Water pump energy-saving control system based on intelligent data model

Abstract

The application relates to the technical field of water pump energy-saving control, and discloses a water pump energy-saving control system based on an intelligent data model, which comprises a data sensing module, a modal analysis module and the like. The data sensing module collects flow, lift and power data and sets a dynamic monitoring window; the modal analysis module divides operation modes and generates an energy efficiency characteristic matrix; a rule construction module extracts core energy efficiency factors and constructs an energy-saving regulation and control strategy; a priority scheduling module analyzes a control priority matrix and evaluates an energy efficiency deviation index; a mapping calibration module simulates an optimal energy efficiency benchmark curve and generates an energy efficiency deviation gradient atlas; and a control execution module generates a regulation and control instruction set according to the atlas. Through multi-modal energy efficiency analysis and dynamic regulation and control, the system realizes fine management and energy efficiency optimization of water pump operation, solves problems such as poor adaptability and low energy-saving efficiency of traditional control working conditions, and is suitable for water pump energy-saving control in the fields of industry and municipal administration and the like, and is accurate and efficient.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water pump energy-saving control, in particular to a water pump energy-saving control system based on an intelligent data model. BACKGROUND

[0002] As the core fluid conveying equipment in the fields of industrial production, municipal water supply, and agricultural irrigation, water pumps account for a large proportion of the overall system energy consumption. Traditional water pump control systems generally have low operating efficiency and serious energy waste, mainly due to the difficulty of existing control strategies to dynamically adapt to complex and variable working conditions, and the lack of fine analysis of water pump operating states and energy efficiency optimization capabilities.

[0003] From the current technical development status, early water pump control mainly relied on manual experience or simple start-stop control, which could not realize real-time monitoring and dynamic adjustment of key parameters such as flow rate, head, and power. With the development of automation technology, variable frequency speed regulation systems based on PID control gradually became popular, but their control logic was relatively fixed, only allowing adjustment of a single parameter, making it difficult to handle multi-variable coupling problems. When facing frequently fluctuating working conditions or multi-modal operating scenarios, it is prone to adjustment lag and low energy efficiency matching.

[0004] In existing technologies, some studies attempt to achieve energy-saving control by establishing water pump operation models, but most are based on static models under fixed working conditions, which cannot effectively cope with the complexity of working conditions caused by changes in equipment models, pipe characteristics, liquid physical properties (such as density), environmental vibrations, etc. For example, different models of water pumps may have significant differences in power consumption under the same flow rate and head conditions. Traditional models fail to fully consider the impact of individual differences in equipment and operating stages (such as the run-in period, stable operation period, and aging period) on energy efficiency, resulting in insufficient universality and precision of energy-saving strategies.

[0005] In addition, existing control systems lack systematicness and dynamics in energy efficiency evaluation and control strategy generation. First, the extraction of energy efficiency factors (such as mechanical loss proportion, ineffective circulation proportion, and pressure fluctuation parameters) is relatively single, and a comprehensive energy efficiency feature matrix cannot be constructed. Second, control strategies are often based on a single objective (such as minimum power), without considering the priority relationship and coupling effect between different adjustment parameters (such as frequency and pressure), which can easily lead to energy efficiency deviation or increased equipment wear during adjustment. For example, in the cooperative control of frequency adjustment and pressure adjustment, traditional methods do not explicitly quantify the relationship between time weight (such as adjustment period) and space weight (such as pressure distribution range), making it difficult to achieve optimal scheduling of multiple parameters.

[0006] With the development of industrial intelligence, how to use data-driven technology to improve the energy efficiency optimization capability of the water pump control system has become a technical problem to be solved. The existing technology lacks deep mining and intelligent modeling capability of water pump operation data, cannot analyze multiple operation modes in real time through dynamic monitoring window, and is difficult to realize intelligent control of the whole process from data perception, mode decoupling, rule construction to regulation execution, so that the energy saving potential of the water pump under complex working conditions cannot be fully released. SUMMARY

[0007] The purpose of the present application is to provide a water pump energy-saving control system based on an intelligent data model to solve the problems raised in the background art.

[0008] To achieve the above purpose, the present application provides the following technical solution: a water pump energy-saving control system based on an intelligent data model, comprising:

[0009] A data perception module is used to collect flow, head and power data in the operation of the water pump, and set a dynamic monitoring window matched with the operation condition of the water pump;

[0010] A mode analysis module is used to divide multiple operation modes within the dynamic monitoring window, decouple the power data of each operation mode, and generate an energy efficiency characteristic matrix corresponding to the operation mode;

[0011] A rule construction module is used to extract core energy efficiency factors from the energy efficiency characteristic matrix, construct energy-saving regulation strategies associated with the operation mode, and extract device adjustment parameters corresponding to the strategies;

[0012] A priority scheduling module is used to analyze the control priority matrix in the device adjustment parameters, schedule the core energy efficiency factors according to the priority matrix, and evaluate the energy efficiency deviation index of each operation mode under different scheduling sequences;

[0013] A mapping calibration module is used to fit an optimal energy efficiency reference curve according to the energy efficiency deviation index, generate an energy efficiency deviation gradient map by matching the current operation energy efficiency value with the optimal energy efficiency reference curve;

[0014] A control execution module is used to analyze the energy efficiency deviation gradient map, convert the energy efficiency deviation gradient map into a water pump energy-saving regulation instruction set based on the energy efficiency deviation direction of the operation mode.

