Fire pump station water supply regulation control system and method based on artificial intelligence

Abstract

This invention discloses an artificial intelligence-based water supply regulation and control system and method for fire pump stations. The system includes: an acquisition module for acquiring real-time water supply parameters and environmental status information of the fire pump station; a strategy generation module for generating an optimal water supply regulation strategy based on a pre-trained artificial intelligence model and the real-time water supply parameters and environmental status information; and a control module for executing the optimal water supply regulation strategy to achieve intelligent regulation and control of the water supply to the fire pump station. This significantly improves the operating efficiency of fire pump stations, reduces operating costs, and ensures their reliability and safety, thereby providing a more stable and efficient guarantee for fire protection systems.

Description

Technical Field

[0001] This invention relates to the field of water supply regulation and control technology, and in particular to an artificial intelligence-based water supply regulation and control system and method for fire pump stations. Background Technology

[0002] With the continuous expansion of urban areas and the sustained increase in the number of high-rise buildings, fire safety has become increasingly prominent in the urban public safety system. As a crucial means of responding to emergencies such as fires, fire pump stations play an indispensable role in urban fire water supply systems. However, existing fire pump station water supply regulation technologies generally suffer from the following shortcomings:

[0003] Most fire pump stations still rely on fixed threshold control or manual intervention for water supply regulation. Due to the highly sudden and unpredictable nature of fires, fixed threshold control often cannot dynamically adapt to the rapidly changing fire situation and water demand. For example, in cases of sudden fire escalation or a large influx of water in a short period, traditional control methods struggle to adjust pump start-up and valve opening in a timely manner, leading to insufficient water supply or excessive fluctuations, affecting firefighting efficiency and safety. Traditional systems have limited information perception capabilities regarding the status of fire pump stations and the surrounding environment. Many fire pump stations only monitor a few parameters such as their own pressure and flow rate, lacking the ability to analyze and comprehensively assess crucial information such as external weather conditions, building fire situations, and water supply network pressure distribution in real time. Furthermore, the collection and feedback of pump operating status is inadequate, resulting in a disconnect between control decisions and actual operating conditions, making it difficult to effectively balance operating energy consumption and equipment wear. Secondly, existing fire pump station water supply regulation systems lack the ability to predict future demand changes. Water demand at fire scenes is highly unpredictable, and simply relying on passive adjustments based on current operating conditions often fails to balance firefighting needs with system load. Without accurate forecasting of water demand, the system is prone to blind water allocation, resource waste, or inability to respond quickly to sudden changes in fire conditions. Finally, in the face of sudden situations or system anomalies at fire scenes, traditional systems lack rapid and effective emergency response capabilities. In the event of equipment failure or drastic changes in the fire situation, multiple manual commands and dispatches are required, making it difficult to meet water supply demands in a timely manner and posing potential safety hazards.

[0004] Therefore, there is an urgent need for an artificial intelligence-based water supply regulation and control system and method for fire pump stations. Summary of the Invention

[0005] This invention provides an artificial intelligence-based water supply regulation and control system and method for fire pump stations to solve the above-mentioned problems existing in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The AI-based fire pump station water supply regulation and control system includes:

[0008] The acquisition module is used to acquire real-time water supply parameters and environmental status information of the fire pump station;

[0009] The strategy generation module is used to generate the optimal water supply regulation strategy based on a pre-trained artificial intelligence model and real-time water supply parameters and environmental status information.

[0010] The control module is used to execute the optimal water supply regulation strategy to achieve intelligent regulation and control of the water supply to the fire pump station.

[0011] The acquisition module includes:

[0012] The data acquisition submodule is used to collect real-time pressure, flow rate, temperature, water level, and operating status data of each pump unit of the fire pump station;

[0013] The environmental monitoring submodule is used to monitor the meteorological conditions around the fire pump station, the fire status of buildings, and the pressure distribution of the water supply network.

[0014] The strategy generation module includes:

[0015] The model building submodule is used to pre-train and build an artificial intelligence model based on the historical operating water supply parameters of the fire pump station and the corresponding environmental status information.

[0016] The scene recognition submodule is used to analyze the acquired real-time water supply parameters and environmental status information based on a pre-trained artificial intelligence model to identify the current water supply scene type.

[0017] The demand forecasting submodule is used to predict the trend of water supply demand changes within a preset time period based on the identified scenario type and historical data.

[0018] The strategy reasoning submodule is used to generate multiple sets of candidate water supply regulation strategies by comprehensively considering factors such as water supply demand, system energy consumption and equipment lifespan.

[0019] The strategy evaluation submodule is used to evaluate multiple candidate strategies and select the strategy with the highest comprehensive score as the optimal water supply regulation strategy.

[0020] The control module includes:

[0021] The strategy decomposition submodule is used to perform complexity analysis on the optimal water supply regulation strategy. When the complexity value is greater than the execution complexity threshold, the strategy is decomposed into multiple sub-strategies that are executed consecutively.

[0022] The instruction conversion submodule is used to convert water supply regulation strategies or sub-strategies into specific control instructions for pump start-up and shutdown, frequency conversion speed regulation, and valve opening.

[0023] The execution monitoring submodule is used to monitor the execution effect of control commands in real time and record changes in key water supply parameters;

[0024] The emergency response submodule is used to quickly activate the corresponding emergency water supply adjustment plan based on the preset emergency plan library when a sudden change in water supply demand or system abnormality is detected at the fire scene.

[0025] The model building submodule includes:

[0026] The acquisition unit is used to acquire historical operating water supply parameters of the fire pump station, corresponding environmental status information, and corresponding water supply efficiency evaluation data.

[0027] The model building unit is used to build multi-scenario water supply demand models based on historical operating water supply parameters and corresponding environmental status information.

[0028] The sample acquisition unit is used to acquire multiple sets of training samples, including water supply demand data under different fire types, different building structures, and different environmental conditions.

[0029] The evaluation unit is used to evaluate each training sample based on the multi-scenario water supply demand model and obtain the water supply efficiency score of each training sample.

[0030] The training unit is used to train an artificial intelligence model based on training samples whose water supply efficiency scores are greater than or equal to a preset efficiency threshold.

[0031] The model building unit includes:

[0032] The data classification subunit is used to classify and process historical water supply parameters and corresponding environmental status information to obtain a multi-dimensional feature parameter set;

[0033] The rule determination subunit is used to determine the water supply rule knowledge corresponding to the multi-dimensional feature parameter set from the water supply rule knowledge base; the water supply rule knowledge includes: multiple sets of one-to-one corresponding scene feature recognition rules and water supply demand calculation rules;

[0034] The scene recognition subunit is used to sequentially traverse the scene feature recognition rules and identify specific scene types from historical operation data based on the traversed scene feature recognition rules.

[0035] The sub-model generation sub-unit is used to generate a scene water supply demand sub-model based on the water supply demand calculation rules corresponding to the scene feature recognition rules that have been traversed, according to a specific scene type.

[0036] The model integration sub-unit is used to integrate the scene water supply demand sub-models generated by each scene feature recognition rule after traversing each scene feature recognition rule, so as to obtain a multi-scene water supply demand model.

[0037] The execution monitoring submodule includes:

[0038] The monitoring unit is used to continuously monitor the operating status of the water supply system;

[0039] The analysis unit is used to extract features from the pressure data of each fluctuation if the water supply pressure fluctuation exceeds the stable range in N consecutive monitoring data, and the pattern of each fluctuation conforms to a predefined standard fluctuation relationship, and the sum of the correlation strength of the conforming standard fluctuation relationships is greater than or equal to the warning threshold, to obtain a fluctuation feature set; the warning threshold is the product of N and the risk coefficient.

[0040] The matching unit is used to match the system risk factors corresponding to the fluctuation feature set from the fault prediction knowledge base;

[0041] The prediction unit is used to predict the operating trend of the water supply system in a preset first time period in the future, based on system risk factors.

[0042] The adjustment unit is used to adjust the water supply parameters in advance based on the prediction results to prevent system failures.

