Fire-fighting water system fault detection and early warning method and system based on fire-fighting internet of things

Abstract

The present application relates to the technical field of fire control monitoring, and more particularly to a fire water system fault detection and early warning method and system based on a fire control Internet of Things, wherein the method comprises: step 1: periodically collecting pressure data through pressure sensors deployed at the outlet of a fire water pump, branch points of a pipe network trunk and the most unfavorable terminal; step 2: generating a theoretical pressure expected value and a dynamic allowable deviation band for each node according to the current valve opening, pump state and historical normal working condition data; step 3: when the measured pressure deviates from the theoretical pressure expected value beyond the dynamic allowable deviation band, marking an abnormal node and constructing a fault propagation tree in reverse along a topological model, and assigning a weight factor to an associated node according to the fault type; and step 4: triggering a hierarchical early warning based on the number of abnormal nodes, the fault propagation path and the cumulative value of the weight factor. The system can automatically take measures when a fault occurs through a device linkage function, ensuring efficient operation of the fire water system and improving the efficiency and safety of fire emergency response.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fire monitoring, in particular to a fire water system fault detection and early warning method and system based on a fire Internet of Things. BACKGROUND

[0002] As a core component of building fire facilities, the reliability of the fire water system is directly related to the effectiveness of fire emergency response. The current mainstream monitoring system deploys pressure sensors at key nodes in the pipe network and uses a fixed threshold value to trigger an alarm mechanism (such as triggering when the pressure is below 0.5 MPa). However, under the complex branched pipe network topology, this mechanism has the following inherent defects:

[0003] High single parameter misjudgment rate:

[0004] Since pipe leakage, pump group failure, and sensor drift are all manifested as pressure drop, the system cannot distinguish the root cause of the fault. For example, in a certain large commercial complex project, sensor false alarms caused an average of 3.2 times of misactivation of the standby pump per month, resulting in an annual maintenance cost increase of 150,000 yuan. The technical rationality of this problem lies in the fact that the fire water system is a closed pressure vessel, and any fault at any location will be transmitted to the entire system through pressure waves, and traditional single-point threshold detection will inevitably confuse the fault type.

[0005] Lack of fault location capability:

[0006] Existing technologies can only detect pressure abnormalities, but cannot trace the fault source. When a branch pipe leakage causes fluctuations in the main pipe pressure, maintenance personnel need to manually check all branch valves, which takes an average of more than 2 hours. The necessity of this problem lies in the fact that the branched pipe network has a unique fault propagation path (the pressure attenuation gradient upstream of the leakage point is significantly higher than that downstream), but existing methods do not establish a topological correlation model, resulting in low fault isolation efficiency.

[0007] Serious response lag:

[0008] The fixed threshold value mechanism needs to wait for the pressure to drop to the critical value before it can alarm, and it cannot detect gradual small leaks of less than 10%. Experimental data show that a DN100 pipe with a 2L / min leakage requires 6-8 hours to trigger an alarm, and the leakage can reach 1 ton during this period. The reasonable cause of the lag lies in the fact that small leaks are initially masked by system noise, and traditional methods lack dynamic deviation analysis capability.

[0009] Therefore, there is an urgent need for a fire water system fault detection and early warning method and system based on a fire Internet of Things to solve the above problems. SUMMARY

[0010] In order to achieve the above purposes, the present application provides a fire water system fault detection and early warning method and system based on a fire Internet of Things, characterized by comprising:

[0011] Step 1: Collecting pressure data periodically through pressure sensors deployed at the outlet of fire water pumps, branch points of pipe network trunk and the most unfavorable terminal, collecting system total flow through electromagnetic flow meters of pump group outlet pipes, collecting branch flow through ultrasonic flow meters after zone control valves, and obtaining valve opening state and pump running state through valve opening degree sensors and pump current sensors respectively;

[0012] Step 2: Establishing pressure conduction relationship based on pipe network topology node-edge relationship model, generating each node theoretical pressure expected value and dynamic allowable deviation band according to current valve opening, pump state and historical normal working condition data;

[0013] Step 3: When the measured pressure deviates from the theoretical pressure expected value beyond the dynamic allowable deviation band, marking abnormal nodes and constructing fault propagation tree along the topology model in reverse, and assigning weight factors to associated nodes according to fault types;

[0014] Step 4: Triggering hierarchical early warning based on abnormal node number, fault propagation path and weight factor cumulative value:

[0015] Single node anomaly and weight not spreading trigger level I alarm;

[0016] Adjacent node anomaly and fault mode matching trigger level II alarm;

[0017] Key node anomaly and weight cumulative value exceeding adaptive threshold trigger level III equipment linkage.

[0018] Preferably, the determination method of the time interval for periodically collecting pressure data in step 1 comprises:

[0019] Obtaining pipe material elastic modulus, fluid density and pipe network maximum segment length;

[0020] Calculating the propagation speed of water hammer wave in the pipe segment, which is proportional to the square root of the pipe material elastic modulus and inversely proportional to the square root of the fluid density;

[0021] Dividing the pipe network maximum segment length by the water hammer wave propagation speed to obtain the critical time of pressure wave transmission;

[0022] Setting the sampling interval to be less than one half of the critical time to avoid distortion of pressure shock signal.

[0023] Preferably, the generation method of the dynamic allowable deviation band in step 2 comprises:

[0024] Extracting pressure-flow data pairs of consecutive time windows under historical normal working conditions;

[0025] Calculating the residual standard deviation of pressure measured value and theoretical value in each time window;

[0026] The boundary coefficient is dynamically adjusted according to the recent false alarm rate statistics based on the moving average of the residual standard deviation:

[0027] When the number of consecutive false alarms increases, the boundary coefficient is expanded by a fixed proportion;

[0028] When the missed alarm event occurs, the boundary coefficient is contracted by the proportion of the fault impact range.