[0015] Preferably, the implementation mode of the mode analysis module comprises: establishing a working condition characteristic library corresponding to the operation mode, and the working condition characteristic library contains a working condition parameter matrix mapped by the flow, head and power data;

[0016] Similar working condition matching is performed on the working condition parameter matrix, and working condition clusters of the working condition parameter matrix are divided according to a matching result.

[0017] Preferably, the dividing the working condition clusters of the working condition parameter matrix further comprises:

[0018] According to the equipment model and the running stage in the working condition parameter matrix, pipeline pressure, liquid density and environmental vibration parameters are extracted, and a working condition feature identifier is generated based on the above parameters;

[0019] The working condition feature identifier is associated with the working condition parameter matrix, the working condition correlation degree between the feature identifiers is calculated, and the working condition parameter matrix with a correlation degree exceeding a preset working condition threshold is screened to form a working condition cluster.

[0020] Preferably, the implementation manner of generating the energy efficiency feature matrix corresponding to the running mode comprises:

[0021] For each running mode, flow deviation data of the mode in a set period are obtained according to the time sequence distribution of the mode in the dynamic monitoring window, and a flow deviation coefficient of the mode is calculated;

[0022] When the flow deviation coefficient exceeds a first deviation threshold, the mode is marked as an abnormal mode, and power data thereof are extracted to construct an energy efficiency feature matrix; when the flow deviation coefficient is lower than the first deviation threshold, the mode is marked as a stable mode, and power data of adjacent modes of the mode are dimensionally aggregated, and the aggregated data are reorganized into an energy efficiency feature matrix.

[0023] Preferably, the implementation manner of the rule construction module comprises:

[0024] Mechanical loss proportion, invalid cycle proportion and pressure fluctuation parameters are separated from the energy efficiency feature matrix, and an energy saving regulation and control strategy of the running mode is generated based on the above parameters;

[0025] If the number of running modes covered by the current energy saving regulation and control strategy is less than a preset mode threshold, the energy efficiency feature matrices of adjacent running modes are traversed, and energy efficiency factors not contained in the strategies of the adjacent modes are added to the current strategy.

[0026] Preferably, the implementation manner of the priority scheduling module comprises: obtaining a time weight of frequency adjustment and a space weight of pressure adjustment in a control priority matrix;

[0027] A scheduling decision tree associated with the time weight and the space weight is constructed, and an energy efficiency deviation index under different scheduling sequences is determined according to path scores of branches in the decision tree.

[0028] Preferably, the construction of the scheduling decision tree further comprises:

[0029] recognize the periodicity of the time weight, if the current periodicity is completely consistent with the preset operation period, the time weight is set as the root node of the scheduling decision tree;

[0030] calculate the coupling degree of the time weight and the space weight, and generate the intermediate nodes and the leaf nodes of the scheduling decision tree in turn according to the coupling degree from high to low;

[0031] perform path backtracking on the leaf nodes, and mark the leaf nodes as the preferred path of the scheduling decision tree when the coupling degree of the leaf nodes is lower than the preset coupling threshold.

[0032] Preferably, the implementation manner of the evaluation of the energy efficiency deviation index comprises:

[0033] statistically calculate the time weight variance of each leaf node in the scheduling decision tree and the space weight extreme value, and calculate the comprehensive standard deviation of all nodes in the tree;

[0034] calculate the time deviation coefficient by taking the time weight variance of a single leaf node as the dividend, the time weight variance of the same level node as the divisor, and multiplying the result by the comprehensive standard deviation; meanwhile, calculate the ratio of the space weight extreme value to the comprehensive standard deviation, and take the geometric mean of the two as the energy efficiency deviation index of the node.

[0035] Preferably, the implementation manner of the fitting of the optimal energy efficiency reference curve comprises:

[0036] call the control record closest to the current energy efficiency deviation index in the historical database, calculate the cosine similarity of the two in the spatial distribution as the first mapping reference value;

[0037] statistically calculate the coincidence degree of the current energy efficiency deviation index and the historical control record in the peak point distribution, and take the reciprocal of the coincidence degree as the second mapping reference value;

[0038] based on the vector composition result of the first mapping reference value and the second mapping reference value, match the optimal energy efficiency reference curve in the preset energy efficiency curve library.

[0039] Preferably, the implementation manner of the control execution module comprises: according to the energy efficiency deviation angle of each mode in the energy efficiency deviation gradient map, dividing the positive adjustment region and the reverse adjustment region;

[0040] extract the adjustment step of the energy efficiency deviation in the positive adjustment region and the compensation amplitude of the energy efficiency deviation in the reverse adjustment region, dynamically match the two according to the working condition weight of the operation mode, and generate the core execution parameter of the water pump energy-saving control instruction set.

