Intelligent fire-fighting equipment instrument management system

Abstract

The invention discloses an intelligent fire-fighting equipment instrument management system, which comprises a data acquisition layer, a data processing layer, an intelligent analysis layer, an application service layer and a user interface layer, the data acquisition layer comprises an Internet of Things data acquisition module, and the Internet of Things data acquisition module is responsible for acquiring various data such as running states, position information and maintenance records of fire-fighting equipment in real time and uploading the data to an edge computing node for primary processing; the data processing layer comprises a data preprocessing module, and the data preprocessing module processes and analyzes the collected data; the intelligent analysis layer comprises a deep learning analysis module, and the deep learning analysis module constructs an equipment fault prediction model, carries out the intelligent analysis of the processed data, and predicts the fault and shortage conditions of the equipment. According to the method, the equipment fault prediction model is constructed, the equipment fault and shortage conditions can be accurately predicted, an early warning mechanism is automatically triggered, and a maintenance plan and a purchase strategy are automatically made according to the early warning information and the equipment state data.

Description

Technical Field

[0001] The present invention relates to the technical field of fire fighting equipment management, and particularly to an intelligent fire fighting equipment management system. Background Art

[0002] Fire fighting equipment is the main force for fire fighting and rescue by fire fighting teams, and is an important factor determining the combat effectiveness of fire fighting teams. The quality of equipment allocation, use and management directly affects the combat effectiveness of fire fighters. Some fires cause heavy property losses and casualties, often due to poor equipment or lack of corresponding fire fighting equipment. Therefore, strengthening the management of fire fighting equipment to ensure its good performance and effective use is of great significance for improving the fire fighting and rescue capabilities of fire fighting teams.

[0003] With the acceleration of the urbanization process and the increasing urgency and complexity of fire fighting tasks, the traditional manual control mode is difficult to meet the needs of modern fire fighting equipment management. Poor equipment management leads to the inability to accurately predict equipment failures and shortages, which may result in the abnormal use of equipment in emergency situations. If fire fighting equipment cannot be used normally, it will seriously affect the efficiency of fire fighting and rescue. Therefore, an intelligent fire fighting equipment management system is proposed. Summary of the Invention

[0004] The purpose of the present invention is to solve the problems existing in the prior art, that is, the traditional manual control mode is difficult to meet the needs of modern fire fighting equipment management. At the same time, it cannot accurately predict equipment failures and shortages, which may lead to the abnormal use of equipment in emergency situations, thus seriously affecting the efficiency of fire fighting and rescue. An intelligent fire fighting equipment management system is proposed.

[0005] In order to achieve the above purpose, the present invention adopts the following technical solutions:

[0006] An intelligent fire fighting equipment management system includes a data acquisition layer, a data processing layer, an intelligent analysis layer, an application service layer and a user interface layer;

[0007] The data acquisition layer is connected to an Internet of Things data acquisition module. The Internet of Things data acquisition module is responsible for real-time acquisition of various data such as the operating status, location information, and maintenance records of fire fighting equipment, and uploading them to an edge computing node for preliminary processing;

[0008] The data processing layer includes a data preprocessing module. The data preprocessing module processes and analyzes the acquired data, including data compression, encryption, cleaning, conversion, etc., to ensure the accuracy and security of the data;

[0009] The intelligent analysis layer includes a deep learning analysis module. The deep learning analysis module constructs an equipment fault prediction model, conducts intelligent analysis on the processed data, predicts equipment faults and shortages, provides a scientific basis for decision-making support. The deep learning analysis module conducts intelligent analysis on real-time or recent equipment operation data, outputs fault prediction results, generates early warning information based on the prediction results, and uses particle swarm optimization to formulate a maintenance plan according to the early warning information and equipment status data;

[0010] The application service layer includes a user-defined interface module and an API interface docking module. The user-defined interface module provides a user-defined interface, allowing users to adjust the system interface layout according to actual needs. The API interface docking module docks with other fire protection systems through the API interface to achieve real-time information sharing and coordinated operations;

[0011] The user interface layer includes a user interface display module responsible for providing a user interface and displaying various contents such as equipment status, early warning information, and maintenance plans;

[0012] The Internet of Things data collection module uploads the data of fire protection equipment and instruments collected to the data preprocessing module. The data preprocessing module transfers the processed data to the deep learning analysis module. The deep learning analysis module transfers the analysis results to the user-defined interface module and the API interface docking module of the application service layer. The modular design module, user-defined interface module, and API interface docking module of the application service layer generate corresponding contents such as equipment status, early warning information, and maintenance plans according to the analysis results and user needs, and transfer this information to the user interface display module.

[0013] The above technical solution further includes:

[0014] Furthermore, the IoT data acquisition module includes an IoT data acquisition unit, an edge computing node unit, and an equipment management database unit. The IoT data acquisition unit is responsible for collecting data of fire-fighting equipment through IoT technology. The IoT data acquisition unit interacts with the edge computing node and uploads the collected data. The edge computing node unit receives the data uploaded by the IoT data acquisition unit and performs preliminary processing on the data, such as data cleaning, format conversion, etc. The equipment management database unit stores various data for the full life cycle management of equipment, such as equipment information, outbound records, maintenance records, scrapping records, etc. The IoT data acquisition unit uploads the collected data to the edge computing node unit. The edge computing node unit receives the data and performs preliminary processing. The edge computing node unit uploads the processed data to the equipment management database unit for storage. The equipment management database unit provides operation interfaces such as data query and update for the edge computing node unit to support real-time data update and synchronization.

