Fire-fighting water supply pump station control system and design method thereof

Abstract

The invention relates to the technical field of automatic control of fire-fighting engineering, and discloses a fire-fighting water supply pump station control system and a design method thereof. According to the method, modal decomposition is carried out on real-time pressure and flow signals of a fire-fighting pipe network, water pressure and flow reference models are established respectively, and fault features are matched from abnormal components so as to carry out preliminary diagnosis. On the basis, a reference model and diagnosis information are fused, a working condition evolution track of the system in a future early warning time period is deduced, and then a predictive control strategy framework is generated. According to the method, mechanical and electrical characteristic archives of a specific water pump unit are called, multi-dimensional fusion verification is carried out on the strategy framework, and it is ensured that control logic is accurately matched with the actual performance of equipment. According to the method, the conversion from passive response to active prediction is realized, the engineering feasibility of a control strategy and the equipment operation safety are guaranteed, and an optimized control system design is finally generated.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology for fire protection engineering, specifically to a fire water supply pump station control system and its design method. Background Technology

[0002] Currently, the design of control systems for fire-fighting water supply pump stations largely relies on static threshold judgments. The commonly used method is to monitor the pressure or flow rate at key nodes in the pipeline network; once the measured value falls below or exceeds a preset fixed threshold, the pump is triggered to start / stop or its speed adjusted. This control logic is essentially a passive, reactive mode. Its control strategies are often based on experience or theoretical hydraulic calculations, forming a fixed set of parameters, lacking forward-looking consideration of the system's dynamic processes.

[0003] Existing design processes typically separate the hydraulic model of the pipeline network from the characteristics of the pump units themselves. Control logic design focuses on meeting the terminal pressure or flow requirements of the pipeline network, while the pump units executing this logic are treated as idealized standard equipment. The pump unit characteristic parameters used in the design are mostly general data or theoretical curves for the specific model, without mandatory closed-loop verification against the actual mechanical performance and electrical response characteristics of the unit to be put into operation. This disconnect may lead to risks such as unit start-up failure, low operating efficiency, or mechanical and electrical overload during actual execution of the designed control commands.

[0004] Fire protection piping networks are dynamic systems, and the occurrence and evolution of faults or anomalies is a process. Fixed threshold control cannot distinguish between normal fluctuations in system pressure and abnormal fluctuations that indicate equipment failure, easily leading to malfunctions or response delays. Furthermore, control strategy design detached from the specific characteristics of each unit cannot guarantee the accuracy of control actions and the safety of equipment operation. A design methodology is needed that can anticipate the evolution trend of system operating conditions and ensure precise matching between the control strategy and the actual equipment characteristics. Summary of the Invention

[0005] The purpose of this invention is to provide a fire-fighting water supply pump station control system and its design method to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a design method for a fire-fighting water supply pump station control system, the method comprising: It receives real-time pressure and flow signals from multiple points in the fire water supply network to form an initial monitoring data stream; Modal decomposition is performed on the initial monitoring data stream, decomposing it into normal trend components and abnormal fluctuation components; Using the aforementioned normal trend components, a water pressure and flow rate baseline state model for the fire water supply pump station system is established; Identify characteristic event fragments related to equipment operation from the abnormal fluctuation components; The characteristic event fragments are matched one by one with a preset fault mode feature library to obtain preliminary anomaly diagnosis results; Based on the water pressure and flow rate baseline state model and the preliminary anomaly diagnosis results, the operating condition evolution trajectory of the fire water supply pump station system within the set early warning time window is deduced. Based on the aforementioned operating condition evolution trajectory, a strategy parameter framework including response thresholds and action logic is formulated for the fire water supply pump station control system. Access the inherent attribute file of the fire pump unit and extract the unit's mechanical characteristic curves and electrical characteristic parameters; The strategy parameter framework, the mechanical characteristic curve, and the electrical characteristic parameters are fused and verified in multiple dimensions. Based on the verification results, the final design method for the fire water supply pump station control system is generated.

[0007] Preferably, the step of performing mode decomposition on the initial monitoring data stream, decomposing it into normal trend components and abnormal fluctuation components, includes: The initial monitoring data stream is time-series aligned and missing values ​​are imputed to form a regular time series; Applying an adaptive noise-based complete set empirical mode decomposition algorithm to the regular time series yields a series of intrinsic mode function components; Calculate the sample entropy value of each intrinsic mode function component, and classify the components with sample entropy values ​​less than a set threshold as components of the normal trend component; Components with sample entropy values ​​greater than or equal to a set threshold are classified as components of the abnormal fluctuation components; All intrinsic mode function components classified as the normal trend components are reconstructed to form the normal trend components. All intrinsic mode function components classified as the anomalous fluctuation components are reconstructed to form the anomalous fluctuation components.

[0008] Preferably, the step of establishing a water pressure and flow reference state model for the fire water supply pump station system using the normal trend component includes: The normal trend component is subjected to a stationarity test. If the test fails, the difference is performed until the data is stationary. Based on the stabilized data, a vector autoregression model is constructed to characterize the dynamic relationship between multiple variables of water pressure and flow rate; The optimal lag order of the vector autoregressive model is determined using the Akaike information criterion. The coefficient matrix of the vector autoregressive model is fitted using the maximum likelihood estimation method; The fitted vector autoregressive model is compared with long-term historical data, and residual white noise test is performed. The vector autoregression model that passed the test was established as the water pressure and flow rate reference state model for the fire water supply pump station system.

[0009] Preferably, identifying feature event segments related to equipment operation from the abnormal fluctuation components includes: Slide a time window over the abnormal fluctuation component and calculate the statistical characteristics of the data within each window, including kurtosis, skewness, and root mean square. The statistical features are input into a pre-trained isolated forest model to identify outliers located at the edge of the data distribution. Using the anomaly point as the center, extend forward and backward for a set time period to capture candidate event segments; Calculate the energy spectral density of each candidate event segment and perform dynamic time warping matching with the energy spectral template of a standard fire pump in start-up, stop, and switching states. Candidate event segments whose dynamic time warp distance is less than the tolerance threshold are identified as feature event segments related to equipment actions.