[0043] The execution monitoring submodule also includes:

[0044] The mutation detection unit is used to obtain information on changes in water supply parameters within a preset second time period before and after the mutation when the system detects a sudden change in water supply demand at a fire scene.

[0045] The pattern analysis unit is used to perform time-series analysis on water supply parameter changes to obtain demand change patterns.

[0046] The matching unit is used to match demand change patterns with multiple predefined standard emergency patterns and obtain matching similarity.

[0047] The scheme selection unit is used to extract key parameters from water supply parameter change information based on the emergency response rules corresponding to the standard emergency situation pattern with the highest matching similarity.

[0048] The condition matching unit is used to match key parameters with the triggering conditions of multiple preset emergency plans. When a match is found, the preset emergency plan that matches the match and its execution priority are obtained.

[0049] The execution judgment unit is used to determine whether the current system status meets the execution conditions of the selected emergency plan;

[0050] The activation unit is used to immediately activate the emergency water supply regulation plan and optimize the on-site water supply efficiency when the system status meets the execution conditions of the selected emergency plan.

[0051] This also includes:

[0052] The knowledge graph building module is used to build a knowledge graph for fire pump station water supply regulation. The knowledge graph includes pump station equipment parameters, water supply network topology, historical fire cases and fire water supply experience.

[0053] The verification module is used to verify the rationality of the water supply regulation strategy generated by the artificial intelligence model based on the knowledge graph of fire pump station water supply regulation.

[0054] The execution module is used to execute the water supply regulation strategy when the verification results show that the strategy reliability is greater than or equal to the reliability threshold.

[0055] The optimization module is used to optimize and adjust the water supply regulation strategy based on the reason for the failure when the verification results show that the strategy reliability is less than the reliability threshold, and then re-verify until the verification passes or the maximum number of verifications is reached.

[0056] One of the methods for regulating and controlling the water supply of a fire pump station based on artificial intelligence includes:

[0057] S101: Obtain real-time water supply parameters and environmental status information of the fire pump station;

[0058] S102: Based on a pre-trained artificial intelligence model, the optimal water supply regulation strategy is generated according to real-time water supply parameters and environmental status information.

[0059] S103: Execute the optimal water supply regulation strategy to achieve intelligent regulation and control of the water supply to the fire pump station.

[0060] Compared with the prior art, the present invention has the following advantages:

[0061] An artificial intelligence-based water supply regulation and control system for fire pump stations includes: an acquisition module for acquiring real-time water supply parameters and environmental status information of the fire pump station; a strategy generation module for generating an optimal water supply regulation strategy based on a pre-trained artificial intelligence model and the real-time water supply parameters and environmental status information; and a control module for executing the optimal water supply regulation strategy to achieve intelligent regulation and control of the water supply to the fire pump station. This system can significantly improve the operating efficiency of fire pump stations, reduce operating costs, and ensure their reliability and safety, thereby providing a more stable and efficient guarantee for fire protection systems.

[0062] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention.

[0063] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0064] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0065] Figure 1 This is a structural diagram of the fire pump station water supply regulation and control system based on artificial intelligence in an embodiment of the present invention;

[0066] Figure 2 This is a structural diagram of the acquisition module in an embodiment of the present invention;

[0067] Figure 3 This is a flowchart of an artificial intelligence-based water supply regulation and control method for fire pump stations, as described in an embodiment of the present invention. Detailed Implementation

[0068] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0069] The embodiments of the present invention provide, as follows Figure 1 As shown, an artificial intelligence-based fire pump station water supply regulation and control system includes:

[0070] The acquisition module is used to acquire real-time water supply parameters and environmental status information of the fire pump station;

[0071] The strategy generation module is used to generate the optimal water supply regulation strategy based on a pre-trained artificial intelligence model and real-time water supply parameters and environmental status information.

[0072] The control module is used to execute the optimal water supply regulation strategy to achieve intelligent regulation and control of the water supply to the fire pump station.

[0073] The working principle of the above technical solution is as follows: The acquisition module collects real-time data on parameters such as pressure, flow rate, and temperature of the water supply system through a sensor network installed at key locations in the fire pump station. Simultaneously, the module also collects external environmental status information such as ambient temperature and humidity, weather conditions, and building fire risk levels. After preprocessing and filtering, this data forms a standardized dataset, which is then transmitted to the strategy generation module for analysis and processing. The system supports multiple data acquisition frequencies and can automatically adjust the sampling interval according to the level of urgency, ensuring the provision of high-frequency, high-precision parameter information at critical moments.

[0074] The strategy generation module receives real-time data from the acquisition module and inputs it into a pre-trained artificial intelligence model. This model is trained on a large amount of historical data and fire water supply cases, and possesses deep learning and predictive capabilities. Upon receiving real-time data, the model comprehensively evaluates various water supply regulation schemes by comparing current parameters with historical optimal parameters and combining environmental status information. Through weight calculation and multi-objective optimization algorithms, it quickly generates the most suitable water supply regulation strategy for the current situation, including specific indicators such as pump start-up and shutdown timing, water supply pressure adjustment values, and flow control parameters, and transmits these strategy instructions to the control module for execution.

[0075] The control module receives the optimal water supply regulation strategy instructions output by the strategy generation module and translates the strategy into specific execution commands through a standardized communication protocol. These commands directly affect various actuators in the fire pump station, including pump start / stop controllers, variable frequency speed controllers, and valve opening controllers. The control module adopts a closed-loop control method, monitoring the deviation between the execution results and the expected targets in real time, and automatically fine-tuning the execution parameters according to the magnitude of the deviation to ensure the stable operation of the water supply system according to the optimal strategy. Simultaneously, the control module also has an emergency intervention mechanism, which can accept manual commands in extreme situations to immediately adjust the system's operating status.

[0076] The beneficial effects of the above technical solution are as follows: Through artificial intelligence strategy optimization, water supply parameters can be accurately matched to the needs of different fire scenarios, reducing unnecessary energy consumption. Intelligent regulation and control reduce the risk of human error and improve the stability and reliability of the water supply system. The optimal water supply strategy can reduce frequent equipment start-ups and shutdowns and overload operation, effectively extending the service life of key equipment in the pumping station.

[0077] In another embodiment, such as Figure 2 As shown, the acquisition module includes:

[0078] The data acquisition submodule is used to collect real-time pressure, flow rate, temperature, water level, and operating status data of each pump unit of the fire pump station;

[0079] The environmental monitoring submodule is used to monitor the meteorological conditions around the fire pump station, the fire status of buildings, and the pressure distribution of the water supply network.

[0080] The working principle of the above technical solution is as follows: The data acquisition submodule continuously collects core parameters of the system operation through a network of multiple high-precision sensors deployed at key locations in the fire pump station. Specifically, pressure sensors are installed on the inlet and outlet pipes of the pump station to monitor changes in water supply pressure in real time; electromagnetic flow meters are installed on the main water supply pipe to accurately record flow data; temperature sensors are distributed at pump motors, bearings, and water supply network nodes to monitor equipment operating temperature; ultrasonic level gauges are installed in the water storage tank to track water reserve status; and the pump unit operating status is comprehensively monitored through current monitors, vibration sensors, and tachometers. These sensors collect data at a preset sampling frequency (5 times / s under normal conditions, which can be increased to 1 time / s in emergencies) and transmit the data to the data processing unit via a fieldbus network. In the data processing unit, the raw data is filtered, calibrated, and standardized to form structured data packets. The system uses edge computing technology to perform preliminary screening of abnormal data and triggers an early warning mechanism for parameters exceeding preset thresholds. Finally, the processed data is transmitted to the central control system in real time, providing a data foundation for subsequent strategy generation.

