Fire hydrant water use state detection system and detection method

Abstract

The invention discloses a fire hydrant water use state detection system and detection method, and relates to the technical field of ship fire fighting, the system comprises the following components: a physical sensing layer, a data transmission layer, a data processing and analysis layer, a digital twinborn layer and an application layer; according to the invention, the digital twinborn body of the application layer is utilized to simulate the water use state response of the fire hydrant under different working conditions and parameter settings in a virtual environment, the optimal system parameter configuration is found by analyzing key indexes such as pressure stability and flow uniformity, and the optimal system parameter configuration is applied to an actual system, so that the field debugging time is greatly shortened, and the debugging efficiency is improved. The debugging cost is reduced, the performance problem caused by unreasonable parameters is avoided, the overall operation stability of the system is improved, meanwhile, in the aspects of fault diagnosis and deduction, when abnormity is detected, the development processes of different fault scenes are simulated through the digital twin, the fault position can be quickly positioned, and the fault diagnosis and deduction efficiency is improved. And fault causes, influence ranges and propagation paths are analyzed.

Description

Technical Field

[0001] This invention relates to the field of marine fire protection technology, specifically to a fire hydrant water usage status detection system and detection method. Background Technology

[0002] During ship operation, fire safety is a critical link concerning the safety of life and property of ships and personnel. As the core facility of the ship's fire protection system, the accurate monitoring and effective management of the water status of fire hydrants is of utmost importance. With the increasingly complex navigation environment of ships, ships may face different sea conditions and marine environmental factors. Fire hydrants need to operate stably under complex conditions such as ship rolling, high temperature and humidity, and salt spray corrosion.

[0003] However, traditional ship fire hydrant water status monitoring systems and methods have significant limitations. On the one hand, during system commissioning, traditional methods rely heavily on manual experience and on-site operation for parameter setting and adjustment, lacking the means to comprehensively simulate and optimize system performance before actual operation. This makes it difficult for system parameters to reach optimal configuration, resulting in a time-consuming and costly commissioning process. Furthermore, the commissioned system may still experience performance instability during actual operation. On the other hand, in terms of fault detection and handling, traditional systems often only respond passively after a fault occurs, unable to predict potential faults in advance, and even less able to simulate the development process and impact range of faults. This prevents ship management personnel from developing effective maintenance plans and emergency measures in advance. In the event of an emergency such as a fire, the failure of the fire hydrant system may prevent the timely provision of fire-fighting water, leading to serious consequences.

[0004] In summary, existing technologies for detecting the water status of ship fire hydrants can no longer meet the growing demands for ship fire safety. Therefore, it is of great importance to develop a fire hydrant water status detection system and method. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a fire hydrant water usage status detection system and method. It can construct a multi-physics coupled virtual model of the fire hydrant by introducing digital twin technology, and combine edge computing and cloud computing to achieve efficient data processing and analysis. This allows for pre-tuning of system parameters, simulating system operation under different parameter configurations in a virtual environment, quickly selecting the optimal parameters, significantly shortening the on-site debugging cycle, reducing debugging costs, and improving system operational stability. Furthermore, with the help of this digital twin model and a fault analysis module based on an improved algorithm, it can also simulate the occurrence process and impact range of potential faults, allowing ship management personnel to develop maintenance and emergency plans in advance, significantly improving the emergency response capability of the fire hydrant system.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: On the one hand, a fire hydrant water status detection system, the system comprising the following components: a physical sensing layer, a data transmission layer, a data processing and analysis layer, a digital twin layer, and an application layer; The physical sensing layer deploys various types of sensors on the fire hydrant body and related pipelines, including pressure sensors to measure water pressure in the pipeline, flow sensors to monitor water flow, temperature sensors to sense ambient temperature, salt spray sensors to detect salt spray concentration in the air, and attitude sensors to monitor the ship's roll and pitch angles in real time; the collected physical quantities are converted into electrical or digital signals and the data is sent to the data transmission layer via wired or wireless transmission. The data transmission layer is responsible for securely and stably transmitting the data collected by the physical sensing layer to the data processing and analysis layer. It employs communication technology, selects the transmission method according to the actual communication environment of the ship, ensures the timeliness and accuracy of data transmission, and encrypts the transmitted data to prevent data leakage and tampering. The data processing and analysis layer uses a combination of edge computing and cloud computing to process and analyze the transmitted data. The edge computing device performs preliminary screening, filtering, and data fusion processing on the sensor data locally to reduce invalid data transmission and improve data processing efficiency. Then, the processed data is uploaded to the cloud server, where big data analysis technology and machine learning algorithms are used to deeply mine the data, establish a fire hydrant water usage status model, and identify normal and abnormal water usage patterns. The digital twin layer constructs a highly realistic digital twin of the actual ship's fire hydrant system. This digital twin integrates multi-physics models of ship swaying, temperature, and salt spray, as well as fluid dynamics and structural mechanics models of the fire hydrants. By receiving data from the physical sensing layer in real time, the models in the digital twin are dynamically updated to keep them synchronized with the actual fire hydrant system, thereby achieving real-time simulation and visualization of the water usage status of the fire hydrants. The application layer provides a user interface for ship managers and firefighters, through which they can intuitively view the real-time water usage status, historical data records, and alarm information of fire hydrants. At the same time, it uses digital twins to pre-debug system parameters, simulate water usage response under different fault conditions, formulate response strategies in advance, and conduct fault simulations to predict the probability and scope of potential faults.