[0010] Preferably, the step of matching the feature event fragments with a preset fault mode feature library one by one to obtain preliminary anomaly diagnosis results includes: Extract time-domain statistical features, frequency-domain wavelet packet energy features, and time-frequency domain features from the feature event segments; The time-domain statistical features, the frequency-domain wavelet packet energy features, and the time-frequency domain features are combined into a high-dimensional feature vector. Calculate the Mahalanobis distance between the high-dimensional feature vector and each fault feature template in the fault mode feature library; The fault mode corresponding to the fault feature template with the smallest Mahalanobis distance is selected as the candidate fault mode; If the minimum Mahalanobis distance is less than the diagnostic distance threshold, the candidate fault mode and its confidence level are output as the preliminary anomaly diagnosis result. If the minimum Mahalanobis distance is greater than or equal to the diagnostic distance threshold, the preliminary abnormality diagnosis result is marked as an unknown abnormality type.

[0011] Preferably, the step of combining the water pressure and flow rate baseline state model with the preliminary anomaly diagnosis results to deduce the operating condition evolution trajectory of the fire water supply pump station system within a set early warning time window includes: The initial state is the output state of the water pressure and flow rate reference state model at the current moment; The preliminary anomaly diagnosis results are quantified as perturbation parameters of the coefficient matrix or residual sequence of the vector autoregressive model; Within the set warning time window, the Monte Carlo simulation method is used to iterate the water pressure and flow rate reference state model that incorporates disturbance parameters; In each simulation iteration, the boundary conditions for safe system operation are dynamically determined based on the model output. By aggregating all the simulation iteration paths, a cloud map of the operating condition evolution trajectory of the fire water supply pump station system is formed, representing multiple possible future states. Key statistics are extracted from the cloud map of the operating condition evolution trajectory, including the distribution of minimum pressure, the distribution of maximum flow, and the probability of contact.

[0012] Preferably, the step of formulating a strategy parameter framework for the fire water supply pump station control system, which includes response thresholds and action logic, based on the evolution trajectory of the operating conditions, includes: Analyze the pressure minimum distribution extracted from the cloud map of the working condition evolution trajectory to determine the pressure safety lower limit under the specified guarantee probability; Analyze the distribution of maximum flow rate extracted from the cloud map of the working condition evolution trajectory to determine the upper limit of flow rate safety under the specified guarantee probability; Set the lower pressure safety limit and the upper flow safety limit as the pressure alarm response threshold and the flow alarm response threshold, respectively. Different warning levels are set based on the aforementioned contact probability, including attention warning, action warning, and emergency warning; Each warning level is configured with preset action logic for fire pumps, and the preset action logic includes at least one of starting the standby pump, frequency conversion speed control command, and closing non-critical branch valves; The pressure alarm response threshold, the flow alarm response threshold, each warning level and its corresponding preset action logic are integrated into the strategy parameter framework of the fire water supply pump station control system.

[0013] Preferably, the multi-dimensional fusion verification of the strategy parameter framework, the mechanical characteristic curve, and the electrical characteristic parameters includes: The preset action logic under different warning levels in the strategy parameter framework is mapped to the expected speed sequence and expected torque sequence of the fire pump unit. On the mechanical characteristic curve, find the operating point corresponding to the desired speed sequence and the desired torque sequence, and determine whether the operating point is within the high-efficiency region of the mechanical characteristic curve; Using the electrical characteristic parameters, calculate the theoretical current and temperature rise curves of the motor windings when the desired speed sequence is achieved; Determine whether the theoretical current exceeds the rated current of the motor, and whether the temperature rise curve is within the allowable range of the motor insulation material; If the operating point is in the high-efficiency zone, the theoretical current does not exceed the rated current, and the temperature rise is within the allowable range, then the verification is considered successful. If any condition is not met, the response threshold or action logic in the strategy parameter framework is adjusted, and the verification is repeated until all conditions are met.

[0014] Preferably, the method for generating the final fire water supply pump station control system based on the verification results includes: Record the strategy parameter framework that has passed the final verification as the core control strategy; The water pressure and flow rate benchmark state model of the fire water supply pump station system is integrated as a comparison benchmark for online status monitoring; The processing flow of the abnormal fluctuation components and the identification and matching rules of the feature event segments are integrated into a real-time fault diagnosis process. The core control strategy, the comparison benchmark for online status monitoring, and the real-time fault diagnosis process are logically and temporally integrated and encapsulated into an independent functional module. The data interface protocol and calling relationship between the core control strategy module, the online status monitoring module, and the real-time fault diagnosis module are specified. Based on the aforementioned functional modules and their interface protocols and calling relationships, a documented design method for a fire-fighting water supply pump station control system is generated.

[0015] Preferably, the present invention also includes a fire-fighting water supply pump station control system, the system including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements the steps of the fire-fighting water supply pump station control system design method described above.

[0016] Compared with the prior art, the beneficial effects of the present invention are: By receiving real-time data streams from the pipeline network and performing modal decomposition, abnormal fluctuation components are extracted and matched with a fault feature database. Based on preliminary diagnosis and combined with an established baseline state model, the system dynamically predicts the pressure and flow change trajectories within future warning time windows. This predictive capability based on the evolution trajectory of operating conditions ensures that control strategies are no longer rigid responses to the current instantaneous state. The system can adjust response thresholds in advance based on predicted trends, forming forward-looking action logic, thereby reducing false alarms and missed alarms and improving the intelligence and reliability of the system response.

