A data processing system, data processing method, and computer-readable storage device connected to a seismometer via IoT

Abstract

A data processing system is provided for processing seismic wave data obtained by a seismometer to identify geothermal resources. The data processing system includes a communication module and a processing module. The communication module is coupled to at least one seismometer via a communication transmission interface to obtain seismic wave data and is also coupled to the processing module. The processing module comprises a debugging unit and a machine learning model. The processing module removes duplicate or erroneous seismic wave data and filters out noise. The machine learning model is coupled to the debugging unit and is used to compare the debugged seismic wave data with multiple known geothermal characteristic data. Additionally, it analyzes the seismic wave spectrum to calculate the probability of a geothermal reservoir, thereby generating an analytical result.

Description

Technical Field

[0001] This disclosure relates to a data processing method, particularly a data processing method for seismic wave data. Prior Technology

[0002] Geothermal development relies heavily on precise geological and seismic data to pinpoint the exact location of geothermal reservoirs. Current seismic monitoring technology primarily depends on seismometers to collect seismic wave data, which is then used to analyze geological structure and potential geothermal activity. However, most existing seismometer systems use independent storage devices (such as memory cards) to record data, making data collection and transmission highly dependent on manual labor. In eastern Taiwan, for example, due to frequent earthquakes and complex terrain, researchers must regularly travel to various seismometer installation sites to manually replace memory cards and retrieve data. This not only increases manpower and time costs but also risks compromising data integrity due to equipment damage, data loss, or insufficient memory card storage space. Furthermore, because data is stored locally, it is difficult to monitor seismic activity in real time, hindering the provision of timely and accurate geothermal exploration information and limiting the efficiency of geothermal development.

[0003] On the other hand, existing seismic data processing methods also present numerous challenges. Because seismograph-sensed data often contains various types of noise and interference, data interpretation is not precise enough, leading to a lack of guarantee of data purity and reliability. Furthermore, the analysis of field data still relies on manual interpretation, which is not only time-consuming but also susceptible to subjective judgment. Therefore, how to more efficiently process the seismic wave information collected by seismographs has become a crucial issue. Summary of the Invention

[0004] In view of this, this disclosure provides a data processing system and method that connects to a seismograph via the Internet of Things, as well as a computer-readable storage device, which can process and analyze seismic wave data collected by the seismograph in real time, and can improve the accuracy and reliability of the data.

[0005] The data processing system disclosed herein is suitable for processing seismic wave data obtained by seismographs to locate geothermal resources. The data processing system includes an Internet of Things (IoT) module and a processing module. The IoT module can be coupled to at least one seismograph via a communication interface to acquire seismic wave data and is also coupled to the processing module. The processing module can be coupled to the IoT module to acquire seismic wave data, and the processing module has an edge computing architecture that can filter out unimportant seismic wave data.

[0006] In some embodiments, the edge computing architecture includes a debugging unit that can remove duplicate or erroneous seismic wave data based on GPS markers in the acquired seismic wave data. Additionally, the processing module includes an artificial intelligence engine that can be coupled to the debugging unit to identify and filter out anomalous seismic wave data with extreme values.

[0007] In other embodiments, the edge computing architecture also includes an artificial intelligence engine coupled to the debugging unit for identifying and filtering out abnormal seismic wave data with extreme values.

[0008] In other embodiments, the data processing system also includes an application unit coupled to a machine learning model, which generates a visual report for the user based on the analysis results.

[0009] The data processing method disclosed herein is suitable for processing seismic wave data collected by seismographs to locate geothermal resources, including acquiring seismic wave data through a communication interface. A preprocessing procedure is then performed to remove duplicate or erroneous seismic wave data, followed by a filtering procedure to remove noise from the seismic wave data. The preprocessed and filtered seismic wave data is then input into a machine learning model to compare the data with multiple known geothermal characteristics and calculate the probability of geothermal reservoirs through seismic wave spectral analysis, generating an analysis result. Furthermore, a visualization report can be generated based on the analysis results.

[0010] In some embodiments, this disclosure involves time synchronization of seismic wave data based on the Global Positioning System (GPS) time stamps in the seismic wave data to remove duplicate or erroneous seismic wave data.

[0011] In other embodiments, the preprocessing procedure in the data processing method disclosed herein includes performing a high-frequency noise filtering process to remove environmental noise, and performing a low-frequency noise filtering process to eliminate long-period interference noise.