[0029] Preferably, the allocation rule of the weight factor in step 3 includes:

[0030] The pump node weight factor is assigned according to the correlation between the pressure drop and the current change:

[0031] If the pressure drop is accompanied by current overrun, assign the first weight level;

[0032] If the pressure slowly drops with current fluctuation, assign the second weight level;

[0033] The pipeline node weight factor is assigned according to the pressure gradient change rate:

[0034] When the upstream pressure gradient steeply drops, assign the value according to the proportion of the gradient change rate and the historical maximum value;

[0035] The valve node weight factor is assigned according to the delay time of the opening command and the flow response:

[0036] When the delay time exceeds the preset threshold multiple, the weight is linearly increased with the delay time.

[0037] Preferably, the verification method of fault mode matching in step 3 includes:

[0038] a: Construct a fault feature mapping rule library, which includes:

[0039] The determination condition of pipeline leakage is that the abnormal node pressure suddenly drops, the associated downstream branch flow shows an upward trend, and the pump group current is within the normal fluctuation range;

[0040] The determination condition of pump group efficiency decline is that the abnormal node pressure appears persistent slow drop, the system total flow remains unchanged, and the pump group current exceeds the rated working interval;

[0041] The determination condition of valve jamming is that the abnormal node pressure fluctuates irregularly, the associated branch flow has no response to the valve opening command, and the pump group current is not abnormal;

[0042] Collect the real-time parameter change direction of the abnormal node and its direct topological associated node, and when the parameter combination matches the determination condition in any fault feature mapping rule library, perform similarity calculation:

[0043] b: Extract the derivative sign of the pressure change trend as the first matching factor;

[0044] c: extracting the Boolean value of the flow response direction as the second matching factor;

[0045] d: extracting the current over-limit status flag as the third matching factor;

[0046] When the joint confidence of the three matching factors exceeds the dynamic similarity threshold, the fault type is confirmed.

[0047] Preferably, the dynamic adjustment method of the adaptive threshold in step 4 comprises:

[0048] The initial threshold is based on the average value of the historical fault weights of the key nodes;

[0049] When a level III false positive occurs, the fault propagation path depth is obtained;

[0050] When the path depth is greater than the set level, the threshold is increased by a value proportional to the depth;

[0051] When the path depth is less than the set level, the threshold is increased by a fixed step value;

[0052] When a level III false negative occurs, a weight compensation value is calculated based on the actual impact range of the fault, and the threshold is adjusted inversely by the compensation value.

[0053] Preferably, the calculation process of the water hammer wave propagation speed comprises:

[0054] The slope of the stress-strain curve is obtained through a pipe material elastic modulus test experiment;

[0055] The real-time medium density is obtained through a fluid density detection device;

[0056] The ratio of the pipe wall thickness to the pipe diameter is input into the elastic deformation correction model, and the wave speed correction coefficient is output;

[0057] The actual propagation speed is calculated in combination with the reference speed of the sound wave in the fluid.

[0058] Preferably, the determination method of the similarity threshold comprises:

[0059] The cosine similarity of the parameter combination in the historical fault sample to the standard mode is counted;

[0060] The 75th percentile of the similarity distribution is taken as the initial threshold;

[0061] Each time a fault sample is added, the threshold is adjusted according to the deviation of the sample from the standard mode:

[0062] If the sample is correctly classified and the similarity is higher than the current threshold, the threshold remains unchanged;

[0063] If the sample is misclassified, the threshold is lowered or raised according to the type of classification error.

[0064] Preferably, the method for constructing the fault propagation tree in step 3 comprises:

[0065] Taking the abnormal node as a root node, expanding child nodes layer by layer along the inverse water flow direction of the topological model;

[0066] Calculating the influence weight of each edge:

[0067] Obtaining the proportion of the pipe segment pressure drop change amount in the total pressure drop change amount of the system;

[0068] Obtaining the proportion of the pipe segment length in the total length of the tracing path;

[0069] Taking the product of the two proportions as the edge weight;

[0070] Selecting the path with the largest edge weight as the main fault propagation chain;

[0071] The execution of the Ⅲ-level equipment linkage in step 4 comprises:

[0072] When starting the standby pump:

[0073] Obtaining the water hammer sensitivity coefficient between the fault pump and the standby pump;

[0074] Calculating the maximum allowable flow rate change rate according to the sensitivity coefficient;

[0075] Controlling the standby pump to be pressurized at the allowable change rate gradient;

[0076] When adjusting the valve:

[0077] Based on the fault node pressure recovery target value, the required flow is backstepped;

[0078] According to the valve flow characteristic curve, the opening adjustment amount is calculated;

[0079] Segmented step adjustment and real-time verification of pressure recovery progress.

[0080] Correspondingly, the embodiment of the present application also provides a fire water system fault detection and early warning system based on a fire fighting Internet of Things, which comprises a memory configured to store instructions, and a processor configured to call the instructions from the memory and capable of realizing the fire water system fault detection and early warning method based on the fire fighting Internet of Things as any of the embodiments of the present application.

[0081] The present application has the following beneficial effects:

[0082] 1. The multi-parameter cooperative detection method proposed by the present application, the system not only relies on pressure data, but also introduces multi-dimensional data such as flow, valve opening, pump group current, etc., combined with dynamic allowable deviation band analysis, when pressure abnormalities are found, the fault type can be accurately distinguished through matching of historical data and system state, thereby greatly reducing the misjudgment rate. Through the accurate construction of weight factor and fault propagation path, false alarms caused by single parameter are avoided, and the system can accurately determine the fault type in the complex fire water pipe network.

[0083] 2. The present application can accurately trace the fault source and mark the abnormal node by constructing a node-edge relationship model based on the topology of the pipe network, real-time monitoring the changes of pressure, flow and pump group state, and constructing a fault propagation tree in reverse according to the abnormal node. This method greatly improves the accuracy and efficiency of fault location, avoids the time-consuming manual inspection problem of existing methods, significantly reduces the work burden of operation and maintenance personnel, and improves the efficiency of fault isolation and processing.

