A method and system for detecting an indoor fire source signal

Abstract

The embodiment of the application provides a kind of indoor fire source signal detection method and system, it is related to fire source signal detection technical field, method includes first obtaining the electrical signal data of multiple parallel power supply branches;The electrical signal data of multiple parallel power supply branches is identified, in the case where at least two electrical signal data of parallel power supply branches is identified as abnormal signal, the correlation index between the abnormal signal is calculated;Upstream common node failure is calculated according to the correlation index, wherein the upstream common node failure represents the abnormal signal with high synchronism;Establish normal power consumption background mode and prevent the abnormal signal from being absorbed into the normal power consumption background mode;The electrical signal data is compared with the normal power consumption background mode, and the alarm of fire source is triggered.The embodiment of the application can improve the accuracy of fire source detection.

Description

Technical Field

[0001] This application relates to the field of fire source signal detection technology, and in particular to an indoor fire source signal detection method and system. Background Technology

[0002] In related technologies, traditional electrical fault detection methods often rely on manual inspections or simple overload and short-circuit protection, making it difficult to effectively identify and locate hidden, progressively developing electrical fire hazards. This is especially true in complex multi-story buildings where electrical wiring is intricate, and some critical electrical connections may be hidden in inaccessible locations. When these hidden connections experience physical degradation such as poor contact, they may initially only produce weak electrical signal anomalies. These anomalies are often insufficient to trigger traditional protection devices in the early stages, but they can gradually worsen over time, eventually leading to electric arcs or even fires. Summary of the Invention

[0003] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes an indoor fire source signal detection method and system, aiming to improve the accuracy of fire source detection.

[0004] In a first aspect, an indoor fire source signal detection method provided in the embodiments of this application includes: Acquire electrical signal data from multiple parallel power supply branches; The electrical signal data of multiple parallel power supply branches are identified, and when the electrical signal data of at least two of the parallel power supply branches are identified as abnormal signals, the correlation index between the abnormal signals is calculated. The upstream shared node fault is calculated based on the correlation index, wherein the upstream shared node fault represents the abnormal signal with high synchronization. Establish a normal power consumption background mode and prevent the abnormal signal from being absorbed into the normal power consumption background mode; The electrical signal data is compared with the normal power consumption background mode to trigger a fire source alarm.

[0005] According to some embodiments of this application, establishing a normal power consumption background mode and preventing the abnormal signal from being absorbed into the normal power consumption background mode includes: An abnormal feature isolation region is established based on the abnormal signal, wherein the abnormal feature isolation region includes the feature markers of the abnormal signal; A normal power consumption background mode is established based on the abnormal feature isolation zone, and the abnormal signal is prevented from being absorbed into the normal power consumption background mode.

[0006] According to some embodiments of this application, upstream shared node faults are calculated based on the correlation index, including: Obtain the power characteristics and harmonic characteristics of the abnormal signal; The upstream shared node fault is determined based on the correlation index, the power characteristics, and the harmonic characteristics.

[0007] According to some embodiments of this application, the step of comparing the electrical signal data with the normal power consumption background mode to trigger a fire source alarm includes: The electrical signal data is compared with the normal power consumption background mode to obtain the comparison results; When the comparison result shows fire source characteristics, the difference signal between the electrical signal data and the normal power consumption background mode is obtained; Obtain the spectral characteristics of the difference signal and the instantaneous rate of change of the electrical signal data; The type of fire source characteristic is determined based on the spectral characteristics and the instantaneous rate of change. The severity of the fire source characteristics is determined based on the amplitude of the difference signal or the amplitude of the instantaneous rate of change. Based on the type and severity of the fire source feature, generate alarm information that includes the type and severity of the fire source feature; The alarm for the fire source is triggered based on the alarm information.

[0008] According to some embodiments of this application, after obtaining the difference signal between the electrical signal data and the normal power consumption background mode, the method further includes: Identify the broadband noise or high-frequency oscillation present in the difference signal, obtain the duration of the broadband noise or high-frequency oscillation, and obtain the energy distribution of the broadband noise or high-frequency oscillation in different high-frequency sub-bands; Obtain load change information for the multiple parallel power supply branches during the period when broadband noise or high-frequency oscillation occurs; Differentiation information is obtained based on the duration, energy distribution, and load change information.

[0009] According to some embodiments of this application, obtaining the distinguishing information based on the duration, the energy distribution, and the load change information includes: Obtain a preset feature library, which includes duration patterns of non-fire source events and fire source features, energy distribution fingerprints, and load change correlation patterns. The duration, energy distribution, and load change information are matched with patterns in the preset feature library to obtain matching results; The distinguishing information is obtained based on the matching results.

[0010] According to some embodiments of this application, determining the type of the fire source characteristic based on the spectral characteristics and the instantaneous rate of change includes: Based on the spectral characteristics and the instantaneous rate of change, a preliminary judgment result is obtained; If the preliminary judgment result is unclear or indicates that the fire source characteristics are weak, high-frequency energy concentration analysis is performed on the spectral characteristics to obtain the high-frequency energy concentration analysis results. Waveform morphology analysis was performed on the instantaneous rate of change to obtain the waveform morphology analysis results; Within a preset time window, the analysis results of the high-frequency band energy concentration and the waveform morphology analysis are continuously evaluated to obtain a continuous evaluation result; Based on the high-frequency energy concentration analysis results, the waveform morphology analysis results, and the continuity assessment results, the type of the fire source characteristics is determined.

[0011] According to some embodiments of this application, the step of performing high-frequency energy concentration analysis on the spectral features to obtain high-frequency energy concentration analysis results includes: The spectral features are divided into multiple high-frequency sub-bands; Calculate the energy within the multiple high-frequency sub-bands based on the multiple high-frequency sub-bands; Based on the distribution of energy in multiple high-frequency sub-bands, the high-frequency band energy concentration analysis results are obtained.