[0015] Furthermore, the data preprocessing module includes a data acquisition unit, a distributed computing unit, a data storage unit, and a data interaction interface unit. The data acquisition unit is responsible for collecting raw data from the data acquisition layer. The distributed computing unit uses multiple computers or servers to work together to perform parallel processing on the data, improving the processing speed and efficiency. The data storage unit is used to store the processed data for subsequent analysis and application. The data interaction interface unit provides interfaces for data interaction with other systems or modules, such as APIs, message queues, etc. The data acquisition unit sends the collected raw data to the distributed computing unit for processing. The processed data is sent by the distributed computing unit to the data storage unit for storage. The data interaction interface unit obtains data from the data storage unit.

[0016] Furthermore, the deep learning analysis module includes a model construction unit, a fault prediction unit, a warning notification unit, a maintenance plan unit, and a decision support unit. The model construction unit constructs an equipment fault prediction model using deep learning algorithms. The fault prediction unit performs intelligent analysis on real-time or recent equipment operation data and outputs a fault prediction result. The warning notification unit generates a warning message based on the prediction result. The maintenance plan unit uses particle swarm optimization to formulate a maintenance plan according to the warning message and equipment status data. The decision support unit provides an equipment fault prediction report to display the prediction result and trend. The data preprocessing module transfers the preprocessed data to the model construction unit and the maintenance plan unit for model training and formulating a maintenance plan. The model construction unit transfers the trained model to the fault prediction unit. The fault prediction unit transfers the prediction result to the warning notification unit. The warning notification unit transfers the sending status and receiving record of the warning message to the decision support unit and the maintenance plan unit. The maintenance plan unit transfers the maintenance plan to the decision support unit.

[0017] Furthermore, the user-defined interface module includes an interface layout management unit, a function setting management unit, a configuration management unit, a permission management unit, and a data storage unit. The interface layout management unit is responsible for handling user-defined operations on the interface layout, such as dragging, adding, deleting elements, etc., saving and loading the interface layout configuration. The function setting management unit is responsible for handling user-defined operations on function settings, such as enabling / disabling functions, adjusting function parameters, etc., validating the effectiveness of the user configuration, and updating the system status. The configuration management unit provides a visual configuration management interface to display various configurations of the system, supports the import, export, validation, and error prompt of configurations. The permission management unit manages the roles and permissions of users, ensures that users can only access and operate the resources they are authorized to, and records the operation logs of users for easy tracking and auditing. The data storage unit stores user-defined configurations, interface layouts, and function settings, and provides a data access interface for other units to read and write data. The interface layout management unit saves the user-defined interface layout configuration to the data storage unit. When the user loads the interface layout, the interface layout management unit reads the configuration from the data storage unit and applies it. The function setting management unit saves the user-defined function settings to the data storage unit. When the user adjusts the function settings, the function setting management unit reads the current configuration from the data storage unit, validates and updates it.

[0018] Furthermore, the API interface docking module includes an API interface management unit, a data synchronization unit, a data format conversion unit, a security authentication unit, an access control unit, and a log recording unit. The API interface management unit is responsible for the creation, maintenance, and management of API interfaces, and processes API call requests and responses with other fire protection systems. The data synchronization unit is responsible for the real-time synchronization of equipment status information, and realizes information sharing and integration with other fire protection systems. The data format conversion unit is responsible for the conversion of data formats between different systems, parses the received data, and extracts useful information. The security authentication unit performs security authentication on the API interface, such as using OAuth, API keys, etc., to verify the identity and permissions of the requester. The access control unit implements access control policies, restricts access to the API interface, monitors and records the access situation of the API interface. The log recording unit records the log information during the API interface docking process, and provides log query and analysis functions. The data synchronization unit obtains the corresponding equipment status information or other shared information from the database according to the request type, and returns it to the API interface management unit. When the received data format does not match the internal format of the system, the API interface management unit passes the data to the data format conversion unit. Before the API call request reaches the API interface management unit, the security authentication unit performs security authentication on the request, verifies the identity and permissions of the requester, and passes the authentication result to the API interface management unit. The access control unit implements the access control policy according to the request of the API interface management unit.

[0019] Furthermore, the user interface display module includes a user interface management unit, a data display unit, a user authentication and permission management unit, a business logic processing unit, and a data visualization unit. The user interface management unit is responsible for the design and layout of the user interface, including page structure, styles, interaction logic, etc., processes user input and interaction events, and passes user operations to the corresponding business logic units. The data display unit is responsible for obtaining equipment status, warning information, and maintenance plans from the back-end system and displaying them on the user interface, providing dynamic updates and refresh functions for the data to ensure the real-time nature of the user interface. The user authentication and permission management unit is responsible for user login authentication and permission management, ensuring that users can only access the content they are authorized to, and recording user login and operation logs for auditing and tracking. The business logic processing unit processes various operation requests initiated by users through the user interface, such as querying equipment status, processing warning information, editing maintenance plans, etc., interacts with the back-end system, executes the corresponding business logic, and returns the processing results to the user interface. The data visualization unit is responsible for converting equipment status data and analysis results into visual charts or dashboards so that users can more intuitively understand the data, provides data export and report generation functions, and supports users in using the data for analysis and decision-making.

