Water Online

MAY 2014

Water Online the Magazine gives Water & Wastewater Engineers and end-users a venue to find project solutions and source valuable product information. We aim to educate the engineering and operations community on important issues and trends.

Issue link: http://wateronline.epubxp.com/i/303843

Contents of this Issue

Navigation

Page 28 of 38

Tutorial wateronline.com ■ Water Online The Magazine 24 D eveloped from a water treatment plant (WTP) in Brazil, a new study aimed to evaluate the use of analytical instrumentation in coagulation process control and to develop models that, once associated with control technique, could be applied to carry out the coagulant dosage automatic control. Once the more suitable instruments for this WTP's water quality were defined, a database was built. Applying non- parametric methods through cross correlation, it was possible to evaluate the relationship among the process variables and identify the most relevant variables for the development of the models. Once the variables were defined, a variety of models were developed using an artificial neural network (ANN) and a neuro-fuzzy network (NFN) of distinct topology. The goal was to come up with a representative model of the operator's behavior when controlling the coagulant dosage, and a model to predict the settled water turbidity. These models were implemented in a program in C++ language. The program was connected to a SCADA system by applying the plant's automation resources, which are what made it possible to acquire the values of the input variables and to record the output of each model online. Analytical Instrumentation For Water Features Measurement The analytical instruments supply data in real time for a great variety of needs in WTPs. Before any modeling technique or process control system could be implemented, we first had to identify the instruments we needed to assess its response, choose the measurement technique best suited to installation conditions, define the needs for processing the output signal, and decide on data validation rules. Considering the errors from the measurement system, we determined that using the raw data without any kind of processing could damage the efficiency and the reliability of the model and the control system. Methodology Used For The Development Of The Work The development of systems that could identify the operator's behavior and predict the turbidity of the settled water involved creating models from a database collected throughout one year. For the construction of these models we used a technique of parametric modeling based on fuzzy logic and an artificial neural networks approach. In the WTP studied, the operator performs the analysis of turbidity of the settled water to evaluate the efficiency of dosage adjustment. The operator's behavior while controlling the coagulation system is similar to an open-loop control because the control action is defined only by the input parameters of the system. In this case this behavior could be represented as shown in Figure 1: In order to close the loop, it is necessary to obtain a prediction of the system output because there is a dead time between the control action and the effective reaction of the system. The alternatives for the identification of the systems involved in the coagulation process (coagulant dosage Figure 1: Identification of the operator's behavior Model Behavior: Evaluating Instrumentation And Control In The Coagulation Process An electrical engineer does the math on coagulation process control, using computational modeling to determine best practices. By Selma Parreira Capanema Figure 2: Prediction of settled water turbidity Turbidity Turbidity Turbidity Turbidity Turbidity 2 4 _ V E R T _ 0 5 1 4 C l e a n w a t e r _ C o p a s a _ D G . i n d d 1 24_VERT_0514 Cleanwater_Copasa_DG.indd 1 4 / 2 2 / 2 0 1 4 2 : 0 1 : 3 6 P M 4/22/2014 2:01:36 PM

Articles in this issue

Links on this page

Archives of this issue

view archives of Water Online - MAY 2014