[0012] According to some embodiments of this application, the step of performing waveform morphology analysis on the instantaneous rate of change to obtain waveform morphology analysis results includes: Obtain the morphological characteristics of the instantaneous rate of change waveform, wherein the morphological characteristics include peak value, slope, periodicity index, and symmetry index; Based on the randomness of the peak value and the slope within a preset time window, the randomness analysis results are obtained; Based on the degree of non-periodicity of the aforementioned periodic indicators, the results of the non-periodicity degree analysis are obtained; Based on the degree of asymmetry of the aforementioned symmetry index, the results of the asymmetry degree analysis are obtained; Based on the results of the randomness analysis, the results of the non-periodicity analysis, and the results of the asymmetry analysis, the waveform morphology analysis results are obtained.

[0013] Secondly, embodiments of this application provide an indoor fire source signal detection system, including: The acquisition module is used to acquire electrical signal data from multiple parallel power supply branches; The identification module is used to identify the electrical signal data of multiple parallel power supply branches, and to calculate the correlation index between the abnormal signals when the electrical signal data of at least two of the parallel power supply branches are identified as abnormal signals. The calculation module is used to calculate the upstream shared node fault based on the correlation index, wherein the upstream shared node fault represents the abnormal signal with high synchronization. A module is established to establish a normal power consumption background mode and prevent the abnormal signal from being absorbed into the normal power consumption background mode. The triggering module is used to compare the electrical signal data with the normal power consumption background mode and trigger an alarm for the fire source.

[0014] According to the technical solution of the embodiments of this application, at least the following beneficial effects are achieved: The embodiments of this invention first acquire electrical signal data of multiple parallel power supply branches; identify the electrical signal data of the multiple parallel power supply branches; when the electrical signal data of at least two of the parallel power supply branches are identified as abnormal signals, calculate the correlation index between the abnormal signals; calculate the upstream shared node fault based on the correlation index, wherein the upstream shared node fault characterizes the abnormal signal with high synchronization; establish a normal power consumption background mode and prevent the abnormal signal from being absorbed into the normal power consumption background mode; compare the electrical signal data with the normal power consumption background mode to trigger a fire source alarm. The embodiments of this application can improve the accuracy of fire source detection.

[0015] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0016] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0017] Figure 1 This is a flowchart illustrating an indoor fire source signal detection method provided in one embodiment of this application. Figure 2 A schematic diagram of the process for establishing a normal power consumption background mode according to an embodiment of this application; Figure 3 This is a schematic diagram illustrating the process of obtaining an upstream shared node fault, provided as an embodiment of this application. Figure 4 This is a schematic diagram of a process for triggering an alarm for a fire source, provided in one embodiment of this application. Figure 5This is a schematic diagram illustrating the subsequent process of obtaining the difference signal between electrical signal data and normal power consumption background mode, as provided in one embodiment of this application. Figure 6 This is a schematic diagram of the process for obtaining differentiation information provided in one embodiment of this application; Figure 7 A flowchart illustrating the process of determining the type of fire source characteristics provided in one embodiment of this application; Figure 8 A flowchart illustrating the process of obtaining high-frequency energy concentration analysis results according to one embodiment of this application; Figure 9 This is a schematic flowchart illustrating the process of obtaining waveform morphology analysis results according to one embodiment of this application; Figure 10 This is a schematic diagram of an indoor fire source signal detection system provided in one embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical methods, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. It should be noted that the meaning of "multiple" (or "more than") in the description of the embodiments of this application refers to two or more, and "greater than," "less than," "exceeding," etc. are understood to exclude the number itself, while "above," "below," "within," etc. are understood to include the number itself. If "first," "second," etc. are used in the description, they are only for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated or the order of the technical features indicated. In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, or the existence of B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: the existence of a alone, the existence of b alone, the existence of c alone, the simultaneous existence of a and b, the simultaneous existence of a and c, the simultaneous existence of b and c, or the simultaneous existence of a, b, and c, where a, b, and c can be single or multiple.

[0019] In the description of this application, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0020] Based on the above, this application proposes an indoor fire source signal detection method and system, aiming to improve the accuracy of fire source detection.

[0021] The indoor fire source signal detection method provided in this application can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms; the software can be an application that implements the indoor fire source signal detection method, but is not limited to the above forms. This application can be applied to numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via communication networks. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices. It should be noted that in various specific embodiments of this invention, when processing is required based on data related to the characteristics of an object (e.g., user attributes or sets of attribute information), permission or consent from the corresponding object is obtained first, and the collection, use, and processing of this data comply with relevant laws and standards. Furthermore, when the embodiments of the present invention need to obtain the attribute information of an object, they will obtain the separate permission or separate consent of the corresponding object through pop-up windows or redirection to a confirmation page. After obtaining the separate permission or separate consent of the corresponding object, they will then obtain the relevant data of the object necessary for the embodiments of the present invention to operate normally.

[0022] See Figure 1 , Figure 1 This is a flowchart illustrating an indoor fire source signal detection method according to an embodiment of this application. The indoor fire source signal detection method provided in this embodiment includes, but is not limited to, steps S110 to S150, which are described below. Step S110: Obtain electrical signal data from multiple parallel power supply branches; Step S120: Identify the electrical signal data of multiple parallel power supply branches. If the electrical signal data of at least two parallel power supply branches are identified as abnormal signals, calculate the correlation index between the abnormal signals. Step S130: Calculate the upstream shared node fault based on the correlation index, wherein the upstream shared node fault represents an abnormal signal with high synchronization. Step S140: Establish a normal power consumption background mode and prevent abnormal signals from being absorbed into the normal power consumption background mode; Step S150: Compare the electrical signal data with the normal power consumption background mode to trigger the fire source alarm.