[0020] Furthermore, the maintenance plan unit uses particle swarm optimization to formulate the specific steps of the maintenance plan based on warning information and equipment status data:

[0021] Initialize the particle swarm: Particle position: Each particle represents a maintenance plan, and its position consists of multiple decision variables, such as maintenance time interval, spare part quantity, maintenance personnel allocation, etc.; Particle velocity: The initial velocity is randomly set; Fitness function: Define a fitness function to evaluate the quality of each maintenance plan.

[0022] Update the particle position and velocity: Personal best position (pBest): Each particle remembers the best position it has found so far; Global best position (gBest): The best position found in the entire particle swarm; The velocity update formula is new velocity = inertia weight * old velocity + cognitive learning factor * random number * (pBest - current position) + social learning factor * random number * (gBest -

[0023] current position), and the position update formula is new position = current position + new velocity.

[0024] Evaluate the fitness: Calculate the fitness value of each particle's new position, and update the personal best position (pBest) and the global best position (gBest).

[0025] Iteration: Repeat updating the particle position and velocity and evaluating the fitness until a predetermined number of iterations is reached or the fitness value no longer improves significantly.

[0026] Further, the specific steps for the model construction unit to construct the equipment fault prediction model using deep learning algorithms are as follows:

[0027] Model architecture selection:

[0028] Select the Transformer model based on the attention mechanism as the architecture of the equipment fault prediction model. The Transformer model captures the dependencies in the sequence data through the self-attention mechanism and positional encoding;

[0029] Construct the Transformer model:

[0030] The Transformer model consists of two parts: an encoder and a decoder. However, in the fault prediction task, only the encoder part is needed because our goal is to predict future equipment faults rather than generate sequences;

[0031] Initialize the model parameters:

[0032] Initialize the parameters such as weights and biases in the Transformer model. These parameters will be optimized through subsequent training processes.

[0033] Determine the loss function:

[0034] For equipment fault prediction, use the Mean Squared Error (MSE) as the loss function. The definition of MSE is: where y i is the true value, is the predicted value, and n is the number of samples.

[0035] Input the training feature set:

[0036] Input the preprocessed training feature set into the encoder of the Transformer model. These features will be converted into high-dimensional vector representations through the Embedding Layer.

[0037] Optimize the parameters:

[0038] Use the Adam optimizer to optimize the parameters of the Transformer model. The Adam optimizer combines the advantages of the Momentum and RMSProp optimizers and can converge to the optimal solution more quickly. The update rule of the Adam optimizer is as follows:

[0039] m i = β1mi-1 +(1 - β1)g i ;

[0040]

[0041] where m i and v i are the first - order moment estimate and second - order moment estimate of the gradient respectively, β1 and β2 are the decay rates, η is the learning rate, and ∈ is a small constant to prevent division by zero.

[0042] Distributed Training and Parallel Computing:

[0043] Utilize multiple computers or GPUs to accelerate the model training process, which can be achieved through data parallelism or model parallelism.

[0044] Adaptive Learning Rate Adjustment:

[0045] Dynamically adjust the learning rate according to the training situation of the model. For example, when the loss value no longer decreases significantly, the learning rate can be reduced to avoid overfitting.

[0046] Model Convergence:

[0047] Converge the model by comparing the loss value between the predicted value and the measured value, and when the loss value is less than the preset threshold, which means the model has found a better parameter configuration and can fit the training data well.

[0048] Cross - Validation:

[0049] Use the cross - validation method to evaluate the performance of the model, which can help us understand the performance of the model on different datasets and avoid overfitting.

[0050] Test Set Evaluation:

[0051] Evaluate the performance of the model on the test set, which can give us an intuitive feeling about the generalization ability of the model.

[0052] Model Adjustment and Optimization:

[0053] Adjust and optimize the model according to the evaluation results. For example, try different model architectures, loss functions, optimizers, etc. to improve the performance of the model.

[0054] Suppose we have a dataset of fire - fighting equipment with 1000 samples, and each sample has 5 features (such as working hours, power, fault alarm, location information, maintenance records). We can build and train a Transformer model according to the following steps:

[0055] Preprocess the dataset and perform feature engineering.

[0056] Build a Transformer model and initialize its parameters.

[0057] Input the training feature set into the model and optimize the parameters using the Adam optimizer.

[0058] Dynamically adjust the learning rate during training and monitor the change of the loss value.

[0059] When the loss value is less than the preset threshold, it is considered that the model has converged.

[0060] Use cross-validation and the test set to evaluate the performance of the model.

[0061] Adjust and optimize the model according to the evaluation results until the optimal equipment fault prediction model is obtained.

[0062] The present invention has the following beneficial effects:

[0063] 1. In the present invention, the Transformer model based on the attention mechanism is used as the equipment fault prediction model. An equipment fault prediction model is constructed. Through the equipment fault prediction model, intelligent analysis is performed on the real-time or recent equipment operation data, and the fault prediction result is output. Warning information is generated according to the prediction result, and according to the warning information and the equipment status data, particle swarm optimization is used to formulate a maintenance plan, providing a scientific decision-making basis for managers, reducing the failure rate, and extending the service life of the equipment.

