Fire detection sensor data fusion method and system

Abstract

The invention discloses a data fusion method and system for a fire detection sensor, relates to the technical field of data fusion, remarkably improves the reliability and accuracy of the system, and solves the problems of false alarm, missing report, environmental interference, sensor fault and the like. By introducing a detailed anomaly detection module and a multi-level threshold evaluation mechanism, the system can accurately distinguish normal fluctuation and real fire signals, and reduces false alarm and missing alarm; in addition, the environment monitoring module calculates an environment interference coefficient in real time and compares the environment interference coefficient with a threshold value, so that the influence of environment factors on the detection accuracy is effectively dealt with, and the detection reliability is improved. The safety fault-tolerant module monitors the state of the sensor and the data transmission process in real time to ensure that a data compensation mechanism is started when the sensor fails or data is lost, so that the fault-tolerant capability of the system is enhanced; the comprehensive measure ensures the stability and accuracy of the system under various interference and fault conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of data fusion, and in particular to a method and system for fusion of fire detection sensor data. Background Art

[0002] Traditional fire detection systems usually rely on a single type of sensor, such as smoke, temperature, flame or carbon monoxide sensors. Each of these sensors has different detection capabilities and limitations. For example, smoke sensors are sensitive to thick smoke, but may not work well in smokeless fire scenes; temperature sensors are sensitive to heat, but are slow to react to fires without obvious temperature rise. In order to improve the accuracy and response speed of fire detection, modern technology has begun to introduce multi-sensor data fusion methods, which integrates data from multiple sensors and combines intelligent algorithms for comprehensive analysis, thereby improving the overall performance of the fire detection system.

[0003] However, the existing fire detection sensor data fusion system has the following technical disadvantages in practical applications:

[0004] 1. There are problems with false alarms and missed reports;

[0005] 2. Environmental interference leads to inaccurate detection;

[0006] 3. Insufficient fault tolerance leads to sensor failure and data loss; Summary of the invention

[0007] In view of the deficiencies in the prior art, the present invention provides a fire detection sensor data fusion method and system, which solves the problems mentioned in the background technology.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a fire detection sensor data fusion method and system, including an environment monitoring module, a data preprocessing module, a fusion algorithm module, an anomaly detection module, a safety fault-tolerant module and a response trigger module;

[0009] The environmental monitoring module is used to monitor the interference factors in the environment where the sensor is located in real time, collect environmental data sets, and generate an environmental interference coefficient Wd. If the environmental interference coefficient Wd exceeds a preset interference threshold T1, an interference signal is transmitted to the data preprocessing module;

[0010] The data preprocessing module is used to filter and standardize the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, and generate a preliminary processed data set, which is then transmitted to the fusion algorithm module;

[0011] The fusion algorithm module is used to fuse the preliminary processed data group based on the preset multi-sensor data fusion strategy, generate the initial fire detection coefficient Fx, and calculate the noise sensitivity Ns and its fault tolerance deviation index ε during the fusion process, and transmit the fusion result to the anomaly detection module;

[0012] The anomaly detection module is used to detect the initial fire detection coefficient Fx output by the fusion algorithm module, and generate an abnormal alarm coefficient Ax according to a preset standard curve, and simultaneously fit the abnormal alarm coefficient Ax and the noise sensitivity Ns to obtain and evaluate the abnormal alarm coefficient Ax, and finally determine and confirm the potential false alarm or missed alarm risk based on the evaluation result of the abnormal alarm coefficient Ax;

[0013] The safety fault-tolerant module is used to monitor the operating status and data transmission process of the sensor in real time, by generating and evaluating the sensor state stability coefficient St, and finally starting the data compensation mechanism according to the evaluation result of the state stability coefficient St;

[0014] The response trigger module is used to receive the abnormal alarm coefficient Ax generated by the abnormal detection module and the sensor state stability coefficient St generated by the safety fault tolerance module, and combine the two to generate and evaluate the final response coefficient Ry, and finally trigger the corresponding fire warning response according to the evaluation result of the final response coefficient Ry.

[0015] In this embodiment,

[0016] Preferably, the environment monitoring module includes an environment acquisition unit and an interference analysis unit;

[0017] The environment acquisition unit is used to continuously monitor the environment in which the sensor is located, and set a monitoring time period, and use multiple types of sensors, including temperature sensors, humidity sensors and gas sensors, to monitor environmental changes within the monitoring time period, summarize and obtain environmental data sets, and extract their key environmental parameter values ​​Zmax, and obtain the average environmental parameter value Zavg in the set time period through the statistical averaging algorithm, and extract features from the information in the environmental data set to establish an environmental fluctuation model;

[0018] The interference analysis unit is used to train and verify the environmental fluctuation model, extract the environmental parameter value Zmax and the average environmental parameter value Zavg, perform dimensionless processing, and calculate the environmental interference coefficient Wd through the following formula:

[0019]

[0020] Where i=1, 2, ..., n, n represents the time sequence monitored in the set time period, Zmax i It is expressed as the maximum environmental parameter value of the i-th time sequence within the monitoring period.

