Licensed by Decision Maker
ADVANCED DATA-ANALYTICS AND EVENT DETECTION SOFTWARE
Smart INSIGHTS is a data analytics software that helps water utility managers, engineers and operators make decisions faster with fewer mistakes, ensuring public safety, reducing costs and improving service.
Smart INSIGHTS analyzes data gathered from water quality instrumentation in the field. The software uses machine learning technology (ML) and proprietary algorithms to analyze historical data to provide valuable alerts about water quality changes and potential pollution events, as well as operational insights such as functionality of instrumentation and communication systems and on-going calibration requirements.
Smart INSIGHTS helps to overcome the challenge of assessing and interpreting the large volume of water quality information that is collected in the field, as well as validating the quality of the information and performance of analyzers in the field.
WHAT DO WE DO AND HOW WE DO IT
Smart INSIGHTS software includes more than 20 different algorithms we call DETECTORS. Some DETECTORS are well-known and others are proprietary and patent protected. The goal of the DETECTORS is to detect specific deviations in behavior [abnormalities] of water quality data that typically cannot be identified by simply “watching the data” or by using conventional SCADA tools.
An example of a DETECTOR is an algorithm that detects if a variable is beyond its statistical limits. Such a Detector “learns” the statistical limits of each variable; when actual data violates these limits Smart INSIGHTS will generate an alert. All DETECTORS have a self-learning capability and a default setup to shorten the system’s setup time and startup.
TYPE OF DETECTORS WE USE
Learning statistical limits; Rare dynamic change; Abnormal rate of change; Long trends; Data with low quality; I/O cards problems; Rare combinations; Similiraty to hazard known combination; Violations of rules; etc.
Smart INSIGHTS can be set up quickly to analyze any continuous water quality data that was generated by field sensors such as chlorine, turbidity, pH, conductivity, temperature, ORP, ammonium, nitrates, chloramines, TOC, and spectrometer sensors, as well as operational parameters.
SMART INSIGHTS GENERATES 4 LEVELS OF ALERTS:
Level 1 – Single sensor violation – Alert about violation of single variable limits – high and low limits, statistical limits, fixed data, long trends, I/O cards malfunctions, rare dynamic change, etc.
Level 2 – Probability for abnormal event – Alert about rare combination between several variables that is not known from the past or was not classified in the past.
Level 3 – Similarity for known hazard event – Alert about specific combination between several variables that was already classified in the past to hazard event.
Level 4 – Rules violation – Alert about specific rules that were created by the engineer to detect specific relations between variables.
WATER NETWORK SPATIAL MODELING
Using statistical model that correlates patterns of water quality fluctuations between stations along the water network, Smart INSIGHTS provides additional information about spatial behavior of both physical and chemical properties of the water network. The spatial model enables to estimate the changes of water quality between stations, as well as learning the distribution and flow of pollution through the water network. This information can be used for real-time decision-making, and for future engineering planning.
TRAINING THE MODEL AND DEMO OPTIONS
Smart INSIGHTS can easily be trained using historical data from scada systems or other data sources. Every CSV file that contains water quality data can be easily loaded to the software to generate a working model. Generating a new model is simple and quick due to self-learning capabilities and default setup. Training a new model enables the customer to profit from Smart INSIGHTS from the very beginning, shortening the learning period for the designated water system.