[0012] In addition, in some embodiments, the filtering procedure of the data processing method disclosed herein includes using an artificial intelligence model engine to determine abnormal seismic wave data with extreme values, so as to filter out the abnormal seismic wave data.

[0013] The computer-readable storage device disclosed herein stores at least one instruction code. When the instruction code is read, the following instructions are executed: acquiring seismic wave data, wherein the seismic wave data is collected by a seismograph; performing a preprocessing procedure to remove duplicate or erroneous seismic wave data; performing a filtering procedure to filter out noise in the seismic wave data; inputting the seismic wave data after the preprocessing and filtering procedures into a machine learning model to compare the seismic wave data with multiple known geothermal characteristic values, and calculating the probability of geothermal reservoirs through seismic wave spectrum analysis, thereby generating an analysis result; and generating a visualization report based on the analysis result.

[0014] Because this analysis utilizes a communication interface to obtain collected seismic wave data from the seismograph, it allows for more immediate processing without human delay. Furthermore, this analysis pre-processes the seismic wave data for error correction and filtering, resulting in more accurate and reliable results. Simple Explanation of the Diagram

[0015] Figure 1 is a functional block diagram of a geothermal analysis system according to one of the preferred embodiments of the present disclosure. Figure 2 illustrates a functional block diagram of a complete seismograph facility according to one of the preferred embodiments of the present disclosure. Figure 3 is a functional block diagram of a data processing system according to a preferred embodiment of the present disclosure. Figure 4 is a functional block diagram of a processing module according to a preferred embodiment of the present disclosure. Figure 5 illustrates a flowchart of a data processing method according to a preferred embodiment of the present disclosure. Implementation

[0016] The following drawings disclose several embodiments of this disclosure. For clarity, many practical details will be described in the following description. However, it should be understood that these practical details should not be used to limit this disclosure. That is, in some embodiments of this disclosure, these practical details are not essential.

[0017] The term "coupled" or "connected" as used herein may refer to two or more components making direct physical or electrical contact with each other, or indirectly making direct physical or electrical contact with each other, or to two or more components operating or moving together. Furthermore, in this document, the term "module" refers to an object consisting of one or more transistors and / or one or more active / passive components connected in a certain manner to process signals.

[0018] Figure 1 illustrates a functional block diagram of a geothermal analysis system according to a preferred embodiment of the present disclosure. Referring to Figure 1, the geothermal analysis system 10 of this embodiment includes at least one seismograph, such as 12 or 14, and a data processing system 16. In this embodiment, seismographs 12 and 14 can be coupled to the data processing system 16 via a communication transmission interface. In some embodiments, the communication transmission interface can be a LoRa or mobile communication network interface. In some embodiments, seismographs 12 and 14 are located at different locations and are used to collect seismic wave data. In addition, seismographs 12 and 14 can transmit the collected seismic wave data to the data processing system 16 via the communication transmission interface.

[0019] Figure 2 illustrates a functional block diagram of a complete seismograph facility according to a preferred embodiment of this disclosure. Referring to Figure 2, in order for seismographs 12 and 14 to operate and to smoothly transmit the collected seismic wave data to the data processing system 16, in this embodiment, each seismograph 12 and 14 may also be equipped with a solar cell module 20 to provide the power required for the operation of seismographs 12 and 14 and the transmission of seismic wave data. This reduces the cost of manual maintenance of the equipment.

[0020] In this embodiment, the solar cell module 20 includes a solar panel 22 and a lithium battery 24, which are coupled to each other. Additionally, the lithium battery 24 is also coupled to seismographs 12 and 14 (Figure 2 illustrates this using seismograph 12 as an example). The solar panel 22 receives sunlight and performs photoelectric conversion, using the generated electricity to charge the lithium battery 24. The lithium battery 24 then supplies power to the seismograph 12. Because the lithium battery 24 can store electricity, the seismograph 12 can still operate normally even in inclement weather.

[0021] In this embodiment, seismographs 12 and 14 are waterproof to prevent them from malfunctioning due to moisture or water intrusion in harsh environments. Additionally, seismographs 12 and 14 may be equipped with a three-axis balancer and a Global Positioning System (GPS) module (not shown) to accurately capture multi-directional seismic motion information and ensure precise matching between the collected seismic wave data and the seismograph's location, thereby enhancing the scientific validity of the monitoring data.