[0081] The environmental monitoring submodule establishes an environmental sensing network around the pumping station to comprehensively monitor external factors affecting fire-fighting water supply. First, meteorological monitoring devices, including temperature and humidity sensors, anemometers, and barometers, are installed at appropriate locations outside the pumping station to collect meteorological data in real time. Second, building fire status monitoring connects to the building's fire protection system via a data interface to obtain the signal status of fire alarms, smoke detectors, and heat detectors, allowing for real-time monitoring of fire development. Third, water supply network pressure distribution is collected by pressure sensor arrays distributed across various nodes of the water supply network, forming a complete network pressure distribution map. After environmental data collection, it undergoes format standardization and time synchronization processing via a dedicated data converter, and is then transmitted to the environmental data analysis platform via a secure communication network. This platform runs specific environmental model algorithms to integrate the dispersed environmental data into an environmental status assessment report, including fire risk level assessment, water supply demand forecasting, and network pressure balance recommendations. This environmental status information, along with the operating parameters of the data acquisition submodule, constitutes the input conditions for strategy generation, supporting the system in making water supply regulation decisions that are more aligned with actual conditions.

[0082] The beneficial effects of the above technical solution are as follows: multi-dimensional and high-frequency data acquisition ensures that the information obtained by the system is accurate and reliable, reducing control errors. Real-time environmental monitoring enables the system to detect changes in fire conditions in advance, pre-adjust water supply parameters, and shorten response time. Based on accurate environmental and equipment data, the system can allocate water resources and energy more rationally, improving utilization efficiency.

[0083] In another embodiment, the policy generation module includes:

[0084] The model building submodule is used to pre-train and build an artificial intelligence model based on the historical operating water supply parameters of the fire pump station and the corresponding environmental status information.

[0085] The scene recognition submodule is used to analyze the acquired real-time water supply parameters and environmental status information based on a pre-trained artificial intelligence model to identify the current water supply scene type.

[0086] The demand forecasting submodule is used to predict the trend of water supply demand changes within a preset time period based on the identified scenario type and historical data.

[0087] The strategy reasoning submodule is used to generate multiple sets of candidate water supply regulation strategies by comprehensively considering factors such as water supply demand, system energy consumption and equipment lifespan.

[0088] The strategy evaluation submodule is used to evaluate multiple candidate strategies and select the strategy with the highest comprehensive score as the optimal water supply regulation strategy.

[0089] The working principle of the above technical solution is as follows: The model building submodule systematically processes and analyzes historical data from fire pump stations to establish an artificial intelligence model for predicting water supply demand. First, this module obtains at least two years of historical operating data from the pump station database, including water pressure change curves, flow fluctuation records, pump start-up and shutdown times, and corresponding environmental status information (such as fire severity, meteorological conditions, and pipeline pressure distribution). Then, the raw data is cleaned, outliers and missing records are removed, and time-series alignment technology is used to standardize the data from different sources. Next, feature engineering methods are used to extract key features from multidimensional data, such as pressure peaks, flow rate changes, and the correlation between pump start-up frequency and environmental status. Finally, using these features as input, a deep learning neural network (mainly employing LSTM and Transformer structures) is used to train the model, establishing a mapping relationship between water supply parameters and environmental status. The training process adopts a phased strategy: first, 80% of the historical data is used for basic training; then, 10% of the data is used for fine-tuning; and finally, the remaining 10% of the data is used for model validation. The goal of model training is to minimize the mean squared error between predicted and actual values, while introducing a regularization term to control model complexity and prevent overfitting. After training, the model will be saved and incrementally updated periodically (once a week by default) using the latest data to adapt to gradual changes in system performance and environmental characteristics.

[0090] The scene recognition submodule, based on a pre-trained artificial intelligence model, analyzes and categorizes the current operating status of the pumping station in real time. Every 10 seconds, this module acquires a set of real-time water supply parameters (pressure, flow, temperature, etc.) from the data acquisition submodule and environmental status information (fire situation, weather, pipeline pressure distribution) from the environmental monitoring submodule. This real-time data is integrated into a feature vector according to a preset format. The feature vector is then input into a pre-trained neural network model, and a probability distribution for different scene types is obtained through forward computation. The system identifies the current situation based on the scene type with the highest probability and records the probability value as a confidence index. Scene types include: normal maintenance scenarios (daily water supply, periodic inspections), small-scale fire scenarios (single-point fire, initial fire), large-scale fire scenarios (multi-point fire, spreading fire), and extreme environment scenarios (extreme cold / heat, strong winds and heavy rain). When the confidence level of a scene falls below a preset threshold (default 75%), the system activates a multi-model voting mechanism, combining rule-based judgment algorithms and historical similarity comparisons to improve the reliability of scene recognition. The identification results will be recorded and displayed in real time on the system monitoring interface, and will also serve as input conditions for the demand prediction submodule.

[0091] The demand forecasting submodule predicts changes in water supply demand within a future time window based on the identified scenario type and historical data. First, the system obtains the current scenario type and its confidence level from the scenario identification submodule and extracts historical development patterns for that scenario type from the database. Next, the system selects a suitable prediction algorithm based on the scenario type: for normal maintenance scenarios, time series analysis methods (such as the ARIMA model) are primarily used; for fire scenarios, a combined prediction method is employed, integrating physical fire spread models and data-driven deep learning prediction; for extreme environment scenarios, external variables from meteorological data are introduced for comprehensive prediction. The prediction process is conducted at three time scales: short-term (30 minutes ahead, 5-minute granularity), medium-term (4 hours ahead, 30-minute granularity), and long-term (24 hours ahead, 2-hour granularity). The system updates the prediction results every 5 minutes using a sliding window technique and calculates the prediction interval (default 95% confidence interval) to represent the uncertainty of the prediction. When the actual water supply demand deviates from the prediction interval, the system triggers an online adjustment mechanism for the prediction model, increasing the weight of the latest deviation sample to improve the model's adaptability to similar situations. The prediction results are output in the form of water supply-time curves, which serve as the key input to the strategy reasoning submodule.

[0092] The strategy reasoning submodule generates a combination of water supply regulation strategies suitable for the current scenario through a multi-objective optimization algorithm. First, the system receives the water supply demand forecast curve output by the demand forecasting submodule and obtains the current operating status of the pumping station equipment from the data acquisition submodule. Then, the system establishes a multi-objective optimization model with water supply guarantee rate, energy consumption, and equipment wear as objectives. The water supply guarantee rate is defined as the ratio of actual water supply to demand, energy consumption considers electricity usage efficiency, and equipment wear is evaluated through the number of start-stop cycles and operating time. The optimization process uses an improved particle swarm optimization algorithm, starting from an initial particle swarm (each particle represents a pump group operation scheme) and iteratively approaching the optimal solution. In each iteration, the system evaluates the fitness of each particle according to the objective function and updates the particle's position and velocity. To improve algorithm efficiency, the system sets up an adaptive weight adjustment mechanism, dynamically adjusting the weights of the three objectives—water supply guarantee rate, energy consumption, and equipment wear—for different scenario types: in fire scenarios, the weight of the water supply guarantee rate is significantly increased; in normal maintenance scenarios, the weights of energy consumption and equipment wear are relatively increased. After a preset number of iterations (default 50), the system selects the Pareto optimal solution set from the final particle swarm, forming 5-8 candidate water supply regulation strategies. Each strategy includes specific pump set on / off states, speed settings, and pressure regulation parameters, as well as the expected water supply curve, energy consumption estimation, and equipment lifespan impact assessment. These candidate strategies will be passed to the strategy evaluation submodule for further screening.

