The objectives of this work package are:

• To adapt automatic neural net control system (ANCS) to the needs of such a large scale process and to demonstrate its benefits
• To implement a new process control system based on artificial neural networks in an existing membrane treatment system in order to optimize the process performance

The following tasks (with deliverables and milestones) are planned:

T2.3.1: Configuration of ANCS

Soft- and hardware will be configured and installed in order to adapt ANCS to the needs of large scale processes. A controller will be delivered in December 2012.

T23.2: Training of the neural net and adjustment of automatic functions of ANCS

ANCS is based on artificial neural networks (ANN), which require a training of the nets in terms of the operational settings. The training is completed in February 2014 A pilot plant equipped with online measurement and remote PLC’s will be operated parallel to the technical plant with original backwash-water at WAG.

T23.3: Demonstration of the feasibility of ANCS

In a six month demonstration, the pilot plant will be operated automatically with ANCS. The results in terms of flux rate, recovery, chemical and energy demand will be compared to the respective results of large technical scale backwash-water treatment plant without ANCS.

T23.4: Implementation of ANCS to the large technical process

After the demonstration of the feasibility, ANCS will be implemented to the large technical scale backwash-water treatment plant. The demonstrations will begin in December 2014 and a report on the demonstration is expected in August 2015.

T23.5: Demonstration of the benefit of ANCS

In a six month operation of ANCS with the large technical scale backwash-water treatment plant, it will be demonstrated that the performance of the system can be improved,  so that the incoming backwash water can be treated without the extension of membrane surface.

To download the deliverables that are already completed, please see the results section.