[0023] It should be noted that the embodiments of this application lay the foundation for comprehensive system status monitoring by continuously acquiring electrical signal data from multiple parallel power supply branches. When the system identifies at least two branches simultaneously exhibiting abnormal signals, it immediately calculates the correlation index between these abnormal signals. Since upstream shared node failures typically lead to highly synchronized abnormal signals in multiple downstream parallel branches, by analyzing this correlation, the system can accurately determine whether these abnormal signals originate from a common upstream fault point. This avoids misjudging multi-point synchronous anomalies as independent regional interference. After identifying the upstream shared node failure, a normal power consumption background pattern is further established, and crucially, signals identified as abnormal are prevented from being absorbed into this normal power consumption background pattern. This mechanism ensures the purity of the normal power consumption background pattern, avoiding the problem of background pattern contamination caused by incorrect learning of fault signals in traditional systems. Ultimately, by comparing real-time electrical signal data with this pure, normal power background mode, the system can accurately detect anomalies that do not conform to the normal mode and indicate a fire source, and trigger corresponding alarms, thus effectively solving the problem of traditional methods failing to detect complex electrical faults.

[0024] In one embodiment, the system collects electrical signal data such as voltage, current, and power in real time through smart meters or sensors deployed on each parallel power supply branch. This data is then transmitted to a central processing unit (CPU). The CPU uses a built-in anomaly detection algorithm, based on statistical thresholds or machine learning models, to analyze the electrical signal data of each branch in real time, identifying abnormal fluctuations that deviate from the normal range. When at least two branches simultaneously exhibit abnormal signals, the system immediately invokes a correlation analysis module to calculate the cross-correlation coefficient or dynamic time warping distance between these abnormal signals to assess their synchronicity and similarity. If the correlation index exceeds a preset threshold, the system determines that there is a fault in an upstream shared node. When establishing a normal power consumption background pattern, the system continuously collects electrical signal data during normal operation and constructs the pattern using a pattern recognition algorithm. Simultaneously, any data identified as an abnormal signal is marked and isolated to ensure that this abnormal data is not used to update or train the normal power consumption background pattern, by maintaining an abnormal feature list or isolation zone. When new electrical signal data enters the system, it is compared in real time with this "purified" normal power consumption background pattern, and the signal differences or pattern matching degree can be calculated. When the comparison results show a significant deviation from the characteristics of the fire source, the system will immediately trigger a fire source alarm and issue an alert to the user via an audible and visual alarm, SMS notification, or network platform.

[0025] It is worth noting that the embodiments of this application can effectively identify and locate common node faults hidden upstream of multiple parallel power supply branches. By preventing abnormal signals from being incorrectly absorbed into the normal power consumption background mode, this application avoids the system's mislearning of fault signals and ensures the accuracy of the normal power consumption background mode. This allows the system to accurately detect fire sources even when there is partial overlap between the actual fire source signal and the previously misjudged "normal" background, significantly improving the accuracy and reliability of indoor fire source signal detection and reducing the risk of electrical fires.

[0026] See Figure 2 , Figure 2 This is a schematic diagram of the process for establishing a normal power consumption background mode according to an embodiment of this application; regarding the above step S140 of establishing a normal power consumption background mode and preventing abnormal signals from being absorbed into the normal power consumption background mode, it includes, but is not limited to, steps S210 to S220, and each step will be described in turn below. Step S210: Establish an abnormal feature isolation zone based on the abnormal signal, wherein the abnormal feature isolation zone includes the feature markers of the abnormal signal; Step S220: Establish a normal power consumption background mode based on the abnormal characteristics isolation zone and prevent abnormal signals from being absorbed into the normal power consumption background mode.

[0027] It should be noted that the abnormal feature isolation area is a logical or physical area used to store and manage abnormal signal feature information. It can be a database table, a memory structure, or a file. Its purpose is to clearly distinguish the features of abnormal signals from normal power consumption data. The feature tag of an abnormal signal is a data label or set used to uniquely identify and describe the characteristics of an abnormal signal. It can be the frequency range, amplitude range, waveform pattern, duration, occurrence frequency, or specific harmonic components of the abnormal signal. Its purpose is to classify and record the identified abnormal signals in detail so that these known abnormal features can be accurately identified and excluded when constructing a normal power consumption background pattern.

[0028] In one embodiment, when the system detects a persistent specific high-frequency harmonic component or a non-periodic transient voltage drop, the system activates an anomaly feature analysis module. This module can perform Fourier transform on these anomaly signals to obtain their spectral characteristics, or perform wavelet analysis to extract time-frequency features. When the detected high-frequency noise is caused by poor contact, the system extracts its energy distribution pattern within a specific high-frequency sub-band and uses it as a feature marker for the anomaly signal. These feature markers, such as a vector containing energy values ​​across multiple frequency bands, are stored in a dedicated "anomaly feature database" as part of an anomaly feature isolation zone. The database can be a relational database, where each record contains a unique identifier for an anomaly event and its corresponding feature marker. Subsequently, when establishing a normal power consumption background pattern, the system periodically collects electrical signal data from each branch. Before processing this data to construct a baseline model of normal power consumption (e.g., average power curve, current waveform template, or statistical distribution model), the system first compares the currently collected data with the feature markers in the anomaly feature isolation zone. The system can run a pattern matching algorithm to calculate the similarity between the spectral characteristics of the current data and all feature markers in the anomaly feature database. If the similarity between a feature in the current data segment and a certain abnormal feature marker in the database exceeds a preset threshold, the data segment will be identified as abnormal data and excluded from the training dataset of the normal power consumption background pattern. When calculating the average current for a certain period, if the data within that period is identified as containing marked abnormal harmonic features, the data points for that period will not be included in the calculation of the average value. In this way, it can be ensured that the final normal power consumption background pattern is based on clean data free from abnormal interference, thus more accurately reflecting the actual normal power consumption situation.