[0093] The strategy evaluation submodule comprehensively evaluates candidate water supply regulation strategies using a comprehensive scoring system. First, the system receives multiple candidate strategies generated by the strategy inference submodule and reads the current scoring weight configuration from the system configuration. Then, the system scores each strategy across three dimensions: safety, reliability, and efficiency. The safety score primarily considers water supply pressure stability and the risk of exceeding limits, calculated through pressure fluctuation simulation. The reliability score focuses on equipment failure probability and maintenance requirements, evaluated based on equipment wear models and historical failure data. The efficiency score combines energy efficiency and economic cost, calculated using energy consumption models and operating costs. The scoring process employs the analytic hierarchy process (AHP), first scoring the sub-indicators under each dimension (0-100 points), then applying the corresponding weight configuration based on the current scenario type to obtain the dimension score, and finally summing them into a comprehensive score. In special cases, the system introduces additional scoring adjustment mechanisms: for example, when the urgency of a fire scenario is high, the weight of the response speed indicator is increased; when equipment is nearing its maintenance cycle, the weight of the equipment protection indicator is increased. After scoring, the system selects the strategy with the highest overall score as the optimal water supply regulation strategy and generates a detailed strategy execution report, including the adjustment values ​​of each parameter, expected effects, and risk warnings. In addition, the system retains a second-best strategy as an alternative, allowing for rapid switching if the optimal strategy fails to perform well. The final strategy is then distributed to the execution equipment via a control interface, enabling intelligent control of the pump station's water supply.

[0094] The beneficial effects of the above technical solution are as follows: by quickly identifying scenarios and predicting future demand, the system can adjust water supply parameters in advance; the multi-objective optimization algorithm balances water supply demand and energy consumption, saving electricity consumption while ensuring water supply; and by generating strategies that take into account equipment wear factors, unnecessary frequent start-ups and shutdowns can be reduced, thus extending the service life of key equipment.

[0095] In another embodiment, the control module includes:

[0096] The strategy decomposition submodule is used to perform complexity analysis on the optimal water supply regulation strategy. When the complexity value is greater than the execution complexity threshold, the strategy is decomposed into multiple sub-strategies that are executed consecutively.

[0097] The instruction conversion submodule is used to convert water supply regulation strategies or sub-strategies into specific control instructions for pump start-up and shutdown, frequency conversion speed regulation, and valve opening.

[0098] The execution monitoring submodule is used to monitor the execution effect of control commands in real time and record changes in key water supply parameters;

[0099] The emergency response submodule is used to quickly activate the corresponding emergency water supply adjustment plan based on the preset emergency plan library when a sudden change in water supply demand or system abnormality is detected at the fire scene.

[0100] The working principle of the above technical solution is as follows: The strategy decomposition submodule ensures system execution efficiency by evaluating the complexity of the optimal water supply regulation strategy. When the complexity of the optimal water supply strategy calculated by the system exceeds a preset threshold, the module will automatically decompose the complex strategy into multiple sequentially executed sub-strategies. Specifically, the module receives the optimal water supply regulation strategy from the decision-making layer; then, it analyzes the strategy using a multi-dimensional complexity evaluation algorithm, with evaluation indicators including pump switching frequency, speed regulation amplitude changes, and valve regulation complexity; finally, when the complexity value exceeds the execution threshold, the module uses a time-series decomposition algorithm to split the strategy into multiple sequentially executed sub-strategies, each with a complexity lower than the system's preset threshold, ensuring smooth execution.

[0101] The instruction conversion submodule is responsible for transforming abstract water supply regulation strategies into concrete execution instructions. This module first receives the water supply regulation strategy or sub-strategy from the strategy decomposition submodule, and then converts the strategy into equipment-level control instructions through the instruction mapping library. These instructions include: pump start-up and shutdown sequences, inverter speed control parameters, and precise control signals for the opening degrees of various valves. During the conversion process, the module considers equipment characteristics, startup sequence, and safety constraints to generate optimized instruction sequences, ensuring the safety and efficiency of instruction execution. For example, in a fire-fighting pressurized water supply scenario, the module automatically generates an instruction sequence that first starts the main pump and then adjusts the valve openings to avoid water hammer effects.

[0102] The monitoring submodule tracks the execution effect of control commands in real time and records changes in key water supply parameters. This module monitors in three ways: first, it collects real-time data such as pipeline pressure, flow rate, and pump operating status through a distributed sensor network; second, it compares and analyzes the collected data with the expected results to calculate the execution deviation; and third, it establishes parameter change curves and records the water supply response characteristics. When the detected execution deviation exceeds the allowable range, the module triggers a strategy fine-tuning mechanism to ensure that the water supply regulation effect meets expectations. For example, when insufficient pipeline pressure is detected in a certain area, the system can promptly adjust the speed of relevant pump units or the opening of valves.

[0103] The emergency response submodule is used for rapid response to sudden changes in water supply demand or system anomalies at fire scenes. This module maintains a pre-designed emergency response plan library, with strategies designed in advance for different types of emergencies (such as main pump failure, pipeline rupture, sudden increase in water demand at the fire scene, etc.). When the system detects an anomaly, the emergency response submodule quickly identifies the anomaly type using a scenario matching algorithm and retrieves the most suitable emergency water supply adjustment plan from the plan library. For example, when a main pump failure is detected, the module can activate a backup pump group and reallocate the flow path within milliseconds to ensure uninterrupted water supply.

[0104] The beneficial effects of the above technical solution are as follows: it constructs an intelligent and highly reliable fire pump station water supply regulation and control system, which can adapt to complex and ever-changing fire scenarios, provide accurate, stable and reliable water supply guarantee, and significantly improve fire rescue efficiency and safety.

[0105] In another embodiment, the model building submodule includes:

[0106] The acquisition unit is used to acquire historical operating water supply parameters of the fire pump station, corresponding environmental status information, and corresponding water supply efficiency evaluation data.

[0107] The model building unit is used to build multi-scenario water supply demand models based on historical operating water supply parameters and corresponding environmental status information.

[0108] The sample acquisition unit is used to acquire multiple sets of training samples, including water supply demand data under different fire types, different building structures, and different environmental conditions.

[0109] The evaluation unit is used to evaluate each training sample based on the multi-scenario water supply demand model and obtain the water supply efficiency score of each training sample.

[0110] The training unit is used to train an artificial intelligence model based on training samples whose water supply efficiency scores are greater than or equal to a preset efficiency threshold.

[0111] The water supply efficiency score for each training sample is obtained, including:

[0112] Obtain a sample set including multiple training samples. The sample set includes the first type of water supply parameter configuration corresponding to the first type of fire scenario and the second type of water supply parameter configuration corresponding to the second type of fire scenario. The first type of water supply parameter configuration and the second type of water supply parameter configuration are initially classified according to the first evaluation standard.

[0113] A multi-scenario water supply demand model is constructed, which includes a water supply demand feature extraction module and a water supply efficiency evaluation module for different fire-fighting scenarios.

[0114] The sample set is input into the multi-scenario water supply demand model, and the water supply parameter configuration of each training sample under different fire-fighting scenarios is simulated and calculated.

[0115] Based on the simulation results, the performance indicators of each training sample in multiple dimensions, including fire extinguishing efficiency, water resource utilization rate, and energy consumption, were obtained.

[0116] Based on the preset comprehensive scoring formula, the water supply efficiency score of each training sample is calculated, where the comprehensive scoring formula is:

[0117] E = α·F + β·W + γ·C

[0118] Where E represents the water supply efficiency score, F represents the fire extinguishing efficiency index, W represents the water resource utilization rate index, C represents the energy consumption index, and α, β, and γ are the weight coefficients of the corresponding dimensions, and α+β+γ=1.

[0119] Display the water supply efficiency score results of each training sample so that fire pump station managers can view the score data under the first type of fire scenario;

[0120] When the evaluation system detects that it has received a scene switching instruction, it adjusts the evaluation parameters of the multi-scene water supply demand model, recalculates and obtains the water supply efficiency score under the second type of fire-fighting scenario.

[0121] The results of the water supply efficiency rating under the second type of fire-fighting scenario are displayed to enable fire pump station technicians to optimize the corresponding water supply parameter configuration.

[0122] The working principle of the above technical solution is as follows: The acquisition unit is responsible for collecting historical operating data of the fire pump station, including water supply parameters such as water pressure, flow rate, and energy consumption. Simultaneously, it acquires environmental status information at corresponding time points (such as meteorological data like temperature, humidity, and wind speed), as well as the corresponding water supply efficiency evaluation results (a comprehensive evaluation index of the fire pump station's water pressure stability, water supply sufficiency, and energy utilization efficiency during firefighting). This data is collected in real time through a sensor network within the pump station and stored in a database, providing a data foundation for subsequent model construction.