[0007] Furthermore, in the physical sensing layer, the pressure sensor adopts a piezoresistive pressure sensor based on MEMS technology, with a measurement accuracy of 0.05%FS. To further improve the accuracy of pressure measurement in the complex temperature environment of a ship, an adaptive temperature compensation mechanism is introduced. By collecting experimental data from the pressure sensor under different temperature-pressure combination conditions, the least squares method is used to fit and determine the temperature compensation coefficients. Based on these coefficients, an adaptive temperature compensation algorithm is constructed. This algorithm can correct the original pressure value measured by the pressure sensor according to the difference between the real-time temperature value measured by the temperature sensor and the reference temperature value, effectively eliminating the influence of temperature changes on pressure measurement, and reducing the error of pressure measurement in the complex temperature environment of a ship by more than 80%.

[0008] Furthermore, in the data transmission layer, to address the issue of dynamic changes in wireless link quality caused by hull structure obstruction and electromagnetic interference during ship navigation, a multi-path data transmission scheme based on dynamic weights is adopted. This scheme first establishes a link quality assessment model, comprehensively considering the bit error rate, received signal strength indication, and transmission delay parameters of each link to quantitatively assess the quality of each link. By analyzing the ratio of successfully transmitted data packets to the total number of transmitted data packets in historical transmission data, and combining the changing trends of link quality parameters, the weight coefficients in the assessment model are iteratively optimized using a reinforcement learning algorithm. During data transmission, the system dynamically allocates the transmission ratio of data on different links based on the assessment quality value of each link, effectively improving the data transmission success rate to 65%, thereby ensuring stable data transmission from the physical sensing layer to the data processing and analysis layer.

[0009] Furthermore, the edge computing portion of the data processing and analysis layer employs a data fusion algorithm based on an improved DS evidence theory, introducing an evidence conflict correction factor. For pressure, flow rate, temperature Using measurement data from each sensor, a basic probability assignment function is constructed. The fused probability allocation function The calculation is as follows: ,in For the event The supplement, The coefficient of conflict of evidence. By analyzing the fluctuation range and correlation of historical sensor data, the Particle Swarm Optimization algorithm is used to search for the optimal value. This algorithm can effectively fuse multi-source heterogeneous sensor data. Compared with traditional algorithms, the accuracy of data fusion is improved by 40%, invalid data is reduced from being uploaded to the cloud server, and the overall data processing efficiency is improved.

[0010] Furthermore, in the cloud computing portion of the data processing and analysis layer, when establishing a fire hydrant water usage anomaly detection model, an algorithm based on an improved Long Short-Term Memory network-attention mechanism is adopted. This involves introducing a multi-head attention mechanism into the LSTM network for the input time-series data. Attention weight ,in , The function uses dot product operation. For the first Step to the first Attention weights for each input data item For the first Step and the first Relevance score of each input data point For LSTM networks The hidden state of the step, To compute the hidden state using dot product operation With input By training on historical water usage data of ship fire hydrants and adjusting network parameters using backpropagation algorithm and adaptive moment estimation optimizer, the model can accurately identify abnormal water usage status of fire hydrants, with an anomaly detection accuracy of over 95%, providing reliable data support for subsequent fault diagnosis and simulation.

[0011] Furthermore, in the digital twin layer, when constructing the ship sway-fire hydrant fluid coupling model, an improved algorithm based on the arbitrary Lagrange-Euler method is adopted, introducing a fluid-structure interaction coefficient. This describes the effect of ship rolling on the fluid in fire hydrant pipes. In Cartesian coordinates, the fluid control equation is: ,in For time, For gradient operators, For fluid pressure, For the Laplace operator, For fluid velocity, For grid speed, For pressure, For dynamic viscosity, For fluid density, It is the acceleration due to gravity. Based on experimental data on the changes in fluid pressure and flow rate in fire hydrant pipes under different sway amplitudes and frequencies of ships, the improved model was optimized using a genetic algorithm. This improved model can more accurately simulate the water usage status of fire hydrants when ships are swaying, and the simulation error is reduced by 35%.

