A multi-sensor-based electrical fire control method and system

Abstract

The present invention relates to the technical field of fire control, and particularly to an electrical fire control method and system based on multiple sensors. The method includes: acquiring relevant data of electrical fire influencing factors and performing preprocessing; extracting current peaks using the AMPD peak detection algorithm based on the preprocessed current data and constructing an arc fault intensity index; combining the arc fault intensity index with temperature data, analyzing the relationship between abnormal heat release of arc faults and temperature changes, constructing an electrical heat dissipation goodness, and comprehensively constructing a fire risk index based on the electrical heat dissipation goodness and the change rules of carbon monoxide and carbon dioxide concentrations to quantify the risk degree of electrical fire occurrence; modifying the SOS anomaly detection algorithm based on the fire risk index to obtain a modified significant anomaly score, and preventing and controlling electrical fire according to the significant anomaly score. The present invention combines multiple sensor data to construct a fire risk index, improving the accuracy of electrical fire early warning and the prevention and control efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of fire control. In particular, it relates to an electrical fire control method and system based on multi-sensors. Background Art

[0002] Electrical fires are fires caused by electrical reasons. Common electrical line faults include short circuits, overloads, and poor contacts, while electrical equipment faults include motor overheating, transformer faults, etc. These faults can cause arcs, electric sparks, or overheating phenomena, which can then ignite surrounding combustibles and lead to fires. Given that electrical fires can cause huge losses and even endanger life safety, it is crucial to prevent and control electrical fires in a timely and effective manner. And sensors, as key tools for detecting fires, play an indispensable role in the process of electrical fire detection.

[0003] Traditional electrical fire detection mostly relies on a single sensor, such as a smoke sensor or a temperature sensor, which can detect abnormal smoke or temperature after a fire occurs. However, this method has obvious limitations: one is the detection delay. Only when the fire develops to a certain stage and the smoke concentration or temperature reaches the threshold can the alarm be triggered, which is likely to miss the initial control opportunity; the second is the high false alarm rate. Non-fire factors in the environment may cause false alarms; the third is the lack of sensitivity to early fires, making it difficult to detect abnormalities in a timely manner when potential hazards first appear.

[0004] With the development of technology, electrical fire detection algorithms based on multi-sensors have emerged. For example, the SOS anomaly detection algorithm integrates multi-source sensor data such as current, voltage, and temperature, and efficiently analyzes electrical system anomalies to achieve fire detection and control. However, in practical applications, since the development process of electrical fires is a dynamically changing process with obvious chronological characteristics, and the SOS anomaly detection algorithm fails to fully mine and utilize the key information in the time series when processing such data, it will lead to deviations in the detection results, resulting in inaccurate detection effects, and may cause missed alarms or false alarms, etc., affecting its reliability and effectiveness in actual electrical fire prevention and control work. Summary of the Invention

[0005] To solve the problems that the traditional SOS anomaly detection algorithm fails to fully mine the time series characteristics, resulting in deviation and insufficient accuracy of the detection results, missed alarms or false alarms, and affecting the reliability and effectiveness of prevention and control, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a multi-sensor-based electrical fire control method includes: obtaining relevant data of factors affecting electrical fires and performing preprocessing; the relevant data include: current data, temperature data, carbon monoxide concentration data and carbon dioxide concentration data; constructing an arc fault intensity index based on the current change characteristics of the current data in the preprocessed relevant data when an arc fault occurs; analyzing the relationship between abnormal heat release and temperature change of arc faults during the operation of electrical equipment according to the arc fault intensity index, and constructing electrical heat dissipation excellence to reflect the quality of heat dissipation effect of electrical equipment; constructing a fire risk index based on the change rules of electrical heat dissipation excellence and carbon monoxide and carbon dioxide concentrations before the occurrence of an electrical fire to reflect the degree of electrical fire; correcting the SOS anomaly detection algorithm based on the fire risk index to obtain a corrected significant anomaly score, and preventing and controlling electrical fires according to the significant anomaly score.