[0041] Compared with the prior art, the present application has the following advantages:

[0042] The data awareness module collects core data such as flow, head and power by setting a dynamic monitoring window matching the operating condition of the water pump, providing accurate real-time data support for subsequent energy efficiency analysis. Compared with the traditional fixed threshold monitoring method, the dynamic monitoring window can adaptively adjust the monitoring range according to the operating condition changes, effectively improving the pertinence and timeliness of data collection, and ensuring that the system can capture the impact of operating condition fluctuations on energy efficiency in a timely manner.

[0043] The modal analysis module divides the water pump operating process into multiple operating modes with clear characteristics by constructing a working condition feature library, matching similar working conditions and clustering, and decoupling the power data of each mode to generate an energy efficiency feature matrix. This process not only achieves fine classification of complex working conditions, but also accurately identifies stable and abnormal modes by extracting the distribution characteristics of power data. For abnormal modes, the power data is directly extracted to construct an energy efficiency feature matrix, which can quickly locate the root cause of energy efficiency anomalies; for stable modes, the power data of adjacent modes is aggregated in dimension to achieve comprehensive representation of energy efficiency characteristics under normal operating conditions, laying a foundation for accurate formulation of subsequent energy-saving control strategies.

[0044] The rule construction module separates core energy efficiency factors such as mechanical loss, ineffective circulation and pressure fluctuation from the energy efficiency feature matrix, constructs energy-saving control strategies associated with operating modes, and dynamically extends the strategy coverage range by traversing adjacent mode energy efficiency feature matrices. This multi-factor analysis-based strategy construction method breaks through the limitations of traditional single-parameter adjustment and can develop individualized control programs for different operating modes. At the same time, by dynamically expanding the strategy coverage range, the system always maintains the integrity and effectiveness of the strategy when the operating condition changes, avoiding the control blind spot caused by insufficient mode coverage.

[0045] The priority scheduling module constructs a scheduling decision tree associated with time weight and space weight by analyzing the control priority matrix, realizes hierarchical scheduling of core energy efficiency factors, and determines the optimal control sequence by evaluating the energy efficiency deviation index of each operating mode under different scheduling sequences. This module fully considers the time characteristics of frequency regulation and the space characteristics of pressure regulation, calculates the coupling degree of the two to generate the node structure of the scheduling decision tree, and selects the optimal path through path backtracking. This multi-dimensional weight analysis-based scheduling mechanism can ensure energy efficiency optimization while considering the stability and reliability of device regulation, avoiding device wear or energy fluctuation problems caused by traditional unordered regulation.

[0046] The mapping calibration module calculates the cosine similarity and peak coincidence degree by calling the historical regulation record based on the energy efficiency deviation index, fits the optimal energy efficiency reference curve, and generates the energy efficiency deviation gradient atlas. This process dynamically maps the real-time energy efficiency data with the historical optimal data, ensures that the system can quickly match the best energy efficiency reference according to the current working condition, and provides a scientific basis for precise regulation. The generation of the energy efficiency deviation gradient atlas directly displays the energy efficiency deviation degree and direction of each operating mode, and provides a clear regulation guide for the control execution module.

[0047] The control execution module divides the positive regulation region and the reverse regulation region according to the energy efficiency deviation gradient atlas, extracts the regulation step and the compensation amplitude, dynamically matches with the working condition weight, and generates the energy saving regulation instruction set. This regulation method based on region division and weight matching realizes the differentiated processing of energy efficiency deviation of different modes, ensures that the regulation process can quickly correct the energy efficiency deviation, and avoids the influence of excessive regulation on the running stability of the equipment. By converting the energy efficiency deviation gradient atlas into specific instruction set, the system can directly drive the water pump equipment to execute precise regulation, form a complete closed loop from data sensing to control execution, and improve the real-time performance and effectiveness of energy saving control. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 The working principle diagram of the water pump energy saving control system based on intelligent data model is described.

[0049] Figure 2 The working principle diagram of the mode analysis module is described.

[0050] Figure 3 The working principle diagram of the working condition cluster division is described.

[0051] Figure 4 The working principle diagram of the energy efficiency characteristic matrix generation is described. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0053] Please refer to Figures 1-4 The present application relates to a water pump energy saving control system based on intelligent data model, which comprises:

[0054] Data sensing module: During the operation of the water pump, real-time collection of flow, head and power data is performed, and a dynamic monitoring window matching the actual operation condition of the water pump is set according to the actual operation condition of the water pump. The dynamic monitoring window can dynamically adjust the monitoring range and frequency according to factors such as water pump load and operation time, so as to ensure that the collected data can accurately reflect the current working condition characteristics.

[0055] Modal analysis module: In the dynamic monitoring window, the water pump operation state is divided into multiple operation modes. Through mode decoupling of the power data of each operation mode, an energy efficiency characteristic matrix corresponding to each operation mode is generated.

[0056] Rule construction module: The core energy efficiency factor is extracted from the energy efficiency characteristic matrix, the energy saving control strategy associated with the operation mode is constructed, and the device adjustment parameter corresponding to the strategy is extracted.

[0057] Priority scheduling module: The control priority matrix in the device adjustment parameter is analyzed, the core energy efficiency factor is scheduled according to the priority matrix, and the energy efficiency deviation index of each operation mode under different scheduling sequences is evaluated.