[0012] Furthermore, in the application layer, during system parameter pre-tuning, a multi-objective optimization algorithm based on the Pareto front is adopted, with the optimization objectives being the stability of fire hydrant water pressure, the uniformity of flow, and the minimization of energy consumption, to construct a set of objective functions. ,in , , These are the function expressions corresponding to the three optimization objectives, and their specific forms are defined according to the actual optimization requirements. As a system parameter vector, the operation of the digital twin is simulated under more than 1,000 different parameter combinations. The Pareto front solution set is obtained by iterative calculation using the non-dominated sorting genetic algorithm II. From the solution set, the weights of each objective are determined by the analytic hierarchy process based on the actual operational needs of the ship, thereby selecting the optimal system parameter configuration, which improves the overall performance of the fire hydrant system by more than 30%.

[0013] Furthermore, in the application layer, the fault propagation prediction algorithm based on Bayesian networks is adopted in the fault simulation module. First, a Bayesian network topology of the fire hydrant system fault nodes and state variables is established. Based on historical fault data and expert experience, the conditional probability distribution between nodes is determined. When an abnormal state is detected, the probability of other node failures is calculated using Bayes' theorem. The formula is: ,in Known abnormal state When it occurs, the node to be predicted as a fault. The posterior probability, Faulty node When it occurs, an abnormal state The conditional probability is determined based on historical failure data and expert experience. Faulty node The prior probability, Abnormal state The prior probability, For the node to be predicted as a fault, For nodes in known abnormal states, this algorithm can predict the propagation path and impact range of faults in the fire hydrant system, providing a scientific basis for ship managers to formulate emergency strategies and reducing emergency response time by 40%.

[0014] Furthermore, in the data transmission layer, to ensure the security of fire hydrant water usage status data during transmission, a hybrid encryption method based on chaotic mapping and elliptic curve encryption is adopted. A chaotic sequence is generated using Logistic chaotic mapping, and this chaotic sequence is XORed with the original data to achieve initial encryption. Then, the ECC algorithm is used to encrypt and transmit the key of the chaotic sequence. The ECC algorithm is based on an elliptic curve equation under specific conditions over a finite field, where the parameters of the elliptic curve equation are generated by a secure random number generator. This hybrid encryption method ensures data security while improving encryption and decryption efficiency by 25% compared to traditional encryption methods, effectively preventing data from being stolen or tampered with during transmission.

[0015] On the other hand, a method for detecting the water usage status of fire hydrants, characterized in that the specific steps of the method are as follows: Data acquisition: Various sensors collect relevant physical parameters of fire hydrants in real time according to the set sampling frequency, such as pressure, flow rate, temperature, salt spray concentration, and ship attitude data, and send the data to the data transmission layer; Data transmission and preprocessing: The data transmission layer transmits the collected data to the data processing and analysis layer. In this layer, the edge computing device first performs noise reduction and filtering preprocessing on the data to remove abnormal data caused by sensor errors and electromagnetic interference. Then, it fuses different types of sensor data to form a unified data format, providing a basis for subsequent analysis. Model training and updating: Using historical data and preprocessed real-time data, deep learning algorithms are used on a cloud server to train and optimize the fire hydrant water usage status model. As new data continues to be added, the model parameters are continuously updated to improve the model's ability to identify and predict the operating status of fire hydrants. Digital twin driving and simulation: Processed and analyzed data is input into the digital twin to drive the multiphysics coupling model and fire hydrant related models to run dynamically, and simulate the water use status of fire hydrants in the current ship environment in real time. Through visualization technology, the simulation results in the digital twin are displayed on the application layer interface in an intuitive way, including information such as water flow velocity distribution, pressure field distribution, and valve opening changes. Fault Diagnosis and Simulation: Based on the simulation results of the digital twin and the analysis of the water status model, it is determined whether the fire hydrant is in normal working condition. If an abnormality is detected, the digital twin is used to simulate the development process of different fault scenarios by changing the model parameters, and the causes, scope of impact and possible propagation paths of the fault are analyzed to provide a basis for formulating maintenance plans and emergency measures. At the same time, alarm information is issued at the application layer to remind relevant personnel to deal with it in a timely manner. System parameter pre-commissioning: After the installation or maintenance of the ship's fire protection system, the system parameters are pre-commissioned using a digital twin. By simulating different working conditions and parameter settings in a virtual environment, the changes in the water usage status of the fire hydrants in the digital twin are observed to find the optimal system parameter configuration. These parameters are then applied to the actual fire hydrant system, reducing on-site commissioning time and costs, and improving the stability and reliability of system operation.