[0021] Preferably, the data preprocessing module includes a filtering unit and a standardization unit;

[0022] The filtering unit is used to preliminarily screen the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, remove noise and abnormal data, and screen out qualified data by setting thresholds and signal strength detection methods, and organize them into a preliminary data set;

[0023] The standardization unit is used to standardize the preliminary data set screened by the filtering unit, including converting the units and adjusting the scales of the data from different sensors to have a unified dimension, and adjusting the data to the same range through linear transformation, and finally generating a standardized data set and transmitting it to the fusion algorithm module.

[0024] Preferably, the fusion algorithm module includes a data fusion unit and a parameter calculation unit;

[0025] The data fusion unit is used to calculate the preliminary processed data group generated by the data preprocessing module based on the preset multi-sensor data fusion strategy to obtain the initial fire detection coefficient Fx. The specific calculation formula is as follows:

[0026]

[0027] Wherein, Wd represents temperature, Ynd represents smoke concentration, Qnd represents harmful gas concentration, and Gqd represents photoelectric signal intensity.

[0028] Preferably, the parameter calculation unit is used to calculate the noise sensitivity Ns and the error tolerance deviation index ε during the data fusion process; the noise sensitivity Ns is used to reflect the influence of the noise in the sensor data on the fusion result, and is calculated according to the following formula:

[0029]

[0030] In the formula, represents the number of sensors considered in calculating noise sensitivity, represents the noise value of the ith sensor, represents the average value of the noise value;

[0031] The tolerance deviation index ε is calculated by the following formula:

[0032]

[0033] In the formula, represents the number of sensors considered when calculating the error tolerance deviation, represents the measurement deviation of the ii-th sensor, Represents the average value of the measurement deviation.

[0034] Preferably, the anomaly detection module includes a standard curve unit and an anomaly alarm generation unit;

[0035] The standard curve unit is used to generate a standard curve Sx of the initial fire detection coefficient Fx according to historical data and the set detection standard; the standard curve Sx is used to evaluate the normal range of the current fire detection coefficient, and the standard curve value is calculated by the following formula:

[0036]

[0037] Where a and b are the coefficients obtained from regression analysis.

[0038] Preferably, the abnormal alarm generating unit is used to compare the standard curve value Sx generated by the standard curve unit with the initial fire detection coefficient Fx output by the fusion algorithm module, and calculate the abnormal alarm coefficient Ax according to the following formula:

[0039]

[0040] In the formula, K represents the standard deviation;

[0041] By presetting the first deviation threshold T1 and the second deviation threshold T2, the abnormal alarm coefficient Ax is evaluated, and the first deviation threshold T1> the second deviation threshold T2, and the specific contents are as follows:

[0042] If the abnormal alarm coefficient Ax> the first deviation threshold T1, it is marked as the first abnormal alarm, indicating that there is a fire risk. At this time, the system makes corresponding alarm response measures according to the actual situation;

[0043] If the first deviation threshold T1 ≥ abnormal alarm coefficient Ax > second deviation threshold T2, it is marked as the second abnormal alarm, and a detailed inspection and confirmation is performed to further verify its authenticity to eliminate the possibility of false alarm;

[0044] If the second deviation threshold T2 ≥ abnormal alarm coefficient Ax, no abnormal mark is made, indicating that the abnormal alarm coefficient Ax is within the preset normal range, indicating that the current fire detection coefficient Fx does not deviate from the standard curve, and no fire or abnormal situation has occurred; at this time, continue to monitor subsequent data.

[0045] Preferably, the safety fault-tolerant module includes a state monitoring unit and a data compensation unit;

[0046] The state monitoring unit collects sensor state-related data by monitoring the operating state of the sensor in real time, and obtains the sensor state value Czt and the sensor state average value The sensor state stability coefficient St is calculated by the following formula:

[0047]

[0048] In the formula, x represents the number of sensors involved in the calculation;

[0049] The data compensation unit is used to evaluate the sensor state stability coefficient St and start the data compensation mechanism; the evaluation is performed by presetting the stability threshold Q and the sensor state stability coefficient St, and the specific contents are as follows:

[0050] If the stability threshold Q is less than the sensor state stability coefficient St, it means that the state stability of the sensor does not meet the requirements. At this time, the data compensation mechanism is activated, including adjusting the collected data according to the actual fluctuation of the sensor state; calibrating or replacing the unstable sensor;

[0051] If the stability threshold Q ≥ sensor state stability coefficient St, it means that the state stability of the sensor meets the requirements, and data compensation is not performed at this time.

[0052] Preferably, the response trigger module obtains the final response coefficient Ry through the following formula:

[0053]

[0054] In the formula, 2 represents a constant;

[0055] The final response coefficient Ry is evaluated by presetting the response threshold W. The specific evaluation contents are as follows:

[0056] If the final response coefficient Ry>response threshold W, it means that the current abnormal alarm and sensor stability are abnormal; therefore, the system will trigger a fire warning response;

[0057] If the final response coefficient Ry≤response threshold W, it means that the current abnormal alarm and sensor stability are normal; at this time, no warning is triggered and the data continues to be monitored.