[0029] See Figure 3 , Figure 3 This is a schematic diagram illustrating the process of obtaining the fault of the upstream shared node according to an embodiment of this application. The above step S130, which calculates the fault of the upstream shared node based on the correlation index and obtains a verification score, includes, but is not limited to, steps S310 to S320. Each step will be described in turn below. Step S310: Obtain the power characteristics and harmonic characteristics of the abnormal signal; Step S320: Based on the correlation indicators, power characteristics, and harmonic characteristics, the upstream shared node fault is obtained.

[0030] It should be noted that power characteristics are the energy distribution features of an abnormal signal in the time or frequency domain, such as its instantaneous power, average power, and power spectral density. These can be analyzed by performing Fourier transform or wavelet transform on the abnormal signal to examine its energy intensity in different frequency ranges, or by calculating the root mean square value and peak power of the signal. The purpose is to quantify the energy intensity and variation pattern of the abnormal signal. Harmonic characteristics are the features of the non-fundamental frequency components contained in the abnormal signal, such as the frequency, amplitude, phase, and total harmonic distortion rate of harmonics. These are identified and quantified by performing spectral analysis on the abnormal signal to identify its energy distribution at integer multiples of the fundamental frequency. The purpose is to identify whether the abnormal signal is caused by a specific type of nonlinear load or fault.

[0031] In one embodiment, for an anomalous signal with a long duration, its average power within a specific time window can be calculated, and its fluctuation pattern observed. Simultaneously, the harmonic characteristics of the anomalous signal can be extracted using Fast Fourier Transform or Wavelet Transform, allowing analysis of the signal's amplitude and phase at integer multiples of the 50Hz fundamental frequency (e.g., 150Hz, 250Hz, etc.), and calculation of the total harmonic distortion rate. Subsequently, a rule-based expert system or a trained classifier comprehensively utilizes the calculated correlation index, power characteristics, and harmonic characteristics to determine whether an upstream shared node fault exists. When anomalous signals from multiple branches exhibit high correlation (e.g., correlation coefficient greater than 0.8), and their power characteristics show continuous power drops or periodic fluctuations, while their harmonic characteristics contain significant odd harmonics or broadband noise components, the system can determine that an upstream shared node fault exists. Conversely, when the correlation is high but the power and harmonic characteristics do not conform to the preset fault mode, it may be classified as other types of interference. Through this multi-dimensional feature fusion, the system can effectively distinguish synchronization anomaly signals from different sources, thereby improving the accuracy of fault identification of upstream shared nodes.

[0032] See Figure 4 , Figure 4 This is a schematic flowchart illustrating the process of triggering a fire source alarm according to an embodiment of this application. The step S310, which compares electrical signal data with the normal power consumption background mode to trigger the fire source alarm, includes, but is not limited to, steps S410 to S470. Each step will be described in turn below. Step S410: Compare the electrical signal data with the normal power consumption background mode to obtain the comparison results; Step S420: When the comparison result shows fire source characteristics, obtain the difference signal between the electrical signal data and the normal power consumption background mode; Step S430: Obtain the spectral characteristics of the difference signal and the instantaneous rate of change of the electrical signal data; Step S440: Determine the type of fire source characteristics based on spectral features and instantaneous rate of change; Step S450: Determine the severity of the fire source characteristics based on the amplitude of the difference signal or the amplitude of the instantaneous rate of change; Step S460: Generate alarm information containing the type and severity of the fire source characteristics based on the type and severity of the fire source characteristics. Step S470: Trigger the alarm for the fire source based on the alarm information.

[0033] It should be noted that fire source characteristics are specific patterns or abnormal manifestations in electrical signal data that significantly deviate from normal power consumption patterns and indicate potential electrical fire risks. These characteristics can be represented in various forms, such as nonlinear distortion of the signal, high-frequency noise, intermittent interruptions, or abnormal temperature rises. The difference signal is the numerical or waveform difference between the real-time acquired electrical signal data and the normal power consumption background pattern. Specifically, it can be obtained by subtracting or dividing the real-time signal from the background pattern signal, aiming to highlight abnormal signal components and filter out normal background interference. The spectral characteristics are the energy distribution, frequency components, and intensity of the difference signal in the frequency domain. Specifically, it can be obtained through signal processing methods such as Fourier transform, wavelet transform, or short-time Fourier transform, aiming to reveal the composition of different frequency components in the signal to distinguish different types of fire source signals. The instantaneous rate of change is the speed at which the electrical signal data changes value within a very short time. Specifically, it can be obtained by performing differentiation operations on the signal or calculating the slope between adjacent sampling points, aiming to capture time-domain characteristics such as signal abrupt changes, spikes, or rapid oscillations. The types of fire source characteristics are distinguished based on the specific manifestations of electrical signals, classifying fire sources into different categories, such as electric arcs, overheating, poor contact, or insulation breakdown. The purpose is to categorize fire sources so that targeted treatment measures can be taken. The severity of fire source characteristics refers to the degree of potential hazard the fire source poses to the electrical system or environment. This can be quantified by assessing indicators such as the amplitude of the differential signal, the amplitude of the instantaneous rate of change, the duration, or the energy level. The purpose is to assess the hazard level of the fire source and guide the priority of alarm responses.

[0034] In one embodiment, when comparing electrical signal data with a normal power consumption background pattern, an anomaly detection method based on a machine learning model can be used. A support vector machine or neural network model is trained to learn the electrical signal characteristics under normal power consumption patterns. When real-time collected electrical signal data is input into the model, the model outputs an anomaly score or classification result as the comparison result. When this comparison result indicates the presence of fire source characteristics, and when the anomaly score exceeds a preset threshold, the system acquires the difference signal between the electrical signal data and the normal power consumption background pattern. The difference signal can be obtained by subtracting the real-time collected voltage or current waveform data points from the corresponding normal background waveform data points point by point.