[0017] After generating a preliminary control strategy framework, the mechanical characteristic curves and electrical characteristic parameters from the specific fire pump unit files are retrieved. These data, representing the actual physical performance and response capability of the equipment, are then fused and verified with the response thresholds and action logic in the control strategy from multiple dimensions. This process essentially involves "pre-executing" and verifying the control commands against the real equipment in a virtual environment. Verification can determine whether the proposed rapid start-stop frequency exceeds the motor's allowable starting current and thermal load, and whether the required water supply operating point falls within the pump's actual high-efficiency zone. This ensures that the finalized control strategy is not only theoretically correct but also feasible and safe in engineering practice, avoiding the risk of execution failure or equipment damage caused by a mismatch between the control logic and the equipment's physical limits. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the working principle of the fire-fighting water supply pump station control system design method described in this invention. Figure 2 This is a flowchart of mode decomposition. Figure 3 A flowchart for identifying characteristic event fragments; Figure 4 Comparison chart of multi-dimensional fusion verification results for fire water supply pump station control strategies; Figure 5 Radar diagram of the failure modes of the fire water supply pump station. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1This invention provides a design method for a fire-fighting water supply pump station control system. The method includes: receiving multi-point real-time pressure and flow signals from a fire-fighting water supply network, which constitute an initial monitoring data stream; performing modal decomposition on the initial monitoring data stream to decompose it into normal trend components and abnormal fluctuation components; establishing a water pressure and flow reference state model for the fire-fighting water supply pump station system using the normal trend components; identifying characteristic event fragments related to equipment actions from the abnormal fluctuation components; matching the characteristic event fragments with a preset fault mode feature library one by one to obtain preliminary anomaly diagnosis results; combining the water pressure and flow reference state model and the preliminary anomaly diagnosis results to deduce the operating condition evolution trajectory of the fire-fighting water supply pump station system within a set warning time window; formulating a strategy parameter framework for the fire-fighting water supply pump station control system that includes response thresholds and action logic based on the operating condition evolution trajectory; calling the inherent attribute files of the fire pump unit to extract the mechanical characteristic curves and electrical characteristic parameters of the unit; performing multi-dimensional fusion verification of the strategy parameter framework, mechanical characteristic curves, and electrical characteristic parameters; and generating the final design method for the fire-fighting water supply pump station control system based on the verification results.

[0021] In one embodiment of the present invention, see [reference] Figure 2 The initial monitoring data stream is subjected to modal decomposition, which decomposes it into normal trend components and abnormal fluctuation components. This includes time-series alignment and missing value imputation of the initial monitoring data stream to form a regular time series. The complete set empirical mode decomposition algorithm with adaptive noise is applied to the regular time series to obtain a series of intrinsic mode function components. The sample entropy value of each intrinsic mode function component is calculated, and the components with sample entropy values ​​less than a set threshold are classified as components of the normal trend components, while the components with sample entropy values ​​greater than or equal to the set threshold are classified as components of the abnormal fluctuation components. All intrinsic mode function components classified as normal trend components are reconstructed to form normal trend components, and all intrinsic mode function components classified as abnormal fluctuation components are reconstructed to form abnormal fluctuation components.

[0022] In practical implementation, in the fire water supply network of a commercial complex, three pressure sensors located at the pump outlet, the middle and end of the network, and an electromagnetic flowmeter on the main pipeline synchronously collect data once per second. The pressure values ​​fluctuate between 0.65 MPa and 0.68 MPa, and the flow rates vary between 12 L / s and 15 L / s. These time-series data collectively form the initial monitoring data stream. The first step in performing modal decomposition on the initial monitoring data stream is to perform time-series alignment and missing value interpolation to form a regular time series. Due to network transmission delays or instantaneous interference, the timestamps of data from different monitoring points have millisecond-level deviations. By resampling and aligning with a unified time reference point, and for individual data gaps caused by transmission packet loss, linear interpolation of adjacent time-time data is used to fill in the gaps, thereby obtaining a continuous and strictly consistent regular time series.

[0023] In practical implementation, an adaptive noise-based complete set empirical mode decomposition algorithm is applied to the regular time series. This algorithm effectively suppresses mode aliasing by repeatedly adding adaptive white noise and performing empirical mode decomposition and overall averaging, decomposing the regular time series into a series of intrinsic mode function components arranged from high frequency to low frequency. In the specific implementation, the sample entropy value of each intrinsic mode function component is calculated. The sample entropy value is used to quantify the complexity of the time series, and its calculation formula involves the probability of matching the template vector. The calculation formula is as follows:

[0024] Where: characters Represents the sample entropy value, character Represents the embedding dimension, character Represents similarity tolerance, characters Represents the length of the time series, characters Represents the dimension of The number of template vectors matched at the time, characters Represents the dimension of The number of template vectors matched at any given time. After calculating the sample entropy value of each intrinsic mode function component, components with sample entropy values ​​less than a set threshold are classified as components of the normal trend component, and components with sample entropy values ​​greater than or equal to the set threshold are classified as components of the abnormal fluctuation component. The set threshold is determined by analyzing the distribution of sample entropy values ​​of each intrinsic mode function component during stable operating periods in historical data. For example, the upper quartile of this distribution is taken as the threshold, with a value of 0.7.

[0025] In practical implementation, all intrinsic mode function components classified as normal trend components are reconstructed. The reconstruction process involves directly adding the corresponding time-time values ​​of these components to form a smooth normal trend component that reflects the long-term operating baseline of the system. Similarly, all intrinsic mode function components classified as abnormal fluctuation components are reconstructed. The reconstruction process also involves directly adding the corresponding time-time values ​​of these components to form an abnormal fluctuation component that includes transients, intermittent events, and noise. It can be understood that the classification reconstruction based on the sample entropy threshold effectively separates the components representing the stable operation of the system from those representing random disturbances, equipment actions, and potential anomalies in the original monitoring data stream.

[0026] In one embodiment of the present invention, see [reference] Figure 3A baseline state model for water pressure and flow rate of a fire-fighting water supply pump station system was established using normal trend components. This included a stationarity test on the normal trend components; if the test failed, differential processing was performed until the data was stationary. Based on the stationary data, a vector autoregression model was constructed to characterize the dynamic relationship between multiple variables of water pressure and flow rate. The optimal lag order of the vector autoregression model was determined using the Akaike information criterion. The coefficient matrix of the vector autoregression model was fitted using the maximum likelihood estimation method. The fitted vector autoregression model was compared with long-term historical operating data for residual white noise testing. The vector autoregression model that passed the test was established as the baseline state model for water pressure and flow rate of the fire-fighting water supply pump station system. Identifying feature event segments related to equipment operation from abnormal fluctuation components involves calculating the statistical characteristics of the data within each window by sliding a time window over the abnormal fluctuation components. These statistical characteristics include kurtosis, skewness, and root mean square. The statistical characteristics are then input into a pre-trained isolated forest model to identify outliers located at the edge of the data distribution. Candidate event segments are obtained by expanding forward and backward by a set time from the outliers. The energy spectral density of each candidate event segment is calculated and dynamically time-warped and matched with the energy spectral templates of standard fire pump start-up, stop, and switching states. Candidate event segments whose dynamic time warping distance is less than the tolerance threshold are identified as feature event segments related to equipment operation.