[0058] A fire detection sensor data fusion method comprises the following steps:

[0059] Step 1: monitor the interference factors in the environment where the sensor is located in real time, collect environmental data sets, and generate an environmental interference coefficient Wd. If the environmental interference coefficient Wd exceeds a preset interference threshold T1, an interference signal is transmitted to a data preprocessing module;

[0060] Step 2: Filter and standardize the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, generate a preliminary processed data set, and then transmit it to the fusion algorithm module;

[0061] Step 3: Based on the preset multi-sensor data fusion strategy, the preliminary processed data group is fused to generate the initial fire detection coefficient Fx, and the noise sensitivity Ns and its fault tolerance deviation index ε are calculated during the fusion process. The fusion result is transmitted to the anomaly detection module;

[0062] Step 4: Detect the initial fire detection coefficient Fx output by the fusion algorithm module, generate the abnormal alarm coefficient Ax according to the preset standard curve, and simultaneously fit the abnormal alarm coefficient Ax and the noise sensitivity Ns to obtain the abnormal alarm coefficient Ax and evaluate it. Finally, based on the evaluation result of the abnormal alarm coefficient Ax, determine and confirm the potential false alarm or missed alarm risk;

[0063] Step 5: monitor the operating status and data transmission process of the sensor in real time, generate and evaluate the sensor state stability coefficient St, and finally start the data compensation mechanism according to the evaluation result of the state stability coefficient St;

[0064] Step 6: Receive the abnormal alarm coefficient Ax generated by the abnormal detection module and the sensor state stability coefficient St generated by the safety fault-tolerant module, combine the two to generate the final response coefficient Ry and evaluate it, and finally trigger the corresponding fire warning response based on the evaluation result of the final response coefficient Ry.

[0065] The present invention provides a fire detection sensor data fusion method and system. It has the following beneficial effects:

[0066] (1) Solving the problem of false alarms and missed alarms: By introducing a detailed anomaly detection module and a strict threshold evaluation mechanism, the problem of false alarms and missed alarms is effectively solved. The anomaly detection module uses a standard curve unit and an abnormal alarm generation unit to evaluate the initial fire detection coefficient, and performs multi-level judgments through the set threshold of the abnormal alarm coefficient. This method accurately distinguishes normal fluctuations from real fire signals, ensuring the high accuracy of the system in judging fire risks, thereby reducing the occurrence of false alarms and missed alarms. This comprehensive evaluation mechanism enables the system to respond more reliably to potential fire risks, improving the overall alarm accuracy.

[0067] (2) Dealing with inaccurate detection caused by environmental interference: The design of the environmental monitoring module significantly improves the system's adaptability to environmental interference; the environmental acquisition unit monitors environmental changes through multiple sensors, calculates the environmental interference coefficient in real time, and compares it with the preset interference threshold; when the interference coefficient exceeds the threshold, the system transmits and processes the interference signal; this method enables the system to dynamically adjust its sensitivity to environmental interference, reduce the impact of environmental factors on fire detection accuracy, and thus improve the reliability and effectiveness of detection; through effective environmental data processing and analysis, the system can more accurately identify and distinguish fire signals from environmental interference;

[0068] (3) Enhance fault tolerance to solve the problems of sensor failure and data loss: The safety fault-tolerant module monitors the status of the sensor and the data transmission process in real time to ensure that the system can effectively deal with the problems of sensor failure and data loss. The status monitoring unit calculates the sensor status stability coefficient and starts the data compensation mechanism based on this coefficient. When the stability does not meet the requirements, the system will adjust the data and calibrate or replace the sensor. This design enhances the system's fault tolerance to sensor failures, ensures the integrity and accuracy of the data, and thus improves the overall reliability and robustness of the system. Through this fault-tolerant mechanism, the system can maintain normal working conditions and provide stable fire detection services in the event of sensor failure or data loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 A schematic diagram of the framework structure of a fire detection sensor data fusion system of the present invention;

[0070] Figure 2 The present invention is a schematic flow chart of the steps of a fire detection sensor data fusion method. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] Example 1

[0073] See also Figure 1 ,The present invention provides a fire detection sensor data fusion method and system, including an environment monitoring module, a data preprocessing module, a fusion algorithm module, an anomaly detection module, a safety fault-tolerant module and a response trigger module;

[0074] The environmental monitoring module is used to monitor the interference factors in the environment where the sensor is located in real time, collect environmental data sets, and generate an environmental interference coefficient Wd. If the environmental interference coefficient Wd exceeds a preset interference threshold T1, an interference signal is transmitted to the data preprocessing module;

[0075] The data preprocessing module is used to filter and standardize the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, and generate a preliminary processed data set, which is then transmitted to the fusion algorithm module;

[0076] The fusion algorithm module is used to fuse the preliminary processed data group based on the preset multi-sensor data fusion strategy, generate the initial fire detection coefficient Fx, and calculate the noise sensitivity Ns and its fault tolerance deviation index ε during the fusion process, and transmit the fusion result to the anomaly detection module;

[0077] The anomaly detection module is used to detect the initial fire detection coefficient Fx output by the fusion algorithm module, and generate an abnormal alarm coefficient Ax according to a preset standard curve, and simultaneously fit the abnormal alarm coefficient Ax and the noise sensitivity Ns to obtain and evaluate the abnormal alarm coefficient Ax, and finally determine and confirm the potential false alarm or missed alarm risk based on the evaluation result of the abnormal alarm coefficient Ax;

[0078] The safety fault-tolerant module is used to monitor the operating status and data transmission process of the sensor in real time, by generating and evaluating the sensor state stability coefficient St, and finally starting the data compensation mechanism according to the evaluation result of the state stability coefficient St;

[0079] The response trigger module is used to receive the abnormal alarm coefficient Ax generated by the abnormal detection module and the sensor state stability coefficient St generated by the safety fault tolerance module, and combine the two to generate and evaluate the final response coefficient Ry, and finally trigger the corresponding fire warning response according to the evaluation result of the final response coefficient Ry.