[0035] See Figure 5 , Figure 5 This is a schematic diagram illustrating the subsequent process of obtaining the difference signal between electrical signal data and the normal power consumption background mode according to an embodiment of this application. Following step S420, which involves obtaining the difference signal between the electrical signal data and the normal power consumption background mode, steps S510 to S530 are included, and will be described in detail below. Step S510: Identify broadband noise or high-frequency oscillations present in the difference signal, obtain the duration of broadband noise or high-frequency oscillations, and obtain the energy distribution of broadband noise or high-frequency oscillations in different high-frequency sub-bands. Step S520: Obtain load change information for multiple parallel power supply branches during periods of broadband noise or high-frequency oscillation; Step S530: Obtain differentiation information based on duration, energy distribution, and load change information.

[0036] It should be noted that identifying broadband noise or high-frequency oscillations in a difference signal refers to using signal processing techniques, such as Fourier transform, wavelet analysis, or digital filtering, to perform spectral or time-frequency analysis on the difference signal. This aims to detect and separate non-periodic or periodic oscillation components with broadband spectral characteristics or concentrated energy in specific high-frequency bands. The purpose is to extract noise or oscillation signals that may interfere with fire source identification from complex electrical signals. The duration of the broadband noise or high-frequency oscillation is the time from the detection of the signal until its disappearance or decay to below a preset threshold. This duration can be obtained using a timer or by counting consecutive sampling points based on a signal amplitude threshold. The energy distribution of the broadband noise or high-frequency oscillation within different high-frequency sub-bands involves dividing the high-frequency range into multiple narrower frequency intervals and calculating the energy or power spectral density of the broadband noise or high-frequency oscillation signal within each sub-band. This can be achieved using a bandpass filter bank or a spectrum analyzer, with the aim of revealing the frequency composition and energy concentration areas of the interference signal. Load change information refers to the fluctuations or trends of electrical parameters such as power, current, or voltage of electrical equipment on multiple parallel power supply branches within a specific time period. This can be obtained by real-time monitoring of meter data or smart socket data on each branch and calculating their rate of change or statistical characteristics over a short period. Distinguishing information, on the other hand, is the basis for judgment or classification results used to distinguish real fire source signals from non-fire source interference signals, derived after comprehensively analyzing the duration, energy distribution, and load change information of broadband noise or high-frequency oscillations. This information can be generated using pattern recognition algorithms, machine learning models, or rule-based expert systems.

[0037] In one embodiment, a Fast Fourier Transform (FFT) can be used to perform spectral analysis on the difference signal, combined with a high-pass filter or band-pass filter bank to identify broadband noise or high-frequency oscillations. A high-frequency threshold can be set, and frequency components exceeding this threshold are considered high-frequency oscillations, which are further divided into multiple high-frequency sub-bands, and the energy within each sub-band is calculated. Simultaneously, by monitoring the continuity of the signal amplitude within a preset time window, the duration of broadband noise or high-frequency oscillations can be obtained. If the signal amplitude exceeds a certain noise threshold for N consecutive sampling points, its duration is recorded. The system can acquire real-time current and voltage data from smart meters or current sensors in each parallel power supply branch and calculate the power change rate or current fluctuation amplitude during the period of broadband noise or high-frequency oscillations. This data serves as load change information, allowing the calculation of the percentage change in total power of each branch within one second before and after the interference signal appears. Subsequently, the system can use a pre-trained classification model, such as a Support Vector Machine (SVM) or neural network, to obtain distinguishing information based on the acquired duration, energy distribution, and load change information. The model can be trained on a large amount of historical data, which includes characteristic parameters of known fire source events and non-fire source interference events (such as welding, motor starting, and switch operation). When the duration is extremely short, the energy distribution is concentrated in a specific high-frequency sub-band, and it is accompanied by a large load change in a certain branch, the model may classify it as "non-fire source interference"; conversely, when the duration is long, the energy distribution exhibits broadband characteristics and is not significantly related to load changes, it may be classified as "potential fire source".

[0038] See Figure 6 , Figure 6 This is a schematic flowchart illustrating the process of obtaining differentiation information according to an embodiment of this application. The process of obtaining differentiation information based on duration, energy distribution, and load change information in step S530 includes, but is not limited to, steps S610 to S630, which will be described in detail below. Step S610: Obtain a preset feature library, which includes duration patterns of non-fire source events and fire source features, energy distribution fingerprints, and load change correlation patterns. Step S620: Match the duration, energy distribution, and load change information with patterns in the preset feature library to obtain the matching results; Step S630: Obtain the distinction information based on the matching results.

[0039] It should be noted that the pre-built feature library refers to a pre-established and stored dataset containing typical feature patterns of known electrical events (including non-fire source events and fire source features). This dataset can be stored in the form of a database, file system, or memory structure. Among these, the duration pattern describes the range, variation, or typical time-series characteristics of different electrical events over time. It can be represented using time series models, statistical distribution parameters, or discrete time intervals, aiming to capture the persistence characteristics of events. The energy distribution fingerprint describes the energy distribution characteristics of different electrical events at different frequencies or frequency bands. It can be represented using spectrum diagrams, frequency band energy ratios, or energy values ​​at specific frequency points, aiming to reflect the event's... Frequency characteristics; among them, load change correlation pattern describes the relationship or synchronicity characteristics of load changes (e.g., current, power) among multiple parallel power supply branches when an electrical event occurs. It can be represented by correlation coefficient matrix, load curve similarity, or load change trend vector, etc., and its purpose is to reveal the common impact of the event on the load of multiple branches; among them, matching is the process of comparing and evaluating the real-time acquired duration, energy distribution, and load change information with various patterns stored in a preset feature library. It can be implemented by distance metrics (e.g., Euclidean distance, cosine similarity), classification algorithms (e.g., support vector machine, neural network), or rule reasoning, etc., and its purpose is to determine which known event pattern is most similar to or matches the real-time information.