[0058] Mapping calibration module: According to the energy efficiency deviation index, the optimal energy efficiency reference curve is fitted, and the energy efficiency deviation gradient atlas is generated by matching the current operation energy efficiency value with the optimal energy efficiency reference curve.

[0059] Control execution module: The energy efficiency deviation gradient atlas is analyzed, the energy efficiency deviation gradient atlas is converted into a water pump energy saving control instruction set based on the energy efficiency deviation direction of the operation mode, and the adjustment action is completed by driving the execution mechanism.

[0060] The application will be further described below in combination with Examples 1 to 5.

[0061] Example 1:

[0062] This embodiment relates to the specific implementation of the modal analysis module. The modal analysis module realizes operation mode division by establishing a working condition characteristic library, which specifically includes working condition characteristic library construction, similar working condition matching, working condition cluster division, operation mode setting and working condition characteristic identification association.

[0063] For different operation modes, corresponding working condition characteristic libraries are constructed, which store working condition parameter matrices formed by mapping flow, head and power data. The working condition parameter matrix is presented in the form of a multidimensional array, and each dimension corresponds to the real-time monitoring value and historical data statistical characteristics of different physical quantities. For example, the flow dimension includes instantaneous flow, flow fluctuation range per unit time, flow average and other parameters; the power dimension includes active power, reactive power, power factor and power change rate and other parameters. These parameters are collected and preprocessed by the data sensing module and stored in the matrix to form a multidimensional description of the water pump operation state.

[0064] Similar operating condition matching is performed based on the operating condition parameter matrix. The dynamic time warping algorithm or the cosine similarity calculation method is used for similar operating condition matching. The dynamic time warping algorithm is suitable for processing non-aligned problems of time series data. The similarity of operating condition characteristics is measured by calculating the time series distance of the operating condition parameter matrix in different time periods. The cosine similarity calculates the cosine value of the angle between two operating condition parameter matrix vectors to determine the direction similarity in the vector space. Regardless of the method used, a similarity judgment threshold needs to be set. When the similarity calculation result of two operating condition parameter matrices is less than the threshold, it is determined that the operating conditions are similar.

[0065] According to the similar operating condition matching result, the operating condition parameter matrix is divided into several operating condition clustering clusters. Each clustering cluster represents a typical operating state, and the characteristic difference of the operating condition parameter matrix in the cluster needs to be controlled within a preset range. For example, for the stable running stage of the water pump, the operating condition data with a flow fluctuation of less than 3%, a head change of less than 5%, and a power fluctuation within ±4% of the rated power can be divided into the same clustering cluster, representing the operating state of the water pump under a certain stable load.

[0066] The distribution feature points of the power data are extracted from each operating condition clustering cluster as the key identification of the operating mode. The extraction of the distribution feature points is based on statistical methods and signal processing techniques, including: calculating the local peak and valley of the power data by the sliding window algorithm, identifying the peak point of the probability density function of the power data by the mean shift algorithm, analyzing the frequency components of the power data by the Fourier transform and extracting the feature points corresponding to the main frequency components, etc. These feature points cover power peak, valley, mean, main frequency and change trend turning point, etc., and each feature point corresponds to an independent operating mode. For example, when the power data has a significant peak, it is defined as a "high load mode"; when it has a valley, it is defined as a "low load mode"; when the main frequency changes by more than a preset threshold, it is defined as a "speed switching mode".

[0067] In addition, the operating condition clustering cluster needs to be divided in combination with the device model and the running stage. The impeller structure, motor power and fluid mechanics characteristics of different models of water pumps are different, resulting in different energy efficiency curves and operating condition characteristics. For example, a large-diameter water pump may experience "surge" at low flow operating conditions, while a small-diameter water pump may run stably at the same operating conditions. The operating condition characteristics of the running stage (such as the starting stage, the stable running stage, and the shutdown stage) also have obvious differences. The power during the starting stage will first increase sharply and then flatten out, and the head during the shutdown stage will gradually decrease. Therefore, the pipeline pressure, liquid density and environmental vibration parameters need to be extracted, and the operating condition characteristic identification information including the device model, the running stage and the environmental parameters needs to be generated.

[0068] The association between the operating condition characteristic identifier and the operating condition parameter matrix is realized through data labeling. Each operating condition parameter matrix is assigned a unique characteristic identifier label. The label content includes the equipment model code, the running stage identifier (such as S1 representing the start-up stage and S2 representing the stable running stage), the pipeline pressure level, the liquid density range, and the environmental vibration intensity, etc. The operating condition parameter matrix is screened by calculating the operating condition correlation degree between the characteristic identifiers. The operating condition parameter matrix corresponding to the operating condition characteristic identifier with the same equipment model, consistent running stage, and small differences in pipeline pressure, liquid density, and environmental vibration parameters less than the preset value is included in the same clustering cluster by using the Euclidean distance or correlation coefficient method. For example, for the water pump with the equipment model P-100 in the stable running stage (S2), only the operating condition parameter matrix with the pipeline pressure of 0.2-0.25 MPa, the liquid density of 900-1000 kg / m³, and the environmental vibration intensity less than 2.8 mm / s is included, so as to ensure that the environmental conditions in the clustering cluster are consistent with the equipment characteristics and improve the accuracy of the running mode division.

