Event detection method

Abstract

The invention discloses a detection method and a system for carrying out the method. The method comprises the followings steps: a) measuring at least two variables which are detected by at least one detection system that is arranged at a decentralised measuring point; b) transferring the measurement values of the at least two variables from the detection system to a data processing unit which is designed to execute a model for evaluating the current environmental condition and / or for predicting the possibility of an environmental event occurring. The method is characterised in that the correlation of the at least two variables, which is based on cross-correlation, is measured.

Description

[0001] Event detection procedure

[0002] The invention relates to a method for detecting an environmental event according to the preamble of claim 1 and a system for carrying out the method.

[0003] Early warning systems provide information about areas vulnerable to specific environmental events such as natural disasters, such as landslides, floods, and earthquakes; unauthorized discharge of raw or inadequately treated wastewater into a prohibited discharge zone; and illegal water abstraction. These environmental events can be life-threatening or environmentally harmful, so accurate forecasting and warning are important.

[0004] Early warning systems often include sensors, event detection, and decision-making subsystems for early identification of hazards. Historical data and predictive models are often used to detect such events. For example, to predict a flood event, precipitation data is continuously fed into a model. The simulation results are then displayed as discharge and water level forecasts at predefined desired locations.

[0005] Typically, these event detection methods are based on a single measured variable. Although some methods are based on more than one variable, they often do not provide the relationship between the variables. Therefore, their warning and prediction / forecasting capabilities are limited.

[0006] The object of the present invention is to provide a more detailed and precise prediction method for environmental events.

[0007] The object is achieved by the method specified in independent claim 1 and the system for carrying out the method described in claim 10.

[0008] With regard to the method, the problem is solved with the following steps: a) measuring at least two variables detected by at least one detection system arranged at a decentralized measuring point; b) transmitting the measured values ​​of the at least two variables from the detection system to a data processing unit configured to execute a model for evaluating the current environmental state and / or predicting the possibility of an environmental event occurring. The method is characterized in that the correlation of the at least two variables is measured, which is based on cross-correlation. Cross-correlation is a measure of the similarity of two data series depending on the displacement relative to each other, which is frequently used in signal processing.

[0009] Advantageously, the at least two variables are measured continuously in a predefined measurement interval.

[0010] In one embodiment, the detection system comprises at least one of: a sensor, a measuring device, a geographic information system (GIS) model, a web service, and / or a database.

[0011] In one embodiment, the data processing unit is part of the recognition system.

[0012] In one embodiment, the data processing unit is located in a central station for remote monitoring to be protected from damage during the environmental event.

[0013] In one embodiment, the model comprises a simulation based on artificial intelligence techniques.

[0014] In one embodiment, the model is trained based on historical measurements of the at least two variables.

[0015] In one embodiment, the at least two variables comprise at least two of soil moisture, water level, groundwater head (or pressure), water temperature, pH, conductivity, turbidity, flow, oxygen content, nitrite content, soil bearing capacity, and / or other ion concentration data, geological data, geographical data, climate data, weather data, or weather forecast data. In one embodiment, the environmental event comprises at least one of a flood, a landslide, an unauthorized discharge / withdrawal of water, and / or an upstream chemical spill.

[0016] This object is further achieved by a system comprising at least one detection system arranged at a decentralized measuring point; and a data processing unit configured to execute a model for evaluating the current environmental state and / or for predicting the possibility of an environmental event occurring in order to carry out the method described above.

[0017] This is explained in more detail with reference to the following figure.

[0018] Figure 1 shows an embodiment of the claimed measuring system. In Figure 1, identical features are marked with the same reference numerals.

[0019] With reference to Figure 1, the method is used to detect / predict a flood event, a landslide event, or other natural disasters. A detection system 1 (e.g., a field device) comprises at least one sensor located at a measuring point near a river 2 (could also be other water bodies, such as a lake, a reservoir) or on a slope 3 (e.g., for detecting a landslide event). The detection system 1 measures two variables (m, n): soil moisture and water level. However, the variables could also be at least one of a pressure head / groundwater head (or pressure), a water temperature, a flow rate, a soil bearing capacity, and / or other geological data, geographical data, climate data, weather data, or weather forecast data (distinguishing between rain and snow). The two variables (m, n) are then continuously measured at a predefined measurement interval.The measured values ​​(mt, nt) of the two variables are then transmitted to a data processing unit 4, which is designed to execute a model for evaluating the current risk of flooding and / or for predicting the possibility of a flooding event occurring.