[0027] In practical implementation, a baseline state model of water pressure and flow rate for the fire-fighting water supply pump station system is established using the normal trend components. First, the stationarity of the normal trend components is tested using the extended Dickey-Fuller test. The test statistic is calculated for a pressure normal trend component sequence containing 1000 continuous sampling points. If the sequence fails the test, it is subjected to first-order differencing, and then the stationarity test is performed again until the data is stationary. In some embodiments, a vector autoregressive (VAR) model is constructed based on the stationary data to characterize the dynamic relationship between multiple variables of water pressure and flow rate. The model uses the pressure and flow rate at key points in the pipeline network as endogenous variables, and its mathematical expression involves a system containing two variables. The optimal lag order of the VAR model is determined using the Akaike Information Criterion (AIC). The Akaike AIC values ​​for VAR models with lags from 1 to 10 are calculated, and the lag order that minimizes the Akaike AIC value is selected as the model order. The coefficient matrix of the VAR model is fitted using the maximum likelihood estimation method, using historical stationary data as input, to solve for the coefficient matrix parameters that maximize the model's likelihood function. The fitted vector autoregression model was compared with long-term historical data to perform residual white noise test, the autocorrelation function and partial autocorrelation function of the model residual sequence were calculated, and the Ljung-Box test statistic was used to determine whether there is significant autocorrelation in the residual sequence.

[0028] In specific implementation, feature event segments related to equipment operation are identified from abnormal fluctuation components. First, a sliding time window with a window length of 10 seconds is applied to the abnormal fluctuation components. Statistical features of the data within each window are calculated, including kurtosis, skewness, and root mean square (RMS). For example, for a sliding window sequence composed of abnormal pressure fluctuation components, the skewness is calculated to be -0.12, the kurtosis to be 4.35, and the RMS to be 0.043 MPa. The statistical feature vector containing kurtosis, skewness, and RMS is input into a pre-trained isolated forest model. The isolated forest model evaluates the isolation degree of data points by constructing multiple isolated trees and outputs an anomaly score for each data point in the sliding window. Points with anomaly scores greater than a set threshold are identified as anomalies located at the edge of the data distribution. In some embodiments, the time is extended forward and backward by 5 seconds from the identified anomaly point to obtain a candidate event segment with a total duration of 10 seconds. The energy spectral density of each candidate event segment is calculated. This energy spectral density is obtained by calculating the signal's frequency domain energy distribution using a Fast Fourier Transform (FFT) and then dynamically time-warped to match the energy spectral density against energy spectral templates for standard fire pump start-up, stop, and switching states. Dynamic time warping is an algorithm used to measure the similarity between two time series of different lengths. It finds the optimal alignment path by bending the time axis and calculating the cumulative distance. Candidate event segments with a dynamic time warping distance less than a tolerance threshold are identified as feature event segments related to equipment action. For example, if a candidate event segment has a dynamic time warping distance of 8.7 with the "fire pump start" standard template, and the preset tolerance threshold is 10.0, then this segment is identified as a fire pump start event.

[0029] Optionally, the calculation of the extended Dickie-Fuller test statistic involves the following relationship:

[0030] Where: characters Represents the difference operator, character Represents time Observations, characters Represents a constant term, character The coefficient representing the time trend term, character Represents the lagged coefficient, character Represents the order of the lag difference term, character Representing the The coefficients of the first-order lag difference term, characters This represents the error term. The null hypothesis being tested is... This means the sequence has a unit root. It can be understood that by constructing a vector autoregression (VAR) model, we can quantitatively describe the historical values ​​of water pressure and flow rate variables themselves, as well as the mutual influence between their historical values. In practice, the VAR model treats water pressure and flow rate in the fire water supply network as a set of interrelated endogenous variables, establishing a linear relationship between the current value of each variable and its own lagged value, as well as the lagged values ​​of other variables. The model uses the Akaike Information Criterion to determine the optimal lag order, thereby quantitatively defining the time range of the impact of historical data on the current state. The coefficient matrix fitted by the maximum likelihood estimation method specifically quantifies how the water pressure variable is affected by its own historical water pressure and historical flow rates, and also quantifies how the flow rate variable is affected by its own historical flow and historical water pressure values.

[0031] In one embodiment of the present invention, a preliminary anomaly diagnosis result is obtained by matching feature event fragments with a preset fault mode feature library one by one. This includes extracting time-domain statistical features, frequency-domain wavelet packet energy features, and time-frequency features from the feature event fragments; combining the time-domain statistical features, frequency-domain wavelet packet energy features, and time-frequency features into a high-dimensional feature vector; calculating the Mahalanobis distance between the high-dimensional feature vector and each fault feature template in the fault mode feature library; selecting the fault mode corresponding to the fault feature template with the smallest Mahalanobis distance as a candidate fault mode; if the smallest Mahalanobis distance is less than the diagnostic distance threshold, the candidate fault mode and its confidence level are output as the preliminary anomaly diagnosis result; if the smallest Mahalanobis distance is greater than or equal to the diagnostic distance threshold, the preliminary anomaly diagnosis result is marked as an unknown anomaly type. By combining the water pressure and flow rate baseline state model with the preliminary anomaly diagnosis results, the operating condition evolution trajectory of the fire water supply pump station system within a set warning time window is deduced. This includes taking the output state of the water pressure and flow rate baseline state model at the current moment as the initial state, quantifying the preliminary anomaly diagnosis results into perturbation parameters of the vector autoregression model coefficient matrix or residual sequence, and iteratively introducing the perturbation parameters into the water pressure and flow rate baseline state model using the Monte Carlo simulation method within the set warning time window. In each simulation iteration, the boundary conditions for safe system operation are dynamically determined based on the model output. All simulation iteration paths are collected to form a cloud map of the operating condition evolution trajectory of the fire water supply pump station system, representing various future states. Key statistics are extracted from the operating condition evolution trajectory cloud map, including the distribution of minimum pressure, the distribution of maximum flow, and the probability of reaching the boundary.

