Multi-pump coordinated variable frequency energy-saving water supply equipment and control method

Abstract

This invention discloses a multi-pump collaborative variable frequency energy-saving water supply equipment and control method, belonging to the field of intelligent control technology for water supply equipment. It aims to solve the problems of high energy consumption and poor adaptability of traditional water supply equipment. The method involves collecting water usage data, extracting features through a sliding window, using principal component analysis to screen parameters and construct a feature parameter set, and combining time attribute clustering to divide feature clusters. Typical water usage patterns are constructed by associating water usage time periods and demand intensity levels. Based on pump performance curves and attenuation indicators, an updated dynamic capability matrix is ​​built. The target water usage pattern is matched, pumps are selected to construct the execution pump group, flow demand is decomposed, load is allocated, and control commands are generated. Commands are executed, and the cumulative deviation time is calculated to determine whether to update the commands. This achieves dynamic adaptation between the pump group and water demand, improving water supply stability and energy efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for water supply equipment, and more specifically to a multi-pump coordinated variable frequency energy-saving water supply equipment and control method. Background Technology

[0002] In the modern water supply sector, multi-pump combined operation has become the mainstream configuration to meet water demand of different scales. However, current water supply systems still face a series of problems that urgently need to be solved in actual operation.

[0003] From the water user's perspective, water consumption is not constant but fluctuates due to various factors such as seasonal changes, daily routines, and unexpected events. These fluctuations include differences in water usage at fixed times each day, weekly and monthly cyclical changes, and occasional unpredictable surges in water consumption. Existing water supply solutions are often inflexible in responding to these changes, struggling to accurately predict water usage trends. This leads to a disconnect between supply and actual demand, resulting in either excessive supply causing waste or insufficient supply affecting normal water use. Regarding equipment operation, water pumps, as the core component of the water supply system, gradually change performance with usage time and environmental variations. For example, prolonged high-load operation exacerbates mechanical wear, seasonal water temperature changes can affect pump efficiency, and frequent start-ups and shutdowns can lead to performance degradation. However, current management and control methods largely rely on the pump's factory performance parameters, lacking dynamic assessment of real-time equipment status. This results in pump operating parameter settings that do not match actual performance, reducing water supply efficiency, potentially shortening equipment lifespan, and increasing operating costs. At the system control level, the key to ensuring the efficient operation of a water supply system lies in how to rationally allocate the operating status of multiple water pumps. Existing methods often employ relatively fixed patterns in pump selection and load distribution, making it difficult to flexibly adjust according to real-time water usage and equipment status. Furthermore, there is a lack of effective tracking and evaluation mechanisms for issued operating commands, failing to promptly identify and correct deviations between commands and actual needs. This makes it difficult for the water supply system to maintain a stable and efficient operating state when facing complex and changing situations. Therefore, to overcome these limitations, this invention proposes a multi-pump coordinated variable frequency energy-saving water supply device and control method. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a multi-pump coordinated variable frequency energy-saving water supply equipment and control method, which solves the problem of how to accurately adapt the water supply equipment to dynamically changing water demand, while ensuring the long-term stable and efficient operation of the water supply system even as the equipment performance degrades during operation.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] Control methods for multi-pump coordinated variable frequency energy-saving water supply equipment include:

[0007] Water usage data is collected, and water usage characteristics are extracted from the water usage data through a sliding window; principal component analysis is performed on the quantitative indicators of the water usage characteristics to select characteristic parameters and construct a characteristic parameter set; based on the time interval attributes of the characteristic parameter set, the characteristic clusters are divided, and typical water usage patterns are constructed by associating them with water usage time periods and demand intensity levels.

[0008] Based on the performance curves of each pump in the water supply equipment's pump group and the quantitative index of performance degradation, a dynamic capability matrix for each pump is constructed and updated.

[0009] Water usage characteristics are extracted from real-time water usage data to perform similarity matching on typical water usage patterns, locate target water usage patterns, and select target water pumps for water supply in the pump group by combining the dynamic capability matrix of water pumps, construct execution pump group, and decompose the flow demand of target water usage pattern into basic load and dynamic adjustment load, perform load allocation, and generate control instructions for water pump group.

[0010] Execute the control command and calculate the cumulative duration of the deviation from the target water use mode, the dynamic capability matrix, and the load allocation to determine whether to trigger the update of the water pump group control command.

[0011] Specifically, the steps for clustering feature groups include:

[0012] Water usage data is collected, and the data is aligned using a time synchronization mechanism based on the data collection timestamps to construct a basic water usage dataset.

[0013] A sliding window is set up, and water use features are extracted from the water use data within the sliding window based on the basic water use dataset. The water use features include trend features, periodic features, and random features.

[0014] The trend feature is used to reflect the changing pattern of water use data within the time range of the sliding window; the periodic feature is used to reflect the regular fluctuation pattern of water use data within a fixed period; and the random feature is used to capture irregular water use changes caused by sudden water use events.

[0015] Based on the timestamp of the sliding window, the water usage characteristics are marked with time interval attributes. The time interval attributes refer to the time range information corresponding to the sliding window, including the start time and end time.

[0016] For the extracted water use characteristics, principal component analysis is performed on the quantitative indicators they contain, parameter screening thresholds are configured, feature parameters are screened, and feature parameters are classified and integrated according to feature type to form a feature parameter set.

[0017] Based on the feature parameter set, a clustering algorithm is used to perform cluster analysis on the feature parameters of attributes in different time intervals and divide them into feature cluster groups.

[0018] Specifically, the steps for constructing a typical water use pattern include:

[0019] Configure time interval clustering parameters and use the time interval clustering algorithm to divide water usage periods, thereby clarifying the distribution range of different feature clusters in the time dimension;

[0020] Configure demand intensity related parameters and assign a corresponding demand intensity level to each feature cluster group to quantify the urgency and scale of water demand represented by different feature cluster groups;

[0021] The characteristic clusters are associated with and integrated with the divided water use periods and the defined demand intensity levels to form typical water use patterns.

[0022] Specifically, the dynamic capability matrix is ​​a two-dimensional structured data table. The horizontal dimension is the division of operating frequency ranges, and the vertical dimension is the core parameter categories that characterize the pump performance. It includes basic performance parameters and state correction parameters. The basic performance parameters are used to reflect the inherent performance characteristics of the pump at different operating frequencies and reflect the basic operating capability of the pump under standard conditions. The state correction parameters are used to reflect the performance deviation and boundary changes of the pump during actual operation and reflect the dynamic characteristics of the pump performance as operating conditions change.

[0023] Specifically, the steps for constructing and updating the dynamic capability matrix of each pump include:

[0024] Based on the pump's factory rated performance curve, the key performance parameters of the pump are discretized at preset frequency intervals. The key performance parameter values ​​at the endpoints of each frequency interval are recorded as basic performance parameter values. Based on the pump's factory performance, various state correction parameters are initialized to construct an initial dynamic capability matrix.

[0025] A full-condition testing platform was built to conduct offline full-condition testing, and basic performance parameter data under different operating frequencies were collected. This data was then used to correct the basic performance parameter values ​​of the initial dynamic capability matrix through a data calibration algorithm.

[0026] During the operation phase, real-time operating data of the water pumps are collected and data on the current typical water use patterns are extracted as correction data to construct a correction dataset;

[0027] Based on the corrected dataset, the deviation analysis algorithm is used to calculate the deviation value of the basic performance parameters, and the deviation is mapped to the corresponding dimension of the basic performance parameters proportionally for point-by-point correction, and the basic performance parameters are dynamically updated.