[0084] 3. The present application can sensitively respond to small changes in the system and timely discover small amplitude leaks or faults through dynamic adjustment of the allowable deviation band and analysis of the fault propagation tree. In the early stage of small leakage, the system can optimize the early warning mechanism through dynamic adjustment of residual standard deviation, comparison of historical data and false alarm rate statistics, and significantly improve the response speed. BRIEF DESCRIPTION OF DRAWINGS

[0085] In order to more clearly illustrate the technical solutions in the present application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, for those skilled in the art, other drawings can also be obtained based on these drawings without creative labor.

[0086] Fig. 1 The step flow chart of the method of the present application;

[0087] Fig. 2 The step flow chart of the weight factor allocation rule in step 3 of the method of the present application;

[0088] Fig. 3 The step flow chart of the calculation process of the water hammer wave propagation speed of the method of the present application. DETAILED DESCRIPTION

[0089] The present application will be described in detail below in combination with the drawings and specific embodiments. It should be noted here that in order to make the embodiments more detailed, the following embodiments are the best, preferred embodiments, and other alternative ways can also be used by those skilled in the art to implement; and the drawings are only used to more specifically describe the embodiments, and are not intended to specifically limit the present application.

[0090] Please refer to Figs. 1-3 The embodiment of the present application provides a fire-fighting water system fault detection and early warning method based on a fire-fighting Internet of Things. In step 1, first, the system deploys pressure sensors at the outlet of a fire-fighting water pump, branch points of a pipe network trunk, and the most unfavorable terminal position, periodically collects pressure data in the pipe network. At the same time, the total flow of the entire system is collected through an electromagnetic flowmeter installed on the pump group outlet header, the flow information of each branch is obtained through an ultrasonic flowmeter installed behind the zone control valve, and the valve opening and pump running state are further obtained through a valve opening sensor and a pump current sensor. This data collection process provides comprehensive real-time information, including pressure, flow, valve state, and pump running state, which helps to comprehensively monitor the running state of the fire-fighting water system.

[0091] In step 2, on the basis of data collection, the system establishes a pressure conduction relationship of the pipe network based on a node-edge relationship model of the pipe network topology. By analyzing the current valve opening and pump state, combining historical normal working condition data, the theoretical pressure expected value of each node is generated, and a dynamic allowable deviation band is set according to the system characteristics. The core of this step is to accurately simulate the pressure change law by using the topological structure relationship of the pipe network, and dynamically adjust the theoretical pressure expected value according to the real-time state and historical data of the system, so that the system can real-time evaluate the pressure deviation of each node.

[0092] In step 3, when the measured pressure deviates from the theoretical expected value beyond the dynamic allowable deviation band, the system will mark the node as an abnormal node, and reversely construct a fault propagation tree along the pipe network topology model. According to the specific type of the fault (such as pipe leakage, pump failure, etc.), the system will assign a weight factor to the associated node, so as to further analyze the propagation path and influence range of the fault. In this way, the system not only can detect pressure abnormalities, but also can quickly identify the fault source by analyzing the pressure wave propagation path, thereby improving the fault positioning capability.

[0093] In step 4, on the basis of fault detection and propagation path construction, the system triggers a hierarchical warning according to the number of abnormal nodes, the fault propagation path, and the weight factor cumulative value of each node. Specifically, if a single node is abnormal and the fault has not spread, the system will trigger level I alarm; if the adjacent nodes are abnormal and the fault mode matches, the system will trigger level II alarm; if the key node is abnormal and the weight cumulative value exceeds the adaptive threshold, the system will trigger level III alarm, and start the equipment linkage response, automatically adjusting the corresponding equipment or starting the backup measures. Through this multi-level warning mechanism, the system can take appropriate response measures in time according to the severity and extent of the fault, avoiding the expansion of potential disasters.

[0094] Through multi-dimensional data collection and analysis, the problems of high single parameter misjudgment rate, difficult fault location, and response lag in existing systems are solved. Through accurate modeling of the pipe network topology model and pressure conduction relationship, the system can accurately monitor small leaks and initial faults, realize rapid positioning of the fault source, and provide a weighted factor-based hierarchical warning mechanism, effectively reducing false alarms and missed alarms. At the same time, the system can automatically take measures to ensure the efficient and reliable operation of the fire water system when a fault occurs, greatly improving the efficiency and safety of fire emergency response.

[0095] In one possible implementation, first, the system needs to obtain the basic physical parameters of the pipe network, including the elastic modulus of the pipe material, the density of the fluid, and the maximum segment length of the pipe network. The elastic modulus describes the ability of the pipe material to resist deformation, while the fluid density affects the inertia of the fluid, and the maximum segment length of the pipe network determines the distance of pressure wave propagation.

[0096] After obtaining the elastic modulus of the pipe material and the fluid density, the system calculates using the water hammer wave propagation speed formula. According to the formula, the propagation speed of the water hammer wave in the pipe segment is proportional to the square root of the elastic modulus of the pipe material and inversely proportional to the square root of the fluid density. Through this relationship, the system can accurately calculate the propagation speed of the water hammer wave in the pipe network, providing a theoretical basis for subsequent sampling time interval setting.

[0097] According to the propagation speed of the water hammer wave, the system divides the maximum segment length of the pipe network by the propagation speed to obtain the critical time of pressure wave transmission. This critical time refers to the time required for the pressure wave to propagate from one end of the pipe network to the other end. This time is a key factor affecting the sampling frequency, because if the sampling interval is too long, it may not be able to capture the changes in the pressure wave in time, resulting in data distortion or missing.

[0098] In order to avoid distortion of the pressure oscillation signal, the system sets the sampling interval to be less than half of the critical time. This setting ensures that the sampling frequency is high enough to accurately record the changes in the pressure wave, thereby avoiding the problem of signal distortion caused by too long sampling interval. In this way, the system can capture subtle pressure fluctuations, ensuring the accuracy and real-time nature of fault detection.