[0032] In practical implementation, the characteristic event segments are matched one by one with a preset fault mode feature library to obtain preliminary anomaly diagnosis results. First, time-domain statistical features, frequency-domain wavelet packet energy features, and time-frequency features are extracted from the characteristic event segments. The time-domain statistical features include mean, standard deviation, kurtosis, and waveform factor. The frequency-domain wavelet packet energy features are obtained by calculating the proportion of energy at each node to the total energy after performing a three-level wavelet packet decomposition of the characteristic event segment signal. The time-frequency features are obtained by calculating the marginal spectral entropy using the Hilbert-Huang transform. In a specific example, a characteristic event segment of abnormal pressure fluctuation lasting 2 seconds has a time-domain standard deviation of 0.08 MPa and a kurtosis of 5.2; the normalized energy proportion of the 5th node after the third-level wavelet packet decomposition is 0.31; and its marginal spectral entropy is 2.45. The time-domain statistical features, frequency-domain wavelet packet energy features, and time-frequency features are combined into a multi-dimensional feature vector. The Mahalanobis distance between the multidimensional feature vector and each fault feature template in the fault mode feature library is calculated. The library pre-stores standard feature templates for various fault modes, such as "valve jamming," "pipeline micro-leakage," and "sensor drift." The Mahalanobis distance calculation considers the correlation between the dimensions of the feature vector. The fault mode corresponding to the fault feature template with the smallest Mahalanobis distance is selected as the candidate fault mode. For example, if the calculated Mahalanobis distance with the "pipeline micro-leakage" template is 2.1 and with the "valve jamming" template is 4.7, then the candidate fault mode is "pipeline micro-leakage." If the smallest Mahalanobis distance is less than the diagnostic distance threshold, the candidate fault mode and its confidence level are output as the preliminary anomaly diagnosis result. The diagnostic distance threshold is set to 3.0, and the confidence level is calculated using a mapping function based on Mahalanobis distance. For example, the output might be "Preliminary anomaly diagnosis result: Pipeline micro-leakage, confidence level 78%." If the minimum Mahalanobis distance is greater than or equal to the diagnostic distance threshold, the preliminary anomaly diagnosis result is marked as an unknown anomaly type. For example, when the minimum Mahalanobis distance between a feature event fragment and all known templates is 5.2, the output is "Preliminary anomaly diagnosis result: unknown anomaly type".

[0033] In practical implementation, the operating trajectory of the fire water supply pump station system within a set warning time window is deduced by combining the water pressure and flow rate baseline state model with the preliminary anomaly diagnosis results. The warning time window length is set to 60 seconds in the future. The output state of the water pressure and flow rate baseline state model at the current moment is taken as the initial state, which includes the current system pressure estimate of 0.66 MPa and the flow rate estimate of 13.5 L / s. The preliminary anomaly diagnosis results are quantified as perturbation parameters on the coefficient matrix or residual sequence of the vector autoregressive model. For example, if the diagnosis result is "micro-leakage in the pipeline," the corresponding quantification method is to multiply the variance matrix of the model residual sequence by a factor greater than 1 to simulate increased uncertainty, or to correct specific coefficients in the vector autoregressive model coefficient matrix that reflect the relationship between pressure and flow rate. Within the set warning time window, the water pressure and flow rate baseline state model with perturbation parameters is iteratively introduced using the Monte Carlo simulation method. The Monte Carlo simulation is performed 1000 times independently and randomly. In each simulation iteration, the model output dynamically determines whether the boundary conditions for safe system operation have been met. Boundary conditions are defined as the pressure at the most unfavorable point in the pipeline network falling below 0.50 MPa or the main pipeline flow exceeding 30 L / s. The paths from all 1000 simulation iterations are aggregated to form a trajectory cloud map representing the operating condition evolution of the fire-fighting water supply pump station system under various future states. Key statistics are extracted from the trajectory cloud map, including the distribution of minimum pressure, maximum flow, and the probability of reaching the boundary. For example, in 1000 simulations, the distribution of the minimum system pressure within the next 60 seconds shows a median of 0.55 MPa, and the distribution of the maximum flow shows a median of 22 L / s. 47 simulations reached the pressure safety boundary, with a probability of 4.7%.

[0034] It is understandable that the formula for calculating Mahalanobis distance is:

[0035] Where: characters Represents Mahalanobis distance, character The feature vector representing the feature event fragment to be diagnosed, character This represents the mean vector of a specific fault mode feature template in the fault mode feature library, and the character... The matrix representing the inverse of the covariance matrix of the fault mode feature template, character This represents the transpose of a matrix. It can be understood that Monte Carlo simulations approximate the future state distribution of a system through extensive random sampling, thereby quantifying the probability of risk in system operation under the influence of preliminarily diagnosed anomalies.

[0036] In one embodiment of the present invention, a strategy parameter framework including response thresholds and action logic is formulated for the fire water supply pump station control system based on the operating condition evolution trajectory. This includes analyzing the distribution of minimum pressure extracted from the operating condition evolution trajectory cloud map to determine the lower limit of pressure safety under a specified guarantee probability, analyzing the distribution of maximum flow rate extracted from the operating condition evolution trajectory cloud map to determine the upper limit of flow rate safety under a specified guarantee probability, setting the lower limit of pressure safety and the upper limit of flow rate safety as pressure alarm response thresholds and flow alarm response thresholds, respectively, and setting different warning levels based on the edge probability. These warning levels include attention warning, action warning, and emergency warning. For each warning level, preset action logic is configured for the fire pump. The preset action logic includes at least one of starting the standby pump, variable frequency speed control command, and closing non-critical branch valves. The pressure alarm response threshold, flow alarm response threshold, each warning level, and their corresponding preset action logic are integrated into the strategy parameter framework of the fire water supply pump station control system. The strategy parameter framework, mechanical characteristic curves, and electrical characteristic parameters are integrated and verified in multiple dimensions. This includes mapping the preset action logic under different warning levels in the strategy parameter framework to the expected speed sequence and expected torque sequence of the fire pump unit; finding the corresponding operating points on the mechanical characteristic curve for the expected speed sequence and expected torque sequence and determining whether the operating points are within the high-efficiency zone of the mechanical characteristic curve; using the electrical characteristic parameters to calculate the theoretical current and temperature rise curve of the motor winding when the expected speed sequence is achieved; and determining whether the theoretical current exceeds the rated current of the motor and whether the temperature rise curve is within the allowable range of the motor insulation material. If the operating point is in the high-efficiency zone, the theoretical current does not exceed the rated current, and the temperature rise is within the allowable range, the verification is considered successful. If any condition is not met, the response threshold or action logic in the strategy parameter framework is adjusted and the verification is repeated until all conditions are met.