[0040] It is worth noting that the system no longer relies solely on simple feature judgment, but instead performs pattern matching between the real-time detected electrical signal features and a pre-set feature library containing rich prior knowledge. This allows the system to perform more refined and comprehensive signal analysis, effectively identifying events that resemble fire source features but are essentially non-fire source events. When an electrical device generates electrical features that are partially similar to fire source signals in duration or energy distribution, the system can accurately identify it as a non-fire source event by matching it with the corresponding non-fire source event patterns in the feature library, thus avoiding false alarms caused by feature similarity. This significantly improves the accuracy and reliability of fire source detection, reduces the false alarm rate, and ensures the effectiveness of alarms.

[0041] See Figure 7 , Figure 7 This is a flowchart illustrating the process of determining the type of fire source characteristics according to an embodiment of this application. The step S440, which determines the type of fire source characteristics based on spectral characteristics and instantaneous change rate, includes, but is not limited to, steps S710 to S750, which will be described in detail below. Step S710: Based on the spectral characteristics and instantaneous rate of change, a preliminary judgment result is obtained; Step S720: When the preliminary judgment result is unclear or the indication of fire source characteristics is weak, perform high-frequency energy concentration analysis on the spectral characteristics to obtain the high-frequency energy concentration analysis results. Step S730: Perform waveform morphology analysis on the instantaneous rate of change to obtain the waveform morphology analysis results; Step S740: Within a preset time window, perform a continuity assessment on the high-frequency band energy concentration analysis results and waveform morphology analysis results to obtain the continuity assessment results; Step S750: Determine the type of fire source characteristics based on the high-frequency energy concentration analysis results, waveform morphology analysis results, and continuity assessment results.

[0042] It should be noted that high-frequency energy concentration analysis examines the energy distribution in the high-frequency portion of the signal spectrum to assess the degree of energy concentration in a specific high-frequency region. This can be achieved by calculating the energy proportion or energy entropy of different high-frequency sub-bands, aiming to extract the distinguishing characteristics of different fire source types in the high-frequency band. Waveform morphology analysis examines the structure of the time-domain waveform of the instantaneous rate of change signal to identify its unique shape, periodicity, or randomness. This can be achieved using Fourier transform, wavelet analysis, or pattern recognition algorithms, aiming to distinguish the instantaneous signal characteristics generated by different fire source events. Continuity assessment judges the persistence or stability of the analysis results over a period of time to eliminate the influence of instantaneous interference or occasional noise. This can be achieved using moving average, state machine, or time series analysis, aiming to improve the reliability of fire source characteristic judgment. The preset time window is a fixed or variable time interval set for continuous assessment. It can be determined according to the actual application scenario and the duration characteristics of the fire source signal, aiming to ensure the timeliness and effectiveness of the assessment process.

[0043] In one embodiment, when the system receives electrical signal data and compares it with a normal power consumption background pattern, if a fire source characteristic is present, it acquires the spectral characteristics of the difference signal and the instantaneous rate of change of the electrical signal data. First, a preliminary judgment module can use machine learning models such as support vector machines or decision trees to quickly generate a preliminary judgment result based on the initial parameters of these spectral characteristics and the instantaneous rate of change. If the confidence level of this preliminary judgment result is lower than a preset threshold, or the strength of the indicated fire source characteristic signal is lower than a specific multiple of the background noise, the system will initiate a refined analysis process. For spectral characteristics, a high-frequency energy analyzer can divide it into multiple high-frequency sub-bands and calculate the energy within each sub-band. Subsequently, by calculating the concentration index of these energies in different sub-bands, the high-frequency energy concentration analysis result is obtained. For the instantaneous rate of change, a waveform analyzer can extract the morphological characteristics of the waveform, such as peak amplitude, the slope of the rising and falling edges, the periodicity index indicating the presence of repeating patterns, and the symmetry index indicating whether the waveform is symmetrical. These morphological characteristics can be acquired using digital signal processing techniques. Next, a continuous evaluation module continuously monitors the high-frequency energy concentration analysis results and waveform morphology analysis results within a preset time window. This module can use methods such as sliding window averaging or state duration counters to determine whether these analysis results continuously meet specific conditions throughout the entire time window. If they are met, a continuous evaluation result is generated.

[0044] See Figure 8 , Figure 8 This is a flowchart illustrating the process of obtaining high-frequency energy concentration analysis results according to an embodiment of this application. The above step S720, which involves performing high-frequency energy concentration analysis on the spectral characteristics to obtain the high-frequency energy concentration analysis results, includes, but is not limited to, steps S810 to S830. Each step will be described in turn below. Step S810: Divide the spectral features into multiple high-frequency sub-bands; Step S820: Calculate the energy within the multiple high-frequency sub-bands based on the multiple high-frequency sub-bands; Step S830: Based on the energy distribution in multiple high-frequency sub-bands, obtain the high-frequency band energy concentration analysis results.

[0045] It should be noted that the distribution of energy in multiple high-frequency sub-bands is a set of energy values ​​calculated in each high-frequency sub-band and their relative magnitudes. It can be represented as an energy spectral density map, an energy bar chart, or an energy percentage list. Its purpose is to intuitively or quantitatively reflect the degree of energy concentration or dispersion of the fire source signal in the high-frequency band.

[0046] In one embodiment, this application can effectively extract the fine features of fire source signals from complex electrical signals. Even when the initial judgment is unclear or the fire source characteristics are weak, it can provide a valid basis for determining the type of fire source. This, combined with the previous overall detection process of acquiring electrical signal data, identifying abnormal signals, calculating correlations, establishing normal power consumption background patterns, and comparing them, allows for accurate identification of the fire source type even in the presence of background noise or weak signals in the spectral characteristics of the difference signal. This is achieved through in-depth analysis of the energy concentration in the high-frequency band, thus avoiding missed detections due to signal ambiguity or submersion, and effectively improving the accuracy and reliability of the entire indoor fire source signal detection method.