[0080] In this embodiment, the environmental monitoring module timely identifies environmental interference factors and transmits them to the data preprocessing module through real-time monitoring and calculation of interference coefficients, thereby effectively reducing the impact of environmental interference on detection accuracy; the data preprocessing module filters and standardizes the collected raw data to ensure the consistency and accuracy of the data, and provides a high-quality data basis for the subsequent fusion algorithm module; the fusion algorithm module combines the multi-sensor data fusion strategy to generate an initial fire detection coefficient Fx, and calculates the noise sensitivity Ns and its fault tolerance deviation index ε during the fusion process, thereby improving the accuracy of fire detection and the fault tolerance of the system; the anomaly detection module generates and evaluates the abnormal alarm coefficient Ax through a standard curve, effectively distinguishes normal fluctuations from fire signals, and reduces the risk of false alarms and missed alarms; the safety fault-tolerant module monitors the sensor status in real time, generates a stability coefficient and starts a data compensation mechanism, thereby enhancing the stability and reliability of the system; the response trigger module combines the abnormal alarm coefficient Ax and the sensor status stability coefficient St, and finally generates a response coefficient Ry, thereby ensuring that the early warning response can be quickly triggered when a fire occurs, thereby improving the response efficiency of the system.

[0081] Example 2

[0082] The environment monitoring module includes an environment acquisition unit and an interference analysis unit;

[0083] The environment acquisition unit is used to continuously monitor the environment in which the sensor is located, and set a monitoring time period, and use multiple types of sensors, including temperature sensors, humidity sensors and gas sensors, to monitor environmental changes within the monitoring time period, summarize and obtain environmental data sets, and extract their key environmental parameter values ​​Zmax, and obtain the average environmental parameter value Zavg in the set time period through the statistical averaging algorithm, and extract features from the information in the environmental data set to establish an environmental fluctuation model;

[0084] The interference analysis unit is used to train and verify the environmental fluctuation model, extract the environmental parameter value Zmax and the average environmental parameter value Zavg, perform dimensionless processing, and calculate the environmental interference coefficient Wd through the following formula:

[0085]

[0086] Where i=1, 2, ..., n, n represents the time sequence monitored in the set time period, Zmax i It is expressed as the maximum environmental parameter value of the i-th time sequence within the monitoring period.

[0087] In this embodiment, the environment acquisition unit continuously monitors the environment in which the sensor is located and uses multiple sensors to comprehensively and carefully collect environmental data. The use of multiple types of sensors improves the ability to detect environmental interference factors. In addition, the environment acquisition unit establishes an environmental fluctuation model by extracting the key environmental parameter value Zmax and calculating the average environmental parameter value Zavg, thereby achieving accurate assessment and modeling of environmental fluctuations. The interference analysis unit trains and verifies the environmental fluctuation model, performs dimensionless processing on the environmental parameter values ​​Zmax and Zavg, and calculates the environmental interference coefficient Wd, thereby providing a quantitative analysis of the impact of environmental interference, enabling the system to more accurately identify interference signals and take corresponding measures, thereby optimizing the reliability and accuracy of data processing.

[0088] Example 3

[0089] The data preprocessing module includes a filtering unit and a standardization unit;

[0090] The filtering unit is used to preliminarily screen the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, remove noise and abnormal data, and screen out qualified data by setting thresholds and signal strength detection methods, and organize them into a preliminary data set;

[0091] The standardization unit is used to standardize the preliminary data set screened by the filtering unit, including converting the units and adjusting the scales of the data from different sensors to have a unified dimension, and adjusting the data to the same range through linear transformation, and finally generating a standardized data set and transmitting it to the fusion algorithm module.

[0092] In this embodiment, the filtering unit performs preliminary screening on the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, thereby removing noise and abnormal data and improving the quality and accuracy of the data; by setting thresholds and signal strength detection methods, it ensures that only qualified data are retained, thereby reducing errors and interference in the data and improving the reliability of data processing; the standardization unit performs standardization processing on the filtered data, including unit conversion and scale adjustment, so that the data from different sensors have a unified dimension and range, which can effectively eliminate the deviation between different data sources and ensure the consistency and comparability of the data. The final generated standardized processed data group provides high-quality input data for the fusion algorithm module, thereby improving the accuracy of fire detection and the overall performance of the system.