[0020] The data processing unit 4 is located in a central remote monitoring station, which is connected to the detection system 1 by wire or wirelessly. The model includes AI (artificial intelligence)-based methods (e.g., anomaly detection) that are trained based on the historical measurements for the two variables (m, n). The model is improved using cross-correlation. In particular, it is used to test whether the two variables (m, n) are correlated and, if so, to test the time lag between the two sets of variables (m, n). High soil moisture can indicate a high probability of flooding or landslide events. When abnormal water levels are detected, this is often a sign of a flood or landslide.The correlation and time delay between the two sets of variables (m, n) could improve the detection process by providing more detailed and accurate prediction.

[0021] A further implementation of the invention is a method for detecting unauthorized discharge / withdrawal of water and / or predicting what has happened upstream. A detection system 1 comprises two sensors arranged at a measuring point in a river that measures two variables (m, n): water level and water temperature. However, other variables could also be used for the method, such as soil moisture, head / groundwater head (or pressure), flow, water temperature, pH, conductivity, turbidity, oxygen content, nitrite content, and / or other ion concentration data, climate data, weather data, or weather forecast data (distinguishing between rain and snow). The two variables (m, n) are then continuously measured at a predefined measuring interval.The measured values ​​(mt, nt) of the two variables are then transmitted to a data processing unit 4, which is designed to execute a model for evaluating the current water condition, for detecting an unauthorized discharge / withdrawal of water and / or for predicting the possibility of an upstream chemical accident.

[0022] The data processing unit 4 could be a computer in a central station for remote monitoring. The model includes AI (artificial intelligence)-based methods (e.g., anomaly detection) that are trained based on the historical measurements of the two variables (m, n). An abnormal water level or temperature may indicate unauthorized discharge / withdrawal of water. The model is then improved using cross-correlation. In particular, it is used to test whether the two variables (m, n) are correlated and, if so, to test the time lag between the two sets of variables (m, n).

[0023] List of reference symbols Detection system River Slope Data processing unit

Claims

Patent claims 1 . A method for detecting an environmental event, comprising the steps of: a) measuring at least two variables (m, n) detected by at least one detection system (1) arranged at a decentralized measuring point; b) transmitting the measured values ​​(mt, nt) of the at least two variables (m, n) from the detection system (1) to a data processing unit (4) designed to execute a model to evaluate the current environmental state and / or to predict the possibility of an environmental event occurring; characterized in that the correlation of the at least two variables (m, n) is measured, which is based on cross-correlation.

2. Method according to claim 1, characterized in that the at least two variables (m, n) are measured continuously in a predefined measuring interval.

3. Method according to at least one of claims 1 to 2, characterized in that the recognition system (1) comprises at least one of: a measuring sensor, a measuring device, a model of a geographical information system (GIS model), a web service and / or a database.

4. Method according to at least one of claims 1 to 3, characterized in that the data processing unit (4) is part of the recognition system (1).

5. Method according to at least one of claims 1 to 3, characterized in that the data processing unit (4) is located in a central station for remote monitoring.

6. Method according to at least one of claims 1 to 5, characterized in that the model comprises a simulation based on artificial intelligence methods.

7. Method according to at least one of claims 1 to 6, characterized in that the model is trained based on historical measured values ​​of the at least two variables (m, n).

8. Method according to at least one of claims 1 to 7, characterized in that the at least two variables (m, n) comprise at least two of a soil moisture content, a water level, a pressure head / groundwater pressure head (or a pressure), a water temperature, a pH value, a conductivity, a turbidity, a flow rate, an oxygen content, a nitrite content, a soil bearing capacity and / or other ion concentration data, geological data, geographical data, climate data, weather data, weather forecast data.

9. Method according to at least one of claims 1 to 8, characterized in that the environmental event comprises at least one of a flood, a landslide, an unauthorized discharge / withdrawal of water and / or an upstream chemical accident.

10. A system for implementing the method according to at least one of claims 1 to 9, comprising at least one detection system (1) arranged at a decentralized measuring point; and a data processing unit (4) configured to execute a model to evaluate the current environmental state and / or predict the possibility of an environmental event occurring.

Images