[0022] Figure 3 illustrates a functional block diagram of a data processing system according to a preferred embodiment of the present disclosure. Referring to Figure 3, the data processing system 16 provided in this embodiment can be a cloud server, which includes an Internet of Things (IoT) module 102, a processing module 104, and a memory device 106. The IoT module 102 is coupled to seismometers 12 and 14 via a communication transmission interface to receive collected seismic wave data. The IoT module 102 is also coupled to the processing module 104, which can be implemented using a central processing unit (CPU), a graphics processing unit (GPU), an embedded system, or a microsystem; the present disclosure is not limiting. Additionally, the processing module 104 can also be coupled to the memory device 106. In this embodiment, the memory device 106 can be non-volatile memory or a solid-state drive, storing at least one instruction code.

[0023] Referring to Figure 3, when the processing module 104 reads the aforementioned instruction code from the memory device 106, it can obtain seismic wave data through the Internet of Things module 102 and process it. Figure 4 illustrates a functional block diagram of a processing module according to a preferred embodiment of this disclosure. Referring to Figure 4, the processing module 104 provided in this embodiment has an edge computing architecture. In some embodiments, this edge computing architecture may include a debugging unit 112, an artificial intelligence engine 114, and a machine learning model 116. Those skilled in the art will know that the debugging unit 112, the artificial intelligence engine 114, and the machine learning model 116 can be implemented using hardware circuits, firmware code, or software, and this disclosure is not limiting. In other embodiments, the artificial intelligence engine 114 may be a deep learning model.

[0024] After the processing module 104 obtains seismic wave data through the IoT module 102, the debugging unit 112 can first perform error correction on the obtained seismic wave data. Specifically, when a seismograph detects an earthquake, the GPS module in each seismograph obtains Coordinated Universal Time (UTC) via satellite, which is accurate to milliseconds. Then, when the seismograph generates seismic wave data, it can attach a UTC time stamp and convert it to the local time zone. When the debugging unit 112 receives the seismic wave data, it aligns the UTC time stamps of each piece of seismic wave data and identifies seismic wave data with the same UTC time stamp as duplicate seismic wave data, retaining only one piece of seismic wave data and filtering out the rest.

[0025] In some embodiments, the debugging unit 112 can also compare the signal similarity (e.g., cross-correlation coefficient) between adjacent seismic wave data. If the similarity is higher than a first preset percentage, such as 99%, it is determined to be duplicate data, and the duplicate part is deleted.

[0026] Additionally, in some embodiments, the edge computing architecture also includes a high-frequency noise filtering unit (not shown) to filter out environmental noise in the seismic wave data, such as vibrations caused by vehicle movement, machinery operation, or strong winds, heavy rain, and lightning strikes. Conversely, the edge computing architecture also includes a low-frequency noise filtering unit (not shown) to filter out long-period interference noise in the seismic wave data, such as interference from Earth's tidal effects, changes in air pressure and temperature, ground deformation and subsidence, and underground fluid movement.

[0027] After the debugging unit 112 completes the aforementioned debugging process on the seismic wave data, it can input the data into the artificial intelligence engine 114. The artificial intelligence engine 114 analyzes each piece of historical data and scores the newly acquired seismic wave data. If the reliability of the newly acquired seismic wave data is less than a second preset percentage, such as 85%, the seismic wave data can be marked as "low-quality data." At this time, the artificial intelligence engine 114 can choose to exclude it directly, or allow manual review to determine whether to retain it, based on user settings.

[0028] In addition, the artificial intelligence engine 114 can also identify seismic wave data with extreme values ​​and remove or correct them. Common extreme values ​​in seismic wave data include instantaneous anomalies from equipment malfunctions, electromagnetic interference, and lightning strikes, typically manifested as sudden spikes or sharp drops in signal points; periodic anomalies from industrial vibrations, ocean waves, and wind pressure changes, exhibiting periodic oscillations rather than true seismic signals; and repetitive or abrupt data from communication errors or abnormal data transmission, which may result in duplicate data or abrupt changes (e.g., unreasonable changes within a short period). In this embodiment, when the artificial intelligence engine 114 determines that the seismic wave data exceeds the measurable range of the seismograph 12 or 14, it determines that the seismic wave data is abnormal and excludes it, or marks it for manual interpretation to determine whether it needs to be excluded.