[0016] Compared with existing technologies, this fire hydrant water usage status detection system and method have the following advantages: I. This invention utilizes a digital twin at the application layer to simulate the water usage response of fire hydrants under different working conditions and parameter settings in a virtual environment. By analyzing key indicators such as pressure stability and flow uniformity, the optimal system parameter configuration is found and applied to the actual system, greatly shortening on-site commissioning time, reducing commissioning costs, avoiding performance problems caused by unreasonable parameters, and improving the overall stability of the system. Furthermore, in terms of fault diagnosis and simulation, when an anomaly is detected, the digital twin simulates the development process of different fault scenarios, enabling rapid location of the fault and analysis of its cause, scope of impact, and propagation path. For example, it accurately determines the specific location of a pipeline leak, the amount of leakage, and its impact on the entire fire protection system, providing a strong basis for ship management personnel to formulate scientific and reasonable maintenance plans and emergency measures in advance. Compared with traditional detection systems, this significantly improves the emergency response capability of ship fire protection systems, providing a more reliable guarantee for ship fire safety.

[0017] Second, this invention constructs a comprehensive data acquisition system by deploying multiple types of sensors, such as pressure, flow, temperature, salt spray, and attitude sensors, at the physical sensing layer. Compared to the traditional method of monitoring only water pressure or flow, it can acquire real-time operational data of fire hydrants under complex environments such as ship swaying, temperature and humidity changes, and salt spray corrosion. At the data processing and analysis layer, by leveraging edge computing and cloud computing technologies, it performs in-depth mining and fusion analysis of multi-source data. Combined with a highly realistic multi-physics coupled digital twin in the digital twin layer, it achieves multi-dimensional and accurate monitoring of the water usage status of fire hydrants. For example, it can accurately analyze the impact of changes in ship sway angle on the opening status of fire hydrant valves and water flow, as well as the impact of changes in salt spray concentration on the corrosion and leakage risks of fire hydrant equipment. This effectively solves the problem that traditional detection methods cannot fully reflect the actual operating status of fire hydrants, and significantly improves the accuracy and reliability of monitoring.

[0018] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0020] Figure 1 This is a schematic diagram of the architecture of a fire hydrant water usage status detection system; Figure 2 This is a flowchart of a method for detecting the water usage status of fire hydrants. Detailed Implementation

[0021] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below. Example 1

[0022] Data acquisition phase: Pressure sensors, flow sensors, and temperature sensors are deployed on the ship's fire hydrant body and connected pipes in the physical sensing layer. The pressure sensor is a piezoresistive pressure sensor based on MEMS technology. The adaptive temperature compensation algorithm introduced in this sensor can monitor the ambient temperature in real time through the temperature sensor and dynamically correct the pressure measurement value by combining the pressure characteristics at different temperatures. This function effectively avoids the interference of the cabin temperature with the navigation area and equipment operating status during the ship's navigation, ensuring that the collected physical parameters such as pressure, water flow velocity, and ambient temperature are accurate and reliable. These sensors continuously collect data at a set frequency, convert the physical quantities into electrical signals, and then transmit them to the data transmission layer wirelessly.

[0023] Data transmission and preprocessing stage: The data transmission layer adopts a multi-path data transmission scheme based on dynamic weights. This scheme addresses the issue of fluctuating wireless link quality caused by hull structure obstruction and electromagnetic interference from equipment operation during ship navigation. First, it evaluates the bit error rate, received signal strength, and transmission delay of each link. Then, it uses reinforcement learning algorithms to optimize link weights and dynamically allocate data transmission ratios, ensuring data can bypass poor-quality links and be stably transmitted to its destination. Simultaneously, a hybrid encryption method based on chaotic mapping and elliptic curve cryptography is used. The raw data is first processed using Logistic chaotic mapping, and then the ECC algorithm is used to encrypt the key. This encryption combination provides double protection for data transmission security, preventing data theft or tampering during transmission. After the data reaches the data processing and analysis layer, the edge computing device first performs noise reduction and filtering to remove outliers caused by sensor errors and external electromagnetic interference, providing a clean data foundation for subsequent analysis. Then, a data fusion algorithm based on improved DS evidence theory is used, introducing an evidence conflict correction factor. For pressure, flow rate, temperature Using measurement data from each sensor, a basic probability assignment function is constructed. The fused probability allocation function The calculation is as follows: ,in For the event The supplement, The coefficient of conflict of evidence. By analyzing the fluctuation range and correlation of historical sensor data, the particle swarm optimization algorithm is used to search for the optimal value and fuse the data of pressure, flow and temperature sensors. The purpose of this algorithm is to integrate information from multiple types of sensors, reduce the limitations of single sensor data, and improve the integrity and reliability of the data.