[0099] By accurately calculating the propagation speed of the water hammer wave and the transmission time of the pressure wave, and reasonably setting the sampling time interval, the system effectively avoids signal distortion or data loss caused by unreasonable sampling frequency. This design ensures that the system can capture the dynamic changes of the pressure wave in a timely and accurate manner when collecting pressure data, providing more accurate data support for fault detection. At the same time, reasonable sampling interval improves the system's sensitivity to small abnormalities, enhances the early detection capability of fault warning, and helps to improve the overall safety and reliability of the fire water system.

[0100] In one possible implementation, first, the pressure-flow data pairs in the continuous time window under normal working conditions are extracted from the fire water system. Each time window contains pressure measurement values and theoretical values within a period of time, which reflect the running state of the water system under normal working conditions. Through these data, the system can establish the pressure-flow characteristics under normal conditions, thereby providing a benchmark for subsequent anomaly detection.

[0101] For each time window, the system calculates the residual error between the measured pressure and the theoretical value, and then calculates the standard deviation of the residual error. The residual error standard deviation reflects the fluctuation between the actual system running pressure and the theoretical expectation, and is an important indicator for measuring system stability and error range. If the system fluctuates less under normal working conditions, the residual error standard deviation is smaller; otherwise, the standard deviation is larger.

[0102] In order to improve the stability of the data, the system will perform moving average processing on the calculated residual error standard deviation. Through the moving average value, the system can smooth the data fluctuation, reduce the impact of short-term mutations, and make the detection result more reliable. The moving average value will serve as the benchmark of the dynamic allowable deviation band, providing a reference for subsequent error adjustment.

[0103] On the basis of normal working conditions, the system will dynamically adjust the boundary coefficient according to the recent false alarm rate. Specifically, when the system detects an increase in the number of consecutive false alarms, it will expand the boundary coefficient of the allowable deviation band by a fixed proportion to avoid frequent false alarms caused by overly strict allowable deviation settings. This adjustment can effectively reduce the false alarm rate and reduce unnecessary alarm events. Conversely, when the system has a missed alarm event, the system will shrink the boundary coefficient according to the proportion of the fault impact range, thereby reducing the occurrence of missed alarms and improving the sensitivity and detection accuracy of the system to faults.

[0104] This dynamic adjustment mechanism can effectively adapt to different states that occur during system operation by flexibly adjusting the boundary coefficient according to historical data and real-time false and missed alarms, avoiding false and missed alarms that may be caused by fixed deviation band settings. By expanding the deviation band when there is a false alarm and shrinking the deviation band when there is a missed alarm, the system can maintain high early warning accuracy and reduce the frequency of false alarms and missed alarms. At the same time, this method can adjust in real time according to changes in the actual operating environment of the fire water system, making fault detection more accurate, improving the reliability and response efficiency of the system, and enhancing the safety protection capability of the entire fire water system.

[0105] In one possible implementation, the allocation rule of the weight factor assigns values according to different node types and specific abnormal signal characteristics. This rule ensures that different types of faults can be accurately identified, and the corresponding warning strength is given according to the severity of the anomaly, helping the system to achieve more intelligent and accurate fault detection.

[0106] Specifically, the fault detection of the pump node relies on the correlation between pressure and current changes:

[0107] If there is a sudden drop in pressure accompanied by an over-limit current, it indicates that the pump may have a serious fault (such as pump idling, blockage, or mechanical failure), which usually leads to a decrease in system efficiency, so a higher alarm level (i.e., the first weight level) should be given. This allocation rule can ensure that the system responds in time when a critical fault occurs, preventing serious accidents.

[0108] If the pressure slowly decreases and is accompanied by current fluctuations, it usually means that the pump is in an unstable or less serious fault state (such as unstable flow or foreign matter in the pump), which has a smaller impact on the system, so the weight factor is set to the second level, indicating that attention is needed but no serious alarm will be triggered immediately.

[0109] Through the above rules, the pump node's fault warning can be accurately classified according to the degree of pressure and current change, thus achieving dynamic alarm for different fault severities.

[0110] The fault of the pipeline node is usually manifested as a change in the pressure gradient, especially a sharp drop in upstream pressure:

[0111] When the upstream pressure gradient suddenly drops, the system determines the weight factor according to the ratio of the gradient change rate to the historical maximum value. That is, when the pressure gradient changes greatly and far exceeds the historical maximum value, the system will assign a higher weight factor. This is usually a warning signal of serious blockage or rupture of the pipeline, which may cause the pipeline system to fail to operate normally if not handled in time.

[0112] This rule flexibly adjusts the alarm level by comparing historical data, effectively avoiding false alarms caused by large fluctuations in historical data.

[0113] If the delay between the opening command of the valve and the flow response exceeds the preset threshold multiple, the system will linearly increase the weight factor according to the delay time. The delay time usually reflects the efficiency problem of the valve response, and a larger delay may indicate that the valve has a fault such as sticking, leakage, or unstable control system, so the weight needs to be dynamically increased according to the severity of the delay. This rule helps the system to more accurately detect valve control problems, so that appropriate repair measures can be taken in time.

[0114] The method provides an accurate, flexible and efficient solution for intelligent fault detection and early warning of the fire water system, effectively improving the safety and reliability of the system.

[0115] In one possible implementation, first, the system defines clear judgment conditions for common fault types by establishing a fault feature mapping rule library. Each fault mode has a set of pre-defined parameter change characteristics, which are obtained through historical data and fault simulation. For example:

[0116] If the pressure of the abnormal node drops suddenly, and the downstream branch flow associated with it shows an upward trend, while the pump set current is within the normal fluctuation range, it can be determined as a pipeline leak. Pipeline leaks usually cause sudden pressure drops and flow changes, but the pump set current changes little.

[0117] When the pressure of the abnormal node shows a persistent slow drop, and the total flow of the system remains unchanged, while the pump set current exceeds the rated working interval, it can be determined as a pump set efficiency decline. This type of fault indicates that the pump set is running, but the efficiency is not high, possibly due to mechanical damage or blockage, etc.