[0028] A multi-factor coupled performance degradation quantification model is constructed to calculate mechanical loss, efficiency degradation correction value and load fluctuation impact factor, and update state correction parameters.

[0029] Specifically, the steps for locating the target water usage pattern include:

[0030] Real-time water usage data of the current pipeline network is collected, and real-time water usage features are extracted and quantified. The real-time water usage features are then transformed into a real-time feature parameter set with the same structure as the feature parameter set in typical water usage patterns through a standardized parameter transformation algorithm.

[0031] Calculate the cosine similarity between the real-time feature parameter set and the feature parameter sets in each typical water use pattern, set a similarity threshold, compare the cosine similarity between the real-time feature parameter set and the feature parameter sets in each typical water use pattern with the similarity threshold, and mark the typical water use pattern with a cosine similarity greater than the similarity threshold as a potential target water use pattern.

[0032] If the number of potential target water use patterns is greater than 1, calculate the time matching degree between each potential target water use pattern and the current actual time, and select the potential target water use pattern as the target water use pattern based on the time matching degree.

[0033] If the number of potential target water use patterns is 1, then the potential target water use pattern is used as the target water use pattern.

[0034] If the number of potential target water use patterns is less than 1, a reference pattern is selected based on cosine similarity, and a temporary water use pattern is generated using a linear interpolation algorithm, which is then used as the target water use pattern.

[0035] Specifically, the steps for constructing candidate pump sets include:

[0036] Obtain the target water usage pattern and extract the feature parameter set, time interval attributes, and demand intensity level;

[0037] The dynamic capability matrix of each pump in the pump group is retrieved, and the basic performance parameters and state correction parameters of each pump are extracted to construct a pump performance parameter library.

[0038] Based on the water pump performance parameter library, a correspondence matrix between the target water use mode and the water pump operating status is constructed to establish a quantitative correlation framework between water demand and water pump performance. The deviation rate of the correspondence matrix is ​​calculated, and dynamic weights are set in combination with the demand intensity level. The initial water pump adaptability score is obtained by weighted summation calculation, and then corrected according to the status correction parameters of each water pump to obtain the water pump adaptability score of each water pump.

[0039] Based on the flow demand value of the target water use mode and the rated flow of the water pump, calculate the target number of water pumps required; sort all water pumps in descending order based on the water pump adaptability score, and select water pumps based on the target number to construct the initial candidate pump group.

[0040] The initial candidate pump sets are subjected to performance verification. In response to failure of performance verification, the number of pumps in the candidate pump sets is increased based on the fit score to construct the execution pump set.

[0041] Specifically, the steps for generating control commands for the water pump set include:

[0042] Obtain candidate pump sets and target water usage patterns, decompose the flow demand value of the target water usage pattern, and divide it into basic load and dynamic adjustment load.

[0043] Based on the dynamic capability matrix of each pump in the candidate pump group, the target operating frequency range of each pump is determined.

[0044] Based on the target operating frequency range and the rated flow ratio of each pump, the basic load component that each pump needs to bear is calculated and the target operating frequency required for each basic load component is determined.

[0045] If the target operating frequency of the water pump is within its target operating frequency range, it is marked as the basic load bearing pump, and the basic load is allocated to it according to the calculated target operating frequency; if the target operating frequency of the water pump exceeds its target operating frequency range, the allocation ratio of the basic load component is readjusted, and the target operating frequency of the water pump is updated.

[0046] The state correction parameters in the dynamic capacity matrix of the water pump are standardized, the dynamic adjustment adaptation index is calculated, the number of pumps bearing the dynamic adjustment load is set, the pumps bearing the dynamic adjustment load are screened and marked according to the dynamic adjustment adaptation index, and the frequency adjustment range of each pump bearing the dynamic adjustment load is determined.

[0047] Based on the fluctuation amplitude and frequency of flow rate in real-time water usage data, the dynamic adjustment load is allocated to each dynamic adjustment load pump according to the proportion of the dynamic adjustment adaptation index.

[0048] Based on the base load and the results of dynamic load adjustment, control commands for the water pump set are generated. These commands include water pump start / stop commands, base operating frequency commands, and dynamic adjustment parameter commands.

[0049] Specifically, the steps to determine whether the pump set control command update has been triggered include:

[0050] Extract water usage characteristics to update the real-time feature parameter set, measure the similarity with the current target water usage pattern feature parameter set, calculate the pattern matching degree, set a similarity judgment threshold, and count the cumulative time of pattern deviation when the pattern matching degree is less than the similarity judgment threshold. If it is greater than the preset pattern deviation time, the water pump group control command update is triggered.

[0051] The pump operating status parameters are compared with the corresponding parameters in the dynamic capability matrix, and the parameter deviation rate is calculated. If the cumulative duration of parameter deviation is greater than the preset parameter deviation threshold and the parameter deviation duration is greater than the preset parameter deviation duration, the pump group control command is updated.

[0052] For the pump that bears the basic load, monitor the cumulative duration of the load deviation between the actual operating frequency and the target operating frequency range. If it exceeds the preset load deviation duration, the pump group control command will be updated.

[0053] For dynamically adjustable load-bearing pumps, monitor the cumulative duration of the actual adjustment amount deviating from the load tolerance range. If the cumulative duration of the deviation exceeds the preset adjustment time, the pump set control command will be updated.

[0054] The multi-pump collaborative variable frequency energy-saving water supply equipment includes: a sensor sensing network, a data processing module, a control module, and an instruction update and determination module;

[0055] The sensor network is used to collect water usage data and pump operation data at key nodes of the pipeline network; the data processing module is used to construct typical water usage patterns and dynamic capability matrices; the control module is used to locate target water usage patterns, screen pumps, and generate control commands; the command update determination module is used to monitor data and comprehensively determine whether to update the control commands.

[0056] The beneficial effects of this invention are:

[0057] This application constructs typical water usage patterns to accurately capture the trends, cycles, and random characteristics of water demand. By combining time intervals and demand intensity levels, the water supply scheme can better match actual water usage changes, avoiding waste or shortages caused by a mismatch between supply and demand. The construction and updating of the dynamic capability matrix can reflect changes in water pumps due to performance degradation in real time, ensuring that water pump operating parameter settings match actual performance, improving water supply efficiency and extending equipment life. By locating target water usage patterns through similarity matching, and combining the dynamic capability matrix to screen water pumps, allocate loads, and generate control commands, flexible adjustments to the operating status of the water pump group are achieved. Furthermore, the statistical analysis of the cumulative deviation time and the command update judgment mechanism can promptly correct command deviations, ensuring that the water supply system maintains stable and efficient operation under complex conditions, thus improving the overall adaptability, energy efficiency, and reliability of the water supply system. Attached Figure Description

[0058] Figure 1 This is a flowchart of the control method for the multi-pump coordinated variable frequency energy-saving water supply equipment of the present invention;

[0059] Figure 2 This is a flowchart illustrating the clustering and segmentation of feature clusters according to the present invention.

[0060] Figure 3 This is a flowchart illustrating the construction of the dynamic capability matrix of this invention;

[0061] Figure 4 This is a flowchart illustrating the target water usage pattern of the present invention.