[0007] By integrating multiple key data such as current, temperature, carbon monoxide and carbon dioxide concentrations, the operating status and fire risk of electrical equipment are evaluated. Compared with the traditional single data detection method, it can more accurately reflect the actual status of electrical equipment, effectively improve detection accuracy, reduce false alarms and missed alarms, and provide a more reliable basis for electrical fire prevention and control; by using the fire risk index to correct the SOS anomaly detection algorithm, it can capture the abnormal trend of electrical equipment in advance and issue an alarm, buying precious time for fire prevention. By quantitatively analyzing the changing patterns of multiple data, an accurate assessment of the degree of abnormality of multiple data before an electrical fire occurs can be achieved, thereby achieving accurate prevention and timely control of electrical fires, avoiding major accidents, and improving the overall prevention and control level of electrical fires.

[0008] Preferably, the pretreatment comprises:

[0009] The median filtering algorithm is used to denoise the relevant data, and the denoised relevant data is normalized using the standard deviation normalization method to obtain the preprocessed relevant data.

[0010] Preferably, the arc fault intensity index includes:

[0011] The current data acquired at the preset frequency is input into the AMPD peak detection algorithm to detect and output the peak value in the current data;

[0012] Take any moment as the moment to be detected, traverse all peak values at the moment to be detected, calculate the average value of the sum of the absolute values of the differences between the peak values of all current data at the moment to be detected and the mean value of the current data at the moment to be detected, and obtain the arc fault intensity index of the current data at the moment to be detected.

[0013] By using the AMPD peak detection algorithm, peaks can be accurately extracted from the current data. These peaks reflect the mutation characteristics of the current. By calculating the average value of the sum of the absolute values of the differences between these peaks and the current mean, the intensity of the arc fault can be quantified, thereby accurately reflecting the degree of the arc fault in the electrical equipment.

[0014] Preferably, the arc fault intensity index includes:

[0015] Input the current data obtained at a preset frequency into the AMPD peak detection algorithm to detect and output the peaks in the current data;

[0016] Taking any moment as the moment to be detected, traverse all the peaks in the moment to be detected, calculate the ratio between the peaks of all the current data in the moment to be detected and the smallest peak among all the peaks in the current data, and take the average value of the sum of the differences between the ratio and 1 as the arc fault intensity index at the moment to be detected.

[0017] By calculating the relative differences between each peak and the smallest peak in the current data and constructing the arc fault intensity index based on this, it can effectively highlight abnormally large peaks, accurately reflect the severity of the arc fault, be more sensitive to abnormal peaks in the current data, contribute to more accurately identifying and quantifying the arc fault intensity, improve the accuracy of electrical system fault diagnosis, and timely discover potential electrical fire risks.

[0018] Preferably, the electrical heat dissipation goodness includes:

[0019] Taking any moment as the moment to be detected, obtain the data of the previous preset number of moments at the moment to be detected as local historical data, use the sigmoid function to traverse the arc fault intensity indexes of two adjacent moments in the local historical data at the moment to be detected, and sum them to obtain the arc fault change trend;

[0020] Traverse the temperature differences between two adjacent moments in the local historical data at the moment to be detected, and use the sigmoid function to map the temperature differences to obtain the temperature change smoothing index; take the average value of the sum of the ratios between all the arc fault change trends and the temperature change smoothing index in the local historical data as the electrical heat dissipation goodness at the moment to be detected.

[0021] By using the sigmoid function to traverse the arc fault intensity indexes in the local historical data, it can smooth the data and capture the change trend of the arc fault, effectively avoid the influence of extreme values on trend judgment, combine the arc fault change trend with the temperature change smoothing index, and obtain the electrical heat dissipation goodness by calculating the average value of the ratios between them, comprehensively reflecting the heat dissipation effect of the electrical equipment, helping to timely discover heat dissipation problems and avoid fires caused by equipment overheating.

[0022] Preferably, the fire risk index includes:

[0023] Taking any moment as the moment to be detected, the sequence composed of the carbon monoxide concentration data at the moment to be detected and the previous preset number of moments is used as the input of the Wilcoxon signed-rank test algorithm, and the values indicating the upward and downward trends of the sequence are obtained. The larger value among the values indicating the upward and downward trends of the selected sequence is taken as the intensity of the change trend of the carbon monoxide concentration; the difference between the carbon dioxide concentration at the moment to be detected and the carbon dioxide concentrations at the previous preset number of moments is calculated and processed using the sigmoid function to obtain the intensity of the change trend of the carbon dioxide concentration. The product of the intensity of the change trend of the carbon monoxide concentration and the intensity of the change trend of the carbon dioxide concentration is divided by the electrical heat dissipation goodness at the moment to be detected to obtain the fire risk index at the moment to be detected.

