ASDECO Project

Editor's Notes

• #2: During these days we have seen many presentations about the on plant water treatment processes on this presentation will move the spatial scale to the ocean in order to analyze the impacts produced by brine discharges into the sea.
• #3: This presentation summarizes the results being obtained in the ASDECO project. ASDECO is the acronym of Automated System for Desalination Dilution Control. The Project has been subsidized by the National Program for Environmental Science and Technology of the Spanish R&D Plan. This program has been coordinated by the Spanish Ministry of Environment. The project is being developed between 2007 and 2010. Tecnoma is the main project coordinator.
• #4: The purpose of this slide is to show the prior circumstances that motivated this research project. Firstly because of the high development of the desalination industry in the Mediterranean Spanish Coastline. Currently there are about 30 desalination plants in different construction phases. All of them will improve agricultural and urban water demand guarantees, with an annual volume close to 700 hm3. The main problem of all these developments is the presence of sensitive ecosystems and protected species. The administrative processes (as the Environmental Impact Assessment and discharge authorization) impose extreme conditions in terms of environmental regulations controlling discharges into the sea. These regulations used the phanerogam seagrass as an indicator species, because of its role in the ecosystem and its sensitivity to high salinity thresholds. The EIAS have established an alarm system for salinity thresholds as a protection measure. The transgression of the salinity threshold over time (mainly 1 week) requires the gradual shutdown of the desalination plant.
• #5: The objective of this project is to combine real-time monitoring with predictive models. Thus one can relate the degree of diffusion of the brine plume in terms of ocean-climatic conditions. This prototype will also allow us to carry out the supervision and monitoring of the discharge, informing the public and the environmental authorities. The project is defined as a modular one. The Instrumental and information Systems are the core system. The obtained information will be collected and categorized according to their quality and analyzed. This information will feed the alarm system whose function is to monitor real-time performance of the threshold values of salinity in the defined vulnerable areas. Additionally, the prototype has a mathematical model that acts as a forecasting system. All these systems compouse a support system for decision making that incorporate adaptive management of the discharge of the desalination plants.
• #6: The prototype is being tested in the Alicante Channel Desalination Plant, which have a combined treatment capacity of 130,000 m3/day. The outfall discharge is done on water surface and it is diluted previously. This dilution is done by sea water. The desalination plant has the necessary infrastructure to provide a 1:6 dilution ratio; the outfall discharge and its salinity depend on that dilution ratio; it is usually a discharge flow between 7-10 m3/s. The brine salinity is about 67.6 psu and the average salinity of seawater is 37.5 psu. On the map the green area represents the location of the Seagrass of Posidonia in front of the plant discharge.
• #7: The ASDECO conceptual framework is composed by 5 main modules: The first one is based on data adquisition and climatic prediction. The second one analyze the information on real time. The other two systems are poited on the alarm generation and the adaptive management.
• #8: Lets talk now about the instrumentation system
• #9: The main variables monitored by the intrumentation system are Directional Waves Wind Air temperature & Pressure Sea level Sea bed and profile currents Salinity And other variables related with water quality
• #10: The variable are shared in three main locations: The discharge point that will control the discharge salinity, The bouy 1 that will control mainly the sea conditions and the sea bed salinity And the second buyo than will control the sea bed salinity as a varible that integrates the alarm system at the vulnerable area.
• #11: On this slide we can see a profile view showing the two bouys and the location of concentration ctds. This ctds will provide extra information about plume distribution during the model calibration.
• #12: The identified instrumentation needs an adaptation to be applied in the pilot plant. As you can see is a superficial discharge that is controlled by a CTD.
• #13: This schema represents the sensor integration in near filed buoy. The picture shows the installed one.
• #14: The bouy two is more simple than the first one becuase is based mainly on the sea bed salinity control. And its focus on the alarm generation on seagrass perimeter.
• #15: The information adquired by bouy are complemented by spatial distributed campaigns during the model calibration phase.
• #16: As the spatial campaings extra CTDS are placed on plume in order to improve its distribution during model calibration.
• #17: The information system uses a SQL server connected to an XML metadata database and to a Geodatabase server. The connection of these database systems will permit the quality preservation of data and also associate each time series with its related geographical information. The system allows centralization of brine discharge environmental information.
• #18: The metadata database system uses an xml standards in order preserve data traceability. This is its main funtion. The sysytem attaches to each time serie the information regarding its quality. As the accuracy, maintanance activities, and also about its calibration and validation.
• #19: A client desktop has been developed as an on-Plant follow up monitor. You can see several screenshots that shown the graphical capacities and the software interface. Its design is pointed to permits the system use and maintenance by non-specialized users. This centralization tool will be useful for Desalinization Plant Control to the environmental authorities.
