - As a central result from combining different estimation steps, a comprehensive supervision and diagnosis architecture was designed using the MF Method. The classical parity space approach has been extended by algebraic estimation steps. Furthermore, application oriented fault detection and isolation was realized for leak localization in water transport systems as well as for detecting membrane fouling inside desalination plant components.
(iii) Smart sensors using exact state observers - The members of the consortium put an emphasis on the concept of distributed and intelligent data processing. To this end, a sensor fusion scheme was developed using real-time capable MF based algorithms that can efficiently be implemented as FIR filters. In addition to that, these state observers guarantee fast convergence while leaving degrees of freedom for adaptation and robustness tuning.
(iv) Validation and testing - A major goal of the cooperation between the application oriented participants within the PURE-WATER project is the development of test benches related to water processes for assessing the performance of the information processing algorithms. First, a simulation environment was built using identification techniques from measurement data that enables digital testing before applying the estimation concepts to real plants. Then, experimental setups were installed for reverse osmosis desalination, water transportation systems and waste water treatment using biochemical reactors.
(v) Networking and training activities - Extensive training of seconded staff members via seminars and personal tutorials led to a strong experience sharing in the fields of non-asymptotic estimators, nonlinear systems theory, experimental setups, sensor design, distributed and efficient data processing, fault diagnosis methods, adaptive control, distributed parameter modeling of water filtration processes, advanced simulation techniques as well as real-time implementation on hardware devices. Planned dissemination activities such as the PURE-WATER workshop in 2021 and a special session at the renowned 2023 IFAC World Congress were pillars of direct contact between the participants and the larger scientific community. Furthermore, members of the consortium demonstrated and used the project results in running bilateral activities.
The following central outcomes of the PURE-WATER project with respect to reasearch and innovation are expected:
(1) Advanced parameter and state estimation techniques for hydrodynamic systems utilizing modulating function approaches;
(2) Accurate modelling and simulation of diffusion processes for water purification;
(3) Real-time capable observer design for distributed membrane systems;
(4) Hardware implementation of modulation based estimation algorithms for water applications;
(5) Smart sensor development for integrated information processing in context of water purification;
(6) Software and test bench validation platform of water filtration processes in an experimental environment.
(1) Advanced parameter and state estimation techniques for hydrodynamic systems utilizing modulating function approaches;
(2) Accurate modelling and simulation of diffusion processes for water purification;
(3) Real-time capable observer design for distributed membrane systems;
(4) Hardware implementation of modulation based estimation algorithms for water applications;
(5) Smart sensor development for integrated information processing in context of water purification;
(6) Software and test bench validation platform of water filtration processes in an experimental environment.