[0037] In practical implementation, a strategy parameter framework including response thresholds and action logic is formulated for the fire water supply pump station control system based on the operating condition evolution trajectory. The pressure minimum distribution extracted from the operating condition evolution trajectory cloud map is analyzed to determine the lower pressure safety limit under a specified guarantee probability. The specified guarantee probability is set to 95%, meaning the pressure value corresponding to the 5th percentile of the future pressure minimum distribution is selected as the pressure safety lower limit; for example, the 5th percentile of the pressure minimum distribution is 0.52 MPa. Similarly, the flow maximum distribution extracted from the operating condition evolution trajectory cloud map is analyzed to determine the upper flow safety limit under a specified guarantee probability. Again, based on a 95% guarantee probability, the flow value corresponding to the 95th percentile of the flow maximum distribution is selected as the flow safety upper limit; for example, the 95th percentile of the flow maximum distribution is 25 L / s.

[0038] The lower pressure safety limit of 0.52 MPa and the upper flow safety limit of 25 L / s are set as the pressure alarm response threshold and flow alarm response threshold, respectively. Different warning levels are set based on the probability of reaching the system's safety boundary conditions. The probability of reaching the boundary conditions is calculated based on the proportion of simulated paths that reach the system's safety boundary conditions in Monte Carlo simulations. A probability of 0% to 2% is set as a warning of caution, 2% to 5% as an action warning, and greater than 5% as an emergency warning. Preset action logic for each warning level is configured for the fire pumps. The preset action logic for the caution warning level is to increase the system monitoring frequency and send warning information to the monitoring center. The preset action logic for the action warning level includes starting one standby pump. The preset action logic for the emergency warning level includes starting all standby pumps, sending a maximum speed command to the frequency converter, and closing some non-critical branch valves. The pressure alarm response threshold, flow alarm response threshold, each warning level, and their corresponding preset action logic are integrated into a strategy parameter framework for the fire water supply pump station control system, as shown in Table 1.

[0039] Table 1: Correspondence between Warning Levels and Action Logic

[0040] In practical implementation, the strategy parameter framework, mechanical characteristic curves, and electrical characteristic parameters are integrated and verified from multiple dimensions. First, the preset action logic under different warning levels in the strategy parameter framework is mapped to the expected speed sequence and expected torque sequence of the fire pump unit. For example, the "inverter full speed operation" action logic under the emergency warning level is mapped to the motor's expected speed linearly increasing from the current power frequency speed to the rated speed of 2950 rpm within the next 30 seconds, and the corresponding expected torque sequence is calculated based on the pump similarity law. The operating points corresponding to the expected speed sequence and expected torque sequence are found on the mechanical characteristic curve, and it is determined whether the operating points are within the high-efficiency zone of the mechanical characteristic curve. The high-efficiency zone of the fire pump's mechanical characteristic curve is usually defined as the area where the efficiency is not lower than 92% of the maximum efficiency. The theoretical current and temperature rise curve of the motor windings when the expected speed sequence is achieved are calculated using electrical characteristic parameters, including the motor stator resistance, inductance, thermal resistance, and thermal capacitance. It is determined whether the theoretical current exceeds the motor's rated current and whether the temperature rise curve is within the allowable range of the motor insulation material. The motor insulation material is Class H, and its allowable temperature rise is 125K. If the operating point is in the high-efficiency zone, the theoretical current does not exceed the rated value, and the temperature rise is within the allowable range, the verification is considered successful. If any condition is not met, the response threshold or action logic in the strategy parameter framework is adjusted and the verification is performed again. For example, if the efficiency of the operating point under the emergency warning is only 85% in the initial verification, which is not in the high-efficiency zone, the "inverter runs at full speed" action logic is adjusted to "inverter runs to 2850 rpm", the desired speed sequence is remapped, and the above verification is performed again until all conditions are met.

[0041] Understandable, lower limit of pressure safety Based on the specified guarantee probability The relationship, calculated from the pressure minimum distribution, satisfies:

[0042] Where: characters Represents the lower limit of safe pressure, character A random variable representing the minimum value of future pressure, character Represents the specified guarantee probability, symbol This represents probability. It is understandable that multi-dimensional fusion verification ensures that the proposed control strategy not only logically responds to system risks but also physically conforms to the actual operating capacity and safety limitations of the pump unit, avoiding equipment overload or damage due to improper control commands. In some embodiments, the calculation of the temperature rise curve involves solving a thermal model of the motor, which is a series of differential equations whose solutions characterize the change of winding temperature over time.

[0043] See Figure 4 This is a comparison chart of the multi-dimensional fusion verification results of the fire water supply pump station control strategy, primarily showing the matching of the values ​​of each verification dimension with the qualified thresholds before and after the adjustment. The chart reflects the multi-dimensional fusion verification and optimization process of the fire water supply pump station control strategy. Before the adjustment, only "speed compliance" met the standard, while the other four dimensions did not meet the safety / efficiency requirements; after the adjustment, all dimensions met the qualified thresholds, achieving adaptation between the control strategy and the mechanical and electrical characteristics of the pump unit. This type of verification ensures that the control strategy not only responds to the needs of fire-fighting operations but also avoids problems such as equipment overload and low efficiency, making it a key verification step for the safe and stable operation of the fire water supply pump station control system.

[0044] In one embodiment of the present invention, a design method for generating the final fire water supply pump station control system based on the verification results includes recording the strategy parameter framework that has passed the final verification as the core control strategy, integrating the water pressure and flow reference state model of the fire water supply pump station system as the comparison benchmark for online status monitoring, integrating the processing flow of abnormal fluctuation components and the identification and matching rules of characteristic event segments as the real-time fault diagnosis flow, logically and temporally integrating the core control strategy, the comparison benchmark for online status monitoring, and the real-time fault diagnosis flow into independent functional modules, defining the data interface protocol and calling relationship between the core control strategy module, the online status monitoring module, and the real-time fault diagnosis module, and generating a documented design method for the fire water supply pump station control system based on the functional modules and their interface protocols and calling relationships.