[0062] Figure 5 A flowchart for constructing the pump unit for this invention. Detailed Implementation

[0063] Please see Figure 1 This embodiment introduces a control method for a multi-pump coordinated variable frequency energy-saving water supply equipment, including:

[0064] Step S1: Deploy a sensor network at key nodes of the pipeline network to collect water usage data, including pressure and flow rate. After time synchronization processing to build a basic water usage dataset, set a sliding window and use time series analysis algorithms to extract water usage characteristics and mark their time interval attributes. Then, use principal component analysis to screen key feature parameters to form a feature parameter set. Based on this, cluster the data into feature clusters. Combine periodic characteristics with the time interval attributes of the feature clusters to divide water usage periods. Determine the demand intensity level based on the feature parameters and other indicators. Finally, integrate the feature clusters with water usage periods and demand intensity levels to form a typical water usage pattern that can provide a standardized reference for pump group control. Specifically, pressure transmitters, electromagnetic flow meters, and smart water meters are installed at the risers, branch nodes, and user terminals at the end of the pipeline network to form a sensor network. This network collects water usage data in real time, including pressure, flow rate, water usage duration, and start and end times. The water usage data is then correlated and integrated according to the timestamp of the collection, forming a basic water usage dataset with time-series tags. Time series analysis algorithms are used to perform sliding window calculations on the flow data in the basic water usage dataset over several consecutive days to obtain trend characteristics reflecting long-term changes. Simultaneously, spectral analysis is conducted on daily and weekly flow fluctuations to extract periodic characteristics that reflect periodic patterns. Statistical identification is performed on flow anomalies that deviate from the periodic trend to identify sudden changes. The random characteristics of water use are identified. Subsequently, a clustering algorithm is used to analyze the three types of water use characteristics extracted. Key quantitative indicators from trend characteristics, periodic characteristics, and random characteristics are selected to form a feature parameter set. Combined with the statistical regularity of water use start and end times, time interval clustering is used to determine the time boundaries of peak, off-peak, and trough periods to construct a time distribution model. Based on the flow peak and pressure fluctuation range in the feature parameter set and combined with the pipeline network design standards, the demand intensity level is defined. Finally, a typical water use pattern containing these three parts is generated. In this way, the scattered water use data is transformed into a structured demand template, providing a standardized reference for subsequent water pump sets to match and control strategies based on real-time water use status, and realizing dynamic adaptation of water pump operation and water demand.

[0065] In this embodiment, by constructing a sensing network covering the entire pipeline network and a systematic data analysis process, the precise transformation of water demand from raw data to a structured pattern is achieved. On the one hand, the multi-dimensional water demand data collected in real time is correlated through a time synchronization mechanism to form a complete and time-series-clear basic dataset, providing high-quality data support for subsequent analysis. On the other hand, by setting a sliding window and using time series analysis algorithms to extract three types of water demand characteristics, the changing patterns, periodic patterns, and sudden characteristics of water demand are comprehensively captured. Typical water demand patterns generated through clustering and other steps are presented in a structured form, including feature parameter sets, time distribution models, and demand intensity levels, intuitively showing the water demand patterns and demand levels under different scenarios. This transformation turns the originally fragmented water demand data into a standardized template with clear guiding significance. It not only provides a directly referable demand benchmark for subsequent pump group control but also ensures that the pump operating status can accurately respond to the dynamic changes in water demand through the deep integration of time interval attributes, laying a key foundation for the stable operation and energy-saving optimization of the water supply system.

[0066] Please see Figure 2 Preferably, the specific steps for clustering feature-based clustering groups include:

[0067] Water usage data is collected in real time through a sensor network. This data includes pressure, flow, duration, and start and end times. Based on the timestamps of the collected water usage data, a time synchronization mechanism is used to align the data and build a basic water usage dataset. This provides continuous and reliable raw data support for subsequent feature extraction.

[0068] A sliding window is set up, and based on the basic water usage dataset, time series analysis algorithms are used to extract water usage features from the water usage data within the sliding window. The water usage features include trend features, periodic features, and random features. Trend features are used to reflect the changing patterns of water usage data within the time range of the sliding window; periodic features are used to reflect the regular fluctuations of water usage data within a fixed period; and random features are used to capture irregular water usage changes caused by sudden water usage events.

[0069] Based on the timestamp of the sliding window, the extracted water use features are marked with time interval attributes. The time interval attribute refers to the time range information corresponding to the sliding window, including the start time and end time. It is used to establish a clear association between the extracted water use features and the actual time dimension, so that the feature data can be mapped to specific time periods, and provide a time series reference for subsequent steps such as feature clustering by time dimension and construction of time distribution models.

[0070] For the extracted water use characteristics, principal component analysis is performed on the included quantitative indicators. Parameter screening thresholds are configured to eliminate redundant parameters and select feature parameters that significantly reflect the original attributes of water demand. The feature parameters are then categorized and integrated according to feature type to form a feature parameter set. The values ​​of the feature parameter set are directly derived from the quantitative results of feature extraction from the basic water use dataset, objectively reflecting the essential attributes of water demand, including flow demand and pressure demand values.

[0071] Based on the generated feature parameter set, a clustering algorithm is used to perform cluster analysis on the feature parameters of different time interval attributes. The number of clusters and similarity thresholds are configured. Feature cluster groups are divided according to the similarity of feature parameters, the correlation of time interval attributes, and the consistency of water use characteristics. Water use characteristics with similar feature parameters, similar time distribution patterns, and matching demand intensity levels are grouped into one category, forming several feature cluster groups. Each cluster group represents a feature set with similar water use patterns, providing a clustering basis for the generation of typical water use patterns.

[0072] Based on the periodic time segmentation in the periodic features, and combined with the time interval attributes corresponding to the feature clustering groups, time interval clustering parameters are configured, and the time interval clustering algorithm is used to divide the water use period. This is used to clarify the distribution range of different feature clustering groups in the time dimension, so that each feature clustering group can correspond to a specific water use period, and to provide a basis for the division of the time dimension for subsequent integration to form a typical water use pattern.

[0073] Based on the core indicators such as flow demand and pressure demand in the feature parameter set, and combined with the feature parameter performance of the feature clusters, demand intensity-related parameters are configured, and a corresponding demand intensity level is defined for each feature cluster. This is used to quantify the urgency and scale of water demand represented by different feature clusters, so that the feature clusters have more explicit demand attributes.

[0074] By associating and integrating feature clusters with the defined water usage periods and demand intensity levels, each feature cluster corresponds to a specific water usage period and demand intensity level, forming a complete typical water usage pattern. This integration organically combines feature parameter sets, time interval attributes, and demand intensity levels, fully reflecting water usage patterns under different scenarios and providing a standardized reference template for subsequent water pump sets to match control strategies based on real-time water usage.

[0075] Step S2: Obtain the pump sets configured for the water supply equipment. Based on the factory performance curve of each pump, discretize key parameters to construct a dynamic capability matrix framework containing basic performance parameters and state correction parameters. Calibrate the dynamic capability matrix by collecting data through offline full-condition testing. After the equipment is operational, integrate real-time data with typical water usage pattern information to form a correction dataset, update the basic performance parameters, and determine the state correction parameters. Finally, construct a dynamic capability matrix that reflects the actual performance of the pumps in real time. Specifically, the dynamic capability matrix is ​​a two-dimensional structured data table. The horizontal dimension represents the operating frequency range, and the vertical dimension represents the core parameter categories characterizing pump performance. It includes two levels: basic performance parameters and state correction parameters. The basic performance parameters reflect the inherent performance characteristics of the pump at different operating frequencies, including core indicators such as rated flow, rated head, and operating efficiency, reflecting the basic operating capability of the pump under standard conditions. The state correction parameters reflect the performance deviation and boundary changes of the pump during actual operation due to factors such as mechanical loss, operating time, and load fluctuations, including indicators such as response speed decay, load tolerance range adjustment, and health status quantification, reflecting the dynamic characteristics of pump performance changing with operating conditions. Together, these two elements constitute a structured parameter set that reflects the performance and variation patterns of the water pump under different operating conditions in real time, providing a precise quantitative basis for the coordinated control of the water pump set. During construction, an initial matrix framework is first formed by discretizing the baseline values ​​based on the water pump's factory rated performance curve. Then, the baseline values ​​are calibrated using data collected from offline full-condition tests to complete the initial construction. During equipment operation, the matrix parameters are dynamically corrected using a performance degradation algorithm, combining real-time operating data collected by sensors and load characteristics of typical water usage patterns. This ensures that both elements together constitute a structured parameter set that reflects the water pump's performance and variation patterns in real time, providing a precise quantitative basis for the coordinated control of the water pump set.