[0069] In the operating condition clustering cluster division process, the timeliness of the data samples also needs to be considered. With the increase of the running time of the water pump, the performance of the equipment may be attenuated (such as impeller wear and seal aging), which may cause changes in the running characteristics under the same operating condition. Therefore, the operating condition characteristic library needs to be updated regularly, and the obsolete data needs to be deleted and the newly collected operating condition parameter matrix needs to be supplemented, so as to ensure that the clustering cluster always reflects the current actual running characteristics of the water pump. The update period can be set according to the equipment maintenance period or the running time, for example, the update is performed once every quarter or the update is triggered after 1000 hours of running.

[0070] The number and granularity of the running modes can be adjusted according to actual needs. For complex operating condition scenarios, the number of running modes can be increased, and the mode division granularity can be refined to more accurately capture the energy efficiency change characteristics; for simple operating condition scenarios, the number of running modes can be reduced to improve the system processing efficiency. For example, in the scenario where the water pump needs to frequently switch the load in the chemical process, 10-15 running modes can be divided; in the scenario of stable load in the civil water supply system, only 3-5 running modes can be divided.

[0071] The output of the mode analysis module provides data basis for the energy efficiency characteristic matrix input interface corresponding to each running mode. Through accurate division of the running mode and extraction of the power data characteristics, reliable energy efficiency analysis basis is provided for the subsequent rule construction module, priority scheduling module, etc., so as to ensure that the entire water pump energy-saving control system can implement accurate energy-saving control strategies for different running states.

[0072] The modal analysis module realizes fine division of the operation mode of the water pump through technical means such as multi-dimensional data association, dynamic similarity matching and feature point extraction, and lays a solid working condition analysis foundation for energy efficiency optimization of the system. The design of the module fully considers factors such as equipment characteristics, operating environment and data timeliness, thereby guaranteeing the scientificity and practicality of the operation mode division.

[0073] Embodiment 2:

[0074] The process of generating the energy efficiency characteristic matrix corresponding to the operation mode is based on the output results of the modal analysis module, and energy efficiency characteristics are extracted and matrixes are constructed for each operation mode, specifically including operation mode time sequence analysis, flow deviation data acquisition, abnormal mode processing, stable mode data aggregation and energy efficiency characteristic matrix generation.

[0075] After the modal analysis module completes the operation mode division, the embodiment first analyzes the time sequence distribution of each operation mode within the dynamic monitoring window. The size of the dynamic monitoring window is set according to the characteristics of the operation mode. For operation modes that change quickly (such as start-stop stage modes), the window time is short (such as 1-5 minutes); for stable operation modes (such as constant speed operation modes), the window time is long (such as 30 minutes-2 hours). Through the analysis of the time sequence distribution, the flow deviation data of the mode in the set period is obtained. The flow deviation data is the difference between the actual flow and the theoretical optimal flow under the mode, and the theoretical optimal flow can be fitted by the design parameters or historical operation data of the water pump. For example, for a centrifugal pump, the theoretical optimal flow can be determined according to the head-flow curve (H-Q curve) of the pump. When the head is a certain fixed value, the flow corresponding to the highest efficiency point on the curve is the theoretical optimal flow.

[0076] The flow deviation coefficient is calculated to judge the stability of the mode. The flow deviation coefficient is the ratio of the flow deviation value to the theoretical optimal flow. When the flow deviation coefficient exceeds the first deviation threshold (such as 5%), it is determined that the mode is an abnormal mode. The appearance of the abnormal mode can be caused by various factors, such as pipe blockage, valve failure, impeller damage, etc. At this time, the power data of the mode is extracted to construct the energy efficiency characteristic matrix, which contains feature parameters such as power fluctuation range, power change rate, proportion of each harmonic component, etc. The power fluctuation range is obtained by calculating the difference between the maximum power and the minimum power in the dynamic monitoring window; the power change rate is the change amount of power per unit time, reflecting the degree of power change; the proportion of harmonic components is obtained by decomposing the power signal into fundamental wave and each harmonic through Fourier transform, and calculating the proportion of the energy of each harmonic component to the total energy. These feature parameters are used to analyze the reasons for energy loss under abnormal state, for example, if the proportion of high-order harmonic components is large, it may indicate that there is a motor fault or power grid interference; if the power change rate is too fast, there may be a problem of load mutation or control system response not timely.

[0077] When the flow deviation coefficient is lower than the first deviation threshold, the mode is marked as a stable mode. For the stable mode, the power data of its adjacent modes need to be dimensionally aggregated. Adjacent modes refer to modes that are continuous in time or have a short time interval with the current mode. For example, the current stable mode is a constant-speed running mode, and its adjacent modes can be start-up modes or speed transition modes. Dimensional aggregation uses dimension reduction algorithms such as principal component analysis or independent component analysis to compress multi-dimensional power data into a few comprehensive feature vectors. Principal component analysis determines the principal component direction of the data by calculating the covariance matrix of the power data, projects the original data into the principal component space, and realizes data dimension reduction; independent component analysis assumes that the power data is mixed by multiple independent source signals, separates the mixed signals into independent components through blind source separation technology, and extracts the main independent components as comprehensive feature vectors.