[0118] If the pressure of the abnormal node fluctuates irregularly, and the associated branch flow has no response to the valve opening command, while the pump set current is not abnormal, it can be determined as a valve jam fault. Valve jamming can cause pressure fluctuations, but will not cause current abnormalities.

[0119] During system operation, the change direction of parameters such as pressure, flow and current of the abnormal node and its directly topologically associated nodes is collected in real time. These real-time data are used to match the judgment conditions in the fault feature mapping rule library, and similarity calculation is performed:

[0120] The system extracts the derivative sign of the pressure change trend as the first matching factor. The derivative sign reflects the change direction of the pressure, which is the key to determining whether the pressure change conforms to the fault mode. For example, a sudden drop or slow drop in pressure is an important basis for determining certain faults such as pipeline leaks or pump set efficiency decline.

[0121] The Boolean value of the flow response direction indicates whether the flow changes in the expected direction. For example, in the case of a pipeline leak, the flow usually shows an upward trend, while in other fault types, the flow may remain unchanged or change in other ways.

[0122] The current over-limit state flag is used to identify whether the pump set is in an abnormal current state. For example, when the pump set efficiency declines, the current may exceed the normal range, while in the case of a pipeline leak or valve jam, the current fluctuates little.

[0123] When the system extracts the above three matching factors, a similarity calculation is performed. These matching factors are combined according to their respective weights to generate a comprehensive confidence value. If the combined confidence of the three matching factors exceeds the dynamically set similarity threshold, the fault type is confirmed.

[0124] Through this feature mapping and similarity calculation-based method, the system can achieve more accurate fault diagnosis, ensure the safe and stable operation of the fire water system, and can early warning potential faults, reduce the risk of accidents.

[0125] In one possible implementation, the initialization of the threshold value is based on the historical fault weight average of the key nodes. This means that when the system starts running, the fault weight average of each key node will be calculated according to the historical fault data. This historical weight average provides a basis for threshold value setting, ensuring that the threshold value in the initial stage is reasonable, thereby avoiding excessive or low false judgments.

[0126] When a level III false positive is detected, the system will obtain the depth of the fault propagation path. The fault propagation path depth refers to the length of the transmission path from the node where the fault occurs to other affected nodes in the system. Path depth is an important indicator of the scope of the fault impact. According to the depth of the fault propagation path, the system will dynamically adjust the threshold value:

[0127] When the path depth is greater than the set level: the threshold value will be increased by the proportion of the depth value. The deeper the path, the more extensive the impact of the fault, and the system needs to improve its sensitivity to identify more potential faults, so the threshold value will increase in proportion to the path depth.

[0128] When the path depth is less than the set level: the threshold value is increased by a fixed step value. For short fault propagation paths, the system adjusts the threshold value with a small amplitude, avoiding excessive sensitivity that affects the stability and accuracy of the system.

[0129] When a level III false negative occurs, it means that the system fails to detect or predict the fault in time. At this time, the system will calculate a weight compensation value according to the actual impact range of the fault. The weight compensation value reflects the actual impact of the false negative fault on the system. Generally, a fault with a wider impact range should be given a higher compensation value. Then, the system will inversely proportionally down-regulate the threshold value according to this compensation value. That is, the larger the fault impact range, the lower the threshold value, and the system will become more sensitive and be able to detect similar faults more quickly.

[0130] By dynamically adjusting the threshold value in real time, a more flexible and accurate fault detection and warning method is provided, which can effectively deal with various complex fault scenarios and improve the overall reliability and safety of the system.

[0131] In one possible implementation, first, the stress-strain curve is obtained by conducting an elastic modulus test on the pipe material. In this experiment, different external pressures (stresses) are applied to the pipe, and the deformation (strain) of the pipe is measured. The slope of the stress-strain curve represents the elastic modulus of the material, which is the degree of deformation of the material under pressure. The elastic modulus is one of the basic parameters for calculating the water hammer wave propagation speed, as it directly affects the inertial and elastic response of the water flow in the pipeline.

[0132] The fluid density detection device is used to obtain the density of the medium (water or other liquid) in the pipeline in real time. Fluid density is one of the important factors affecting the water hammer wave propagation speed. The greater the density, the greater the inertia of the fluid, which affects the propagation speed of the water hammer wave. Therefore, it is very important to obtain accurate fluid density data.

[0133] The ratio of pipe wall thickness to pipe diameter is input into the elastic deformation correction model, which will be corrected according to the geometry of the pipeline, and the wave speed correction coefficient will be output. The wall thickness and diameter of the pipeline affect the flow rate of the water flow in the pipeline and the propagation mode of its fluctuations, which in turn affect the propagation speed of the water hammer wave. The correction model can adjust the calculated propagation speed according to these parameters to adapt to the specific conditions of different pipelines.

[0134] Combined with the reference speed of sound waves in the fluid, the actual propagation speed of the water hammer wave in the pipeline is calculated by the known fluid density, pipe material elastic modulus and correction coefficient. The reference speed of sound waves in the fluid usually depends on the physical properties of the fluid (such as density and compressibility), while the propagation speed of the water hammer wave is closely related to the elasticity and geometry of the pipeline.

[0135] This calculation process is dynamic and can be continuously updated according to real-time data such as fluid density, pipeline stress, etc., ensuring that the calculation of the water hammer wave propagation speed is consistent with the actual situation. This means that the system can adapt to real-time changes to ensure the accuracy of the detection.

[0136] This method of calculating the water hammer wave propagation speed not only improves the accuracy of fault detection, but also provides more efficient early warning functions for the system, ensuring the stability and safety of the fire water system in complex operating environments.

[0137] In one possible implementation, first, the system quantifies the similarity between the sample and the standard mode by calculating the cosine similarity between each parameter combination in the historical fault sample and the standard mode. Cosine similarity measures the similarity between two vectors by calculating the cosine of the angle between them, with a value closer to 1 indicating higher similarity. Parameter combinations usually include multiple indicators such as pressure, flow, temperature, etc. By calculating the cosine similarity of these data, we can identify which fault modes are similar to the standard mode and which differences are larger.