[0047] See Figure 9 , Figure 9 This is a schematic flowchart illustrating the process of obtaining waveform morphology analysis results according to an embodiment of this application. Regarding step S730, which involves performing waveform morphology analysis on the instantaneous rate of change to obtain waveform morphology analysis results, this includes, but is not limited to, steps S910 to S950, which will be described in detail below. Step S910: Obtain the morphological characteristics of the instantaneous rate of change waveform, wherein the morphological characteristics include peak value, slope, periodicity index and symmetry index; Step S920: Based on the randomness of the peak value and slope within the preset time window, obtain the randomness analysis results; Step S930: Based on the degree of non-periodicity of the periodic indicators, obtain the results of the non-periodicity degree analysis; Step S940: Based on the degree of asymmetry of the symmetry index, obtain the asymmetry degree analysis results; Step S950: Based on the results of randomness analysis, aperiodicity analysis, and asymmetry analysis, the waveform morphology analysis results are obtained.

[0048] It should be noted that the morphological characteristics of the instantaneous rate of change waveform refer to the specific shape and structural properties of the instantaneous rate of change signal in the time domain, which can be achieved using digital signal processing techniques. The peak value is the maximum or minimum amplitude reached by the instantaneous rate of change waveform within a specific time period, which can be implemented using extremum detection algorithms in signal processing. The slope is the rate of change of the instantaneous rate of change waveform at a specific point in time or within a time period, which can be implemented using difference operations or derivative calculations, aiming to reflect the steepness of the signal's rise or fall. The periodicity index is a quantitative parameter used to measure the regularity of the recurrence of the instantaneous rate of change waveform, which can be calculated using methods such as autocorrelation functions, Fourier transforms, or wavelet analysis, aiming to assess whether the signal has a regular repetitive pattern. The symmetry index is a quantitative parameter used to measure the degree of evenness of the distribution of the instantaneous rate of change waveform on both sides of a reference point (e.g., the zero axis or the average value), which can be calculated using methods such as statistical moments (e.g., skewness) or the energy ratio of the upper and lower halves of the waveform. Its purpose is to identify deviations or imbalances in the signal in different directions.

[0049] In one embodiment, when acquiring the morphological characteristics of the instantaneous rate of change waveform, a digital signal processor can be used to process the acquired instantaneous rate of change data in real time. For example, for the peak value, a threshold can be set, and when the instantaneous rate of change exceeds the threshold, its maximum value is recorded; for the slope, the difference between adjacent sampling points can be calculated and divided by the sampling time interval to obtain the instantaneous slope. The periodicity index can be obtained by performing a Fast Fourier Transform (FFT) on the spectrum of the instantaneous rate of change signal, judging periodicity by observing whether there are obvious fundamental frequency and harmonic components in the spectrum, or by using an autocorrelation function to detect the repetition pattern of the signal. The symmetry index can be obtained by calculating the energy ratio or area ratio of the positive and negative half-cycles of the waveform, or by calculating the skewness coefficient of the waveform. Then, when obtaining the randomness analysis results based on the randomness of the peak value and slope within a preset time window, a sliding time window, such as 50 milliseconds, can be set, and the frequency and distribution of the peak value and slope can be statistically analyzed within this time window. If the location and magnitude of the peaks and slopes exhibit high irregularity within a time window, such as not conforming to the characteristics of a Gaussian or Poisson distribution, then their degree of randomness can be determined to be high.

[0050] In one embodiment, when obtaining the aperiodicity analysis result based on the degree of aperiodicity of a periodicity index, the decay rate of the signal's autocorrelation function at different delay times can be calculated. If the autocorrelation function decays rapidly and has no obvious periodic peaks, it indicates a high degree of aperiodicity. Alternatively, the spectral entropy of the signal can be calculated; a higher entropy value generally indicates stronger aperiodicity. Next, when obtaining the aperiodicity analysis result based on the degree of aperiodicity of symmetry index, the ratio of the area of ​​the instantaneous rate of change waveform above and below the zero axis can be calculated. If this ratio significantly deviates from 1, it can be determined that the degree of aperiodicity is high. For arc signals, the positive or negative instantaneous impacts may be more significant, leading to waveform asymmetry. Finally, the results of randomness analysis, aperiodicity analysis, and aperiodicity analysis are combined to obtain the waveform morphology analysis result. This can be achieved using multi-feature fusion algorithms, such as machine learning models like support vector machines (SVM), neural networks, or decision trees. These analysis results can be used as input features to train the model and output the final waveform morphology analysis result. This result can be a classification label (e.g., "high randomness, high non-periodicity, high asymmetry") or a comprehensive score to characterize the overall morphological features of the instantaneous rate of change waveform, thereby providing a more refined basis for judging the type of fire source characteristics.

[0051] See Figure 10 , Figure 10 This is a schematic diagram of an indoor fire source signal detection system according to an embodiment of this application. The indoor fire source signal detection system 1000 includes: The acquisition module 1010 is used to acquire electrical signal data from multiple parallel power supply branches; The identification module 1020 is used to identify the electrical signal data of multiple parallel power supply branches. When the electrical signal data of at least two parallel power supply branches are identified as abnormal signals, the correlation index between the abnormal signals is calculated. Calculation module 1030 is used to calculate the upstream shared node fault based on the correlation index, wherein the upstream shared node fault represents an abnormal signal with high synchronization. Module 1040 is established to establish a normal power consumption background mode and prevent abnormal signals from being absorbed into the normal power consumption background mode. The trigger module 1050 is used to compare electrical signal data with the normal power consumption background mode and trigger an alarm for the fire source.

[0052] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0053] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically include computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium. The foregoing has provided a detailed description of the preferred embodiments of this application. However, this application is not limited to the above-described embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined in this application.