[0123] The model building unit, based on historical data collected by the acquisition unit, uses data mining and machine learning algorithms to analyze the correlation between water supply parameters and water supply efficiency under different environmental conditions. This unit categorizes data according to environmental characteristics (such as high temperature, low temperature, strong wind, etc.) and building types (such as high-rise buildings, underground buildings, chemical plants, etc.), and builds a dedicated water supply demand model for each scenario to adapt to the optimal water supply strategy under different scenarios. Multi-scenario water supply demand models: a set of mathematical models established for water supply demand characteristics under different types of fires, different building structures, and different environmental conditions.

[0124] The sample acquisition unit collects multiple sets of training samples through simulation exercises, analysis of historical fire cases, and summaries of expert experience. Each set of samples contains water supply demand data under specific fire types (such as electrical fires, flammable liquid fires, etc.), building structures (such as reinforced concrete structures, wooden structures, etc.), and environmental conditions (such as high temperatures in summer, low temperatures in winter, etc.), forming a rich training dataset.

[0125] The evaluation unit inputs each set of training samples collected by the sample acquisition unit into the multi-scenario water supply demand model generated by the model building unit for evaluation. The system simulates and calculates the water supply parameter configurations for each set of samples under different scenarios, and obtains the corresponding water supply efficiency score. The scoring dimensions include multiple aspects such as fire extinguishing efficiency, water resource utilization rate, and energy consumption, comprehensively evaluating the water supply efficiency of each set of samples.

[0126] The training unit selects training samples with water supply efficiency scores greater than or equal to a preset efficiency threshold as high-quality samples. These high-quality samples are used to train artificial intelligence models (such as deep neural networks, random forests, or support vector machines) so that the models can identify the optimal water supply parameter configurations in different scenarios. The training process includes steps such as feature extraction, model parameter optimization, and cross-validation to ensure the model's generalization ability and accuracy. The preset efficiency threshold is a pre-set standard line for water supply efficiency scores; samples exceeding this line are considered to have good water supply performance and are suitable for AI model training.

[0127] The beneficial effects of the above technical solution are as follows: It improves fire extinguishing efficiency, as the system can select the optimal water supply strategy based on the fire type and building structural characteristics, reducing fire extinguishing time. It optimizes water resource utilization, avoiding insufficient or excessive water supply and achieving scientific allocation of water resources. It reduces energy consumption by intelligently adjusting the pump unit's operation mode, reducing electricity consumption while meeting fire extinguishing needs.

[0128] In another embodiment, the model building unit includes:

[0129] The data classification subunit is used to classify and process historical water supply parameters and corresponding environmental status information to obtain a multi-dimensional feature parameter set;

[0130] The rule determination subunit is used to determine the water supply rule knowledge corresponding to the multi-dimensional feature parameter set from the water supply rule knowledge base; the water supply rule knowledge includes: multiple sets of one-to-one corresponding scene feature recognition rules and water supply demand calculation rules;

[0131] The scene recognition subunit is used to sequentially traverse the scene feature recognition rules and identify specific scene types from historical operation data based on the traversed scene feature recognition rules.

[0132] The sub-model generation sub-unit is used to generate a scene water supply demand sub-model based on the water supply demand calculation rules corresponding to the scene feature recognition rules that have been traversed, according to a specific scene type.

[0133] The model integration sub-unit is used to integrate the scene water supply demand sub-models generated by each scene feature recognition rule after traversing each scene feature recognition rule, so as to obtain a multi-scene water supply demand model.

[0134] The working principle of the above technical solution is as follows: The data classification subunit first processes historical operating data, pairs and analyzes water supply parameters (such as water pressure, flow rate, and temperature) with environmental status information (such as weather conditions and peak water usage periods), and extracts a multi-dimensional feature parameter set. These parameter sets reflect the water supply characteristic patterns under different conditions.

[0135] Based on this, the rule determination subunit accesses the water supply rule knowledge base and retrieves water supply rule knowledge that matches the current feature parameter set. Each rule knowledge contains two parts: scene feature recognition rule (used to determine what kind of water supply scene is currently in) and corresponding water supply demand calculation rule (determining how to calculate water supply demand under this scene).

[0136] The scene recognition subunit sequentially traverses each scene feature recognition rule. For each rule traversed, it is applied to historical operation data to identify specific scene types, such as "fire emergency status", "routine maintenance status" or "energy-saving operation status".

[0137] For each identified scenario type, the sub-model generation unit applies the corresponding water supply demand calculation rules to generate a water supply demand sub-model for that specific scenario. For example, the sub-model for fire emergency conditions may prioritize quickly reaching high pressure and high flow rate, while the sub-model for energy-saving operation conditions may focus on minimizing energy consumption while meeting basic needs.

[0138] After traversing all scene feature recognition rules, the model integration sub-unit integrates the various scene water supply demand sub-models to form a unified multi-scene water supply demand model. This model can automatically switch to the most suitable water supply mode based on real-time monitored conditions.

[0139] The beneficial effects of the above technical solution are as follows: Through multi-dimensional parameter analysis, the system can more accurately predict and meet water supply needs in different scenarios, reducing situations of insufficient or excessive water supply. In fire emergencies, the system can quickly identify the scenario and switch to emergency mode to ensure that water pressure and flow meet fire protection requirements at critical moments. In non-emergency situations, the system can automatically switch to energy-saving mode to reduce unnecessary energy consumption and lower operating costs.

[0140] In another embodiment, the monitoring submodule includes:

[0141] The monitoring unit is used to continuously monitor the operating status of the water supply system;

[0142] The analysis unit is used to extract features from the pressure data of each fluctuation if the water supply pressure fluctuation exceeds the stable range in N consecutive monitoring data, and the pattern of each fluctuation conforms to a predefined standard fluctuation relationship, and the sum of the correlation strength of the conforming standard fluctuation relationships is greater than or equal to the warning threshold, to obtain a fluctuation feature set; the warning threshold is the product of N and the risk coefficient.

[0143] The matching unit is used to match the system risk factors corresponding to the fluctuation feature set from the fault prediction knowledge base;

[0144] The prediction unit is used to predict the operating trend of the water supply system in a preset first time period in the future, based on system risk factors.

[0145] The adjustment unit is used to adjust the water supply parameters in advance based on the prediction results to prevent system failures.

[0146] The working principle of the above technical solution is as follows: the monitoring unit continuously collects real-time operational status data of the water supply system, including key parameters such as water pressure, flow rate, temperature, and vibration. This data is collected through a sensor network and transmitted to the system's central processing module at preset time intervals (such as per second or per minute), forming a continuous time-series data stream.

[0147] The analysis unit performs intelligent analysis on the collected monitoring data. When the system detects that the water supply pressure fluctuations exceed the stable range for N consecutive times (N is a configurable parameter, such as 5 consecutive times), it further analyzes the relationship patterns between these fluctuations. The analysis unit calculates the degree of conformity between each fluctuation and a predefined standard fluctuation relationship (such as linear increase, periodic fluctuation, step-like decrease, etc.) and calculates the sum of the correlation strengths. If this sum of correlation strengths is greater than or equal to the warning threshold (warning threshold = N × risk coefficient, the risk coefficient is preset according to the importance of the system and environmental conditions), the system will extract features from the pressure data of each fluctuation to obtain a fluctuation feature set containing information such as fluctuation amplitude, frequency, and duration.

[0148] The matching unit matches the extracted fluctuation feature set with historical cases and theoretical models stored in the fault prediction knowledge base. The knowledge base contains various possible system risk factors (such as pump wear, pipeline blockage, valve failure, etc.) and their corresponding pressure fluctuation feature patterns. Through pattern recognition and similarity calculation, the matching unit can identify the system risk factor most likely corresponding to the current fluctuation feature.