[0138] According to the similarity distribution of historical failure samples, the system determines the 75th percentile as the initial similarity threshold. The percentile reflects the distribution of similarity data, and the 75th percentile usually represents a higher similarity value, meaning that only highly similar samples will be considered normal. This setting helps filter out failure samples that are far from the standard mode, avoiding misjudgment of false modes as normal failures.

[0139] With the addition of new failure samples, the system will dynamically adjust the threshold based on the deviation of each new sample from the standard mode. If the similarity of the new sample to the standard mode is high and the classification is correct, the system will maintain the current similarity threshold to ensure system stability. However, if the similarity of the new sample to the standard mode is low and the classification is incorrect, the threshold will be adjusted:

[0140] If the sample is misclassified: According to the type of classification error, the system will correspondingly increase or decrease the threshold. For example, if the classification error is due to a too strict threshold causing a false negative, the system may lower the threshold; conversely, if it is due to a too loose threshold causing a false positive, the threshold may be increased.

[0141] Through this dynamic adjustment, the system can continuously optimize the threshold setting based on real-time data, adapting to different failure types and operating environments. Each threshold adjustment can make the system's response to new failure samples more sensitive and accurate, preventing false positives or false negatives due to unreasonable threshold settings.

[0142] Through the dynamic adjustment of the similarity threshold, the fault detection system can more intelligently and accurately identify fault patterns, enhancing the system's robustness and adaptability, and helping to improve the overall operational safety of the fire water system.

[0143] In one possible implementation, when constructing the fault propagation tree, first identify the abnormal node where the fault occurs, which represents the key location in the system where the fault occurs. Then, according to the topology of the pipe network, expand outward along the opposite direction of the water flow direction layer by layer, find the pipes, valves, pumps and other devices connected to the fault node, and form a fault propagation tree. Each node represents a device or pipe segment, and each child node represents a device or area that may be affected by the fault.

[0144] For each pipe edge connecting the nodes, calculate its influence weight. This process is achieved through two important parameters:

[0145] The proportion of the change in pressure drop of the pipe segment to the total pressure drop change of the system: This measures the proportion of the pressure drop change of a pipe segment in the total pressure drop change of the system, reflecting the impact of the fault of that pipe segment on the system.

[0146] Ratio of pipe segment length to total length of the traceability path: This parameter reflects the relative length of the pipe segment in the entire traceability path during fault propagation. Longer pipe segments generally have a greater impact on fault propagation.

[0147] The product of these two ratio values is used as the edge weight, with larger weight edges indicating that the pipe segment or device has a higher impact on fault propagation.

[0148] By comparing the edge weights of different paths, the path with the largest weight is selected, representing the main fault propagation chain. This path indicates the propagation process of the fault from the initial node to other areas, providing a key reference for subsequent fault diagnosis and response.

[0149] When starting the backup pump:

[0150] Obtain the water hammer sensitivity coefficient between the faulty pump and the backup pump: The water hammer sensitivity coefficient measures the response of the fluid in the pipeline to pressure fluctuations. Before starting the backup pump, the water hammer sensitivity coefficient between the faulty pump and the backup pump needs to be obtained, which will be used for subsequent calculation of the maximum allowed flow rate change.

[0151] Calculate the maximum allowed flow rate change: According to the water hammer sensitivity coefficient and the characteristics of the pipeline, calculate the maximum allowed flow rate change when starting the backup pump. This value determines the startup speed of the backup pump, avoiding excessive water hammer effects caused by starting too quickly, thereby reducing pressure fluctuations in the pipeline.

[0152] Control the backup pump to gradually increase pressure according to the allowed change rate: According to the calculated flow rate change, control the backup pump to gradually increase pressure to avoid water hammer phenomena impacting the pipeline.

[0153] When adjusting the valve

[0154] Back-calculate the required flow based on the fault node pressure recovery target value: When a fault occurs, the system calculates the target value for recovering the pressure of the fault node, and calculates the required flow based on this target value. This is to ensure that the pressure level can be restored to the predetermined level when adjusting the valve.

[0155] Calculate the opening adjustment amount based on the valve flow characteristic curve: According to the flow characteristic curve of the valve, calculate the opening adjustment amount to achieve the required flow. The opening adjustment of the valve determines the change in flow, ensuring accurate adjustment of the system.

[0156] Segmented step adjustment and real-time verification of pressure recovery progress: To avoid instability caused by large adjustments, the valve adjustment usually adopts a segmented step adjustment method, adjusting a small amount each time and monitoring the pressure change in real time. The system verifies the progress of pressure recovery based on real-time pressure data to ensure that the pressure is restored to the target value.

[0157] Based on the construction of the fault propagation tree and the implementation of the Ⅲ-level equipment linkage, the accuracy and efficiency of the fire water system fault detection and early warning can be greatly improved, and the rapid response and safe operation of the fire fighting system in emergency situations can be ensured.

[0158] Correspondingly, the embodiment of the application also provides a fire water system fault detection and early warning system based on a fire fighting Internet of Things, which comprises a memory configured to store instructions, and a processor configured to call the instructions from the memory and capable of realizing the fire water system fault detection and early warning method based on the fire fighting Internet of Things as described in any of the embodiments of the application when executing the instructions.

[0159] The following is described in detail through examples:

[0160] The application relates to a fire water system fault diagnosis and control method, in particular to a system fault processing method based on fault propagation tree model construction and Ⅲ-level equipment linkage control. The specific application scenario is a fire water system of a large industrial park, which is composed of multiple water pumps, pipelines, valves and fire fighting equipment and has a relatively complex topological structure. The application is implemented to quickly locate the system fault source through fault propagation tree construction and edge weight analysis, and to avoid water hammer phenomenon and guarantee the efficient operation of the fire fighting system through Ⅲ-level equipment linkage control.

[0161] The fire water system is composed of the following equipment:

[0162] Water pump: three, model JYW-1000, flow rate 1000 m 3 / h, rated power 75kW.