[0093] Example 4

[0094] The fusion algorithm module includes a data fusion unit and a parameter calculation unit;

[0095] The data fusion unit is used to calculate the preliminary processed data group generated by the data preprocessing module based on the preset multi-sensor data fusion strategy to obtain the initial fire detection coefficient Fx. The specific calculation formula is as follows:

[0096]

[0097] Wherein, Wd represents temperature, Ynd represents smoke concentration, Qnd represents harmful gas concentration, and Gqd represents photoelectric signal intensity.

[0098] The parameter calculation unit is used to calculate the noise sensitivity Ns and the error tolerance deviation index ε during the data fusion process; the noise sensitivity Ns is used to reflect the impact of noise in the sensor data on the fusion result, and is calculated according to the following formula

[0099]

[0100] In the formula, represents the number of sensors considered in calculating noise sensitivity, represents the noise value of the ith sensor, represents the average value of the noise value;

[0101] The tolerance deviation index ε is calculated by the following formula:

[0102]

[0103] In the formula, represents the number of sensors considered when calculating the error tolerance deviation, represents the measurement deviation of the ii-th sensor, Represents the average value of the measurement deviation.

[0104] In this embodiment, the data fusion unit calculates the preliminary processed data group based on the multi-sensor data fusion strategy to generate an initial fire detection coefficient Fx, thereby integrating data from different sources, providing a comprehensive fire detection result, and improving the accuracy and reliability of the system; the parameter calculation unit calculates the noise sensitivity Ns and the fault tolerance deviation index ε during the data fusion process to reflect the impact of noise in the sensor data on the fusion result and evaluate the fault tolerance of the data; the noise sensitivity Ns reveals how the noise in the data affects the final result by considering the number of sensors and their noise values, thereby ensuring the stability of data fusion; the fault tolerance deviation index ε evaluates the fault tolerance of the system in the face of sensor failures by considering the number of sensors and their measurement deviations, thereby improving the robustness and reliability of the system and reducing fire detection errors caused by sensor problems.

[0105] Example 5

[0106] The anomaly detection module includes a standard curve unit and an anomaly alarm generation unit;

[0107] The standard curve unit is used to generate a standard curve Sx of the initial fire detection coefficient Fx according to historical data and the set detection standard; the standard curve Sx is used to evaluate the normal range of the current fire detection coefficient, and the standard curve value is calculated by the following formula:

[0108]

[0109] Where a and b are the coefficients obtained from regression analysis.

[0110] The abnormal alarm generation unit is used to compare the standard curve value Sx generated by the standard curve unit with the initial fire detection coefficient Fx output by the fusion algorithm module, and calculate the abnormal alarm coefficient Ax according to the following formula:

[0111]

[0112] In the formula, K represents the standard deviation;

[0113] By presetting the first deviation threshold T1 and the second deviation threshold T2, the abnormal alarm coefficient Ax is evaluated, and the first deviation threshold T1> the second deviation threshold T2, and the specific contents are as follows:

[0114] If the abnormal alarm coefficient Ax> the first deviation threshold T1, it is marked as the first abnormal alarm, indicating that there is a fire risk. At this time, the system makes corresponding alarm response measures according to the actual situation;

[0115] If the first deviation threshold T1 ≥ abnormal alarm coefficient Ax > second deviation threshold T2, it is marked as the second abnormal alarm, and a detailed inspection and confirmation is performed to further verify its authenticity to eliminate the possibility of false alarm;

[0116] If the second deviation threshold T2 ≥ abnormal alarm coefficient Ax, no abnormal mark is made, indicating that the abnormal alarm coefficient Ax is within the preset normal range, indicating that the current fire detection coefficient Fx does not deviate from the standard curve, and no fire or abnormal situation has occurred; at this time, continue to monitor subsequent data.

[0117] In this embodiment, the standard curve unit generates a standard curve Sx of the initial fire detection coefficient Fx through historical data and set detection standards. This process enables the normal range of the current fire detection coefficient to be accurately evaluated, thereby improving the accuracy of fire detection; the abnormal alarm generation unit compares the standard curve value Sx with the initial fire detection coefficient Fx, calculates the abnormal alarm coefficient Ax, and evaluates it through the set deviation thresholds T1 and T2, so as to accurately distinguish different levels of fire risks; if Ax exceeds the first deviation threshold T1, it is marked as the first abnormal alarm, indicating that there is a high fire risk, and the system will take corresponding alarm response measures; if Ax is between T1 and T2, it is marked as the second abnormal alarm, and further inspection and confirmation are required to eliminate false alarms; if Ax is lower than the second deviation threshold T2, it means that the detection result is within the normal range, and the system continues to monitor, avoiding false alarms and ensuring stable operation of the system.

[0118] Example 6

[0119] The safety fault-tolerant module includes a state monitoring unit and a data compensation unit;

[0120] The state monitoring unit collects sensor state related data by real-time monitoring of the sensor state, and obtains the sensor state value Czt and the sensor state average value. The sensor state stability coefficient St is calculated by the following formula:

[0121]

[0122] In the formula, x represents the number of sensors involved in the calculation;

[0123] The data compensation unit is used to evaluate the sensor state stability coefficient St and start the data compensation mechanism; the evaluation is performed by presetting the stability threshold Q and the sensor state stability coefficient St, and the specific contents are as follows:

[0124] If the stability threshold Q is less than the sensor state stability coefficient St, it means that the state stability of the sensor does not meet the requirements. At this time, the data compensation mechanism is activated, including adjusting the collected data according to the actual fluctuation of the sensor state; calibrating or replacing the unstable sensor;

[0125] If the stability threshold Q ≥ sensor state stability coefficient St, it means that the state stability of the sensor meets the requirements, and data compensation is not performed at this time.