[0024] Analyze the intensity of the change trends of the carbon monoxide concentration and the carbon dioxide concentration, accurately capture the dynamic changes of the concentrations of the two gases, and provide a key basis for the assessment of the fire wind direction.

[0025] Preferably, the SOS anomaly detection algorithm corrected based on the fire risk index includes:

[0026] The standard deviation normalization method is used to normalize the fire risk index at each moment. The relevant data of the electrical fire influencing factors at each moment are used as the input of the SOS anomaly detection algorithm, and the normalized fire risk index at each moment is used as the time series anomaly weight to improve the SOS anomaly detection algorithm.

[0027] Preferably, the significant anomaly score includes:

[0028] Taking any moment as the moment to be detected, the sum of the time series anomaly weight at the moment to be detected and the initial weight of the SOS anomaly detection algorithm is used as the comprehensive weight, and the product of the comprehensive weight and the anomaly score in the original SOS anomaly detection algorithm is used as the significant anomaly score at the moment to be detected.

[0029] Preferably, the prevention and control of electrical fires based on the significant anomaly score includes:

[0030] In response to the significant anomaly score being greater than the anomaly threshold, there is a risk of electrical fire at this time, and an alarm should be issued in a timely manner for handling to avoid the occurrence of open flames or reduce the fire intensity.

[0031]

[0032] ​Second aspect, an electrical fire control system based on multi-sensors, comprising: a processor and a memory, where the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned electrical fire control method based on multi-sensors is implemented.

[0033] The present invention has the following effects:

[0034] 1. The present invention constructs an arc fault intensity index and an electrical heat dissipation goodness by combining various sensor data, and then forms a fire risk index, which more accurately reflects the actual operating state of electrical equipment and the fire risk, and effectively reduces the missed alarms or false alarms caused by the traditional SOS anomaly detection algorithm's failure to fully mine time series features.

[0035] 2. The present invention corrects the SOS anomaly detection algorithm through the fire risk index, so that the improved algorithm can more sensitively capture the fire occurrence trend of electrical equipment in the time series, issue an alarm in time, provide more sufficient time for the early prevention and control of fires, and enhance the reliability and effectiveness of electrical fire prevention and control. Description of the Drawings

[0036] Figure 1 is a flowchart of the method from step S1 to step S5 in an electrical fire control method based on multi-sensors according to an embodiment of the present invention.

[0037] Figure 2 is a structural block diagram of an electrical fire control system based on multi-sensors according to an embodiment of the present invention. Detailed Embodiments

[0038] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention.

[0039] Referring to Figure 1 , an electrical fire control method based on multi-sensors includes steps S1 to S5, specifically as follows:

[0040] S1: Obtain relevant data of electrical fire influencing factors and perform preprocessing; the relevant data includes: current data, temperature data, carbon monoxide concentration data, and carbon dioxide concentration data.

[0041] The current data is collected by a current sensor, and the sampling frequency is taken as 1 kHz in this application and can be selected according to the situation; the temperature data is collected by a temperature sensor, and the carbon monoxide and carbon dioxide gas concentration data are collected by a fixed gas detector respectively, and the sampling interval is taken as 1 s in this application and can be selected according to the situation.

[0042] The preprocessing includes:

[0043] The relevant data is denoised using a median filtering algorithm, and the denoised relevant data is normalized using the standard deviation normalization method to obtain the preprocessed relevant data.

[0044] Upon further analysis, in an electrical system, an arc fault refers to the phenomenon where an electric arc is generated when current passes through air or other media due to reasons such as insulation damage, poor contact, equipment aging, and external force damage. An electric arc is a high-energy discharge phenomenon that can generate high temperatures and sparks, with the generated temperature capable of reaching several thousand degrees Celsius, sufficient to ignite surrounding combustibles such as cable insulation layers and plastics, thereby triggering a fire. Therefore, it is necessary to monitor arc faults. When an arc fault occurs, the current will suddenly increase near the fault point, generating current components.

[0045] Since the occurrence time of the electric arc is very short, usually at the millisecond level, and in order to be consistent with the rest of the data, in this application, taking 1 second as the interval, the current data within 1 second is used as an example for analysis to construct an arc fault intensity index to reflect the degree of occurrence of the arc fault. The construction process of the arc fault intensity index is as follows:

[0046] S2: Construct an arc fault intensity index based on the current change characteristics when an arc fault occurs in the current data in the preprocessed relevant data.