[0078] The aggregated data is reorganized into an energy efficiency feature matrix in time sequence or feature importance order. If reorganized in time sequence, the row vectors of the matrix represent comprehensive feature vectors at different times, and the column vectors represent the values of each feature parameter at different times; if reorganized in feature importance, the row vectors of the matrix represent different feature parameters, and the column vectors represent the statistical values (such as mean, variance, maximum, etc.) of each feature parameter in different time windows. The matrix reflects the energy efficiency distribution law and energy transfer characteristics between modes in the stable running state. For example, by analyzing the correlation of each element in the matrix, it can be found that there is a clear linear relationship between some feature parameters (such as flow and power), and there can be a nonlinear relationship between other feature parameters (such as temperature and efficiency).

[0079] When constructing the energy efficiency feature matrix, the spatiotemporal characteristics of the data need to be considered. For spatial characteristics, the power data of different monitoring points can be different, for example, the power data measured by the pressure sensors at the inlet and outlet of the pump can be different due to different pipe resistances. Therefore, spatial calibration of data from different monitoring points is needed, and by establishing a spatial coordinate transformation model or introducing a weight coefficient, the data from each monitoring point is mapped to the same coordinate system. For time characteristics, since there can be periodic fluctuations (such as power frequency fluctuations of the motor) during pump operation, time-frequency analysis methods (such as wavelet transform) are used to decompose the power data into different frequency components, extract the feature frequency components related to energy efficiency, and filter out noise interference.

[0080] The dimension of the energy efficiency feature matrix is determined according to the complexity of the operating mode. For a simple operating mode (such as stable operation at a single speed), the matrix dimension is low (such as 3-5 dimensions), containing basic parameters such as flow rate, head, and power. For a complex operating mode (such as parallel operation of multiple pumps or frequent speed regulation operation), the matrix dimension is high (such as 10-20 dimensions), containing parameters such as coupling coefficients between pump groups, speed regulation response time, and pressure fluctuation frequency in addition to basic parameters. The selection of matrix dimension needs to balance the data expression ability and calculation efficiency, avoiding the increase of calculation complexity due to high dimension, while ensuring that the matrix can fully represent the energy efficiency characteristics of the operating mode.

[0081] After the energy efficiency feature matrix is generated, the matrix needs to be normalized to eliminate the influence of different characteristic parameter dimensions and orders of magnitude. The normalization method can use minimum-maximum normalization, Z-score normalization, or decimal scaling normalization.

[0082] Minimum-maximum normalization linearly maps the characteristic parameter value to the [0, 1] interval, and the calculation formula is:

[0083]

[0084] wherein, is the original value, and are the minimum and maximum values of the characteristic parameter, respectively;

[0085] Z-score normalization converts the characteristic parameter to a standard normal distribution by calculating the mean and standard deviation, and the calculation formula is:

[0086]

[0087] wherein, is the mean, is the standard deviation;

[0088] Decimal scaling normalization maps the characteristic parameter value to the [-1, 1] interval by moving the decimal point position.

[0089] The update frequency of the energy efficiency feature matrix is related to the stability of the operating mode. For stable operating modes, the matrix can be updated every hour or every day; for frequently changing operating modes, the matrix needs to be updated in real time or quasi-real time. During the update process, historical matrix data needs to be preserved to form a matrix sequence for analyzing the trend of energy efficiency characteristics over time. For example, by comparing the energy efficiency feature matrices at different time points, changes in device performance degradation or energy efficiency optimization effects can be found.

[0090] The generated energy efficiency feature matrix serves as an input of the rule construction module, providing a data basis for formulating energy-saving control strategies. By analyzing each feature parameter in the matrix, key factors affecting energy efficiency can be identified, energy-saving potential points under each operating mode can be determined, and targeted control strategies can be constructed. For example, if the matrix analysis shows that the power fluctuation is large under a certain operating mode, a control strategy can be formulated to smooth the power curve; if a certain feature parameter (such as head) has a strong correlation with the energy efficiency index, it can be used as the main control target.

[0091] This embodiment realizes the scientific construction of the energy efficiency feature matrix corresponding to the operating mode through time sequence analysis of the operating mode, flow deviation judgment, differential processing of abnormal and stable modes, and data dimension aggregation and other technical means. The matrix comprehensively reflects the energy efficiency characteristics of the water pump under different operating states, providing accurate and reliable data support for the subsequent formulation of energy-saving control strategies.

[0092] Embodiment 3:

[0093] The core function of the rule construction module is to extract core energy efficiency factors from the energy efficiency feature matrix, construct energy-saving control strategies associated with operating modes, and extract device adjustment parameters corresponding to the strategies. This process includes energy efficiency factor separation, strategy generation, mode coverage judgment, and strategy expansion, and realizes accurate formulation and dynamic optimization of energy-saving strategies through multi-dimensional data analysis.