[0024] Model training and update phase: The cloud server receives data processed by edge computing, combines it with historical water usage data, and trains a fire hydrant water usage status model using an algorithm based on an improved long short-term memory network-attention mechanism. A multi-head attention mechanism is introduced for the input time-series data. Attention weight ,in , The function uses dot product operation. For the first Step to the first Attention weights for each input data item For the first Step and the first Relevance score of each input data point For LSTM networks The hidden state of the step, To compute the hidden state using dot product operation With input The correlation was assessed by training the model on historical water usage data of ship fire hydrants. The backpropagation algorithm and adaptive moment estimation optimizer were used to adjust the network parameters. The algorithm effectively captures long-term dependencies in time series data and focuses on water usage characteristics in key time periods through an attention mechanism, thereby improving the model's accuracy in identifying normal and abnormal water usage states. As real-time data is continuously fed in during navigation, the model continuously updates its parameters. This process allows the model to adapt to changes in the water usage state of fire hydrants at different stages of ship navigation, ensuring that the model always maintains high recognition accuracy.

[0025] Digital Twin Driving and Simulation Phase: The processed data analysis results are input into the digital twin layer, driving the operation of a highly realistic digital twin of the ship's fire hydrant system. This digital twin integrates multiphysics models of ship swaying, temperature, and salt spray, as well as fluid dynamics and structural mechanics models of the fire hydrants. A ship swaying-fire hydrant fluid coupling model is constructed using an improved algorithm based on the arbitrary Lagrange-Euler method, introducing a fluid-structure interaction coefficient. This describes the effect of ship rolling on the fluid in fire hydrant pipes. In Cartesian coordinates, the fluid control equation is: ,in For time, For gradient operators, For fluid pressure, For the Laplace operator, For fluid velocity, For grid speed, For pressure, For dynamic viscosity, For fluid density, The function of gravity acceleration is to accurately simulate the impact of ship swaying on the fluid flow state in the fire hydrant pipe, so that the digital twin can realistically reflect the operation of the fire hydrant in the actual environment. Through visualization technology, the simulation results are displayed intuitively on the application layer interface. Its function is to allow ship managers and firefighters to clearly understand the real-time water use status of the fire hydrant without on-site inspection, including details such as water flow velocity distribution and pressure changes in the pipe.

[0026] Fault Diagnosis and Simulation Phase: Using a water usage model combined with digital twin simulation results, when fire hydrant pressure consistently falls below normal levels or flow is abnormally low, an abnormal leak can be quickly identified. This identification helps to promptly detect potential faults and prevent the leak from escalating. The application layer immediately issues an alarm to alert relevant personnel and facilitate immediate action on the anomaly. Simultaneously, a Bayesian network-based fault propagation prediction algorithm is used, with the following formula: ,in Known abnormal state When it occurs, the node to be predicted as a fault. The posterior probability, Faulty node When it occurs, an abnormal state The conditional probability is determined based on historical failure data and expert experience. Faulty node The prior probability, Abnormal state The prior probability, For the node to be predicted as a fault, Given known abnormal state nodes, with the current water leakage abnormality as a known node, the algorithm calculates the probability of occurrence of potential fault nodes such as loose pipe joints and valve sealing failure. The purpose of this algorithm is to scientifically infer the source of the fault and its possible development direction; and to predict the possible expansion range of the fault, such as causing a sudden drop in local pipeline pressure and affecting the normal use of nearby fire hydrants. Its role is to provide relevant personnel with a prediction of the impact of the fault and assist in formulating targeted countermeasures.

[0027] System parameter pre-tuning phase: Although this was a sudden failure during navigation, the system parameters can be pre-tuned using a digital twin during subsequent maintenance. By using a multi-objective optimization algorithm based on the Pareto front, a set of functions including objectives such as leak detection sensitivity and system energy consumption can be constructed. The role of this algorithm is to find the optimal balance among multiple mutually constraining objectives, avoiding performance degradation caused by optimizing a single objective. The state of fire hydrants under different parameter settings is simulated in a virtual environment to find the optimal parameter configuration. Its purpose is to optimize system performance in advance without affecting the normal navigation of the ship, improve the system's ability to detect similar failures, and reduce the possibility of future failures. Example 2