[0149] Based on identified system risk factors, the prediction unit combines historical data and machine learning models to predict the operational trends of the water supply system within a preset first time period (e.g., the next 30 minutes to 2 hours). The predictions include the changing trends of key parameters (such as pressure and flow rate), as well as the probability and severity of system failures.

[0150] Based on the prediction results, the adjustment unit automatically calculates and executes the optimal parameter adjustment strategy, such as adjusting pump speed, switching to standby equipment, and adjusting valve opening, to intervene in the system's operating status in advance, preventing faults or reducing their impact. The adjustment strategy balances the relationship between water supply demand and equipment protection, ensuring that normal water supply function is not affected while preventing faults.

[0151] The beneficial effects of the above technical solution are as follows: By identifying potential risk factors early, the system can take preventative measures before a failure develops into a severe stage, avoiding the failure of the fire pump station at critical moments. Predictive maintenance, rather than reactive response, reduces the abnormal operating conditions that equipment is subjected to, significantly extending the service life of critical components. Early identification and handling of minor problems prevents them from developing into major failures requiring large-scale repairs or replacements, thereby reducing overall maintenance costs.

[0152] In another embodiment, the monitoring submodule further includes:

[0153] The mutation detection unit is used to obtain information on changes in water supply parameters within a preset second time period before and after the mutation when the system detects a sudden change in water supply demand at a fire scene.

[0154] The pattern analysis unit is used to perform time-series analysis on water supply parameter changes to obtain demand change patterns.

[0155] The matching unit is used to match demand change patterns with multiple predefined standard emergency patterns and obtain matching similarity.

[0156] The scheme selection unit is used to extract key parameters from water supply parameter change information based on the emergency response rules corresponding to the standard emergency situation pattern with the highest matching similarity.

[0157] The condition matching unit is used to match key parameters with the triggering conditions of multiple preset emergency plans. When a match is found, the preset emergency plan that matches the match and its execution priority are obtained.

[0158] The execution judgment unit is used to determine whether the current system status meets the execution conditions of the selected emergency plan;

[0159] The activation unit is used to immediately activate the emergency water supply regulation plan and optimize the on-site water supply efficiency when the system status meets the execution conditions of the selected emergency plan.

[0160] The working principle of the above technical solution is as follows: the mutation detection unit is responsible for monitoring the water supply system parameters in real time. When a sudden change in water supply demand is detected at the fire scene, the system will automatically capture key water supply parameters such as pressure changes, flow fluctuations, and temperature changes within a preset time period before and after the mutation, providing a basis for subsequent analysis.

[0161] The pattern analysis unit receives captured information on changes in water supply parameters and performs time-series analysis. This unit uses time-series algorithms to analyze the trends, magnitudes, and rates of parameter changes, thereby identifying specific demand change patterns. For example, a sharp drop in pressure accompanied by a sudden increase in flow rate in a certain area indicates a pipe burst pattern.

[0162] The matching unit compares the identified demand change patterns with various pre-stored standard emergency scenarios in the system. By calculating a similarity matrix, it determines which predefined scenario the current situation most closely resembles. Different scenarios, such as the initial fire spread mode and the controlled combustion mode, each have their own combination of characteristic parameters.

[0163] The scheme selection unit extracts key decision parameters, such as the required pressure range and minimum guaranteed flow rate, from the currently acquired water supply parameter change information based on the standard pattern with the highest matching similarity and the corresponding emergency response rules.

[0164] The condition matching unit matches the extracted key parameters with the trigger conditions of various preset emergency plans in the system. When the trigger condition of an emergency plan is met, the system identifies one or more emergency plans that are most suitable for the current situation and determines their execution priority according to a preset algorithm.

[0165] The decision-making unit assesses the pump station's current operating status, equipment health, and available resources to determine whether the conditions for executing the selected emergency plan are met. This unit ensures that executing the emergency plan will not lead to more serious problems for the system.

[0166] Once the execution conditions are confirmed, the activation unit immediately initiates the corresponding emergency water supply regulation plan according to priority. This unit is responsible for translating decisions into specific control commands, adjusting pump and valve combinations, changing water supply pressure and flow parameters, and optimizing on-site water supply efficiency.

[0167] The beneficial effects of the above technical solution are as follows: the system can quickly identify emergencies in the early stages of sudden changes in water supply demand, shortening the time lag between the occurrence of a problem and the implementation of countermeasures. By intelligently identifying the type and stage of a fire, the system can provide the most suitable water supply parameters for the current situation, avoiding resource waste or insufficient supply. A multi-level matching and judgment mechanism ensures that the system can accurately distinguish between different types of sudden changes in water supply, reducing false alarms and incorrect responses.

[0168] In another embodiment, it further includes:

[0169] The knowledge graph building module is used to build a knowledge graph for fire pump station water supply regulation. The knowledge graph includes pump station equipment parameters, water supply network topology, historical fire cases and fire water supply experience.

[0170] The verification module is used to verify the rationality of the water supply regulation strategy generated by the artificial intelligence model based on the knowledge graph of fire pump station water supply regulation.

[0171] The execution module is used to execute the water supply regulation strategy when the verification results show that the strategy reliability is greater than or equal to the reliability threshold.

[0172] The optimization module is used to optimize and adjust the water supply regulation strategy based on the reason for the failure when the verification results show that the strategy reliability is less than the reliability threshold, and then re-verify until the verification passes or the maximum number of verifications is reached.

[0173] The working principle of the above technical solution is as follows: the knowledge graph building module collects and integrates various information from fire pump stations to construct a complete knowledge graph. The system first collects pump station equipment parameters (such as pump flow rate, head, and power), then maps the water supply network topology (including pipe length, diameter, and connection points), and records historical fire cases and fire water supply experience. After structured processing, this data forms a multi-dimensional knowledge network, enabling the system to understand the complex relationships within the pump station's operating environment. For example, the knowledge graph contains related information such as "the third water pump is connected to the high-altitude water distribution tank," "the maximum flow rate of the DN200 pipe is X cubic meters per hour," and "historical data shows that increasing the water supply pressure in Zone B during a fire in the commercial area can improve fire extinguishing efficiency."

[0174] Once the AI ​​model generates a water supply regulation strategy, the verification module verifies its rationality based on an established knowledge graph. The system checks multiple dimensions, including whether the strategy complies with equipment parameter limitations, is suitable for the current pipeline network status, and is consistent with historical successful cases. For example, if the AI ​​suggests "increasing the operating frequency of pump 1 to 45Hz and closing the valve on pipeline 2," the verification module will check: whether pump 1 can safely operate at that frequency, whether closing pipeline 2 will cause insufficient water supply to other areas, and whether this combination has been used in similar fire situations. The system calculates a strategy reliability score and compares it with a preset threshold.

[0175] When the verification results indicate that the strategy's reliability is greater than or equal to the reliability threshold (e.g., 85%), the execution module will send instructions to the relevant equipment control systems. The system transmits precise control parameters to equipment such as pump inverters and electric valves via standard communication protocols (e.g., MODBUS, OPC UA) to achieve automatic strategy execution. For example, after confirming the reliability of the strategy "increase the outlet pressure of pump No. 3 to 0.6 MPa and start the high-zone booster pump," the system will send specific control signals to the relevant equipment, causing the equipment to operate according to the strategy requirements.

[0176] When the verification results show that the strategy reliability is less than the reliability threshold, the optimization module analyzes the reasons for the failure. The system identifies specific unreasonable points (such as "Pump No. 4 cannot simultaneously meet pressure and flow requirements"), and then adjusts the original strategy locally or entirely based on alternative solutions and empirical rules in the knowledge graph. For example, when it is found that "closing the ring network valve A3 will cause insufficient water supply in the south area," the system adjusts it to "partially close valve A3 and simultaneously open the backup pipeline B5." The modified strategy re-enters the verification process until it passes verification or reaches the maximum number of verifications (e.g., 5 times).