Claims

1. A method for detecting indoor fire source signals, characterized in that, include: Acquire electrical signal data from multiple parallel power supply branches; The electrical signal data of multiple parallel power supply branches are identified, and when the electrical signal data of at least two of the parallel power supply branches are identified as abnormal signals, the correlation index between the abnormal signals is calculated. The upstream shared node fault is calculated based on the correlation index, wherein the upstream shared node fault represents the abnormal signal with high synchronization. Establish a normal power consumption background mode and prevent the abnormal signal from being absorbed into the normal power consumption background mode; The electrical signal data is compared with the normal power consumption background mode to trigger a fire source alarm; The establishment of a normal power consumption background mode and prevention of abnormal signals from being absorbed into the normal power consumption background mode includes: An abnormal feature isolation region is established based on the abnormal signal, wherein the abnormal feature isolation region includes the feature markers of the abnormal signal; A normal power consumption background mode is established based on the abnormal feature isolation zone, and the abnormal signal is prevented from being absorbed into the normal power consumption background mode.

2. The method according to claim 1, characterized in that, The calculation of upstream shared node faults based on the correlation index includes: Obtain the power characteristics and harmonic characteristics of the abnormal signal; The upstream shared node fault is determined based on the correlation index, the power characteristics, and the harmonic characteristics.

3. The method according to claim 1, characterized in that, The step of comparing the electrical signal data with the normal power consumption background mode to trigger a fire source alarm includes: The electrical signal data is compared with the normal power consumption background mode to obtain the comparison results; When the comparison result shows fire source characteristics, the difference signal between the electrical signal data and the normal power consumption background mode is obtained; Obtain the spectral characteristics of the difference signal and the instantaneous rate of change of the electrical signal data; The type of fire source characteristic is determined based on the spectral characteristics and the instantaneous rate of change. The severity of the fire source characteristics is determined based on the amplitude of the difference signal or the amplitude of the instantaneous rate of change. Based on the type and severity of the fire source feature, generate alarm information that includes the type and severity of the fire source feature; The alarm for the fire source is triggered based on the alarm information.

4. The method according to claim 3, characterized in that, After acquiring the difference signal between the electrical signal data and the normal power consumption background mode, the method further includes: Identify the broadband noise or high-frequency oscillation present in the difference signal, obtain the duration of the broadband noise or high-frequency oscillation, and obtain the energy distribution of the broadband noise or high-frequency oscillation in different high-frequency sub-bands; Obtain load change information for the multiple parallel power supply branches during the period when broadband noise or high-frequency oscillation occurs; Differentiation information is obtained based on the duration, energy distribution, and load change information.

5. The method according to claim 4, characterized in that, The distinction information obtained based on the duration, energy distribution, and load change information includes: Obtain a preset feature library, which includes duration patterns of non-fire source events and fire source features, energy distribution fingerprints, and load change correlation patterns. The duration, energy distribution, and load change information are matched with patterns in the preset feature library to obtain matching results; The distinguishing information is obtained based on the matching results.

6. The method according to claim 3, characterized in that, Determining the type of fire source characteristic based on the spectral characteristics and the instantaneous rate of change includes: Based on the spectral characteristics and the instantaneous rate of change, a preliminary judgment result is obtained; If the preliminary judgment result is unclear or indicates that the fire source characteristics are weak, high-frequency energy concentration analysis is performed on the spectral characteristics to obtain the high-frequency energy concentration analysis results. Waveform morphology analysis was performed on the instantaneous rate of change to obtain the waveform morphology analysis results; Within a preset time window, the analysis results of the high-frequency band energy concentration and the waveform morphology analysis are continuously evaluated to obtain a continuous evaluation result; Based on the high-frequency energy concentration analysis results, the waveform morphology analysis results, and the continuity assessment results, the type of the fire source characteristics is determined.

7. The method according to claim 6, characterized in that, The high-frequency energy concentration analysis of the spectral characteristics, to obtain the high-frequency energy concentration analysis results, includes: The spectral features are divided into multiple high-frequency sub-bands; Calculate the energy within the multiple high-frequency sub-bands based on the multiple high-frequency sub-bands; Based on the distribution of energy in multiple high-frequency sub-bands, the high-frequency band energy concentration analysis results are obtained.

8. The method according to claim 6, characterized in that, The waveform morphology analysis of the instantaneous rate of change, to obtain the waveform morphology analysis results, includes: Obtain the morphological characteristics of the instantaneous rate of change waveform, wherein the morphological characteristics include peak value, slope, periodicity index, and symmetry index; Based on the randomness of the peak value and the slope within a preset time window, the randomness analysis results are obtained; Based on the degree of non-periodicity of the aforementioned periodic indicators, the results of the non-periodicity degree analysis are obtained; Based on the degree of asymmetry of the aforementioned symmetry index, the results of the asymmetry degree analysis are obtained; Based on the results of the randomness analysis, the results of the non-periodicity analysis, and the results of the asymmetry analysis, the waveform morphology analysis results are obtained.

9. An indoor fire source signal detection system, characterized in that, include: The acquisition module is used to acquire electrical signal data from multiple parallel power supply branches; The identification module is used to identify the electrical signal data of multiple parallel power supply branches, and to calculate the correlation index between the abnormal signals when the electrical signal data of at least two of the parallel power supply branches are identified as abnormal signals. The calculation module is used to calculate the upstream shared node fault based on the correlation index, wherein the upstream shared node fault represents the abnormal signal with high synchronization. A module is established to establish a normal power consumption background mode and prevent the abnormal signal from being absorbed into the normal power consumption background mode. The establishment of a normal power consumption background mode and prevention of abnormal signals from being absorbed into the normal power consumption background mode includes: An abnormal feature isolation region is established based on the abnormal signal, wherein the abnormal feature isolation region includes the feature markers of the abnormal signal; A normal power consumption background mode is established based on the abnormal feature isolation zone, and the abnormal signal is prevented from being absorbed into the normal power consumption background mode. The triggering module is used to compare the electrical signal data with the normal power consumption background mode and trigger an alarm for the fire source.

Images