[0047] The arc fault intensity index includes:

[0048] Input the current data obtained at a preset frequency into the AMPD peak detection algorithm to detect and output the peaks in the current data;

[0049] Taking any moment as the moment to be detected, traverse all the peaks at the moment to be detected, and calculate the average value of the sum of the absolute values of the differences between the peaks of all the current data at the moment to be detected and the mean value of the current data at the moment to be detected, to obtain the arc fault intensity index of the current data at the moment to be detected.

[0050] It should be noted that the AMPD peak detection algorithm is a well-known technology and will not be elaborated here.

[0051] Specifically, the arc fault intensity index satisfies the following relational expression:

[0052] ;

[0053] In the formula, represents the arc fault intensity index of the current data at the th moment, represents the number of current data wave peaks at the th moment, represents the th in the current data at the The peak value of a wave crest, represents the mean value of the current data at the moment.

[0054] That is to say, it reflects the degree of difference between the peak value of the current data and the average current. If the peak value in the current data at the moment is significantly greater than the mean value, that is, the more likely the current suddenly increases at this time, the more in line with the characteristics of an arc fault occurrence. Therefore, the calculated arc fault intensity index is larger.

[0055] In addition, another embodiment further includes:

[0056] Input the current data obtained at a preset frequency into the AMPD peak detection algorithm to detect and output the peak value in the current data;

[0057] Taking any moment as the moment to be detected, traverse all the peak values in the moment to be detected, calculate the ratio between the peak value of all the current data in the moment to be detected and the minimum peak value among all the wave crests in the current data, and take the average value of the sum of the differences between the ratio and 1 as the arc fault intensity index at the moment to be detected.

[0058] Specifically, the arc fault intensity index satisfies the following relational expression:

[0059] ;

[0060] In the formula, represents the arc fault intensity index of the current data at the moment, represents the number of wave crests of the current data at the moment, represents the th peak value of the wave crest in the current data at the moment, represents the minimum peak value among all the wave crests in the current data at the

[0061] That is to say, by using the minimum peak value as a reference, this formula can standardize the measurement of each wave crest, enabling the comparison of the sizes of different wave crests. The relative peak value of each wave crest reflects the size of this wave crest relative to the minimum wave crest. This helps to highlight larger wave crests because they may be related to arc faults.

[0062] According to an embodiment of calculating two arc fault intensity indices, in the first embodiment, by using the mean value as a benchmark, the overall fluctuation of the current data can be better reflected. This is very effective for detecting changes in the overall current level. The calculation of the absolute deviation enables the formula to capture the difference between each peak and the overall average level, which is applicable to detecting abnormal fluctuations in the current data. In the second embodiment, by using the minimum peak value as a benchmark, the relative size of the larger peaks can be more prominently reflected. It is beneficial for detecting extreme values in the current data; the calculation of the relative deviation enables the formula to better capture the relative differences between the peaks, which is applicable to detecting abnormal peaks in the current data.

[0063] S3: Analyze the relationship between abnormal heat release and temperature change of the arc fault during the operation of the electrical equipment according to the arc fault intensity index, and construct the electrical heat dissipation goodness to reflect the quality of the heat dissipation effect of the electrical equipment.

[0064] The electrical heat dissipation goodness includes:

[0065] Taking any moment as the moment to be detected, obtaining the moment data of the previous preset number of moments as local historical data at the moment to be detected, traversing the arc fault intensity indices of two adjacent moments in the local historical data at the moment to be detected by using the sigmoid function, and summing them to obtain the arc fault change trend;

[0066] Traversing the temperature differences between two adjacent moments in the local historical data at the moment to be detected, and using the sigmoid function to map the temperature differences to obtain the temperature change smoothing index; taking the average value of the sum of the ratios between all the arc fault change trends and the temperature change smoothing index in the local historical data as the electrical heat dissipation goodness at the moment to be detected.

[0067] Specifically, the electrical heat dissipation goodness satisfies the following relational expression:

[0068] ;

[0069] In the formula, represents the electrical heat dissipation goodness at the th moment, represents the number of historical moments selected, represents the sigmoid function, represents the arc fault intensity index at the th moment and the previous moments, represents the arc fault intensity index at the th moment before the th moment, represents the temperature at the th moment before the th moment, represents the Before a moment The temperature at a moment.