[0177] The beneficial effects of the above technical solution are as follows: the system can quickly generate and verify water supply regulation strategies, significantly shortening the time from the occurrence of a fire to the implementation of the optimal water supply strategy, and improving fire rescue efficiency. Through knowledge graphs and verification mechanisms, errors that may arise from human decision-making in emergency situations are reduced, improving the accuracy and safety of water supply regulation. Expert experience and historical fire rescue data are structured and stored in the knowledge graph, ensuring that valuable fire water supply experience is preserved and applied, and not lost due to personnel changes.

[0178] In another embodiment, such as Figure 3 As shown, an artificial intelligence-based method for regulating and controlling the water supply of a fire pump station includes:

[0179] S101: Obtain real-time water supply parameters and environmental status information of the fire pump station;

[0180] S102: Based on a pre-trained artificial intelligence model, the optimal water supply regulation strategy is generated according to real-time water supply parameters and environmental status information.

[0181] S103: Execute the optimal water supply regulation strategy to achieve intelligent regulation and control of the water supply to the fire pump station.

[0182] The working principle of the above technical solution is as follows: Step S101 includes: S1011: Collect real-time pressure, flow rate, temperature, water level and operating status data of each pump group of the fire pump station;

[0183] S1012: Monitor the meteorological conditions around the fire pump station, the fire status of buildings, and the pressure distribution of the water supply network.

[0184] Step S102 includes:

[0185] S1021: Pre-training is performed on the historical operating water supply parameters of the fire pump station and the corresponding environmental status information to build an artificial intelligence model;

[0186] S1022: Based on a pre-trained artificial intelligence model, analyze the acquired real-time water supply parameters and environmental status information to identify the current water supply scenario type;

[0187] S1023: Based on the identified scenario type and combined with historical data, predict the trend of water supply demand changes within a preset time period in the future;

[0188] S1024: Taking into account factors such as water supply demand, system energy consumption and equipment lifespan, generate multiple sets of candidate water supply regulation strategies;

[0189] S1025: Evaluate multiple candidate strategies and select the strategy with the highest comprehensive score as the optimal water supply regulation strategy.

[0190] Step S103 includes:

[0191] S1031: Perform complexity analysis on the optimal water supply regulation strategy. When the complexity value is greater than the execution complexity threshold, decompose the strategy into multiple sub-strategies that are executed consecutively.

[0192] S1032: Convert the water supply regulation strategy or sub-strategy into specific control commands for pump start-up and shutdown, frequency conversion speed regulation, and valve opening.

[0193] S1033: Real-time monitoring of the execution effect of control commands and recording changes in key water supply parameters;

[0194] S1034: When a sudden change in water supply demand or system anomaly is detected at a fire scene, the corresponding emergency water supply adjustment plan shall be quickly activated based on the preset emergency plan library.

[0195] Building artificial intelligence models, including:

[0196] Obtain historical operating water supply parameters, corresponding environmental status information, and corresponding water supply efficiency evaluation data of fire pump stations;

[0197] Based on historical water supply parameters and corresponding environmental status information, a multi-scenario water supply demand model is constructed.

[0198] Obtain multiple sets of training samples, including water supply demand data under different fire types, different building structures, and different environmental conditions;

[0199] Based on the multi-scenario water supply demand model, each training sample is evaluated to obtain a water supply efficiency score for each training sample.

[0200] The artificial intelligence model is trained based on training samples whose water supply efficiency scores are greater than or equal to a preset efficiency threshold.

[0201] Construct a multi-scenario water supply demand model, including:

[0202] Historical water supply parameters and corresponding environmental status information are classified and processed to obtain a multi-dimensional feature parameter set.

[0203] The water supply rule knowledge corresponding to the multi-dimensional feature parameter set is determined from the water supply rule knowledge base; the water supply rule knowledge includes: multiple sets of one-to-one corresponding scene feature recognition rules and water supply demand calculation rules;

[0204] The scene feature recognition rules are iterated through in sequence, and the specific scene type is identified from the historical operation data based on the iterated scene feature recognition rules.

[0205] Based on the water supply demand calculation rules corresponding to the scene feature recognition rules that have been traversed, a scene water supply demand sub-model is generated according to the specific scene type.

[0206] After traversing the feature recognition rules for each scenario, the sub-models of water supply demand for each scenario generated by the feature recognition rules are integrated to obtain a multi-scenario water supply demand model.

[0207] Real-time monitoring of the execution effect of control commands, including:

[0208] Continuously monitor the operating status of the water supply system;

[0209] If N consecutive monitoring data show that the water supply pressure fluctuations exceed the stable range, and the patterns of each fluctuation conform to a predefined standard fluctuation relationship, and the sum of the correlation strengths of the conforming standard fluctuation relationships is greater than or equal to the warning threshold, feature extraction is performed on the pressure data of each fluctuation to obtain a fluctuation feature set; the warning threshold is the product of N and the risk coefficient.

[0210] Match system risk factors corresponding to fluctuation feature sets from the fault prediction knowledge base;

[0211] Based on system risk factors, the operating trend of the water supply system in the first preset time period is predicted;

[0212] Based on the forecast results, water supply parameters are adjusted in advance to prevent system failures.

[0213] Real-time monitoring of the execution effect of control commands also includes:

[0214] When the system detects a sudden change in water supply demand at a fire scene, it acquires information on the changes in water supply parameters within a preset second time period before and after the change.

[0215] Time-series analysis of water supply parameter changes is performed to obtain demand change patterns;

[0216] The demand change pattern is matched with multiple predefined standard emergency patterns to obtain the matching similarity.

[0217] Based on the emergency response rules corresponding to the standard emergency situation pattern with the highest matching similarity, key parameters are extracted from the water supply parameter change information.

[0218] The key parameters are matched with the triggering conditions of multiple preset emergency plans. When a match is found, the preset emergency plan that matches and its execution priority are obtained.

[0219] Determine whether the current system status meets the execution conditions of the selected emergency plan;

[0220] When the system status meets the execution conditions of the selected emergency plan, the emergency water supply adjustment plan will be activated immediately to optimize the on-site water supply efficiency.

[0221] The water supply regulation and control methods for fire pump stations also include:

[0222] Establish a knowledge graph for fire pump station water supply regulation, which includes pump station equipment parameters, water supply network topology, historical fire cases and fire water supply experience;

[0223] Based on the knowledge graph of water supply regulation in fire pump stations, the rationality of water supply regulation strategies generated by artificial intelligence models is verified.

[0224] When the verification results show that the reliability of the strategy is greater than or equal to the reliability threshold, the water supply regulation strategy is executed.

[0225] When the verification results show that the reliability of the strategy is less than the reliability threshold, the water supply regulation strategy is optimized and adjusted based on the reason for the verification failure, and then verified again until the verification is passed or the maximum number of verifications is reached.

[0226] The beneficial effects of the above technical solution are as follows: By acquiring real-time water supply parameters and environmental status information of the fire pump station, the current water supply situation can be accurately grasped. With the help of an artificial intelligence model, the water supply strategy can be adjusted in real time to ensure that the fire pump station's water supply system is always in optimal working condition, thereby improving water supply efficiency. Based on a pre-trained artificial intelligence model, it can automatically analyze and process large amounts of complex water supply data to generate the optimal water supply regulation strategy adapted to the current situation. This intelligent regulation and control can effectively reduce human error and improve the system's response speed and accuracy. By precisely regulating the water supply system, unnecessary energy waste is avoided, and the operating load of the pump station is reduced, thereby effectively reducing energy consumption and operating costs, achieving the goal of energy conservation and emission reduction.