[0094] The rule construction module separates core energy efficiency factors such as mechanical loss ratio, invalid cycle ratio, and pressure fluctuation parameters from the energy efficiency feature matrix. The mechanical loss ratio is determined by analyzing the constant loss component in the power data, which includes motor iron loss, mechanical friction loss, etc. This type of loss is not directly related to the load size and can be estimated through no-load test data or low-load interval power values in historical operation data. The invalid cycle ratio reflects the energy waste caused by pipe backflow, valve leakage, or poor impeller sealing during pump operation. It is calculated by comparing the difference between the actual flow and the theoretical flow, combined with the head data - for example, when the actual flow is significantly lower than the theoretical flow and the head does not increase synchronously, there may be invalid cycles. The pressure fluctuation parameter is represented by the statistical characteristics of the head data, such as the standard deviation, range, or root mean square value of the head within the dynamic monitoring window. The larger the standard deviation, the more severe the pressure fluctuation, which may be caused by water flow impact, pump set vibration, or load mutation.

[0095] Based on the above core energy efficiency factors, corresponding energy-saving control strategies are generated for different operating modes. The forms of the strategies include device parameter adjustment suggestions, maintenance operation guidelines, or operating mode switching instructions, etc. For example, when the mechanical loss proportion of a certain operating mode exceeds the preset threshold (such as 15% of the rated power), the strategy may suggest adjusting the gap between the impeller and the pump shell, replacing the lubricating oil, or checking the bearing wear; when the invalid circulation proportion is high (such as more than 10% of the theoretical flow), the strategy may trigger automatic adjustment of the valve opening, start the pipeline leakage detection program, or switch to the standby pump operation; when the pressure fluctuation parameter exceeds the safety range, the strategy may reduce the motor speed through the frequency converter, adjust the outlet valve opening to smooth the water flow, or enable the damping device.

[0096] After generating the energy-saving control strategy, the system needs to determine whether the number of operating modes covered by the current strategy is less than the preset mode threshold (such as 70% of the total number of modes). The preset mode threshold can be set according to the actual working condition complexity of the water pump. For scenes with single working conditions, the threshold can be appropriately reduced; for scenes with frequent multi-mode switching, the threshold needs to be increased accordingly to ensure the comprehensiveness of the strategy. If the current strategy is insufficient, the system automatically starts the adjacent mode traversal mechanism, retrieves the energy efficiency characteristic matrix of the adjacent operating mode, and extracts the energy efficiency factors not included in the current strategy. The determination of adjacent modes is based on the spatial distance or time sequence continuity of the modes in the working condition parameter matrix, for example, the first 3 modes with the smallest Euclidean distance from the current mode in the working condition clustering cluster, or the previous and next 2 modes that appear continuously in time.

[0097] In the traversal of adjacent modes, the system identifies the energy efficiency factors not involved in the current strategy by comparing the element differences of the energy efficiency characteristic matrix. These factors may include temperature parameters (such as motor winding temperature, bearing temperature), vibration frequency (such as specific frequency vibration caused by impeller imbalance), fluid viscosity (affecting the volumetric efficiency of the pump), etc. For example, if the current strategy is only formulated for power and flow parameters, and the energy efficiency characteristic matrix of the adjacent mode shows that temperature rise causes liquid viscosity to decrease, thereby causing head loss, then temperature and viscosity parameters are included as new energy efficiency factors in the current strategy, supplementing the control content of "when the temperature exceeds the set value, adjust the cooling system flow or switch to high-temperature-resistant sealing materials".

[0098] In the strategy expansion process, the factor correlation principle needs to be followed, that is, the newly added energy efficiency factors need to have physical correlation or energy efficiency influence path with existing factors. For example, vibration frequency has a direct correlation with mechanical loss (increased vibration will increase friction loss), so it can be included as a derivative factor of mechanical loss in the strategy; while the correlation between environmental humidity and energy efficiency is low, if it does not show significant influence in the energy efficiency characteristic matrix, it is not included. In addition, redundant factors in the strategy need to be avoided, and highly correlated factors are removed through principal component analysis or correlation analysis to ensure the simplicity and executability of the strategy.

[0099] When the extraction strategy corresponds to the device adjustment parameter, the abstract control target needs to be converted into specific execution instruction parameters. For example, if the strategy is "reduce the motor speed to reduce the invalid cycle", the device adjustment parameters include the target speed value, the speed adjustment rate, the adjustment time window, etc.; if the strategy is "adjust the valve opening to stabilize the pressure", the parameters include the target value of the valve opening, the opening adjustment step, the adjustment period, etc. These parameters are matched through the historical control record database or calculated based on the water pump operation characteristic model, for example, the corresponding relationship between the target speed and the flow reduction amount is determined according to the motor speed-flow characteristic curve, and the attenuation coefficient of the opening adjustment and the pressure fluctuation is determined according to the valve opening-pressure characteristic curve.