[0076] In this embodiment, a comprehensive characterization of pump performance, from static baseline to dynamic changes, is achieved through a technical approach combining full-condition testing and real-time correction. An initial matrix constructed based on factory parameters and offline testing provides a standardized quantitative benchmark. Combined with real-time monitoring data and dynamic correction using performance degradation algorithms, it accurately captures performance deviations during pump operation, ensuring that the parameter set matches the actual state in real time. This structured dynamic capability matrix clearly presents the core performance and state boundaries under different operating conditions, providing a directly callable quantitative basis for pump group load allocation. Furthermore, through the dynamic updating of state correction parameters, the collaborative control strategy adapts to the pump performance degradation process, laying a core technical foundation for the long-term stable and energy-saving operation of the water supply system.

[0077] Please see Figure 3 Preferably, the specific steps for constructing the dynamic capability matrix include:

[0078] Based on the pump's factory rated performance curve, key performance parameters such as rated flow, rated head, and operating efficiency are discretized at preset frequency intervals. The performance parameter values ​​corresponding to the endpoints of each frequency interval are recorded as basic performance parameter values. At the same time, based on the pump's factory performance parameters such as mechanical performance parameters, design service life, and load tolerance threshold under standard operating conditions, and according to the pump's optimal operating state without performance degradation and rated load characteristics, initial reference values ​​for various state correction parameters are determined. Based on these, the state correction parameters are initialized to construct an initial dynamic capability matrix.

[0079] A test platform covering various operating conditions such as flow rate, head, and pressure was built to conduct offline full-condition testing, simulating the working state of water pumps under various possible operating scenarios. For each water pump, basic performance parameter data were collected at different operating frequencies. The measured data were compared and corrected with the discretized benchmark values ​​of basic performance parameters in the initial framework using data calibration algorithms such as the least squares method. This overcame the defect in existing technologies where static parameters were difficult to match actual operating conditions, and completed the initial construction of the dynamic capability matrix.

[0080] Once the equipment enters the actual operation phase, sensors installed at the pump inlet and outlet, motor, and other parts are used to collect real-time pump operation data, including actual flow rate, pressure, operating frequency, motor current, and operating time. At the same time, correction data such as the characteristic parameter set, time interval attributes, and demand intensity level corresponding to the current time period are extracted from the current typical water use pattern. The correction data is then correlated and integrated to form a correction dataset for correction calculation.

[0081] Based on the corrected dataset, the deviation analysis algorithm is used to calculate the deviation value between the actual operating data and the corresponding basic performance parameters in the initial dynamic capability matrix. The deviation value is mapped to the corresponding dimension of the basic performance parameter proportionally. The discretized benchmark values ​​of parameters such as flow rate, head, and efficiency are corrected point by point. At the same time, the sliding window algorithm is combined to smooth the deviation trend of continuous multi-time periods and eliminate the deviation interference caused by instantaneous abnormal fluctuations. This allows for dynamic updates of the basic performance parameters in the dynamic capability matrix, overcoming the defect of fixed parameters in the existing technology.

[0082] Simultaneously, a multi-factor coupled performance degradation quantification model is constructed. Using operating frequency, flow deviation, and operating time as input parameters, the mechanical loss per unit time is calculated through a preset loss coefficient formula, quantifying the performance degradation caused by mechanical wear. Based on the attenuation coefficient table corresponding to the operating time interval, the attenuation coefficient matching the current operating time is determined. This coefficient is multiplied by the efficiency parameter in the basic performance parameters to obtain the efficiency degradation correction value, reflecting the cumulative impact of operating time on efficiency. Combining the deviation of actual flow from the flow benchmark value in the characteristic parameter set and the demand intensity level, a weighted calculation is performed according to the weight corresponding to the demand intensity level to obtain the load fluctuation impact factor, which is used to dynamically adjust the load tolerance range parameter.

[0083] Based on the calculated mechanical losses, efficiency decay correction values, and load fluctuation impact factors, the various state correction parameters in the dynamic capability matrix are updated accordingly, enabling the dynamic updates of state correction parameters and basic performance parameters to work in synergy. For example, when the flow rate is detected to be consistently outside the normal range and the operating frequency is high, the specific values ​​of the state correction parameters are derived using a performance decay quantification model, combined with the demand intensity level for the corresponding time period, and then entered into the matrix. This achieves real-time and accurate mapping of the dynamic capability matrix to the actual operating state of the pump, solving the problem that existing technologies cannot dynamically adapt to changes in equipment performance and improving the accuracy of the matrix's support for pump operation control.

[0084] Step S3: Extract water usage features from real-time water usage data to perform similarity matching on typical water usage patterns, locate the target water usage pattern, and select target water pumps for water supply from the pump group based on the dynamic capability matrix of the water pumps. Construct the execution pump group, and decompose the flow demand of the target water usage pattern into basic load and dynamic adjustment load for load allocation, generating control instructions for the pump group. Specifically, similarity matching refers to determining the typical water usage pattern corresponding to the current water usage state by calculating the similarity between the real-time water usage features and the feature parameter sets in the typical water usage patterns. The cosine similarity algorithm is used to quantify the similarity between the current water usage state and each typical water usage pattern. The target water usage pattern is the typical water usage pattern with the highest similarity to the real-time water usage features, serving as the demand benchmark for current water supply control. The execution pump group is a set of several water pumps selected from the pump group that have the ability to meet the current water demand and have the optimal energy consumption. The basic load refers to the relatively stable flow demand portion of the target water usage pattern, which is borne by the fixed water pumps in the execution pump group. The dynamic adjustment load refers to the flow demand portion that fluctuates with real-time water usage, which is dynamically borne by the adjustable water pumps in the execution pump group according to the actual situation.

[0085] In this embodiment, by dynamically matching real-time water usage characteristics with typical water usage patterns, and combining this with pump performance status to select and optimize pump allocation, precise matching between water demand and pump operation is achieved. On one hand, the similarity matching mechanism can quickly locate the target water usage pattern, providing a clear demand benchmark for regulation. When no match is found, an interpolation algorithm can be used to generate a suitable temporary water usage pattern, ensuring the continuity of regulation. On the other hand, the pump selection and load decomposition strategy based on the dynamic capability matrix can flexibly determine the number of target pumps according to the demand scale. Through the reasonable allocation of basic load and dynamically adjusted load, the pumps operate in the high-efficiency range, minimizing energy consumption and providing support for the energy-saving and stable operation of the water supply system.

[0086] Please see Figure 4 The preferred steps for determining the target water usage pattern include:

[0087] The sensor network collects real-time water usage data of the current pipeline network, including real-time pressure, real-time flow, real-time water usage duration, and real-time water usage start and end times. This data will serve as the original basis for extracting real-time water usage characteristics.