[0070] That is to say, in an electrical device, if there is an abnormal heat release due to an arc fault while the temperature does not increase significantly, that is And Is larger, while Is smaller, it indicates that the heat dissipation effect of the electrical device is better at this time, and the risk of electrical fire is lower. Therefore, the calculated electrical heat dissipation goodness is larger.

[0071] And Respectively represent the smoothed results of the arc fault intensity at the th moment and the previous moments, as well as the previous moments. Adding these two smoothed values together can be understood as a comprehensive consideration of the arc fault intensity at two adjacent historical moments. It helps to capture the change trend of the arc fault intensity over time, rather than just the intensity at a single moment.

[0072] Reflects the smoothness of the temperature change, avoiding the influence of too large or too small temperature changes on the result. The smoothed processing of the temperature change helps to capture the dynamic characteristics of the temperature change over time. A larger temperature change may indicate a better heat dissipation effect, while a smaller temperature change may indicate a worse heat dissipation effect.

[0073] Exemplarily, , it can be adjusted according to the specific implementation situation.

[0074] Further analysis shows that in the smoldering stage of an electrical fire, since the temperature has not reached the ignition point of the combustible, the intensity of the chemical reaction between the carbon atoms in the combustible and the oxygen atoms in the air is relatively low at this time, and the reaction products are mainly carbon monoxide and less carbon dioxide. Therefore, in the early stage of an electrical fire, the concentration of carbon monoxide increases rapidly, while the concentration of carbon dioxide increases slowly; in the open fire combustion stage of an electrical fire, at this time the combustion is relatively complete, the intensity of the chemical reaction between the carbon atoms and the oxygen atoms is relatively high, and the reaction products are mainly carbon dioxide. At the same time, carbon monoxide can also burn. Therefore, in the open fire combustion stage, the concentration of carbon monoxide will gradually decrease, and the concentration of carbon dioxide will increase rapidly. Therefore, according to the change of the concentration data during an electrical fire, the analysis is as follows:

[0075] S4: Construct a fire risk index based on the electrical heat dissipation goodness, the change rules of carbon monoxide and carbon dioxide concentrations before an electrical fire, to reflect the degree of electrical fire occurrence.

[0076] The fire risk index includes:

[0077] Taking any moment as the moment to be detected, the sequence composed of the carbon monoxide concentration data at the moment to be detected and the previous preset moments is used as the input of the Wilcoxon signed-rank test algorithm, and the value indicating the presence of upward and downward trends in the sequence is obtained. The larger value among the values indicating the presence of upward and downward trends in the selected sequence is taken as the intensity of the change trend of the carbon monoxide concentration; the difference between the carbon dioxide concentration at the moment to be detected and the carbon dioxide concentrations at the previous preset moments is calculated and processed using the sigmoid function to obtain the intensity of the change trend of the carbon dioxide concentration;

[0078] The product of the intensity of the change trend of the carbon monoxide concentration and the intensity of the change trend of the carbon dioxide concentration is divided by the electrical heat dissipation goodness at the moment to be detected to obtain the fire risk index at the moment to be detected.

[0079] That is to say, the Wilcoxon signed-rank test algorithm is a well-known technology in the art and will not be described in detail. The hypothesis testing condition 1 is that the sequence has an upward trend, and the hypothesis testing condition 2 is that the sequence has a downward trend. The outputs are the value of the sequence having an upward trend and the value of the sequence having a downward trend respectively.

[0080] Exemplarily, assume that at a certain moment , we select the carbon dioxide concentration data of historical moments for analysis. Calculate to obtain the short-term change of the carbon dioxide concentration, and then process it through the sigmoid function to obtain . If the result is close to 1, it indicates that the carbon dioxide concentration has increased significantly and the fire risk may rise; if the result is close to 0, it indicates that the concentration change is small or has decreased, and the fire risk is relatively low.

[0081] Assume that at a certain moment , the significance of the upward trend of the carbon monoxide concentration monitoring sequence is 0.7, and the significance of the downward trend is 0.3. Then , indicating that the upward trend of the current carbon monoxide concentration is more significant and the fire risk may increase.

[0082] In this way, the change trends of the carbon monoxide concentration and the carbon dioxide concentration can be better captured, so as to more accurately evaluate the fire risk.