[0029] In a separate embodiment, when the artificial intelligence engine 114 locates anomalous seismic wave data, data interpolation can be performed, i.e., for missing data or anomaly points, interpolation technique is used to make up. Interpolation techniques include Linear Interpolation, which uses the mean of adjacent data points to fill in, Spline Interpolation, which uses curve fitting to fill in missing data, and K-Nearest Neighbors Interpolation, which uses patterns of adjacent data to fill outlier values. Additionally, if the data is slightly abnormal but can still be corrected, dynamic adjustment can be used. For example, if a data point deviates from the mean by more than 3 times the standard deviation, each data is added to the corrected data value after adding 3 to the standard deviation. In addition, historical data trends are also used for data smoothing processing (such as exponential moving average).

[0030] Specifically, the artificial intelligence engine 114 can use LSTM (Long Short-Term Memory) timing anomaly detection to determine anomalous seismic wave data. The LSTM model can learn normal seismic waveforms and detect anomalous data points that do not conform to the predicted pattern. If the prediction error of a data point is higher than the 95% confidence interval, it is considered an anomalous seismic wave data. In some other embodiments, the AI ​​engine 114 may also employ an isolation forest to detect anomalous seismic wave data. In such a model, a decision tree is created to identify isolated points (anomalies). Among them, data points that appear at low frequencies and are isolated are labeled as anomalous seismic wave data.

[0031] The art has commonly known knowledge that although the above debugging unit 112 and the artificial intelligence engine 114 appear to be two different functional blocks, in some embodiments the debugging unit 112 can also be implemented using the artificial intelligence model, which is not limited by the present disclosure.

[0032] Referring to Figure 4, the artificial intelligence engine 114 inputs the processed seismic wave data into the machine learning model 116. The machine learning model 116 compares the input seismic wave data with multiple known geothermal features and analyzes the seismic wave spectrum to calculate the probability of the presence of geothermal reservoirs, generating an analysis result. Specifically, the machine learning model 116 can be a supervised learning model, such as Random Forest, Support Vector Machine (SVM), or Convolutional Neural Network (CNN); a semi-supervised learning model, such as Graph Convolutional Networks (GCN); or an unsupervised learning model, such as K-Means Clustering.

[0033] Taking a convolutional neural network (CNN) as a machine learning model as an example, it can first convert the time-series data of seismic waves into a spectral image. Then, it trains a CNN model to identify spectral patterns related to geothermal activity. In this way, possible geothermal areas can be labeled based on the model's classification.

[0034] Referring back to Figure 3, in some embodiments, the data processing system 16 further includes a storage device 108 and an application unit (not shown). The application unit can be implemented using firmware or software and can be stored in the memory device 106 or the storage device 108. It is coupled to a machine learning model 116 to generate a visual report for the user based on the analysis results produced by the machine learning model 116, facilitating geothermal assessment and decision-making. In this embodiment, the storage device 108 can be a hard drive or flash memory, which can be coupled to the processing module 104 to store processed seismic wave data and analysis results.

[0035] Figure 5 illustrates a flowchart of a data processing method according to a preferred embodiment of the present disclosure. Referring to Figure 5, the data processing method provided in this embodiment first, as described in step S502, acquires collected seismic wave data from at least one seismograph through a communication interface. Next, as described in step S504, a preprocessing procedure is performed to remove duplicate or erroneous seismic wave data. For example, the aforementioned GPS time stamp should be applied to remove duplicate seismic wave data. Furthermore, the data processing method provided in this embodiment can also perform step 506, which is a filtering procedure to remove noise from the seismic wave data. For example, the aforementioned high- and low-frequency filtering is applied to remove environmental noise and long-period noise. In addition, as described above, anomalous seismic wave data with extreme values ​​can also be filtered out.

[0036] After the preprocessing and filtering procedures, the data processing method of this embodiment can proceed as described in step S508, by inputting the preprocessed and filtered seismic wave data into a machine learning model to compare the seismic wave data with multiple known geothermal characteristic values, and calculating the probability of geothermal reservoirs through seismic wave spectrum analysis, thereby generating an analysis result. Then, step S510 can be performed to generate a visual report for the user based on this analysis result.

[0037] In summary, this embodiment, through LoRa or mobile communication technology, can transmit seismic wave data to a cloud-based data processing system in real time for processing. Furthermore, the data processing system disclosed herein includes an error-correction unit that can remove environmental noise and long-period interference through high-frequency and low-frequency filtering techniques, and combines an artificial intelligence engine to automatically detect abnormal data and eliminate duplicate data or erroneous readings. Therefore, this disclosure offers high accuracy and reliability.