[0088] Based on the collected real-time water usage data, a sliding window with the same size as the window used to construct the typical water usage pattern in step S1 is set. The real-time water usage data within the sliding window is processed using a time series analysis algorithm to extract real-time water usage features, including real-time trend features, real-time periodic features, and real-time random features.

[0089] The extracted real-time water use features are quantified, and a standardized parameter conversion algorithm is used to transform the real-time water use features into a real-time feature parameter set with the same structure as the feature parameter set in the typical water use pattern. The standardized parameter conversion algorithm is an algorithm that standardizes and adjusts the numerical range and dimensional division of the real-time water use features based on the parameter type, dimension definition and quantization scale of the feature parameter set in the typical water use pattern, so as to ensure that the two are completely matched in parameter structure and quantization standard and have a basis for direct comparison.

[0090] Calculate the cosine similarity between the real-time feature parameter set and the feature parameter sets in each typical water use pattern. Cosine similarity refers to the degree of similarity between two vectors by calculating the cosine of the angle between them. The value ranges from [-1, 1], and the closer the value is to 1, the higher the similarity.

[0091] A similarity threshold is set. The similarity threshold is the critical value used to determine whether real-time water usage characteristics match typical water usage patterns. It is set based on the stability requirements of the water supply system and historical matching data.

[0092] The cosine similarity of the real-time feature parameter set with the feature parameter sets in each typical water use pattern is compared with a similarity threshold. Typical water use patterns with a cosine similarity greater than the similarity threshold are marked as potential target water use patterns.

[0093] If the number of potential target water use patterns is greater than 1, the time matching degree between each potential target water use pattern and the current actual time is calculated. The time matching degree refers to the degree of overlap between the time interval attribute of the potential target water use pattern and the current actual time. The value range is between [0,1]. The closer the value is to 1, the higher the time matching degree. The potential target water use pattern with the highest time matching degree is selected as the final target water use pattern.

[0094] If the number of potential target water use patterns is equal to 1, then the potential target water use pattern is directly determined as the final target water use pattern;

[0095] If the number of potential target water use patterns is less than 1, the two typical water use patterns with the highest cosine similarity are selected as reference patterns. A temporary water use pattern is generated using a linear interpolation algorithm. The linear interpolation algorithm is a method of calculating the value of the intermediate point by connecting two known data points with a straight line. Here, it is used to calculate the corresponding parameters of the temporary water use pattern based on the feature parameter set, time interval attribute and demand intensity level of the two reference patterns, and the temporary water use pattern is determined as the final target water use pattern.

[0096] Please see Figure 5 Preferably, the specific steps for constructing the execution pump set include:

[0097] The target water usage pattern obtained from the location is obtained, and the feature parameter set, time interval attributes and demand intensity level are extracted. The feature parameter set includes flow demand value and pressure demand value.

[0098] The dynamic capability matrix of each pump in the pump set is retrieved, and the basic performance parameters and condition correction parameters of each pump are extracted from it. The basic performance parameters include the rated flow rate, rated head, and operating efficiency at different operating frequencies. The condition correction parameters include the response speed attenuation value, the load tolerance range adjustment amount, and the health status quantification value, forming a pump performance parameter library.

[0099] Based on a water pump performance parameter library, a correspondence matrix between target water usage patterns and water pump operating states is constructed. This matrix is ​​a two-dimensional data matrix that arranges key parameters characterizing the scale and characteristics of water demand in the target water usage pattern with key parameters characterizing the output capacity and performance level of the water pump within its high-efficiency operating range, forming a one-to-one correspondence. Each element in the matrix represents a matching node between a set of demand parameters and performance parameters, used to establish a quantitative correlation framework between water demand and water pump performance. The degree of matching between the two can be calculated and analyzed through the numerical relationships of the matrix elements. The deviation rate of the correspondence matrix is ​​calculated, and dynamic weights are set in conjunction with the demand intensity level. An initial water pump suitability score is obtained through weighted summation, and then corrected according to the state correction parameters of each water pump to obtain the water pump suitability score for each pump. The specific steps are as follows:

[0100] Using the flow and pressure demand values ​​of the target water use mode as the column vectors of the matrix, and the rated flow and rated head of a single water pump at the high-efficiency operating frequency as the row vectors of the matrix, a demand-performance correspondence matrix is ​​formed.

[0101] For each corresponding element in the correspondence matrix, the flow deviation rate and pressure deviation rate are obtained by calculating the ratio of the absolute difference between the pump parameters and the demand parameters to the demand parameters. The smaller the deviation rate, the higher the degree of matching between the pump parameters and the demand parameters.

[0102] The weights of flow deviation rate and pressure deviation rate are set according to the demand intensity level of the target water use pattern. The demand intensity level is divided into three levels: high, medium and low. The weights of flow deviation rate and pressure deviation rate are preset according to the demand intensity level. When the demand intensity is high, the weight of flow deviation rate is higher than that of pressure deviation rate, so as to prioritize the flow matching accuracy. When the demand intensity is low, the weight of pressure deviation rate is higher than that of flow deviation rate, so as to focus on pressure stability. When the demand intensity is medium, the weights of the two are balanced.

[0103] Add the product of the flow deviation rate and its corresponding weight, and the product of the pressure deviation rate and its corresponding weight, and then subtract the sum from 1 to obtain the basic matching score. The basic matching score ranges from [0,1], and the closer the value is to 1, the better the basic performance matching.

[0104] The health status quantification value in the status correction parameter is used as the health correction coefficient, and 1 minus the response speed decay value is used as the response correction coefficient. The two are multiplied to obtain the comprehensive correction coefficient, which reflects the impact of the pump's current health status and response capability on adaptability.

[0105] The basic matching score is multiplied by the comprehensive correction coefficient to obtain the final pump adaptability score. This score is used to comprehensively integrate the matching degree between pump performance and water demand, as well as the pump's own condition, providing an accurate and dynamic quantitative basis for pump screening and ranking. This ensures that the selected pumps can meet current water demand and adapt to their actual operating conditions.

[0106] Based on the flow demand of the target water use pattern and the rated flow of the water pumps, calculate the target number of water pumps required to ensure that the total flow output capacity of the candidate pump sets is not lower than the current demand. The number is determined by rounding up during the calculation.

[0107] All pumps are sorted in descending order based on their compatibility scores, with pumps scoring higher being prioritized for inclusion in the candidate pool. During the sorting process, for pumps with the same score, a runtime balancing factor is introduced, prioritizing pumps with shorter cumulative runtimes to balance the usage frequency of each pump.

[0108] Pumps are selected sequentially from the sorted pumps to form an initial candidate pump group. The number of pumps in the initial candidate pump group is the calculated target number.

[0109] The initial candidate pump sets are subjected to performance verification. The verification includes: whether the total flow output range of the candidate pump sets covers the flow demand value of the target water use mode, whether the total head output range covers the pressure demand value of the target water use mode, and whether the load tolerance range adjustment of each pump can adapt to the load fluctuation range corresponding to the demand intensity level.

[0110] If the initial candidate pump group passes the performance verification, it will be determined as the final execution pump group; if it fails the verification, one pump will be added, and the pump with the second highest compatibility score will be selected to reassemble the candidate pump group, and the performance verification will be carried out again. This process will be repeated until the execution pump group passes the verification.