[0045] In practical implementation, the design method for the final fire water supply pump station control system is generated based on the verification results. The final verified strategy parameter framework is recorded as the core control strategy. The final verified strategy parameter framework includes the pressure alarm response threshold of 0.52MPa and the flow alarm response threshold of 25L / s, determined through multi-dimensional fusion verification, as well as the warning levels and specific action logic defined in Table 1. The core control strategy is stored in the form of an executable configuration file. The water pressure and flow rate benchmark state model of the fire water supply pump station system is integrated as the comparison benchmark for online status monitoring. The water pressure and flow rate benchmark state model is a vector autoregression model that has undergone parameter estimation and verification. Its coefficient matrix, lag order, and residual variance matrix are encapsulated as a model parameter file, which is used to calculate the predicted value of the system status online and compare it with the actual measured value in real time. The process of integrating the processing flow of abnormal fluctuation components and the identification and matching rules of feature event segments are used as a real-time fault diagnosis process. The processing flow includes modal decomposition of real-time data stream to obtain abnormal fluctuation components, extracting statistical features by sliding window on abnormal fluctuation components, identifying anomalies by isolated forest model, extracting candidate event segments and performing dynamic time warping matching. The matching rules include calculating the Mahalanobis distance between the feature vector of the candidate event segment and the template in the fault mode feature library and comparing it with the diagnostic distance threshold.

[0046] In practical implementation, the core control strategy, the online status monitoring benchmark, and the real-time fault diagnosis process are logically and temporally integrated and encapsulated into independent functional modules. The core control strategy is encapsulated as a strategy execution module, whose input is the system's real-time status and warning level, and whose output is a specific sequence of equipment control commands. The online status monitoring benchmark is encapsulated as a status assessment module, whose input is real-time pressure and flow data from multiple points in the pipeline network, and whose output is a judgment on whether the system status deviates from the benchmark model and a quantified value of the deviation. The real-time fault diagnosis process is encapsulated as a fault diagnosis module, whose input is abnormal fluctuation component data, and whose output is a preliminary anomaly diagnosis result and its confidence level. The data interface protocol and calling relationship between the core control strategy module, the online status monitoring module, and the real-time fault diagnosis module are specified. The data interface protocol adopts a lightweight data exchange standard based on JSON format, defining the structure and field types of data packets transmitted between modules. The calling relationship stipulates that the status assessment module and the fault diagnosis module process sensor data in parallel, and the outputs of the status assessment module and the fault diagnosis module are jointly input to the strategy execution module, which makes a comprehensive decision.

[0047] A documented design method for a fire-fighting water supply pump station control system is generated based on functional modules, their interface protocols, and calling relationships. The design method document uses a unified modeling language (UML) to describe the system architecture and module interaction sequence. For example, the system architecture diagram shows the hierarchical relationship between the sensor layer, data processing layer, control decision layer, and equipment execution layer; the interaction sequence diagram describes the complete time sequence from data acquisition, parallel execution of status assessment and fault diagnosis, to the comprehensive processing by the strategy execution module and the issuance of control commands. In some embodiments, the definition of the interface data packet format follows a specific paradigm. A data packet format sent from the fault diagnosis module to the strategy execution module includes a timestamp, diagnostic event identifier, fault mode type, confidence level, and feature vector fragments, and its structure is described as follows:

[0048] Where: characters Represents data packets, characters Represents an event timestamp, a character Represents a unique identifier for a diagnostic event, a character Represents the fault mode type, character Represents diagnostic confidence level, character This represents the feature vector of the event. It's understandable that decomposing complex system designs into highly cohesive, loosely coupled functional modules, connected through explicit interface protocols, facilitates system implementation, testing, and maintenance. It's also understandable that a documented design methodology includes not only software logic but also hardware deployment recommendations, such as sensor accuracy requirements, controller performance metrics, and network communication redundancy configurations. In some embodiments, the design methodology document also includes unit test cases and integration test scenarios for the modules to verify the correctness and interoperability of each functional module. Optionally, the appendix to the design methodology document provides detailed field descriptions for the core control strategy configuration file, model parameter file, and fault feature library file.

[0049] See Figure 5 This is a radar chart of fault modes in a fire-fighting water supply pump station, used to compare the degree of abnormality of different fault types across multiple operational indicators. This radar chart is a core tool for fault diagnosis in fire-fighting water supply pump stations, quickly distinguishing fault types through abnormal combinations of various indicators; it also visually displays the impact range of different faults. This type of visualization helps maintenance personnel quickly locate fault types and is a key component of real-time monitoring and fault tracing systems for fire-fighting water supply pump stations.

[0050] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0051] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method of designing a fire service water pumping station control system, characterized by, The method comprises: Receiving multi-point real-time pressure and flow signals of a fire water supply network to form an initial monitoring data stream; Performing modal decomposition on the initial monitoring data stream to decompose into a normal trend component and an abnormal fluctuation component; Using the normal trend component to establish a water pressure and flow reference state model of a fire water supply pump station system; Identifying feature event segments related to equipment action from the abnormal fluctuation component; Matching the feature event segments with a preset fault mode feature library piece by piece to obtain preliminary abnormal diagnosis results; Combining the water pressure and flow reference state model and the preliminary abnormal diagnosis results to deduce an operating condition evolution trajectory of the fire water supply pump station system within a set warning time window; Formulating a strategy parameter framework including a response threshold and an action logic for a fire water supply pump station control system according to the operating condition evolution trajectory; Calling inherent attribute archives of a fire water pump unit to extract mechanical characteristic curves and electrical characteristic parameters of the unit; Fusing and verifying the strategy parameter framework, the mechanical characteristic curves and the electrical characteristic parameters in multiple dimensions; Generating a final design method of the fire water supply pump station control system according to a verification result.

2. The method of designing a control system for a fire service water pumping station as claimed in claim 1, wherein, The modal decomposition on the initial monitoring data stream to decompose into a normal trend component and an abnormal fluctuation component comprises: Performing time series alignment and missing value interpolation on the initial monitoring data stream to form a regular time series; Applying an adaptive noise complete ensemble empirical mode decomposition algorithm to the regular time series to obtain a series of intrinsic mode function components; Calculating sample entropy values of each intrinsic mode function component, and classifying components with sample entropy values less than a set threshold as parts of the normal trend component; Classifying components with sample entropy values greater than or equal to the set threshold as parts of the abnormal fluctuation component; Reconstructing all intrinsic mode function components classified as the normal trend component to form the normal trend component; Reconstructing all intrinsic mode function components classified as the abnormal fluctuation component to form the abnormal fluctuation component.