[0038] 10: Geothermal Analysis System 12, 14: Seismograph 16: Data Processing System 102: Internet of Things (IoT) Module 104: Processing Module 106: Memory device 108: Storage device 112: Debugging Unit 114: Artificial Intelligence Engine 116: Machine Learning Model 20: Solar cell module 22: Solar panels 24: Lithium batteries S502, S504, S506, S508, S510: Data processing method flow

Claims

1. A data processing system connected to a seismograph via the Internet of Things (IoT), suitable for processing seismic wave data obtained by a seismograph (12, 14) to locate geothermal activity, the data processing system (16) comprising: An Internet of Things (IoT) module (102) is coupled to the seismograph (12, 14) via a communication transmission interface to acquire the seismic wave data; a processing module (104) is coupled to the IoT module (102) and has an edge computing architecture to filter out unimportant seismic wave data; and the edge computing architecture has an error correction unit (112) that removes duplicate or erroneous seismic wave data based on the Global Positioning System (GPS) markers in the seismic wave data.

2. The data processing system as described in claim 1, wherein the processing module (104) further comprises an artificial intelligence engine (114) coupled to the debugging unit (112) for identifying abnormal seismic wave data with extreme values ​​and filtering out the abnormal seismic wave data.

3. The data processing system as described in claim 2, wherein the artificial intelligence engine (114) is a deep learning model.

4. The data processing system as described in claim 2, wherein the edge computing architecture has a machine learning model (116) coupled to the artificial intelligence engine (114) to compare the filtered seismic wave data with multiple known geothermal feature values ​​and calculate the probability of geothermal reservoirs through seismic wave spectrum analysis, thereby generating an analysis result.

5. The data processing system as described in claim 1, wherein the edge computing architecture includes: a high-frequency filtering unit for removing environmental noise from the seismic wave data; and a low-frequency filtering unit for removing long-period interference noise from the seismic wave data.

6. The data processing system as described in claim 1, wherein the communication interface is a LoRa or mobile communication network interface.

7. The data processing system as described in claim 1 further includes: A storage module (108) is coupled to the processing module (104) to store the processed seismic wave data and an analysis result generated by the processing module (104) through seismic wave spectrum analysis to calculate the probability of geothermal reservoirs.

8. A data processing method connected to a seismograph via the Internet of Things (IoT), suitable for processing seismic wave data collected by a seismograph (12, 14) to search for geothermal energy, the data processing method comprising at least the following steps: acquiring the seismic wave data through a communication interface; performing an edge computing procedure to filter out unimportant seismic wave data and noise; inputting the filtered seismic wave data into a machine learning model to compare the seismic wave data with multiple known geothermal characteristic values, and calculating the probability of geothermal reservoirs through seismic wave spectrum analysis to generate an analysis result; and generating a visualization report based on the analysis result; wherein the edge computing procedure further comprises at least the following steps: performing time synchronization processing on the seismic wave data based on the Global Positioning System (GPS) time stamp in the seismic wave data to remove duplicate or erroneous seismic wave data.

9. The data processing method as described in claim 8, wherein the communication interface is a LoRa or mobile communication network interface.

10. The data processing method as described in claim 8, wherein the edge computing procedure further includes at least the following steps: performing a high-frequency noise filtering process to remove environmental noise; performing a low-frequency noise filtering process to eliminate long-period interference noise; and using an artificial intelligence model engine to determine abnormal seismic wave data with extreme values, so as to filter out the abnormal seismic wave data.

11. A computer-readable storage device storing at least one instruction code, which, when read, performs at least the following steps: acquiring seismic wave data; performing an edge computing procedure to filter out unimportant seismic wave data and noise; inputting the filtered seismic wave data into a machine learning model to compare the seismic wave data with multiple known geothermal feature values, and calculating the probability of geothermal reservoirs through seismic wave spectrum analysis, thereby generating an analysis result, wherein performing the edge computing procedure further includes at least the following steps: performing time synchronization processing on the seismic wave data based on the Global Positioning System (GPS) time stamp in the seismic wave data to remove duplicate or erroneous seismic wave stamps.

12. The computer-readable storage device as described in claim 11, wherein the edge computing procedure further includes at least the following steps: performing a high-frequency noise filtering process to remove ambient noise; performing a low-frequency noise filtering process to eliminate long-period interference noise; and using an artificial intelligence engine to identify anomalous seismic wave data with extreme values, thereby filtering out the anomalous seismic wave data.

Images