[0028] Data Acquisition Phase: During ship berthing, the pressure, flow, and temperature sensors in the physical sensing layer continue to operate at set frequencies. The pressure sensor employs a piezoresistive pressure sensor based on MEMS technology. Its adaptive temperature compensation algorithm effectively addresses the temperature difference between the dock environment and the cabin. By capturing temperature changes in real time through the temperature sensor, the pressure measurement value is dynamically corrected, avoiding interference from temperature fluctuations on data accuracy. This ensures that the collected data on pressure, flow, and ambient temperature of fire hydrants and pipelines are accurate and reliable. This data is transmitted to the data transmission layer via wired transmission. Compared to wireless transmission, wired transmission further ensures the stability of data transmission in scenarios with less interference, such as ship berthing.

[0029] Data transmission and preprocessing stage: The data transmission layer adopts a multi-path data transmission scheme based on dynamic weights. At this stage, the stability of wired links is primarily evaluated. By comprehensively analyzing parameters such as link bit error rate, received signal strength, and transmission delay, reinforcement learning algorithms are used to optimize link weight allocation. Its role is to select the optimal transmission path among multiple wired links, ensuring efficient and stable data transmission to the data processing and analysis layer. Simultaneously, a hybrid encryption method based on chaotic mapping and elliptic curve cryptography is used during transmission. The original data is first processed using Logistic chaotic mapping, and then the key of the chaotic sequence is encrypted using the ECC algorithm. This dual encryption mechanism effectively prevents data from being transmitted in error. To prevent unauthorized acquisition or tampering during data transmission, the security of maintenance and debugging data is ensured. After the data reaches the data processing and analysis layer, the edge computing device first performs noise reduction and filtering preprocessing on the data. This process removes minor error data generated by sensors in a static state, such as outliers caused by sensor drift, providing a clean data foundation for subsequent analysis. Subsequently, a data fusion algorithm based on an improved DS evidence theory is used to fuse data from multiple types of sensors, such as pressure, flow, and temperature. This algorithm, by introducing an evidence conflict correction factor, can effectively resolve conflicts between different sensor data, improve data consistency and reliability, and provide a more comprehensive data analysis basis for maintenance and debugging.

[0030] Model training and update phase: The cloud server receives data processed by edge computing and combines it with relatively stable fire hydrant water usage data during ship berthing (such as historical data from periodic water supply tests) and past maintenance records. An algorithm based on an improved long short-term memory network-attention mechanism is used to optimize and train the fire hydrant water usage status model. The algorithm captures the time-series features of water usage data through the long short-term memory network, and focuses on key water usage features in maintenance and debugging scenarios (such as pressure changes under low load conditions and flow stability under static conditions) using a multi-head attention mechanism. It focuses on updating the normal state identification parameters of the system under low load and static conditions, so that the model can more accurately adapt to the working state of fire hydrants during berthing and improve the accuracy of judging the normal and abnormal states of the system during maintenance and debugging.

[0031] Digital Twin Drive and Simulation Phase: The processed data analysis results are input into the digital twin layer, driving the operation of a highly realistic digital twin of the ship's fire hydrant system. Since there is no significant swaying when the ship is docked, the digital twin mainly relies on multi-physics field models such as temperature and salt spray, as well as the fluid dynamics and structural mechanics models of the fire hydrant for simulation. Its function is to realistically reproduce the operating state of the fire hydrant under static conditions, such as the stability of water flow in the pipeline, the uniformity of pressure distribution, and the valve sealing performance. Through visualization technology, the simulation results are intuitively displayed on the application layer interface, allowing maintenance personnel to clearly observe the internal operating state of the fire hydrant without disassembling the equipment. This provides a convenient and intuitive virtual observation method for maintenance and debugging, reducing the damage to the equipment caused by on-site disassembly.

[0032] Fault Diagnosis and Simulation Phase: During commissioning, maintenance personnel artificially simulate a fault scenario (such as closing a control valve in a section of pipeline). The digital twin receives real-time data from sensors, such as sudden pressure drops and zero flow, and drives the model to dynamically simulate the system state changes under this operation. By combining the water state model with the simulation results, abnormal states downstream of the closed pipeline can be quickly identified. Its function is to promptly discover the impact range and characteristics of the simulated fault. The application layer immediately displays the abnormal state, reminding maintenance personnel to pay close attention. Simultaneously, using a fault propagation prediction algorithm based on Bayesian networks, taking the known abnormal state of the current closed control valve as a node, the probability of other node faults that this fault may cause is calculated, such as pressure imbalance in adjacent pipelines and response delays of related valves. Its function is to scientifically deduce the propagation path and potential impact of the fault, helping maintenance personnel to fully understand the consequences of the fault, such as predicting the range of fire hydrants in the area that may not be able to supply water normally in the event of a sudden fire, providing a basis for developing targeted maintenance plans.