[0083] Specifically, the fire risk index satisfies the following relational expression:

[0084] ;

[0085] In the formula, represents the fire risk index at the moment, represents the sigmoid function, represents the carbon dioxide concentration at the moment, represents the carbon dioxide concentration at the moments before the moment, represents the function for the maximum value, value indicating an upward trend in the carbon monoxide concentration monitoring sequence, value indicating a downward trend in the carbon monoxide concentration monitoring sequence, value, represents the electrical heat dissipation goodness at the moment.

[0086] That is to say, in the smoldering stage of an electrical fire, the concentration of carbon monoxide has an upward trend, and at this time is larger. In the open-flame combustion stage of an electrical fire, the carbon monoxide concentration decreases, and at this time is larger. Therefore, when an electrical fire occurs, is larger. At the same time, the more the carbon dioxide concentration increases compared to that at the moments before the moment, the more serious the electrical fire situation is. If the value of the electrical heat dissipation goodness is small at this time, it indicates that the ventilation and heat dissipation effect of the electrical equipment is poor, and it is more likely to cause the electrical fire to develop in a more serious direction. Therefore, the calculated fire risk index is larger.

[0087] S5: Modify the SOS anomaly detection algorithm based on the fire risk index to obtain the modified significant anomaly score, and prevent and control electrical fires according to the significant anomaly score.

[0088] Use the standard deviation normalization method to normalize the fire risk index at each moment. Use the relevant data of the influencing factors of electrical fires at each moment as the input of the SOS anomaly detection algorithm, and use the normalized fire risk index at each moment as the time series anomaly weight to improve the SOS anomaly detection algorithm.

[0089] The significant anomaly score includes:

[0090] Taking any moment as the moment to be detected, the sum of the time series anomaly weight at the moment to be detected and the initial weight of the SOS anomaly detection algorithm is used as the comprehensive weight, and the product of the comprehensive weight and the anomaly score in the original SOS anomaly detection algorithm is used as the significant anomaly score at the moment to be detected.

[0091] Specifically, the significant anomaly score satisfies the following relational expression:

[0092] ;

[0093] In the formula, represents the significant anomaly score at the th moment after improving the SOS anomaly detection algorithm, 1 represents the initial weight of the SOS anomaly detection algorithm, represents the th moment's time series anomaly weight, represents the anomaly score at the th moment when using the original SOS anomaly detection algorithm.

[0094] That is to say, during the operation of the electrical equipment, the more likely there is a trend of a fire occurring at the th moment in the time series, and the more the data of the influencing factors related to the electrical fire deviate from the normal range, it indicates that the electrical equipment is more likely to catch fire at this time. Therefore, the calculated anomaly score is larger.

[0095] In response to the significant anomaly score being greater than the anomaly threshold, there is a risk of an electrical fire at this time, and an alarm should be issued in a timely manner for processing to avoid the occurrence of an open flame or reduce the fire intensity.

[0096] Exemplarily, the anomaly threshold is 0.2, and the implementer can adjust it according to the specific situation.

[0097] The present invention also provides a multi-sensor-based electrical fire control system. As Figure 2 shown, the system includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a multi-sensor-based electrical fire control method according to the first aspect of the present invention is implemented. The system also includes a communication bus and a communication interface and other components well known to those skilled in the art. Their settings and functions are known in the art, so they will not be elaborated here.

[0098] It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the appended claims.

Claims

1. An electrical fire control method based on multi-sensors, characterized in that, Including: Obtain relevant data on electrical fire influencing factors and perform preprocessing; the relevant data includes: current data, temperature data, carbon monoxide concentration data, and carbon dioxide concentration data; Construct an arc fault intensity index based on the current change characteristics when an arc fault occurs in the current data among the preprocessed relevant data; Analyze the relationship between abnormal heat release and temperature change of the arc fault during the operation of electrical equipment according to the arc fault intensity index, and construct an electrical heat dissipation goodness to reflect the quality of the heat dissipation effect of electrical equipment; Construct a fire risk index based on the change rules of electrical heat dissipation goodness, carbon monoxide, and carbon dioxide concentration before an electrical fire occurs to reflect the degree of occurrence of an electrical fire; Modify the SOS anomaly detection algorithm based on the fire risk index to obtain a modified significant anomaly score, and prevent and control electrical fires according to the significant anomaly score.