[0163] Valve: 10, model DVC-50, flow control range 50~200 m 3 / h, maximum opening degree 90°.

[0164] Pipeline: 10 kilometers of steel pipe, diameter 200 mm, maximum flow rate 1200 m 3 / h.

[0165] Sensor: pressure sensor installed at the inlet and outlet of the pipeline and key nodes, sampling frequency 1Hz, measurement accuracy ±0.5%.

[0166] The parameters of these equipment are monitored in real time and used as input for fault diagnosis and control.

[0167] At a certain moment of the embodiment of the application, a pipeline rupture event occurs in the fire water system. Through the topological structure of the system, a fault propagation tree can be constructed. The construction algorithm of the fault propagation tree is as follows:

[0168] System topology: Each node in the graph represents a device (such as a pump, valve, pipe, etc.), and each edge represents the connection between devices.

[0169] Faulty node: Suppose a break occurs in pipe 3 (numbered C3).

[0170] Using the Breadth-First Search (BFS) algorithm, start from the node where the break occurs (C3) and expand outward until all affected devices are marked as fault-related.

[0171] The weight of each edge is calculated according to the following formula:

[0172] ;

[0173] where : Change in pressure drop of the pipe segment (unit: Pa). : Total change in pressure drop of the system (unit: Pa). : Length of the pipe segment (unit: m). : Total length of the fault path (unit: m).

[0174] Suppose the total change in pressure drop of the system is 300 Pa, the change in pressure drop of pipe 3 is 50 Pa, the length of pipe 3 is 100 m, and the total length of the fault path is 500 m. Then the weight of this edge is:

[0175] ;

[0176] According to the above calculation, the weights of all related edges are calculated. The system selects the propagation chain with the largest weight as the main fault path. Suppose the calculated main fault propagation chain is:

[0177] Main propagation chain: .

[0178] The total weight of this propagation chain is 0.1, indicating that this path has the greatest impact on the system and needs to be restored first.

[0179] To avoid water hammer phenomenon and ensure system recovery, it is necessary to adjust the system through the linkage control of level III devices (such as backup pumps and valves). The specific process is as follows:

[0180] Condition for starting backup pump: When the pipe pressure is lower than the set threshold (for example, the set value is 1.0 MPa), the system starts the backup pump.

[0181] Flow calculation formula when backup pump starts:

[0182] ;

[0183] where : current pipe pressure. : minimum pressure. : maximum pressure. : maximum flow.

[0184] In the embodiment of the application, the current pressure is 0.8 MPa, the minimum pressure is 0.5 MPa, the maximum pressure is 1.2 MPa, the maximum flow is 1000 m 3 / h, and the flow of the standby pump is:

[0185] ;

[0186] When starting the standby pump, the system needs to close the valve V5 to reduce the sharp change of the pipe pressure. The flow adjustment of the valve is carried out through the following formula:

[0187] ;

[0188] The flow of the pump is 428.57 m 3 / h, the current pressure is 0.8 MPa, the maximum pressure is 1.2 MPa, and the adjusted flow of the valve is:

[0189] ;

[0190] The formula ensures that the valve adjusts the flow according to the actual needs to avoid the water hammer phenomenon.

[0191] Through the above linkage control, the system can adjust the valve opening degree at the same time when the water pump starts to keep the pipe pressure stable and prevent the generation of water hammer effect. Compared with the traditional manual adjustment method, the automatic control system of the application has faster response speed, higher control precision, and can be dynamically adjusted in real time, reducing the error of human operation.

[0192] Before the implementation, the system recovery time is 30 minutes and the water hammer phenomenon occurs at a rate of 20% using the traditional manual valve adjustment method. After the implementation of the application, the automatic control system can recover the system within 10 minutes and completely avoid the water hammer phenomenon. Through the comparison experiment, the automatic control system has a significant improvement in fault recovery speed and system stability compared with the traditional method.

[0193] The embodiment describes in detail how to use the fault propagation tree model to construct and implement the linkage control method of the Ⅲ-level equipment for fault diagnosis and control. Through the application of specific algorithms, formulas and parameter values, the system can quickly locate the problem and efficiently recover the operation when a fault occurs. Compared with the traditional method, the application significantly improves the response speed and stability of the system, and has outstanding practical application effect.

[0194] The present application encompasses any alternatives, modifications, equivalent methods and solutions made to the essence and scope of the present application. In order to make the public have a thorough understanding of the present application, specific details are described in the following preferred embodiments of the present application, and the present application can also be fully understood without the description of these details to those skilled in the art. In addition, in order to avoid unnecessary confusion to the essence of the present application, well-known methods, processes, procedures, elements and circuits, etc. are not described in detail.

[0195] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can also be made, which should be considered as the protection scope of the present application.

Claims

1. A fire water system fault detection and early warning method based on a fire Internet of Things, characterized in that, Comprise: Step 1: periodically collect pressure data through pressure sensors deployed at the outlet of the fire pump, the branch point of the pipe network main pipe and the most unfavorable end, collect the total flow of the pump group outlet pipe through the electromagnetic flowmeter, collect the branch flow through the ultrasonic flowmeter after the partition control valve, and obtain the valve opening state and pump operation state through the valve opening sensor and pump current sensor respectively; Step 2: based on the node-edge relationship model of the pipe network topology, the pressure conduction relationship is established, and the theoretical pressure expectation value and dynamic allowable deviation band of each node are generated according to the current valve opening, pump state and historical normal working condition data; Step 3: when the measured pressure deviates from the theoretical pressure expectation value beyond the dynamic allowable deviation band, the abnormal node is marked and the fault propagation tree is reversely constructed along the topology model, and the weight factor is assigned to the associated nodes according to the fault type; Step 4: based on the number of abnormal nodes, the fault propagation path and the weight factor cumulative value, trigger the hierarchical early warning: Single node anomaly and weight not spread trigger level I alarm; Adjacent node anomaly and fault mode matching trigger level II alarm; Key node anomaly and weight cumulative value exceeding adaptive threshold trigger level III equipment linkage; The verification method of fault mode matching in step 4 includes: a: construct a fault feature mapping rule library, wherein: The determination condition of pipeline leakage is that the pressure of the abnormal node suddenly drops, at the same time, the flow of the associated downstream branch shows an upward trend, and the pump group current is within the normal fluctuation range; The determination condition of pump group efficiency decline is that the pressure of the abnormal node appears sustained slow decline, the system total flow remains unchanged, and the pump group current exceeds the rated working interval; The determination condition of valve jamming is that the pressure of the abnormal node fluctuates irregularly, the flow of the associated branch does not respond to the valve opening command, and the pump group current does not appear abnormal; Collect the real-time parameter change direction of the abnormal node and its directly topologically associated nodes, when the parameter combination matches the determination condition in any fault feature mapping rule library, perform similarity calculation: b: extract the derivative sign of pressure change trend as the first matching factor; c: extract the Boolean value of flow response direction as the second matching factor; d: extract the current out-of-limit state flag as the third matching factor; When the joint confidence of the three matching factors exceeds the dynamic similarity threshold, the fault type is confirmed.