[0126] In this embodiment, the status monitoring unit monitors the operating status of the sensor in real time, and calculates the sensor status stability coefficient St by collecting the sensor status value Czt and its average value. This process can timely identify the operating stability of the sensor, thereby ensuring the reliability and accuracy of the data; the data compensation unit evaluates the sensor status stability coefficient St and the preset stability threshold Q to determine whether to start the data compensation mechanism; if St is lower than the threshold Q, it means that the sensor status is unstable, and the system will start the data compensation mechanism, adjust the data, and calibrate or replace the unstable sensor, thereby effectively solving the sensor failure and data loss problems; if St is higher than the threshold Q, the system considers that the sensor status is stable and no compensation is required, which not only reduces unnecessary maintenance work but also maintains the normal operation of the system.

[0127] Example 7

[0128] The response trigger module obtains the final response coefficient Ry through the following formula:

[0129]

[0130] In the formula, 2 represents a constant;

[0131] The final response coefficient Ry is evaluated by presetting the response threshold W. The specific evaluation contents are as follows:

[0132] If the final response coefficient Ry>response threshold W, it means that the current abnormal alarm and sensor stability are abnormal; therefore, the system will trigger a fire warning response;

[0133] If the final response coefficient Ry≤response threshold W, it means that the current abnormal alarm and sensor stability are normal; at this time, no warning is triggered and the data continues to be monitored.

[0134] In this embodiment, the final response coefficient Ry is calculated and compared with the preset response threshold W, so as to realize the intelligent triggering of the fire warning; when the final response coefficient Ry is higher than the response threshold W, the system will identify that there are problems with the current abnormal alarm and sensor stability, and then trigger the fire warning response, which can effectively deal with potential fire risks; if the final response coefficient Ry is lower than or equal to the response threshold W, it means that the current abnormal alarm and sensor status are normal, and the system will continue to monitor the data to avoid unnecessary warnings, which can not only reduce false alarms but also ensure the efficient operation of the system.

[0135] Example 8

[0136] See also Figure 2 , a fire detection sensor data fusion method, comprising the following steps:

[0137] Step 1: monitor the interference factors in the environment where the sensor is located in real time, collect environmental data sets, and generate an environmental interference coefficient Wd. If the environmental interference coefficient Wd exceeds a preset interference threshold T1, an interference signal is transmitted to a data preprocessing module;

[0138] Step 2: Filter and standardize the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, generate a preliminary processed data set, and then transmit it to the fusion algorithm module;

[0139] Step 3: Based on the preset multi-sensor data fusion strategy, the preliminary processed data group is fused to generate the initial fire detection coefficient Fx, and the noise sensitivity Ns and its fault tolerance deviation index ε are calculated during the fusion process. The fusion result is transmitted to the anomaly detection module;

[0140] Step 4: Detect the initial fire detection coefficient Fx output by the fusion algorithm module, generate the abnormal alarm coefficient Ax according to the preset standard curve, and simultaneously fit the abnormal alarm coefficient Ax and the noise sensitivity Ns to obtain the abnormal alarm coefficient Ax and evaluate it. Finally, based on the evaluation result of the abnormal alarm coefficient Ax, determine and confirm the potential false alarm or missed alarm risk;

[0141] Step 5: monitor the operating status and data transmission process of the sensor in real time, generate and evaluate the sensor state stability coefficient St, and finally start the data compensation mechanism according to the evaluation result of the state stability coefficient St;

[0142] Step 6: Receive the abnormal alarm coefficient Ax generated by the abnormal detection module and the sensor state stability coefficient St generated by the safety fault-tolerant module, combine the two to generate the final response coefficient Ry and evaluate it, and finally trigger the corresponding fire warning response based on the evaluation result of the final response coefficient Ry.

[0143] In this embodiment, step 1 ensures that the impact of environmental interference on data is identified and processed in time by real-time monitoring of environmental interference and generating environmental interference coefficient Wd, thereby reducing the risk of false alarms caused by interference; step 2 filters and standardizes sensor data, optimizes data quality and consistency, and ensures that the basic data for subsequent fusion and analysis is accurate and reliable; step 3 generates an initial fire detection coefficient Fx through a multi-sensor data fusion strategy, and calculates noise sensitivity Ns and fault tolerance deviation index ε, thereby enhancing the system's adaptability to noise and deviation and improving the accuracy of fire detection; step 4 further reduces the possibility of false alarms and missed alarms and ensures the reliability of fire alarms by generating a standard curve of the anomaly detection module and evaluating the anomaly alarm coefficient Ax; step 5 generates a sensor state stability coefficient St by monitoring the sensor state and data transmission process, and starts a data compensation mechanism when necessary, thereby enhancing the system's fault tolerance in the event of sensor failure; step 6 generates a final response coefficient Ry in combination with the anomaly alarm coefficient Ax and the sensor state stability coefficient St, and triggers a fire warning response according to the evaluation results, thereby reducing unnecessary warnings while ensuring safety. This method effectively solves the problems of false alarms and missed alarms in traditional fire detection systems through multi-level and multi-dimensional detection and optimization, and improves the overall performance of the fire warning system.