[0111] Preferably, the specific steps for generating control commands for the water pump unit include:

[0112] Candidate pump sets and target water usage patterns are obtained. The flow demand value of the target water usage pattern is decomposed into basic load and dynamic adjustment load. The basic load is the relatively stable part of the flow demand value, while the dynamic adjustment load is the part that changes with real-time water usage fluctuations. The proportion of the two is determined according to the demand intensity level. The higher the demand intensity level, the larger the proportion of the basic load, in order to ensure water supply stability.

[0113] Based on the dynamic capability matrix of each pump in the candidate pump group, the basic performance parameters of each pump at different operating frequencies are extracted to determine the target operating frequency range of each pump. The target operating frequency range refers to the frequency range corresponding to the pump operating efficiency at a high level, providing an energy efficiency reference for load allocation.

[0114] Based on the target operating frequency range and the rated flow ratio of each pump, the basic load component to be borne by each pump is calculated. Then, based on the flow-frequency correspondence of the pumps, the target operating frequency required for each basic load component is determined. Finally, the target operating frequency is compared with the target operating frequency range of each pump.

[0115] If the target operating frequency of the water pump is within its target operating frequency range, it is marked as the basic load bearing pump, and the basic load is allocated to it according to the calculated target operating frequency; if the target operating frequency of the water pump exceeds its target operating frequency range, the allocation ratio of the basic load component is readjusted, the basic load component borne by the water pump exceeding the target operating frequency range is reduced, and the reduced component is allocated to other water pumps with target operating frequencies within the range according to the rated flow ratio.

[0116] The state correction parameters in the dynamic capacity matrix of the water pump are standardized, the dynamic adjustment adaptation index is calculated according to the preset weight, the number of pumps bearing the dynamic adjustment load is set according to the ratio of the total dynamic adjustment load to the maximum dynamic adjustment capacity of a single pump and the rounding rule, and the pumps bearing the dynamic adjustment load are screened and marked according to the dynamic adjustment adaptation index.

[0117] By combining the fluctuation range of the dynamically adjusted load and the load tolerance range adjustment amount of the pump bearing the dynamically adjusted load, the frequency adjustment range of each pump bearing the dynamically adjusted load is determined. The frequency adjustment range must be within the safe operating frequency range of the pump and connect with the target operating frequency range to form a continuous frequency adjustment range.

[0118] Based on the fluctuation amplitude and frequency of flow in real-time water usage data, the dynamic adjustment load is allocated to each dynamic adjustment load-bearing pump according to the proportion of the dynamic adjustment adaptation index. That is, according to the proportion of the dynamic adjustment adaptation index of each dynamic adjustment load-bearing pump in the total dynamic adjustment adaptation index of all dynamic adjustment load-bearing pumps, the corresponding proportion of dynamic adjustment load is allocated. During the allocation process, it is necessary to ensure that the adjustment amount of each pump does not exceed the adjustment amount of its load tolerance range, and complements the output of the basic load-bearing pump to jointly meet the flow requirements of the target water usage mode.

[0119] Based on the allocation results of the base load and the dynamically adjusted load, control instructions for the water pump group are generated. These instructions include pump start / stop instructions, basic operating frequency instructions, and dynamic adjustment parameter instructions. The pump start / stop instructions specify which pumps in the candidate pump group need to be started to participate in water supply, including the base load pumps and the dynamically adjusted load pumps. This controls the activation and idling of pumps, avoiding ineffective energy consumption, and ensuring that the number of pumps participating in water supply meets the current load demand. The basic operating frequency instructions set a fixed operating frequency value for each base load pump, ensuring that the base load pumps operate efficiently and stably, continuously outputting the flow rate corresponding to the base load. The dynamic adjustment parameter instructions set the upper and lower limits of frequency adjustment, the step size for each adjustment, and the response delay time for each dynamically adjusted load pump. This guides the dynamically adjusted load pumps to flexibly adjust their operating frequency according to real-time water usage fluctuations, meeting the dynamic load demand while avoiding increased equipment wear due to frequent adjustments.

[0120] Step S4: During the coordinated control process based on the pump group control instructions, continuously monitor the water usage data of the pipeline network and the operating status of the pumps. Extract real-time water usage characteristics and calculate the similarity with the feature parameter set of the current target water usage mode. Simultaneously, evaluate the matching degree between the actual operating status parameters of each pump in the pump group and the corresponding parameters in the dynamic capability matrix to comprehensively determine whether the control instructions of the pump group need to be updated. Specifically, collect water usage data such as pressure and flow rate of the pipeline network in real time, as well as operating parameters such as pump operating frequency, output flow rate, and head. Extract real-time trend, periodic, and random features based on the water usage data and convert them into a real-time feature parameter set. Calculate the cosine similarity with the feature parameter set of the current target water usage mode to measure the degree of matching of water demand. Compare the actual operating parameters of the pumps with the basic performance parameters and state correction parameters of the corresponding frequency range in the dynamic capability matrix, calculate the deviation rate, and evaluate the equipment's operational adaptability. If the similarity is lower than the set threshold, the deviation rate exceeds the allowable range, or the operating frequency of the basic load-bearing pump deviates from the target range, or the adjustment amount of the dynamically adjusted load-bearing pump exceeds the limit, it is determined that the control instructions need to be updated; otherwise, the current instructions are maintained.

[0121] In this embodiment, dynamic adaptation of pump group control is achieved. On the one hand, continuous monitoring of water usage data and pump operating status ensures timely capture of fluctuations in water demand and changes in equipment performance, providing comprehensive and real-time data support for subsequent judgment. On the other hand, cosine similarity calculation quantifies the matching degree of water demand, and deviation rate is used to assess the adaptability of equipment operation and dynamic capability matrix. Simultaneously, multiple dimensions, such as the frequency range compliance of the base load-bearing pump and the rationality of the adjustment amount of the dynamically adjusted load-bearing pump, form a scientific judgment basis. This multi-dimensional comprehensive judgment method avoids water supply instability caused by untimely response to sudden changes in water demand, prevents energy efficiency degradation or increased losses caused by equipment performance deviations, and reduces unnecessary command adjustments. While ensuring efficient and stable operation of the water supply system, it extends the service life of equipment and provides crucial closed-loop feedback support for the intelligent and refined control of the entire water supply system.

[0122] Preferably, the specific steps for determining whether to update the pump set control command include:

[0123] When the water pump group control command is implemented, the real-time feature parameter set is updated based on the extracted water use characteristics, and the similarity is measured with the current target water use mode feature parameter set to calculate the mode matching degree. A similarity judgment threshold is set according to the demand intensity level of the target water use mode and the historical matching accuracy. During the implementation of the current water pump group control command, the cumulative duration of mode deviation when the mode matching degree is lower than the similarity judgment threshold is counted. If the cumulative duration exceeds the preset mode deviation duration, the update of the water pump group control command is triggered.

[0124] The pump operation status parameters are monitored synchronously and compared with the corresponding parameters in the dynamic capability matrix. The parameter deviation rate is calculated. If the cumulative duration of parameter deviation is greater than the preset parameter deviation threshold and the parameter deviation is greater than the preset parameter deviation duration, the pump group control command is updated.

[0125] For the pump that bears the basic load, monitor the cumulative duration of the load deviation between the actual operating frequency and the target operating frequency range. If it exceeds the preset load deviation duration, the pump group control command will be updated.

[0126] For dynamically adjustable load-bearing pumps, monitor the cumulative duration of the actual adjustment amount deviating from the load tolerance range. If the cumulative duration of the deviation exceeds the preset adjustment time, the pump set control command will be updated.