3. The method of designing a control system for a fire service water pumping station as claimed in claim 1, wherein, The establishment of the water pressure and flow reference state model of the fire water supply pump station system using the normal trend component comprises: Performing stationarity test on the normal trend component, and performing difference processing until the data is stationary if the test is not passed; Based on the stationary data, constructing a vector autoregression model to depict the dynamic relationship between water pressure and flow multivariate; Using the Akaike information criterion to determine the optimal lag order of the vector autoregression model; Using the maximum likelihood estimation method to fit the coefficient matrix of the vector autoregression model; Comparing the fitted vector autoregression model with long-term historical operation data to perform residual white noise test; The vector autoregression model passing the test is established as the water pressure and flow reference state model of the fire water supply pump station system.

4. The method of designing a control system for a fire service water pumping station as claimed in claim 3, wherein, The identification of feature event segments related to equipment action from the abnormal fluctuation component comprises: Sliding a time window on the abnormal fluctuation component, calculating statistical features of data in each window, including kurtosis, skewness and root mean square; Inputting the statistical features into a pre-trained isolation forest model to identify abnormal points located at the edge of data distribution; Centering the abnormal point, the set duration is extended forward and backward to obtain a candidate event segment; The energy spectrum density of each candidate event segment is calculated, and is matched with an energy spectrum template in a standard fire pump starting, stopping and switching state through dynamic time warping; The candidate event segment with a dynamic time warping distance less than a tolerance threshold is determined as a feature event segment related to equipment action.

5. The method of designing a fire service water pumping station control system of claim 1 wherein, The feature event segment is matched with a preset fault mode feature library piece by piece to obtain a preliminary abnormal diagnosis result, including: Time domain statistical features, frequency domain wavelet packet energy features and time-frequency domain features are extracted from the feature event segment; The time domain statistical features, the frequency domain wavelet packet energy features and the time-frequency domain features are combined into a high-dimensional feature vector; Mahalanobis distances between the high-dimensional feature vector and each fault feature template in the fault mode feature library are calculated; A fault mode corresponding to a fault feature template with the smallest Mahalanobis distance is selected as a candidate fault mode; If the smallest Mahalanobis distance is less than a diagnosis distance threshold, the candidate fault mode and a confidence degree thereof are output as the preliminary abnormal diagnosis result; If the smallest Mahalanobis distance is greater than or equal to the diagnosis distance threshold, the preliminary abnormal diagnosis result is marked as an unknown abnormal type.

6. A method of designing a fire service pumping station control system as claimed in claim 5 wherein, The water pressure and flow reference state model is combined with the preliminary abnormal diagnosis result to deduce a working condition evolution trajectory of the fire water supply pump station system within a set early warning time window, including: An output state of the water pressure and flow reference state model at a current time is taken as an initial state; The preliminary abnormal diagnosis result is quantified as a disturbance parameter of a vector autoregressive model coefficient matrix or a residual sequence; Within the set early warning time window, the water pressure and flow reference state model with the disturbance parameter is iterated by using a Monte Carlo simulation method; In each simulation iteration, whether a boundary condition of safe system operation is reached is dynamically determined according to a model output; Paths of all simulation iterations are collected to form a working condition evolution trajectory cloud map of the fire water supply pump station system representing multiple possible future states; Key statistics including a pressure minimum value distribution, a flow maximum value distribution and a boundary touch probability are extracted from the working condition evolution trajectory cloud map.

7. A fire service water pumping station control system design method as claimed in claim 6 wherein, According to the working condition evolution trajectory, a strategy parameter framework including a response threshold and an action logic is formulated for a fire water supply pump station control system, including: A pressure safety lower limit under a specified guarantee probability is determined by analyzing the pressure minimum value distribution extracted from the working condition evolution trajectory cloud map; A flow safety upper limit under a specified guarantee probability is determined by analyzing the flow maximum value distribution extracted from the working condition evolution trajectory cloud map; The pressure safety lower limit and the flow safety upper limit are respectively set as a pressure alarm response threshold and a flow alarm response threshold; Different early warning levels including a notice early warning, an action early warning and an emergency early warning are set according to the boundary touch probability; A preset action logic for a fire pump is configured for each early warning level, and the preset action logic includes at least one of starting a backup pump, a variable frequency speed regulation instruction and closing a non-key branch valve. The pressure alarm response threshold, the flow alarm response threshold, each early warning level and its corresponding preset action logic are integrated into a strategy parameter framework of the fire water supply pump station control system.

8. The method of designing a fire service water pumping station control system of claim 7, wherein, The multi-dimensional fusion and verification of the strategy parameter framework, the mechanical characteristic curve and the electrical characteristic parameter includes: The preset action logic under different early warning levels in the strategy parameter framework is mapped into a desired speed sequence and a desired torque sequence of the fire water pump unit; On the mechanical characteristic curve, the running working point corresponding to the desired speed sequence and the desired torque sequence is found, and it is determined whether the running working point is in the high-efficiency zone of the mechanical characteristic curve; The theoretical current and temperature rise curve of the motor winding when the desired speed sequence is achieved are calculated using the electrical characteristic parameter; It is determined whether the theoretical current exceeds the rated current of the motor and whether the temperature rise curve is within the allowable range of the motor insulation material; If the running working point is in the high-efficiency zone, the theoretical current does not exceed the rated current, and the temperature rise is within the allowable range, it is determined that the verification is passed; If any condition is not met, the response threshold or action logic in the strategy parameter framework is adjusted, and the verification is performed again until all conditions are met.

9. A fire service water pumping station control system design method as claimed in claim 8, wherein, The final design method of the fire water supply pump station control system is generated according to the verification result, including: The strategy parameter framework that finally passes the verification is recorded as the core control strategy; The water pressure and flow reference state model of the fire water supply pump station system is integrated as the comparison reference of online state monitoring; The processing flow of abnormal fluctuation components, the identification and matching rules of feature event segments are integrated as the real-time fault diagnosis process; The core control strategy, the comparison reference of online state monitoring and the real-time fault diagnosis process are logically and time-series integrated to encapsulate into independent function modules; The data interface protocol and calling relationship between the core control strategy module, the online state monitoring module and the real-time fault diagnosis module are specified; Based on the function modules and their interface protocol and calling relationship, a documented fire water supply pump station control system design method is generated.

10. A fire service water supply pump station control system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein, The processor, when executing the computer program, implements the steps of the fire water supply pump station control system design method of any one of claims 1 to 9.

Images