[0144] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A fire detection sensor data fusion method and system, characterized in that: It includes environment monitoring module, data preprocessing module, fusion algorithm module, anomaly detection module, safety fault tolerance module and response trigger module; The environmental monitoring module is used to monitor the interference factors in the environment where the sensor is located in real time, collect environmental data sets, and generate an environmental interference coefficient Wd. If the environmental interference coefficient Wd exceeds a preset interference threshold T1, an interference signal is transmitted to the data preprocessing module; The data preprocessing module is used to filter and standardize the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, and generate a preliminary processed data set, which is then transmitted to the fusion algorithm module; The fusion algorithm module is used to fuse the preliminary processed data group based on the preset multi-sensor data fusion strategy, generate the initial fire detection coefficient Fx, and calculate the noise sensitivity Ns and its fault tolerance deviation index ε during the fusion process, and transmit the fusion result to the anomaly detection module; The anomaly detection module is used to detect the initial fire detection coefficient Fx output by the fusion algorithm module, and generate an abnormal alarm coefficient Ax according to a preset standard curve, and simultaneously fit the abnormal alarm coefficient Ax and the noise sensitivity Ns to obtain and evaluate the abnormal alarm coefficient Ax, and finally determine and confirm the potential false alarm or missed alarm risk based on the evaluation result of the abnormal alarm coefficient Ax; The safety fault-tolerant module is used to monitor the operating status and data transmission process of the sensor in real time, by generating and evaluating the sensor state stability coefficient St, and finally starting the data compensation mechanism according to the evaluation result of the state stability coefficient St; The response trigger module is used to receive the abnormal alarm coefficient Ax generated by the abnormal detection module and the sensor state stability coefficient St generated by the safety fault tolerance module, and combine the two to generate and evaluate the final response coefficient Ry, and finally trigger the corresponding fire warning response according to the evaluation result of the final response coefficient Ry.

2. A fire detection sensor data fusion method and system according to claim 1, characterized in that: The environment monitoring module includes an environment acquisition unit and an interference analysis unit; The environment acquisition unit is used to continuously monitor the environment in which the sensor is located, and set a monitoring time period, and use multiple types of sensors, including temperature sensors, humidity sensors and gas sensors, to monitor environmental changes within the monitoring time period, summarize and obtain environmental data sets, and extract their key environmental parameter values ​​Zmax, and obtain the average environmental parameter value Zavg in the set time period through the statistical averaging algorithm, and extract features from the information in the environmental data set to establish an environmental fluctuation model; The interference analysis unit is used to train and verify the environmental fluctuation model, extract the environmental parameter value Zmax and the average environmental parameter value Zavg, perform dimensionless processing, and calculate the environmental interference coefficient Wd through the following formula: ; Where i=1, 2, ..., n, n represents the time sequence monitored in the set time period, Zmax i It is expressed as the maximum environmental parameter value of the i-th time sequence within the monitoring period.

3. A fire detection sensor data fusion method and system according to claim 1, characterized in that: The data preprocessing module includes a filtering unit and a standardization unit; The filtering unit is used to preliminarily screen the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, remove noise and abnormal data, and screen out qualified data by setting thresholds and signal strength detection methods, and organize them into a preliminary data set; The standardization unit is used to standardize the preliminary data set screened by the filtering unit, including converting the units and adjusting the scales of the data from different sensors to have a unified dimension, and adjusting the data to the same range through linear transformation, and finally generating a standardized data set and transmitting it to the fusion algorithm module.

4. A fire detection sensor data fusion method and system according to claim 1, characterized in that: The fusion algorithm module includes a data fusion unit and a parameter calculation unit; The data fusion unit is used to calculate the preliminary processed data group generated by the data preprocessing module based on the preset multi-sensor data fusion strategy to obtain the initial fire detection coefficient Fx. The specific calculation formula is as follows: ; Wherein, Wd represents temperature, Ynd represents smoke concentration, Qnd represents harmful gas concentration, and Gqd represents photoelectric signal intensity.

5. A fire detection sensor data fusion method and system according to claim 1, characterized in that: The parameter calculation unit is used to calculate the noise sensitivity Ns and the error tolerance deviation index ε during the data fusion process; the noise sensitivity Ns is used to reflect the impact of noise in the sensor data on the fusion result, and is calculated according to the following formula ； In the formula, represents the number of sensors considered in calculating noise sensitivity, represents the noise value of the ith sensor, represents the average value of the noise value; The tolerance deviation index ε is calculated by the following formula: ; In the formula, represents the number of sensors considered when calculating the error tolerance deviation, represents the measurement deviation of the ii-th sensor, Represents the average value of the measurement deviation.