[0227] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An artificial intelligence-based water supply regulation and control system for fire pump stations, comprising: The acquisition module is used to acquire real-time water supply parameters and environmental status information of the fire pump station; The strategy generation module is used to generate the optimal water supply regulation strategy based on a pre-trained artificial intelligence model and real-time water supply parameters and environmental status information. The control module is used to execute the optimal water supply regulation strategy and realize intelligent regulation and control of the water supply to the fire pump station; The control module includes: a strategy decomposition submodule, used to perform complexity analysis on the optimal water supply regulation strategy, and decompose the strategy into multiple continuously executed sub-strategies when the complexity value exceeds the execution complexity threshold; an instruction conversion submodule, used to convert the water supply regulation strategy or sub-strategies into specific control instructions for pump start / stop, frequency conversion speed regulation, and valve opening; an execution monitoring submodule, used to monitor the execution effect of control instructions in real time and record changes in key water supply parameters; and an emergency handling submodule, used to quickly activate the corresponding emergency water supply regulation scheme based on a preset emergency scheme library when a sudden change in water supply demand or system anomaly is detected at a fire scene. The monitoring submodule includes: a monitoring unit for continuously monitoring the operating status of the water supply system; an analysis unit for extracting features from the pressure data of each fluctuation if N consecutive monitoring data show that the water supply pressure fluctuations exceed the stable range, and the patterns of each fluctuation conform to a predefined standard fluctuation relationship, and the sum of the correlation strengths of the conforming standard fluctuation relationships is greater than or equal to the warning threshold, to obtain a fluctuation feature set; the warning threshold is the product of N and the risk coefficient; a matching unit for matching the system risk factors corresponding to the fluctuation feature set from the fault prediction knowledge base; a prediction unit for predicting the operating trend of the water supply system in a future preset first time period based on the system risk factors; and an adjustment unit for adjusting the water supply parameters in advance based on the prediction results to prevent system failures.

2. The fire pump station water supply regulation and control system based on artificial intelligence according to claim 1, characterized in that, The acquisition module includes: The data acquisition submodule is used to collect real-time pressure, flow rate, temperature, water level, and operating status data of each pump unit of the fire pump station; The environmental monitoring submodule is used to monitor the meteorological conditions around the fire pump station, the fire status of buildings, and the pressure distribution of the water supply network.

3. The fire pump station water supply regulation and control system based on artificial intelligence according to claim 1, characterized in that, The strategy generation module includes: The model building submodule is used to pre-train and build an artificial intelligence model based on the historical operating water supply parameters of the fire pump station and the corresponding environmental status information. The scene recognition submodule is used to analyze the acquired real-time water supply parameters and environmental status information based on a pre-trained artificial intelligence model to identify the current water supply scene type. The demand forecasting submodule is used to predict the trend of water supply demand changes within a preset time period based on the identified scenario type and historical data. The strategy reasoning submodule is used to generate multiple sets of candidate water supply regulation strategies by comprehensively considering factors such as water supply demand, system energy consumption and equipment lifespan. The strategy evaluation submodule is used to evaluate multiple candidate strategies and select the strategy with the highest comprehensive score as the optimal water supply regulation strategy.

4. The fire pump station water supply regulation and control system based on artificial intelligence according to claim 3, characterized in that, The model building submodule includes: The acquisition unit is used to acquire historical operating water supply parameters of the fire pump station, corresponding environmental status information, and corresponding water supply efficiency evaluation data. The model building unit is used to build multi-scenario water supply demand models based on historical operating water supply parameters and corresponding environmental status information. The sample acquisition unit is used to acquire multiple sets of training samples, including water supply demand data under different fire types, different building structures, and different environmental conditions. The evaluation unit is used to evaluate each training sample based on the multi-scenario water supply demand model and obtain the water supply efficiency score of each training sample. The training unit is used to train an artificial intelligence model based on training samples whose water supply efficiency scores are greater than or equal to a preset efficiency threshold.

5. The fire pump station water supply regulation and control system based on artificial intelligence according to claim 4, characterized in that, The model building units include: The data classification subunit is used to classify and process historical water supply parameters and corresponding environmental status information to obtain a multi-dimensional feature parameter set; The rule determination subunit is used to determine the water supply rule knowledge corresponding to the multi-dimensional feature parameter set from the water supply rule knowledge base; the water supply rule knowledge includes: multiple sets of one-to-one corresponding scene feature recognition rules and water supply demand calculation rules; The scene recognition subunit is used to sequentially traverse the scene feature recognition rules and identify specific scene types from historical operation data based on the traversed scene feature recognition rules. The sub-model generation sub-unit is used to generate a scene water supply demand sub-model based on the water supply demand calculation rules corresponding to the scene feature recognition rules that have been traversed, according to a specific scene type. The model integration sub-unit is used to integrate the scene water supply demand sub-models generated by each scene feature recognition rule after traversing each scene feature recognition rule, so as to obtain a multi-scene water supply demand model.

6. The fire pump station water supply regulation and control system based on artificial intelligence according to claim 1, characterized in that, The monitoring submodule also includes: The mutation detection unit is used to obtain information on changes in water supply parameters within a preset second time period before and after the mutation when the system detects a sudden change in water supply demand at a fire scene. The pattern analysis unit is used to perform time-series analysis on water supply parameter changes to obtain demand change patterns. The matching unit is used to match demand change patterns with multiple predefined standard emergency patterns and obtain matching similarity. The scheme selection unit is used to extract key parameters from water supply parameter change information based on the emergency response rules corresponding to the standard emergency situation pattern with the highest matching similarity. The condition matching unit is used to match key parameters with the triggering conditions of multiple preset emergency plans. When a match is found, the preset emergency plan that matches the match and its execution priority are obtained. The execution judgment unit is used to determine whether the current system status meets the execution conditions of the selected emergency plan; The activation unit is used to immediately activate the emergency water supply regulation plan and optimize the on-site water supply efficiency when the system status meets the execution conditions of the selected emergency plan.

7. The fire pump station water supply regulation and control system based on artificial intelligence according to claim 1, characterized in that, Also includes: The knowledge graph building module is used to build a knowledge graph for fire pump station water supply regulation. The knowledge graph includes pump station equipment parameters, water supply network topology, historical fire cases and fire water supply experience. The verification module is used to verify the rationality of the water supply regulation strategy generated by the artificial intelligence model based on the knowledge graph of fire pump station water supply regulation. The execution module is used to execute the water supply regulation strategy when the verification results show that the strategy reliability is greater than or equal to the reliability threshold. The optimization module is used to optimize and adjust the water supply regulation strategy based on the reason for the failure when the verification results show that the strategy reliability is less than the reliability threshold, and then re-verify until the verification passes or the maximum number of verifications is reached.

8. A method for regulating and controlling the water supply of a fire pump station based on artificial intelligence, characterized in that, include: S101: Obtain real-time water supply parameters and environmental status information of the fire pump station; S102: Based on a pre-trained artificial intelligence model, the optimal water supply regulation strategy is generated according to real-time water supply parameters and environmental status information. S103: Execute the optimal water supply regulation strategy to achieve intelligent regulation and control of the water supply to the fire pump station; S103 includes: performing complexity analysis on the optimal water supply regulation strategy; when the complexity value exceeds the execution complexity threshold, decomposing the strategy into multiple continuously executed sub-strategies; converting the water supply regulation strategy or sub-strategies into specific control commands for pump start / stop, frequency conversion speed regulation, and valve opening; monitoring the execution effect of the control commands in real time and recording changes in key water supply parameters; and quickly activating the corresponding emergency water supply regulation scheme based on a preset emergency scheme library when a sudden change in water supply demand or system anomaly is detected at the fire scene. Continuously monitor the water supply system's operating status; if N consecutive monitoring data show that the water supply pressure fluctuations exceed the stable range, and the patterns of each fluctuation conform to a predefined standard fluctuation relationship, and the sum of the correlation strengths of the conforming standard fluctuation relationships is greater than or equal to the warning threshold, extract features from the pressure data of each fluctuation to obtain a fluctuation feature set; the warning threshold is the product of N and the risk coefficient; match the system risk factors corresponding to the fluctuation feature set from the fault prediction knowledge base; based on the system risk factors, predict the operating trend of the water supply system in the future within a preset first time period; based on the prediction results, adjust the water supply parameters in advance to prevent system failures.

Images