[0064] 2. In the present invention, seamless connection with other fire protection systems is achieved through the API interface, realizing the real-time sharing of equipment status information and coordinated operations. This greatly improves the efficiency and accuracy of fire protection work, enabling the fire protection agency to respond and handle emergencies such as fires more quickly. Description of the Drawings

[0065] Figure 1 It is a system block diagram of an intelligent fire protection equipment management system proposed by the present invention. Detailed Embodiments

[0066] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0067] Please refer to Figure 1As shown in the figure, the present invention is an intelligent management system for fire-fighting equipment and instruments, including a data acquisition layer, a data processing layer, an intelligent analysis layer, an application service layer, and a user interface layer;

[0068] The data acquisition layer is connected to an Internet of Things data acquisition module. The Internet of Things data acquisition module is responsible for real-time collecting various data such as the operating status, location information, and maintenance records of fire-fighting equipment and instruments, and uploading them to an edge computing node for preliminary processing;

[0069] The data processing layer includes a data preprocessing module. The data preprocessing module processes and analyzes the collected data, including data compression, encryption, cleaning, conversion, etc., to ensure the accuracy and security of the data;

[0070] The intelligent analysis layer includes a deep learning analysis module. The deep learning analysis module constructs an equipment failure prediction model, conducts intelligent analysis on the processed data, predicts equipment failures and shortages, provides a scientific basis for decision-making support. The deep learning analysis module conducts intelligent analysis on real-time or recent equipment operation data, outputs failure prediction results, generates early warning information according to the prediction results, and uses particle swarm optimization to formulate a maintenance plan based on the early warning information and equipment status data;

[0071] The application service layer includes a user-defined interface module and an API interface docking module. The user-defined interface module provides a user-defined interface, allowing users to adjust the system interface layout according to actual needs. The API interface docking module docks with other fire-fighting systems through the API interface to achieve real-time information sharing and coordinated operations;

[0072] The user interface layer includes a user interface display module responsible for providing a user interface and displaying various contents such as equipment status, early warning information, and maintenance plans;

[0073] The Internet of Things data acquisition module uploads the collected data of fire-fighting equipment and instruments to the data preprocessing module. The data preprocessing module transfers the processed data to the deep learning analysis module. The deep learning analysis module transfers the analysis results to the user-defined interface module and the API interface docking module of the application service layer. The modular design module, user-defined interface module, and API interface docking module of the application service layer generate corresponding contents such as equipment status, early warning information, and maintenance plans according to the analysis results and user needs, and transfer this information to the user interface display module.

[0074] The working principle of an intelligent management system for fire-fighting equipment and apparatus proposed by the present invention is that the data acquisition layer, through Internet of Things technology, real-time collects data such as the operating status, location information, and maintenance records of fire-fighting equipment and apparatus. These data are the basis for subsequent analysis and decision-making. The collected data is first uploaded to the edge computing node for preliminary processing, such as data compression, encryption, etc., to ensure the transmission efficiency and security of the data. Distributed computing technology is used to deeply analyze the data, including data cleaning, transformation, etc., to extract valuable information;

[0075] Intelligent analysis constructs an equipment failure prediction model. The failure prediction model conducts intelligent analysis on the processed data to predict possible equipment failures and shortages. Through continuous learning and optimization, the prediction accuracy of the model will gradually improve. When it is predicted that there may be equipment failures or shortages, the system will automatically trigger an early warning mechanism. The warning information will be sent to the management personnel in a timely manner to prompt them to take corresponding measures, such as maintenance, replacement, or procurement of new equipment;

[0076] The application service layer provides modular design and user-defined interfaces to meet the highly customized requirements of fire-fighting agencies for system functions. At the same time, it seamlessly docks with other fire-fighting systems through API interfaces to achieve real-time information sharing and coordinated operations. According to the warning information and equipment status data, the intelligent decision-making engine will automatically formulate maintenance plans and procurement strategies. These suggestions will be based on the data analysis results and provide a scientific decision-making basis for the management personnel;

[0077] The user interface layer provides an intuitive and friendly user interface, facilitating management personnel to conduct real-time monitoring and operations. The interface will display equipment status, warning information, maintenance plans, etc., enabling management personnel to comprehensively understand the operating conditions of the equipment. Through the user interface layer, management personnel can real-time monitor the operating status of fire-fighting equipment and apparatus. Once an abnormal situation is found, they can immediately take measures to handle it to ensure the normal operation of the equipment and fire safety.

[0078] In one embodiment, for the above-mentioned Internet of Things data acquisition module, the Internet of Things data acquisition module includes an Internet of Things data acquisition unit, an edge computing node unit, and an equipment management database unit. The Internet of Things data acquisition unit is responsible for collecting data of fire-fighting equipment through Internet of Things technology. The Internet of Things data acquisition unit interacts with the edge computing node and uploads the collected data. The edge computing node unit receives the data uploaded by the Internet of Things data acquisition unit and performs preliminary processing on the data, such as data cleaning, format conversion, etc. The equipment management database unit stores various data for the full life cycle management of equipment, such as equipment information, outbound records, maintenance records, scrapping records, etc. The Internet of Things data acquisition unit uploads the collected data to the edge computing node unit. The edge computing node unit receives the data and performs preliminary processing. The edge computing node unit uploads the processed data to the equipment management database unit for storage. The equipment management database unit provides operation interfaces such as data query and update for the edge computing node unit to support real-time update and synchronization of data;

[0079] Full life cycle management of equipment:

[0080] Registration and warehousing: Detailed records of various information of fire-fighting equipment are made to realize the warehousing management of equipment and provide basic data for subsequent management;

[0081] Outbound management: Record relevant information of equipment outbound to ensure the traceability of equipment use and facilitate subsequent management and analysis;

[0082] Maintenance record: Record the maintenance situation of equipment to form a complete maintenance file and provide a basis for equipment status evaluation and maintenance decision-making;

[0083] Scrapping process: Manage the scrapping process of equipment, including scrapping application, approval, processing and other links to ensure the standardization and safety of equipment management.