6. A fire detection sensor data fusion method and system according to claim 1, characterized in that: The anomaly detection module includes a standard curve unit and an anomaly alarm generation unit; The standard curve unit is used to generate a standard curve Sx of the initial fire detection coefficient Fx according to historical data and the set detection standard; the standard curve Sx is used to evaluate the normal range of the current fire detection coefficient, and the standard curve value is calculated by the following formula: ; Where a and b are the coefficients obtained from regression analysis.

7. A fire detection sensor data fusion method and system according to claim 1, characterized in that: The abnormal alarm generation unit is used to compare the standard curve value Sx generated by the standard curve unit with the initial fire detection coefficient Fx output by the fusion algorithm module, and calculate the abnormal alarm coefficient Ax according to the following formula: ; In the formula, K represents the standard deviation; By presetting the first deviation threshold T1 and the second deviation threshold T2, the abnormal alarm coefficient Ax is evaluated, and the first deviation threshold T1> the second deviation threshold T2, and the specific contents are as follows: If the abnormal alarm coefficient Ax> the first deviation threshold T1, it is marked as the first abnormal alarm, indicating that there is a fire risk. At this time, the system makes corresponding alarm response measures according to the actual situation; If the first deviation threshold T1 ≥ abnormal alarm coefficient Ax > second deviation threshold T2, it is marked as the second abnormal alarm, and a detailed inspection and confirmation is performed to further verify its authenticity to eliminate the possibility of false alarm; If the second deviation threshold T2 ≥ abnormal alarm coefficient Ax, no abnormal mark is made, indicating that the abnormal alarm coefficient Ax is within the preset normal range, indicating that the current fire detection coefficient Fx does not deviate from the standard curve, and no fire or abnormal situation has occurred; at this time, continue to monitor subsequent data.

8. A fire detection sensor data fusion method and system according to claim 1, characterized in that: The safety fault-tolerant module includes a state monitoring unit and a data compensation unit; The state monitoring unit collects sensor state-related data by monitoring the operating state of the sensor in real time, and obtains the sensor state value Czt and the sensor state average value The sensor state stability coefficient St is calculated by the following formula: ; In the formula, x represents the number of sensors involved in the calculation; The data compensation unit is used to evaluate the sensor state stability coefficient St and start the data compensation mechanism; the evaluation is performed by presetting the stability threshold Q and the sensor state stability coefficient St, and the specific contents are as follows: If the stability threshold Q is less than the sensor state stability coefficient St, it means that the state stability of the sensor does not meet the requirements. At this time, the data compensation mechanism is activated, including adjusting the collected data according to the actual fluctuation of the sensor state; calibrating or replacing the unstable sensor; If the stability threshold Q ≥ sensor state stability coefficient St, it means that the state stability of the sensor meets the requirements, and data compensation is not performed at this time.

9. A fire detection sensor data fusion method and system according to claim 1, characterized in that: The response trigger module obtains the final response coefficient Ry through the following formula: ; In the formula, 2 represents a constant; The final response coefficient Ry is evaluated by presetting the response threshold W. The specific evaluation contents are as follows: If the final response coefficient Ry>response threshold W, it means that the current abnormal alarm and sensor stability are abnormal; therefore, the system will trigger a fire warning response; If the final response coefficient Ry ≤ the response threshold W, it means that the current abnormal alarm and sensor stability are normal; No warning is triggered at this time, and the data continues to be monitored.

10. A fire detection sensor data fusion method, according to claim 1, characterized in that: The following steps are involved: Step 1: monitor the interference factors in the environment where the sensor is located in real time, collect environmental data sets, and generate an environmental interference coefficient Wd. If the environmental interference coefficient Wd exceeds a preset interference threshold T1, an interference signal is transmitted to a data preprocessing module; Step 2: Filter and standardize the raw data collected by the sensor and the environmental data set transmitted by the environmental monitoring module, generate a preliminary processed data set, and then transmit it to the fusion algorithm module; Step 3: Based on the preset multi-sensor data fusion strategy, the preliminary processed data group is fused to generate the initial fire detection coefficient Fx, and the noise sensitivity Ns and its fault tolerance deviation index ε are calculated during the fusion process. The fusion result is transmitted to the anomaly detection module; Step 4: Detect the initial fire detection coefficient Fx output by the fusion algorithm module, generate the abnormal alarm coefficient Ax according to the preset standard curve, and simultaneously fit the abnormal alarm coefficient Ax and the noise sensitivity Ns to obtain the abnormal alarm coefficient Ax and evaluate it. Finally, based on the evaluation result of the abnormal alarm coefficient Ax, determine and confirm the potential false alarm or missed alarm risk; Step 5: monitor the operating status and data transmission process of the sensor in real time, generate and evaluate the sensor state stability coefficient St, and finally start the data compensation mechanism according to the evaluation result of the state stability coefficient St; Step 6: Receive the abnormal alarm coefficient Ax generated by the abnormal detection module and the sensor state stability coefficient St generated by the safety fault-tolerant module, combine the two to generate the final response coefficient Ry and evaluate it, and finally trigger the corresponding fire warning response based on the evaluation result of the final response coefficient Ry.

Images