2. The electrical fire control method based on multi-sensors according to claim 1, wherein, The preprocessing includes: Denoise the relevant data using a median filtering algorithm, and normalize the denoised relevant data using a standard deviation normalization method to obtain the preprocessed relevant data.

3. A multi-sensor-based electrical fire control method according to claim 1, characterized in that, The arc fault intensity index includes: Input the current data obtained at a preset frequency into the AMPD peak detection algorithm to detect and output the peaks in the current data; Taking any moment as the moment to be detected, traverse all the peaks in the moment to be detected, calculate the average value of the sum of the absolute values of the differences between the peaks of all the current data in the moment to be detected and the average value of the current data in the moment to be detected, and obtain the arc fault intensity index of the current data at the moment to be detected.

4. A multi-sensor-based electrical fire control method according to claim 1, characterized in that, The arc fault intensity index includes: Input the current data obtained at a preset frequency into the AMPD peak detection algorithm to detect and output the peaks in the current data; Taking any moment as the moment to be detected, traverse all the peaks in the moment to be detected, calculate the ratio of the peaks of all the current data in the moment to be detected to the smallest peak among all the peaks in the current data, and take the average value of the sum of the differences between the ratio and 1 as the arc fault intensity index at the moment to be detected.

5. A multi-sensor-based electrical fire control method according to claim 1, characterized in that, The electrical heat dissipation goodness includes: Taking any moment as the moment to be detected, obtain the moment data of the previous preset number of moments as local historical data at the moment to be detected, use the sigmoid function to traverse the arc fault intensity indexes between two adjacent moments in the local historical data at the moment to be detected, and sum them to obtain the arc fault change trend; Traverse the temperature differences between two adjacent moments in the local historical data at the moment to be detected, and use the sigmoid function to map the temperature differences to obtain a temperature change smoothing index; take the average value of the sum of the ratios between all the arc fault change trends and the temperature change smoothing index in the local historical data as the electrical heat dissipation goodness at the moment to be detected.

6. The electrical fire control method based on multiple sensors according to claim 1, wherein, The fire risk index includes: Taking any moment as the moment to be detected, the sequence composed of the carbon monoxide concentration data at the moment to be detected and the previous preset number of moments is used as the input of the Wilcoxon signed-rank test algorithm, and the value indicating the upward and downward trends of the sequence is obtained. The larger value among the values indicating the upward and downward trends of the selected sequence is taken as the intensity of the change trend of the carbon monoxide concentration; the difference between the carbon dioxide concentration at the moment to be detected and the carbon dioxide concentrations at the previous preset number of moments is calculated and processed using the sigmoid function to obtain the intensity of the change trend of the carbon dioxide concentration; Calculate the product of the carbon monoxide concentration change trend intensity and the carbon dioxide concentration change trend intensity and divide it by the electrical heat dissipation goodness at the moment to be detected to obtain the fire risk index at the moment to be detected.

7. A multi-sensor-based electrical fire control method according to claim 1, characterized in that, The modifying the SOS anomaly detection algorithm based on the fire risk index includes: Normalize the fire risk index at each moment using the standard deviation normalization method. Use the relevant data of the electrical fire influencing factors at each moment as the input of the SOS anomaly detection algorithm, and use the normalized fire risk index at each moment as the time series anomaly weight to improve the SOS anomaly detection algorithm.

8. A multi-sensor-based electrical fire control method according to claim 1, characterized in that The significant anomaly score includes: Taking any moment as the moment to be detected, use the sum of the time series anomaly weight at the moment to be detected and the initial weight of the SOS anomaly detection algorithm as the comprehensive weight, and use the product between the comprehensive weight and the anomaly score in the original SOS anomaly detection algorithm as the significant anomaly score at the moment to be detected.

9. A multi-sensor-based electrical fire control method according to claim 1, characterized in that, Preventing and controlling electrical fires according to the significant anomaly score includes: In response to the significant anomaly score being greater than the anomaly threshold, there is a risk of electrical fire at this time, and an alarm should be issued in a timely manner for processing to avoid open flames or reduce the fire intensity.

10. An electrical fire control system based on multi-sensors, characterized in that, Including: A processor and a memory, where the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the multi-sensor-based electrical fire control method according to any one of claims 1-9 is implemented.

Images