[0084] In one embodiment, for the above-mentioned data preprocessing module, the data preprocessing module includes a data acquisition unit, a distributed computing unit, a data storage unit, and a data interaction interface unit. The data acquisition unit is responsible for collecting raw data from the data acquisition layer. The distributed computing unit uses multiple computers or servers to work together to perform parallel processing on the data, improving the processing speed and efficiency. The data storage unit is used to store the processed data for subsequent analysis and application. The data interaction interface unit provides interfaces for data interaction with other systems or modules, such as APIs, message queues, etc. The data acquisition unit sends the collected raw data to the distributed computing unit for processing. The processed data is sent by the distributed computing unit to the data storage unit for storage. The data interaction interface unit obtains data from the data storage unit.

[0085] In one embodiment, for the above-mentioned deep learning analysis module, the deep learning analysis module includes a model construction unit, a fault prediction unit, a warning notification unit, a maintenance plan unit, and a decision support unit. The model construction unit constructs an equipment fault prediction model using deep learning algorithms. The fault prediction unit performs intelligent analysis on real-time or recent equipment operation data and outputs a fault prediction result. The warning notification unit generates a warning message based on the prediction result. The maintenance plan unit uses particle swarm optimization to formulate a maintenance plan according to the warning message and equipment status data. The decision support unit provides an equipment fault prediction report to display the prediction result and trend. The data preprocessing module transfers the preprocessed data to the model construction unit and the maintenance plan unit for model training and formulating a maintenance plan. The model construction unit transfers the trained model to the fault prediction unit. The fault prediction unit transfers the prediction result to the warning notification unit. The warning notification unit transfers the sending status and receiving record of the warning message to the decision support unit and the maintenance plan unit. The maintenance plan unit transfers the maintenance plan to the decision support unit.

[0086] In one embodiment, for the above-mentioned user-defined interface module, the user-defined interface module includes an interface layout management unit, a function setting management unit, a configuration management unit, a permission management unit, and a data storage unit. The interface layout management unit is responsible for handling user-defined operations on the interface layout, such as dragging, adding, and deleting elements, saving and loading interface layout configurations. The function setting management unit is responsible for handling user-defined operations on function settings, such as enabling / disabling functions, adjusting function parameters, verifying the validity of user configurations, and updating the system status. The configuration management unit provides a visual configuration management interface to display various configurations of the system, supports import, export, verification, and error prompts of configurations. The permission management unit manages user roles and permissions to ensure that users can only access and operate the resources they are authorized to, and records user operation logs for easy tracking and auditing. The data storage unit stores user-defined configurations, interface layouts, and function settings, and provides a data access interface for other units to read and write data. The interface layout management unit saves the user-defined interface layout configuration to the data storage unit. When the user loads the interface layout, the interface layout management unit reads the configuration from the data storage unit and applies it. The function setting management unit saves the user-defined function settings to the data storage unit. When the user adjusts the function settings, the function setting management unit reads the current configuration from the data storage unit, verifies it, and updates it.

[0087] In one embodiment, for the above API interface docking module, the API interface docking module includes an API interface management unit, a data synchronization unit, a data format conversion unit, a security authentication unit, an access control unit, and a log recording unit. The API interface management unit is responsible for the creation, maintenance, and management of API interfaces, and processes API call requests and responses with other fire protection systems. The data synchronization unit is responsible for the real-time synchronization of equipment status information, and realizes information sharing and integration with other fire protection systems. The data format conversion unit is responsible for the conversion of data formats between different systems, parses the received data, and extracts useful information. The security authentication unit performs security authentication on the API interface, such as using OAuth, API keys, etc., to verify the identity and permissions of the requester. The access control unit implements access control policies, restricts access to the API interface, monitors and records the access situation of the API interface. The log recording unit records the log information during the API interface docking process, and provides log query and analysis functions. The data synchronization unit obtains the corresponding equipment status information or other shared information from the database according to the request type, and returns it to the API interface management unit. When the received data format does not match the internal format of the system, the API interface management unit passes the data to the data format conversion unit. Before the API call request reaches the API interface management unit, the security authentication unit performs security authentication on the request. The security authentication unit verifies the identity and permissions of the requester, and passes the authentication result to the API interface management unit. The access control unit implements access control policies according to the request of the API interface management unit.

