Utilizing conventional primary detection systems, it takes a substantial financial investment in both CAPEX and OPEX to achieve a suitable level of discrimination between accuracy and sensitivity over large perimeters. Additionally, cross-asset utilization is challenging due to the total division of the security and operations departments. Apart from the previously mentioned drawbacks, traditional systems also have a low equipment availability ratio and a sensitive reliance on the installer’s expertise.
With DAS technology, a single electronic piece of equipment can transform tens of kilometers of optical fiber into a set of thousands of microphones for simultaneous listening. After processing the data, real-time interferences from outside sources are identified, their locations are pinpointed to within a few meters, and they are categorized into distinct events. Which of these occurrences should be reported as alerts will depend on the needs of the client.
The DAS solution for long-perimeter protection reduces investment requirements by half when compared to traditional solutions, and because it uses fiber optics as a sensor instead of field-installed electronics, operational costs are essentially nonexistent. Furthermore, the sensor cable has the ability to share fibers with the telecom network, opening up new applications like patrol monitoring.
Ultimately, an almost total degree of availability is made possible by the simplicity with which a redundant network topology can be incorporated and the lack of electronics in the field. Our equipment is less dependent on the installer thanks to artificial intelligence that is integrated into its learning and environment adaptation processes.
Our customers expect efficient alarms for the protection of their assets, but not all assets require the same level of accuracy or the same alarms, so they are the ones who determine which alerts are required and the degree of discrimination needed.
This means that the client must define the alerts he is interested in knowing and establish the degree of discrimination required, i.e. the compromise between accuracy (minimizing false alarms) and sensitivity (minimizing undetected alarms). To achieve this, our engineers define the topology of the optimal solution (number of rings, position and structure of the sensor cable) and the equipment classifies and groups the events detected using algorithms and AI techniques, informing the customer only of the alerts previously defined. This is how we get really efficient alarms that will be informed to the customer by the requested means: our user interface, SMS, email or integration in the customer’s VMS or PSIM.
Listen
The interrogation unit transforms the optical fiber into a sensor with thousands of microphones that constantly listens to what is going on around it.
Analyze
The processing unit analyzes the information received and classifies it into the various basic events perfectly localized with an accuracy of several meters.
Report
The system groups the events and informs the user only of the alarms that are of interest to him according to his needs
We only sell solutions that really work
Our protection of critical infrastructures of the electric power system solution employs distributed acoustic sensing (DAS) technology and allows operators to detect, locate and classify human intrusions (PID) to protected perimeters before these intrusions damage the electrical installation itself and its operation.
The acquired signals of the HDAS are traces that represent the strain variation along the optical fibre. We accumulate the traces in a waterfall plot, that allows us to analyse the vibration both in time and in space. We use digital image processing techniques in order to enhance the significative patterns of interest, and to reduce the noise.
Then, depending on the application, we apply a number of image analysis techniques using algorithms and mathematical models to detect and recognise patterns. For some use cases, we have developed supervised machine learning models, for other use cases “classic” algorithms give very good results.
EN
ES