[0033] System parameter pre-tuning phase: Maintenance personnel use a digital twin to pre-tune system parameters. A multi-objective optimization algorithm based on the Pareto front is employed to construct a set of functions encompassing objectives such as water supply pressure stability, valve response speed, and system energy consumption. This algorithm aims to find the optimal balance among multiple mutually constraining objectives, avoiding the degradation of other performance indicators caused by optimizing a single objective. By simulating different pipeline pressure settings, valve opening degrees, sensor sampling frequencies, and other parameters in a virtual environment, the water usage status of the fire hydrants in the digital twin is observed (e.g., testing for minor leaks in the pipeline under different pressures, smooth valve operation, and timely reflection of system status by sensor data). Ultimately, the most suitable parameter configuration for the system after maintenance is determined. This process optimizes system performance in advance without affecting normal ship operations, ensuring the fire hydrant system operates stably and efficiently when the ship resumes navigation, reducing the probability of malfunctions during voyage.

[0034] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A fire hydrant water usage status detection system, characterized in that, The system comprises the following components: physical sensing layer, data transmission layer, data processing and analysis layer, digital twin layer, and application layer. The physical sensing layer deploys various types of sensors on the fire hydrant body and related pipelines; converts the collected physical quantities into electrical or digital signals, and sends the data to the data transmission layer via wired or wireless transmission. The data transmission layer is responsible for securely and stably transmitting the data collected by the physical sensing layer to the data processing and analysis layer; it uses communication technology, selects the transmission method according to the actual communication environment of the ship, and encrypts the transmitted data. The data processing and analysis layer uses a combination of edge computing and cloud computing to process and analyze the transmitted data. The edge computing device performs preliminary screening, filtering and data fusion processing on the sensor data locally. Then, the processed data is uploaded to the cloud server, and big data analysis technology and machine learning algorithms are used to deeply mine the data, establish a fire hydrant water use status model, and identify normal and abnormal water use status patterns. The digital twin layer constructs a highly realistic digital twin of the ship's actual fire hydrant system. This digital twin integrates multi-physics models of ship swaying, temperature, and salt spray, as well as fluid dynamics and structural mechanics models of the fire hydrants. By receiving data from the physical sensing layer in real time, the models in the digital twin are dynamically updated. The application layer provides a user interface for ship managers and firefighters, through which they can intuitively view the real-time water usage status, historical data records, and alarm information of fire hydrants. At the same time, it uses digital twins to pre-debug system parameters, simulate water usage response under different fault conditions, formulate response strategies in advance, and conduct fault simulations to predict the probability and scope of potential faults.

2. The fire hydrant water usage status detection system according to claim 1, characterized in that, In the physical sensing layer, the pressure sensor adopts a piezoresistive pressure sensor based on MEMS technology and introduces an adaptive temperature compensation mechanism. By collecting experimental data from the pressure sensor under different temperature-pressure combination conditions, the temperature compensation coefficient is determined by fitting using the least squares method. Based on these coefficients, an adaptive temperature compensation algorithm is constructed. This algorithm can correct the original pressure value measured by the pressure sensor according to the difference between the real-time temperature value measured by the temperature sensor and the reference temperature value.

3. The fire hydrant water usage status detection system according to claim 1, characterized in that, In the data transmission layer, to address the problem of dynamic changes in wireless link quality caused by hull structure obstruction and electromagnetic interference during ship navigation, a multi-path data transmission scheme based on dynamic weights is adopted. This scheme first establishes a link quality assessment model, comprehensively considering the bit error rate, received signal strength indication, and transmission delay parameters of each link to quantitatively assess the quality of each link. By analyzing the ratio of successfully transmitted data packets to the total number of transmitted data packets in historical transmission data, and combining the changing trends of link quality parameters, the weight coefficients in the assessment model are iteratively optimized using a reinforcement learning algorithm. During data transmission, the system dynamically allocates the transmission ratio of data on different links based on the assessment quality value of each link.

4. The fire hydrant water usage status detection system according to claim 1, characterized in that, The edge computing portion of the data processing and analysis layer employs a data fusion algorithm based on an improved DS evidence theory, and introduces an evidence conflict correction factor. For pressure, flow rate, temperature Using measurement data from each sensor, a basic probability assignment function is constructed. The fused probability allocation function The calculation is as follows: ,in For the event The supplement, This represents the conflict coefficient of evidence.