[0088] In one embodiment, for the above-mentioned user interface display module, the user interface display module includes a user interface management unit, a data display unit, a user authentication and permission management unit, a business logic processing unit, and a data visualization unit. The user interface management unit is responsible for the design and layout of the user interface, including page structure, styles, interaction logic, etc., processes user input and interaction events, and passes user operations to the corresponding business logic unit. The data display unit is responsible for obtaining equipment status, warning information, and maintenance plans from the background system and displaying them on the user interface, providing dynamic updates and refresh functions for the data to ensure the real-time nature of the user interface. The user authentication and permission management unit is responsible for user login authentication and permission management, ensuring that users can only access the content they are authorized to, and recording user login and operation logs for auditing and tracking. The business logic processing unit processes various operation requests initiated by users through the user interface, such as querying equipment status, processing warning information, editing maintenance plans, etc., interacts with the background system, executes the corresponding business logic, and returns the processing results to the user interface. The data visualization unit is responsible for converting equipment status data and analysis results into visual charts or dashboards so that users can more intuitively understand the data, provides data export and report generation functions, and supports users in using the data for analysis and decision-making.

[0089] In one embodiment, for the above-mentioned maintenance plan unit, the specific steps for formulating a maintenance plan by the maintenance plan unit according to warning information and equipment status data using particle swarm optimization are as follows:

[0090] Initialize the particle swarm: Particle position: Each particle represents a maintenance plan, and its position is composed of multiple decision variables, such as maintenance time interval, spare part quantity, maintenance personnel allocation, etc. Particle velocity: The initial velocity is randomly set. Fitness function: Define a fitness function to evaluate the quality of each maintenance plan.

[0091] Update the particle position and velocity: Personal best position (pBest): Each particle remembers the best position it has found so far. Global best position (gBest): The best position found in the entire particle swarm. The velocity update formula is the new velocity = inertia weight * old velocity + cognitive learning factor * random number * (pBest - current position) + social learning factor * random number * (gBest -

[0092] current position), and the position update formula is the new position = current position + new velocity.

[0093] Evaluate the fitness: Calculate the fitness value of the new position of each particle, and update the personal best position (pBest) and the global best position (gBest).

[0094] Iteration: Repeatedly update the particle positions and velocities and evaluate the fitness until a predetermined number of iterations is reached or the fitness value no longer improves significantly.

[0095] In one embodiment, for the above-mentioned model construction unit, the specific steps for the model construction unit to construct an equipment fault prediction model using a deep learning algorithm are as follows:

[0096] Model architecture selection:

[0097] Select a Transformer model based on the attention mechanism as the architecture of the equipment fault prediction model. The Transformer model captures the dependencies in the sequence data through the self-attention mechanism and positional encoding.

[0098] Construct the Transformer model:

[0099] The Transformer model consists of two parts: an encoder and a decoder. However, in the fault prediction task, only the encoder part is needed because our goal is to predict future equipment faults rather than generate sequences.

[0100] Initialize model parameters:

[0101] Initialize parameters such as weights and biases in the Transformer model. These parameters will be optimized through subsequent training processes.

[0102] Determine the loss function:

[0103] For equipment fault prediction, use the mean squared error (MSE) as the loss function. The definition of MSE is: where y i is the true value, is the predicted value, and n is the number of samples.

[0104] Input the training feature set:

[0105] Input the preprocessed training feature set into the encoder of the Transformer model. The features will be converted into high-dimensional vector representations through the embedding layer.

[0106] Optimize the parameters:

[0107] The Adam optimizer is used to optimize the parameters of the Transformer model. The Adam optimizer combines the advantages of Momentum and RMSProp optimizers and can converge to the optimal solution more quickly. The update rule of the Adam optimizer is as follows:

[0108] m i = β1m i-1 + (1 - β1)g i ;

[0109]

[0110]

[0111] where m i and v i are the first-order moment estimate and second-order moment estimate of the gradient respectively, β1 and β2 are the decay rates, η is the learning rate, and ∈ is a small constant to prevent division by zero.

[0112] Distributed training and parallel computing:

[0113] The training process of the model is accelerated by using multiple computers or GPUs, which can be achieved through Data Parallelism or Model Parallelism.

[0114] Adaptive learning rate adjustment:

[0115] The learning rate is dynamically adjusted according to the training situation of the model. For example, when the loss value no longer decreases significantly, the learning rate can be reduced to avoid overfitting.

[0116] Model convergence:

[0117] The model is converged by comparing the loss value between the predicted value and the measured value, and when the loss value is less than the preset threshold, it means that the model has found a better parameter configuration and can fit the training data well.

[0118] Cross-validation:

[0119] The cross-validation method is used to evaluate the performance of the model, which can help us understand the performance of the model on different datasets and avoid overfitting.

[0120] Test set evaluation:

[0121] The performance of the model is evaluated on the test set, which can give us an intuitive feeling about the generalization ability of the model.

[0122] Model adjustment and optimization:

[0123] Adjust and optimize the model according to the evaluation results. For example, try different model architectures, loss functions, optimizers, etc. to improve the performance of the model.

[0124] Suppose we have a dataset of fire-fighting equipment with 1000 samples, and each sample has 5 features (such as working hours, power, fault alarm, location information, maintenance records). We can build and train a Transformer model according to the following steps:

[0125] Preprocess the dataset and perform feature engineering.

[0126] Build a Transformer model and initialize the parameters.

[0127] Input the training feature set into the model and use the Adam optimizer to optimize the parameters.

[0128] Dynamically adjust the learning rate during training and monitor the change of the loss value.