• #20: The implementation of the information and alarm system will allow the continuous monitoring of the threshold salinity values. The monitoring will be based on mobile exceedance percentages applied at time intervals (daily and weekly). In this graph the green line represents the hourly measures of salinity between Feb and Dec 2007. The columns show the persistence, the red ones represent the values higher than 25%. The considered salinity threshold value is 38.5 psu. We can compare the daily and weekly time periods (left and right graphs). The daily alarms show increased noise effects due to the daily salinity variance. The use of this mobile percentage will allow us to define Alarm status levels. Low percentages will define a dilution status of “no effect”. The values bigger than 25% will establish the temporary closure of the Plant. Intermediate levels will define a “caution” status that will need a detailed follow up to avoid negative impacts. These alarm levels will be transmited throught SMS or email to the Plant managers. The alarms will be centralized by the information system to the public administration. These alarm levels are based on data received in real time. The brine plume inertia over the sea bed will cause difficulties in avoiding negative effects. However changes in discharge based on this data motivate the development of predictive models and combining it with the IS we can facilitate a more adaptive management. These predictive models will allow us to evaluate future diffusion of salinity and its accumulation over time. La implementación del sistema de información y alarma permitirá el seguimiento automatizado de la superación de los umbrales de salinidad en el tiempo. Este seguimiento se realiza mediante el uso de percentiles de superación aplicados a dos escalas temporales (diaria y semanal). Como se ha explicado en la introducción la administración española a impuesto umbrales de salinidad en función de la especie que varían entre 38.5-40 psu. Y unos porcentajes de excedencia que varían entre el 25% y el 5%. Las gráficas muestran la serie temporal de salinidad entre Febrero y Diciembre de 2007 (Linea verde). También muestra el porcentaje movil de excedencia aplicado sobre mediciones horarias. Las columna rojas corresponden a aquellos que supera el valor el umbral del 25 %. El umbral de salinidad evaluado en esta aplicación corresponde a 38.5 psu. En la comparación de las dos gráficas apreciamos el mayor ruido producido por la acumulación diaria. La expresión de estas alarmas como porcentajes móviles permite realizar un seguimiento continuo del nivel alarma. Analizando el valor del procentaje es posible fijar niveles de alarma: una primera propuesta podría por ejemplo asociar a estados de no afección cuando el percentil es practicamente cero. Un segundo nivel de precaución estaría fijado en valores por ejemplo entre el 5 y el 25%. El tercer nivel que activaría protocolos de parada establecidos en la DIA. Las alarmas son transmitidas vía SMS y email a una lista de contactos. También son monitorizadas en la propia planta y de forma centralizada en la administración pública.i Este nivel de seguimiento propuesto implica un mayor alcance que las condiciones impuestas en la autorización de vertido, aunque dadas la inercia de la pluma de salmuera el seguimiento en tiempo real no evitaría la afección a las zonas vulnerables. Esto marca la importancia de desarrollar las herramientas predictivas enlazadas con la monitorización en tiempo real. Esta predicción debe ayudar a identificar futuros niveles de alarma de forma que podamos aplicar de forma una gestión adaptativa del vertido.
• #21: After give an overview about the information and instrumentation system. we are gona talk about the decision support system. This system is composed by several tools as the predictive module that combines the neuronal networks and the oceanographic modeling. The predictive alarm will interpretate the prediction. And the module pointed in the adaptive management evaluation that will recommend to managers the discharge alternative to avoid negative effects.
• #22: This phase describes the predictive model development that will be integrated within a Decision Support System. This tool has been created combining different mathematical tools such as Multivariate Statistics, Variable normalization and Fuzzy logic combined with Neuronal networks (ANFI). The predictive model allows the integration of adaptive management of the brine discharge in two main ways: The first one consists of the identification of discharge and near field salinity values that can be transported to the vulnerable areas according to the ocean conditions at that time. The second one will evaluate changes in sea energy status and the incidence of saline brine dispersal. Identification of energy changes will be provided by the Spanish Sea Port Authority that make predictions between 6 and 72 h from their own ocean meteorological models. This predictive system will allow us to perform a risk evaluation of outer-limit threshold values. Using these predictions the DSS will activate the implementation of preventive measures such as the increase of pre-discharge dilution, opening of new outlets or the increase of outflow velocity on diffusers.
• #23: The predictive system calibration using acquired data between Feb and Dec 2007 has produce good results. The graph shows a comparison between measured and predicted data on a salinity sensor. The salinity variation has been represented in the model by the incorporation of only a few variables such as the number of the day for that year, which indicates the specific season, the excess salt load discharged and the maximum wave height that includes in the model the possible effect of weather and ocean conditions. Nowadays we are improving the model adjustment by the incorporation of new variables that increase the daily variance observed on measured data. First results shows that variation in sea water levels will provide a better fit than now.
• #24: Continuing with the results, shown in the graph on the right we can see the correlation between measured and predicted data . This correlation shows the best fit for salinity values higher than 38 psu. The bigger dispersal for lower values were influenced by the low salinity happening between august and October 2007. The graph on the left side shows the correlation between salinity values and maximum wave height. As we expected the higher sea energy status increased the saline brine dispersal. These results show that the model predictions are fulfilled.
• #25: As has been shown earlier in the instrumental alarms graphs the model can use these tools to evaluate alarms. This slide will compare the alarms produced with our predictive model with the real data in order to provide a certainty indication. The left side shows the alarms produced by measured data while on the right we can see the predicted ones. The comparison has been made on daily and weekly time intervals. Just remember that the graphs uses a mobile exceedance percentage. The weekly accumulation looks more appropriate due to less dependence on the daily variance that can create quick alarm activation and deactivation in daily time intervals. The measured and predicted results show similar disruption events. The lower variance inherent in models produces differences on the smaller alarms values. These differences will be improved with the ongoing incorporation of new variables.
• #27: Finally I would like to invite you to the diffusion workshop to be held in Valencia on 5-6 October. The workshop will join experts on brine discharge modeling and is organized to disseminate the project results and to provide a forum for discussion on the best techniques for minimizing the discharge of desalination. More information, and the registration form can be found on the project webpage. At last, only to say the interest of ASDECO partners are open to share data and reports as well to future collaborations. I would like to thank the organizers for the invitation to this workshop. Thank you very much for your attention.