5. The fire hydrant water usage status detection system according to claim 1, characterized in that, The cloud computing portion of the data processing and analysis layer, when establishing a fire hydrant water usage anomaly detection model, employs an algorithm based on an improved Long Short-Term Memory (LSTM) network-attention mechanism. This involves introducing a multi-head attention mechanism into the LSTM network for the input time-series data. Attention weight ,in , The function uses dot product operation. For the first Step to the first Attention weights for each input data item For the first Step and the first Relevance score of each input data point For LSTM networks The hidden state of the step, To compute the hidden state using dot product operation With input The correlation was determined by training the network on historical water usage data of ship fire hydrants, and then adjusting the network parameters using the backpropagation algorithm and the adaptive moment estimation optimizer.

6. The fire hydrant water usage status detection system according to claim 1, characterized in that, In the digital twin layer, when constructing the ship sway-fire hydrant fluid coupling model, an improved algorithm based on the arbitrary Lagrange-Euler method is used, introducing a fluid-structure interaction coefficient. This describes the effect of ship rolling on the fluid in fire hydrant pipes. In Cartesian coordinates, the fluid control equation is: ,in For time, For gradient operators, For fluid pressure, For the Laplace operator, For fluid velocity, For grid speed, For pressure, For dynamic viscosity, For fluid density, This is the acceleration due to gravity.

7. The fire hydrant water usage status detection system according to claim 1, characterized in that, In the application layer, during system parameter pre-tuning, a multi-objective optimization algorithm based on the Pareto front is used to construct a set of objective functions. ,in , , The function expressions correspond to the three optimization objectives respectively. This is the system parameter vector.

8. The fire hydrant water usage status detection system according to claim 1, characterized in that, In the application layer, the fault propagation prediction algorithm based on Bayesian networks is used in the fault simulation module. First, a Bayesian network topology of the fire hydrant system's fault nodes and state variables is established. When an abnormal state is detected, the probability of other nodes failing is calculated using Bayes' theorem. The formula is: ,in Known abnormal state When it occurs, the node to be predicted as a fault. The posterior probability, Faulty node When it occurs, an abnormal state The conditional probability, Faulty node The prior probability, Abnormal state The prior probability, For the node to be predicted as a fault, These are nodes in a known abnormal state.

9. A fire hydrant water usage status detection system according to claim 1, characterized in that, In the data transmission layer, a hybrid encryption method based on chaotic mapping and elliptic curve encryption is adopted. A chaotic sequence is generated using Logistic chaotic mapping, and the chaotic sequence is XORed with the original data. Then, the key of the chaotic sequence is encrypted and transmitted using the ECC algorithm. The ECC algorithm is based on the elliptic curve equation under specific conditions over a finite field, where the parameters of the elliptic curve equation are generated by a secure random number generator.

10. A method for detecting the water usage status of a fire hydrant, applicable to the fire hydrant water usage status detection system according to any one of claims 1-9, characterized in that, The specific steps of this method are as follows: Data acquisition: Various sensors collect relevant physical parameters of fire hydrants in real time according to the set sampling frequency and send the data to the data transmission layer; Data transmission and preprocessing: The data transmission layer transmits the collected data to the data processing and analysis layer. In this layer, the edge computing device first performs noise reduction and filtering preprocessing on the data to remove abnormal data caused by sensor errors and electromagnetic interference. Then, it performs fusion processing on different types of sensor data. Model training and updating: Using historical data and preprocessed real-time data, deep learning algorithms are used on a cloud server to train and optimize the fire hydrant water usage status model. As new data continues to be added, the model parameters are continuously updated. Digital twin driving and simulation: Processed and analyzed data is input into the digital twin to drive the multiphysics coupling model and fire hydrant related models to run dynamically, and simulate the water use status of fire hydrants in the current ship environment in real time. Through visualization technology, the simulation results in the digital twin are displayed on the application layer interface in an intuitive way. Fault Diagnosis and Simulation: Based on the simulation results of the digital twin and the analysis of the water status model, it is determined whether the fire hydrant is in normal working condition. If an abnormality is detected, the digital twin is used to simulate the development process of different fault scenarios by changing the model parameters, and the causes, scope of impact and possible propagation paths of the fault are analyzed to provide a basis for formulating maintenance plans and emergency measures. At the same time, alarm information is issued at the application layer to remind relevant personnel to deal with it in a timely manner. System parameter pre-tuning: After the installation or maintenance of the ship's fire protection system, the system parameters are pre-tuned using a digital twin. By simulating different working conditions and parameter settings in a virtual environment, the changes in the water usage status of fire hydrants in the digital twin are observed to find the optimal system parameter configuration.

Images