[0129] When the loss value is less than the preset threshold, it is considered that the model has converged.

[0130] Use cross-validation and the test set to evaluate the performance of the model.

[0131] Adjust and optimize the model according to the evaluation results until the optimal equipment fault prediction model is obtained.

[0132] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent fire fighting equipment management system, characterized in that: It includes data collection layer, data processing layer, intelligent analysis layer, application service layer and user interface layer; The data collection layer includes an IoT data collection module, which is responsible for collecting various data such as the operating status, location information, and maintenance records of firefighting equipment in real time, and uploading them to the edge computing node for preliminary processing; The data processing layer includes a data preprocessing module, which processes and analyzes the collected data; The intelligent analysis layer includes a deep learning analysis module, which constructs an equipment failure prediction model, performs intelligent analysis on the processed data, and predicts equipment failures and shortages. The deep learning analysis module performs intelligent analysis on real-time or recent equipment operation data, outputs failure prediction results, generates warning information based on the prediction results, and uses particle swarm optimization to formulate maintenance plans based on the warning information and equipment status data; The application service layer includes a user-defined interface module and an API interface docking module. The user-defined interface module provides a user-defined interface, allowing the user to adjust the system interface layout according to actual needs. The API interface docking module docks with other fire protection systems through the API interface to achieve real-time information sharing and collaborative operations; The user interface layer includes a user interface display module, which is responsible for providing a user interface to display equipment status, warning information, maintenance plan and other contents; The Internet of Things data acquisition module uploads the collected data of fire-fighting equipment to the data preprocessing module, and the data preprocessing module passes the processed data to the deep learning analysis module. The deep learning analysis module selects the Transformer model based on the attention mechanism as the equipment failure prediction model, and the deep learning analysis module passes the analysis results to the user-defined interface module and the API interface docking module of the application service layer. The user-defined interface module and the API interface docking module of the application service layer generate corresponding equipment status, warning information, and maintenance plan according to the analysis results and user needs, and pass the generated information to the user interface display module.

2. According to claim 1, the intelligent firefighting equipment management system is characterized in that: The Internet of Things data collection module includes an Internet of Things data collection unit, an edge computing node unit and an equipment management database unit. The Internet of Things data collection unit is responsible for collecting data on fire-fighting equipment through Internet of Things technology. The Internet of Things data collection unit interacts with the edge computing node and uploads the collected data. The edge computing node unit receives the data uploaded by the Internet of Things data collection unit and performs preliminary processing on the data. The equipment management database unit stores various types of data for the entire life cycle management of the equipment. The Internet of Things data collection unit uploads the collected data to the edge computing node unit. The edge computing node unit receives the data and performs preliminary processing. The edge computing node unit uploads the processed data to the equipment management database unit for storage.

3. The intelligent firefighting equipment management system according to claim 1 is characterized in that: The data preprocessing module includes a data acquisition unit, a distributed computing unit, a data storage unit and a data interaction interface unit. The data acquisition unit is responsible for collecting original data from the data acquisition layer. The distributed computing unit uses multiple computers or servers to work together to process data in parallel. The data storage unit is used to store processed data. The data interaction interface unit provides an interface for data interaction with other systems or modules. The data acquisition unit sends the collected original data to the distributed computing unit for processing. The processed data is sent by the distributed computing unit to the data storage unit for storage. The data interaction interface unit obtains data from the data storage unit.

4. The intelligent firefighting equipment and apparatus management system according to claim 1, characterized in that: The deep learning analysis module includes a model building unit, a fault prediction unit, an early warning notification unit, a maintenance planning unit and a decision support unit. The model building unit uses a deep learning algorithm to build an equipment fault prediction model. The fault prediction unit performs intelligent analysis on real-time or recent equipment operation data and outputs a fault prediction result. The early warning notification unit generates early warning information according to the prediction result. The maintenance planning unit uses particle swarm optimization to formulate a maintenance plan based on the early warning information and equipment status data. The decision support unit provides an equipment fault prediction report to display the prediction results and trends. The data preprocessing module passes the preprocessed data to the model building unit and the maintenance planning unit for model training and formulating maintenance plans. The model building unit passes the trained model to the fault prediction unit. The fault prediction unit passes the prediction result to the early warning notification unit. The early warning notification unit passes the sending status and receiving status records of the early warning information to the decision support unit and the maintenance planning unit. The maintenance planning unit passes the maintenance plan to the decision support unit.

5. The intelligent firefighting equipment and apparatus management system according to claim 1, characterized in that: The user-defined interface module includes an interface layout management unit, a function setting management unit, a configuration management unit, a permission management unit and a data storage unit. The interface layout management unit is responsible for processing the user's custom operations on the interface layout, saving and loading the interface layout configuration. The function setting management unit is responsible for processing the user's custom operations on the function settings, verifying the validity of the user configuration, and updating the system status. The configuration management unit provides a visual configuration management interface, displays various configurations of the system, supports the import, export, verification and error prompts of the configuration, the permission management unit manages the user's roles and permissions, ensures that the user can only access and operate the resources for which he is authorized, and records the user's operation log. The data storage unit stores the user's custom configuration, interface layout, and function settings. The interface layout management unit saves the user's custom interface layout configuration to the data storage unit. When the user loads the interface layout, the interface layout management unit reads the configuration from the data storage unit and applies it. The function setting management unit saves the user's custom function settings to the data storage unit. When the user adjusts the function settings, the function setting management unit reads the current configuration from the data storage unit for verification and update.