2. The fire water system fault detection and early warning method based on fire Internet of Things according to claim 1, characterized in that, The determination method of the time interval of periodically collecting pressure data in step 1 includes: Obtain the elastic modulus of the pipe material, the fluid density and the maximum segment length of the pipe network; Calculate the propagation speed of water hammer wave in the pipe segment, which is proportional to the square root of the elastic modulus of the pipe material and inversely proportional to the square root of the fluid density; Divide the maximum segment length of the pipe network by the propagation speed of the water hammer wave to get the critical time of pressure wave transmission; Set the sampling interval to be less than one half of the critical time to avoid distortion of pressure shock signals. 3.The fire water system fault detection and early warning method based on fire Internet of Things according to claim 1, characterized in that, The generation method of dynamic allowable deviation band in step 2 includes: Extract the pressure-flow data pairs in the continuous time window under the historical normal working condition; Calculate the residual standard deviation of the measured pressure value and the theoretical value in each time window; Take the moving average of the residual standard deviation as the benchmark, and dynamically adjust the boundary coefficient according to the recent false alarm rate: When the number of consecutive false positives increases, the boundary coefficient is expanded by a fixed ratio; When a missed event occurs, the boundary coefficient is contracted by the proportion of the fault impact range.

4. The fire water system fault detection and early warning method based on fire Internet of Things according to claim 1, characterized in that, The allocation rules of the weight factor in step 3 include: The pump node weight factor is assigned according to the correlation between the pressure drop amplitude and the current change amount: If the pressure drop is accompanied by current overrun, assign the first weight level; If the pressure drops slowly accompanied by current fluctuations, assign the second weight level; The pipeline node weight factor is assigned according to the pressure gradient change rate: When the upstream pressure gradient steeply drops, assign the weight according to the ratio of the gradient change rate to the historical maximum value; The valve node weight factor is assigned according to the delay time of the opening degree command and the flow response: When the delay time exceeds the preset threshold multiple, the weight is linearly increased by the delay time. 5.The fire water system fault detection and early warning method based on fire Internet of Things according to claim 1, characterized in that, The dynamic adjustment method of the adaptive threshold in step 4 includes: The initial threshold is based on the average value of the historical fault weight of the key node; When a level III false positive occurs, the fault propagation path depth is obtained; When the path depth is greater than the set level, the threshold is increased by the depth value proportion; When the path depth is less than the set level, the threshold is increased by a fixed step value; When a level III missed event occurs, calculate the weight compensation value according to the actual impact range of the fault, and lower the threshold by the compensation value in inverse proportion.

6. The fire water system fault detection and early warning method based on fire Internet of Things according to claim 2, characterized in that, The calculation process of the water hammer wave propagation speed includes: Obtain the stress-strain curve slope through the pipe material elastic modulus test experiment; Obtain the real-time medium density through the fluid density detection device; Input the pipe wall thickness to diameter ratio into the elastic deformation correction model, and output the wave speed correction coefficient; Combine the reference speed of the sound wave in the fluid to calculate the actual propagation speed. 7.The fire water system fault detection and early warning method based on fire Internet of Things according to claim 1, characterized in that, The determination method of the similarity threshold includes: Statistical parameter combination and standard mode cosine similarity in historical fault samples; The 75th percentile of the similarity distribution is used as the initial threshold; Each time a new fault sample is added, adjust the threshold according to the deviation of the sample from the standard mode: If the sample is correctly classified and the similarity is higher than the current threshold, maintain the threshold unchanged; If the sample is misclassified, lower or raise the threshold according to the classification error type. 8.The fire water system fault detection and early warning method based on fire Internet of Things according to claim 1, characterized in that, The construction method of the fault propagation tree in step 3 includes: Take the abnormal node as the root node, and expand the child nodes layer by layer along the inverse water flow direction of the topology model; Calculate the influence weight of each edge: Obtain the proportion of the pipe segment pressure drop change amount to the total pressure drop change amount of the system; Obtain the proportion of the pipe segment length to the total length of the tracing path; The product of the two proportions is used as the edge weight; Select the path with the largest edge weight as the main fault propagation chain. The execution of level III equipment linkage in step 4 includes: When starting the standby pump: Obtain the water hammer sensitivity coefficient between the fault pump and the standby pump; Calculate the maximum allowed flow rate change rate according to the sensitivity coefficient; Control the standby pump to pressurize at the allowed change rate gradient; When adjusting the valve: Backpropagate the required flow based on the fault node pressure recovery target value; Calculate the opening adjustment amount according to the valve flow characteristic curve; Segmented step adjustment and real-time verification of pressure recovery progress.

9. A fire water system fault detection and early warning system based on a fire internet of things, characterized in that, The memory is configured to store instructions, and the processor is configured to call the instructions from the memory and implement the fire fighting water system fault detection and early warning method based on the fire fighting Internet of Things as claimed in any one of claims 1-8 when executing the instructions.

Images