[0127] This embodiment also introduces a multi-pump collaborative variable frequency energy-saving water supply equipment, including: a sensor sensing network, a data processing module, a control module and an instruction update determination module;

[0128] The sensor network is used to collect water usage and pump operation data at key nodes of the pipeline network; the data processing module is used to construct typical water usage patterns and dynamic capability matrices; the control module is used to locate target water usage patterns, screen pumps, and generate control commands; the command update determination module is used to monitor data and comprehensively determine whether to update the control commands.

[0129] The sensor network is deployed at key nodes in the pipeline network. Specifically, pressure transmitters, electromagnetic flow meters, and smart water meters are installed at risers, branch nodes, and user terminals at the end of the network. These devices work together to collect water usage data and pump operation data from the pipeline network. Water usage data includes pressure data, flow data, water usage duration data, and water usage start and end times; operation data includes the actual operating frequency, actual output flow, actual head, motor current, and cumulative operating time of the pumps.

[0130] The data processing module connects to the sensor network, receiving and processing water usage and operational data from the sensors. It constructs a basic water usage dataset and a real-time monitoring dataset, extracting water usage characteristics and forming a set of feature parameters. Based on this set, it identifies feature clusters, water usage periods, and demand intensity levels, then integrates these data to form typical water usage patterns. Simultaneously, this module is responsible for constructing and dynamically updating the dynamic capability matrix of the water pumps. This matrix is ​​a two-dimensional structured data table; the horizontal dimension represents the operating frequency range, and the vertical dimension represents the core parameter categories characterizing pump performance, including basic performance parameters and condition correction parameters.

[0131] The control module is connected to the data processing module. Its main function is to locate the target water usage pattern based on the similarity matching between real-time water usage characteristics and typical water usage patterns. Then, combined with the dynamic capacity matrix, it selects target water pumps from the pump group to construct the execution pump group. Next, it decomposes the flow demand of the target water usage pattern into basic load and dynamic adjustment load and allocates them, finally generating control instructions for the pump group. These instructions include pump start / stop instructions, basic operating frequency instructions, and dynamic adjustment parameter instructions. When generating control instructions, the control module determines the proportion of basic load and dynamic adjustment load according to the demand intensity level; the higher the demand intensity level, the larger the proportion of basic load.

[0132] The instruction update determination module is connected to the sensor network, data processing module, and control module to continuously monitor water usage and operational data. By calculating pattern matching similarity and parameter deviation rate, it monitors the operating status of the base load pump and the dynamically adjustable load pump, comprehensively determining whether an update to the control instruction is needed. The determination criteria include the cumulative duration of pattern matching similarity below a set threshold, the number of consecutive times the parameter deviation rate exceeds the set threshold, the cumulative duration of the base load pump's operating frequency deviating from the target range, the cumulative duration of the dynamically adjustable load pump's adjustment exceeding its tolerance range, and situations where a typical water usage pattern update causes the current target water usage pattern to fail.

[0133] Working principle and its effects:

[0134] This invention achieves energy-saving water supply through multi-pump coordinated variable frequency, the core of which lies in accurately matching dynamic water demand with the real-time performance of the pumps, ensuring water supply stability while improving system energy efficiency.

[0135] First, water usage data and pump operation data are collected at key nodes of the pipeline network. Trends, periods, and randomness in water usage characteristics are extracted, and typical water usage patterns are constructed by combining time interval attributes and demand intensity levels. Simultaneously, a dynamic capability matrix reflecting the actual performance of the pumps is built based on their factory performance and real-time degradation data. This process transforms fragmented water usage data into a structured reference model, providing accurate basis for regulation, and dynamically tracks changes in pump performance, ensuring that relevant parameters conform to actual conditions. This solves the problems of inaccurate understanding of water usage patterns and static evaluation of pump performance in traditional water supply systems. Second, the current target water usage pattern is determined through similarity matching. Suitable pumps are selected to form the execution pump group based on the dynamic capability matrix. Flow demand is decomposed into basic load and dynamically adjustable load and rationally allocated, generating corresponding control commands. Simultaneously, the cumulative duration of various deviations is continuously monitored, and control commands are updated promptly. This dynamic control method can flexibly respond to water usage fluctuations, allowing pumps to operate within their efficient range, avoiding energy waste and unstable water supply under fixed control modes. Furthermore, multi-dimensional monitoring of deviations ensures long-term stable system operation.

[0136] In summary, this invention achieves real-time adaptation between water demand and pump operation through systematic water use pattern construction, dynamic pump performance evaluation, and precise load allocation and control. While improving the stability and reliability of water supply, it significantly reduces energy consumption and extends equipment lifespan, providing an effective solution for the intelligent and energy-saving development of multi-pump collaborative water supply systems.

[0137] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A control method for a multi-pump coordinated variable frequency energy-saving water supply equipment, characterized in that, include: Collect water usage data and extract water usage characteristics from the data using a sliding window; Principal component analysis is performed on the quantitative indicators of the water use characteristics to screen characteristic parameters and construct a characteristic parameter set; based on the time interval attributes of the characteristic parameter set, the characteristic clusters are divided into clusters, and typical water use patterns are constructed by associating them with water use time periods and demand intensity levels. The specific steps for constructing a typical water use pattern include: Configure time interval clustering parameters and use the time interval clustering algorithm to divide water usage periods, thereby clarifying the distribution range of different feature clusters in the time dimension; Configure demand intensity related parameters and assign a corresponding demand intensity level to each feature cluster group to quantify the urgency and scale of water demand represented by different feature cluster groups; The characteristic clusters are associated and integrated with the divided water use periods and the defined demand intensity levels to form typical water use patterns; Based on the performance curves of each pump in the water supply equipment's pump group and the quantitative index of performance degradation, a dynamic capability matrix for each pump is constructed and updated. Water usage characteristics are extracted from real-time water usage data to perform similarity matching on typical water usage patterns, locate target water usage patterns, and select target water pumps for water supply in the pump group by combining the dynamic capability matrix of water pumps, construct execution pump group, and decompose the flow demand of target water usage pattern into basic load and dynamic adjustment load, perform load allocation, and generate control instructions for water pump group. Execute the control command and calculate the cumulative duration of the deviation from the target water use mode, the dynamic capability matrix, and the load allocation to determine whether to trigger the update of the water pump group control command.

2. The control method for the multi-pump coordinated variable frequency energy-saving water supply equipment as described in claim 1, characterized in that, The specific steps for clustering and dividing the feature clusters include: Water usage data is collected, and the data is aligned using a time synchronization mechanism based on the data collection timestamps to construct a basic water usage dataset. A sliding window is set up, and water use features are extracted from the water use data within the sliding window based on the basic water use dataset. The water use features include trend features, periodic features, and random features. The trend feature is used to reflect the changing pattern of water use data within the time range of the sliding window; the periodic feature is used to reflect the regular fluctuation pattern of water use data within a fixed period; and the random feature is used to capture irregular water use changes caused by sudden water use events. Based on the timestamp of the sliding window, the water usage characteristics are marked with time interval attributes. The time interval attributes refer to the time range information corresponding to the sliding window, including the start time and end time. For the extracted water use characteristics, principal component analysis is performed on the quantitative indicators they contain, parameter screening thresholds are configured, feature parameters are screened, and feature parameters are classified and integrated according to feature type to form a feature parameter set. Based on the feature parameter set, a clustering algorithm is used to perform cluster analysis on the feature parameters of attributes in different time intervals and divide them into feature cluster groups.