6. The intelligent firefighting equipment and apparatus management system according to claim 1, characterized in that: The API interface docking module includes an API interface management unit, a data synchronization unit, a data format conversion unit, a data format conversion unit, a security authentication unit, an access control unit and a log recording unit. The API interface management unit is responsible for the creation, maintenance and management of the API interface, and processes API call requests and responses with other fire protection systems. The data synchronization unit is responsible for the real-time synchronization of equipment status information and realizes information sharing and integration with other fire protection systems. The data format conversion unit is responsible for the conversion of data formats between different systems, parsing the received data and extracting useful information. The security authentication unit performs security authentication on the API interface and verifies the identity and authority of the requester. The access control unit implements access control strategies, restricts access to the API interface, monitors and record the access status of the API interface. The log recording unit records the log information in the process of API interface docking and provides log query and analysis functions. The data synchronization unit obtains the corresponding equipment status information or other shared information from the database according to the request type, and returns it to the API interface management unit. When the received data format does not match the internal format of the system, the API interface management unit passes the data to the data format conversion unit. Before the API call request reaches the API interface management unit, the security authentication unit performs security authentication on the request. The security authentication unit verifies the identity and authority of the requester and passes the authentication result to the API interface management unit. The access control unit implements the access control policy according to the request of the API interface management unit.

7. The intelligent firefighting equipment and apparatus management system according to claim 1, characterized in that: The user interface display module includes a user interface management unit, a data display unit, a user authentication and permission management unit, a business logic processing unit and a data visualization unit. The user interface management unit is responsible for the design and layout of the user interface, processing user input and interaction events, and passing user operations to the corresponding business logic unit. The data display unit is responsible for obtaining equipment status, warning information, and maintenance plans from the background system, and displaying them on the user interface, providing dynamic update and refresh functions for data. The user authentication and permission management unit is responsible for user login authentication and permission management, ensuring that users can only access authorized content, and recording user login and operation logs for auditing and tracking. The business logic processing unit processes various operation requests initiated by users through the user interface, interacts with the background system, executes corresponding business logic, and returns the processing results to the user interface. The data visualization unit is responsible for converting equipment status data and analysis results into visual charts or dashboards, providing data export and report generation functions, and supporting users to use data for analysis and decision-making.

8. The intelligent firefighting equipment and apparatus management system according to claim 4, characterized in that: The maintenance planning unit uses particle swarm optimization to formulate specific steps of the maintenance plan based on the warning information and equipment status data: Initialize the particle swarm: Particle position: Each particle represents a maintenance plan, and its position is composed of multiple decision variables; Particle speed: The initial speed is set randomly; Fitness function: Define a fitness function to evaluate the quality of each maintenance plan; Update particle position and speed: Individual optimal position: each particle will remember the best position it has found so far; Global optimal position: the best position found in the entire particle group; Speed ​​update formula is new speed = inertia weight * old speed + cognitive learning factor * random number * (pBest-current position) + social learning factor * random number * (gBest-current position), position update formula is new position = current position + new speed; Evaluate fitness: calculate the fitness value of each particle's new position, and update the individual optimal position and the global optimal position; Iteration: Repeatedly update particle positions and velocities and evaluate fitness until a predetermined number of iterations is reached or the fitness value no longer improves significantly.

9. The intelligent firefighting equipment and apparatus management system according to claim 4, characterized in that: The specific steps of the model building unit using the deep learning algorithm to build the equipment failure prediction model are as follows: Model architecture selection: The Transformer model based on the attention mechanism is selected as the architecture of the equipment fault prediction model. The Transformer model captures the dependencies in the sequence data through the self-attention mechanism and position encoding. Build the Transformer model: The Transformer model consists of two parts: encoder and decoder; Initialize model parameters: Initialize the weights and biases in the Transformer model, which will be optimized through subsequent training processes; Determine the loss function: For equipment failure prediction, the mean square error is used as the loss function, and the MSE is defined as: Among them, y i is the true value, is the predicted value, n is the number of samples; Input training feature set: The preprocessed training feature set is input into the encoder of the Transformer model, and the features are converted into high-dimensional vector representations through the embedding layer; Optimization parameters: The Adam optimizer is used to optimize the parameters of the Transformer model. The Adam optimizer combines the advantages of the Momentum and RMSProp optimizers and can converge to the optimal solution more quickly. The update rule of the Adam optimizer is as follows: m i =β1m i-1 +(1-β1)g i ; Among them, m i and v i are the first-order moment estimate and the second-order moment estimate of the gradient, β1 and β2 are the decay rates, η is the learning rate, and ∈ is a small constant to prevent division by zero; Distributed training and parallel computing: Use multiple computers or GPUs to speed up the model training process; Adaptive learning rate adjustment: Dynamically adjust the learning rate based on the model's training status; Model convergence: By comparing the loss value between the predicted value and the measured value, the model converges when the loss value is less than the preset threshold; Cross Validation: Use cross-validation methods to evaluate the performance of the model; Test set evaluation: Evaluate the performance of the model on the test set; Model adjustment and optimization: Adjust and optimize the model based on the evaluation results.

Images