3. The control method for the multi-pump coordinated variable frequency energy-saving water supply equipment as described in claim 1, characterized in that, The dynamic capability matrix is ​​a two-dimensional structured data table. The horizontal dimension is the division of operating frequency ranges, and the vertical dimension is the core parameter categories that characterize the pump performance. It includes basic performance parameters and state correction parameters. The basic performance parameters are used to reflect the inherent performance characteristics of the pump at different operating frequencies and reflect the basic operating capability of the pump under standard conditions. The state correction parameters are used to reflect the performance deviation and boundary changes of the pump during actual operation and reflect the dynamic characteristics of the pump performance as operating conditions change.

4. The control method for the multi-pump coordinated variable frequency energy-saving water supply equipment as described in claim 3, characterized in that, The specific steps for constructing and updating the dynamic capability matrix of each pump include: Based on the pump's factory rated performance curve, the key performance parameters of the pump are discretized at preset frequency intervals. The key performance parameter values ​​at the endpoints of each frequency interval are recorded as basic performance parameter values. Based on the pump's factory performance, various state correction parameters are initialized to construct an initial dynamic capability matrix. A full-condition testing platform was built to conduct offline full-condition testing, and basic performance parameter data under different operating frequencies were collected. This data was then used to correct the basic performance parameter values ​​of the initial dynamic capability matrix through a data calibration algorithm. During the operation phase, real-time operating data of the water pumps are collected and data on the current typical water use patterns are extracted as correction data to construct a correction dataset; Based on the corrected dataset, the deviation analysis algorithm is used to calculate the deviation value of the basic performance parameters, and the deviation is mapped to the corresponding dimension of the basic performance parameters proportionally for point-by-point correction, and the basic performance parameters are dynamically updated. A multi-factor coupled performance degradation quantification model is constructed to calculate mechanical loss, efficiency degradation correction value and load fluctuation impact factor, and update state correction parameters.

5. The control method for the multi-pump coordinated variable frequency energy-saving water supply equipment as described in claim 1, characterized in that, The specific steps for defining the target water usage pattern include: Real-time water usage data of the current pipeline network is collected, and real-time water usage features are extracted and quantified. The real-time water usage features are then transformed into a real-time feature parameter set with the same structure as the feature parameter set in typical water usage patterns through a standardized parameter transformation algorithm. Calculate the cosine similarity between the real-time feature parameter set and the feature parameter sets in each typical water use pattern, set a similarity threshold, compare the cosine similarity between the real-time feature parameter set and the feature parameter sets in each typical water use pattern with the similarity threshold, and mark the typical water use pattern with a cosine similarity greater than the similarity threshold as a potential target water use pattern. If the number of potential target water use patterns is greater than 1, calculate the time matching degree between each potential target water use pattern and the current actual time, and select the potential target water use pattern as the target water use pattern based on the time matching degree. If the number of potential target water use patterns is 1, then the potential target water use pattern is used as the target water use pattern. If the number of potential target water use patterns is less than 1, a reference pattern is selected based on cosine similarity, and a temporary water use pattern is generated using a linear interpolation algorithm, which is then used as the target water use pattern.

6. The control method for the multi-pump coordinated variable frequency energy-saving water supply equipment as described in claim 1, characterized in that, The specific steps for constructing the pump set include: Obtain the target water usage pattern and extract the feature parameter set, time interval attributes, and demand intensity level; The dynamic capability matrix of each pump in the pump group is retrieved, and the basic performance parameters and state correction parameters of each pump are extracted to construct a pump performance parameter library. Based on the water pump performance parameter library, a correspondence matrix between the target water use mode and the water pump operating status is constructed to establish a quantitative correlation framework between water demand and water pump performance. The deviation rate of the correspondence matrix is ​​calculated, and dynamic weights are set in combination with the demand intensity level. The initial water pump adaptability score is obtained by weighted summation calculation, and then corrected according to the status correction parameters of each water pump to obtain the water pump adaptability score of each water pump. Based on the flow demand value of the target water use mode and the rated flow of the water pump, calculate the target number of water pumps required; sort all water pumps in descending order based on the water pump adaptability score, and select water pumps based on the target number to construct the initial candidate pump group. The initial candidate pump sets are subjected to performance verification. In response to failure of performance verification, the number of pumps in the candidate pump sets is increased based on the fit score to construct the execution pump set.

7. The control method for the multi-pump coordinated variable frequency energy-saving water supply equipment as described in claim 6, characterized in that, The specific steps for generating the control command for the water pump set include: Obtain candidate pump sets and target water usage patterns, decompose the flow demand value of the target water usage pattern, and divide it into basic load and dynamic adjustment load. Based on the dynamic capability matrix of each pump in the candidate pump group, the target operating frequency range of each pump is determined. Based on the target operating frequency range and the rated flow ratio of each pump, the basic load component that each pump needs to bear is calculated and the target operating frequency required for each basic load component is determined. If the target operating frequency of the water pump is within its target operating frequency range, it is marked as the basic load bearing pump, and the basic load is allocated to it according to the calculated target operating frequency; if the target operating frequency of the water pump exceeds its target operating frequency range, the allocation ratio of the basic load component is readjusted, and the target operating frequency of the water pump is updated. The state correction parameters in the dynamic capacity matrix of the water pump are standardized, the dynamic adjustment adaptation index is calculated, the number of pumps bearing the dynamic adjustment load is set, the pumps bearing the dynamic adjustment load are screened and marked according to the dynamic adjustment adaptation index, and the frequency adjustment range of each pump bearing the dynamic adjustment load is determined. Based on the fluctuation amplitude and frequency of flow rate in real-time water usage data, the dynamic adjustment load is allocated to each dynamic adjustment load pump according to the proportion of the dynamic adjustment adaptation index. Based on the base load and the results of dynamic load adjustment, control commands for the water pump set are generated. These commands include water pump start / stop commands, base operating frequency commands, and dynamic adjustment parameter commands.

8. The control method for the multi-pump coordinated variable frequency energy-saving water supply equipment as described in claim 1, characterized in that, The specific steps for determining whether the water pump group control command update has been triggered include: Extract water usage characteristics to update the real-time feature parameter set, measure the similarity with the current target water usage pattern feature parameter set, calculate the pattern matching degree, set a similarity judgment threshold, and count the cumulative time of pattern deviation when the pattern matching degree is less than the similarity judgment threshold. If it is greater than the preset pattern deviation time, the water pump group control command update is triggered. The pump operating status parameters are compared with the corresponding parameters in the dynamic capability matrix, and the parameter deviation rate is calculated. If the cumulative duration of parameter deviation is greater than the preset parameter deviation threshold and the parameter deviation duration is greater than the preset parameter deviation duration, the pump group control command is updated. For the pump that bears the basic load, monitor the cumulative duration of the load deviation between the actual operating frequency and the target operating frequency range. If it exceeds the preset load deviation duration, the pump group control command will be updated. For dynamically adjustable load-bearing pumps, monitor the cumulative duration of the actual adjustment amount deviating from the load tolerance range. If the cumulative duration of the deviation exceeds the preset adjustment time, the pump set control command will be updated.

9. A multi-pump coordinated variable frequency energy-saving water supply equipment, implemented based on the control method of the multi-pump coordinated variable frequency energy-saving water supply equipment according to any one of claims 1-8, characterized in that, include: Sensor network, data processing module, control module and instruction update determination module; The sensor network is used to collect water usage data and pump operation data at key nodes of the pipeline network; the data processing module is used to construct typical water usage patterns and dynamic capability matrices; the control module is used to locate target water usage patterns, screen pumps, and generate control commands; the command update determination module is used to monitor data and comprehensively determine whether to update the control commands.

Images