[0100] The rule construction module also needs to have the strategy conflict detection and priority management function. When there are multiple control strategies for the same operation mode, the system automatically generates a strategy execution sequence by analyzing the target priority of the strategy (such as the safety class strategy priority higher than the energy efficiency optimization class strategy), the execution time window (such as the real-time adjustment strategy prior to the periodic maintenance strategy), and the parameter influence range (such as the global parameter adjustment strategy prior to the local parameter adjustment strategy), to avoid control conflicts caused by parallel execution of strategies. For example, when there are "increase the speed to maintain the head" and "reduce the speed to reduce the mechanical loss" strategies at the same time, the system needs to determine which strategy to execute first according to the current energy efficiency deviation direction and the working condition urgency, and suspend the execution of the other strategy.

[0101] The output result of the rule construction module, the energy-saving control strategy and the device adjustment parameter, is the input of the priority scheduling module, which provides the basis for the hierarchical scheduling of the core energy efficiency factor. Through continuous monitoring of the change of the operation mode and the update of the energy efficiency characteristics, this module realizes the dynamic optimization of the energy-saving strategy, ensuring that the system can provide targeted energy efficiency improvement schemes in the whole life cycle of the water pump operation.

[0102] This embodiment builds an energy-saving control strategy system covering the whole operation mode of the water pump through the mechanisms of energy efficiency factor separation, strategy generation and expansion, parameter extraction, and conflict management. The design of this module fully combines the multi-dimensional analysis of energy efficiency characteristics and the engineering practice needs, ensuring the scientificity, comprehensiveness and operability of the strategy.

[0103] Embodiment 4:

[0104] The embodiment relates to specific implementation modes of a priority scheduling module and a mapping calibration module. A working process of the priority scheduling module and the mapping calibration module is based on an energy-saving regulation and control strategy and device adjustment parameters output by a rule construction module, multi-dimensional data analysis and mapping matching are performed, hierarchical scheduling of a core energy efficiency factor and dynamic calibration of an energy efficiency benchmark are realized, and the specific implementation modes include control priority matrix analysis, scheduling decision tree construction, energy efficiency deviation index evaluation, optimal energy efficiency benchmark curve fitting, and energy efficiency deviation gradient atlas generation.

[0105] The priority scheduling module obtains a frequency adjustment time weight and a pressure adjustment space weight in the control priority matrix. The frequency adjustment time weight reflects the time sensitivity of frequency adjustment to energy efficiency optimization. For example, in a load fast-changing working condition, frequency adjustment needs to respond quickly to maintain the water pump in an efficient area, so the time weight is high; in a stable load working condition, the time sensitivity of frequency adjustment is low, and the time weight is correspondingly reduced. The pressure adjustment space weight reflects the influence range of pressure adjustment in the spatial distribution. For example, in a long-distance water conveying pipeline system, the pressure adjustment at different positions has different influences on the overall energy efficiency. The pressure adjustment close to the pump outlet may affect the energy consumption distribution of the entire pipeline network, and the space weight is large; and the local pressure adjustment far away from the pump outlet only affects a specific area, and the space weight is large.

[0106] The scheduling decision tree is constructed based on the time weight and the space weight. In the decision tree construction process, the periodic characteristics of the time weight are first identified. The periodic characteristics of the time weight can be obtained by analyzing the load change law in the historical operation data. For example, some industrial scenes have obvious daily load cycles (high load in the daytime and low load at night) or seasonal cycles (large cooling water demand in summer and small demand in winter). If the current periodic characteristics are completely consistent with the preset operation period (such as a typical load cycle statistically obtained in the past), the time weight is set as the root node of the scheduling decision tree, the time dimension is taken as the primary basis for scheduling priority, and the frequency adjustment task with high time sensitivity is preferentially processed.

[0107] The coupling degree of the time weight and the space weight is calculated. The coupling degree reflects the synergy strength of the two weight parameters in the energy efficiency optimization process. A high coupling degree indicates that frequency adjustment and pressure adjustment need to be simultaneously performed to achieve the best energy efficiency improvement effect. For example, in the variable frequency speed regulation process, if the outlet valve opening degree is not simultaneously adjusted to balance the pressure, the water pump operating point may deviate from the high efficiency area. A low coupling degree indicates that the two parameters can be independently adjusted and do not affect each other. The intermediate nodes and the leaf nodes of the scheduling decision tree are generated in turn according to the coupling degree from high to low. The parameters with high coupling degrees have a synergistic effect on energy efficiency, and need to be preferentially scheduled. The parameters with low coupling degrees can be used as independent adjustment factors and are separately generated into decision branches.

[0108] The path of the leaf node is traced back, and when the coupling degree of the leaf node is lower than the preset coupling threshold, the path is marked as the preferred path. The preset coupling threshold is set according to the characteristics of the water pump system and historical control experience. For example, for a centrifugal pump system, when the coupling degree of the time weight and the space weight is lower than 0.3, it is considered that the two can be independently adjusted, and at this time, the scheduling sequence corresponding to the path has the smallest influence on the energy efficiency deviation. The preferred path is selected by path tracing, which provides a basis for subsequent evaluation of the energy efficiency deviation index.

[0109] When evaluating the energy efficiency deviation index, the time weight variance and the space weight extreme value of each leaf node in the scheduling decision tree are counted. The time weight variance reflects the fluctuation degree of time sensitivity in different scheduling schemes, and the greater the variance, the more unstable the scheduling scheme depends on the time factor; the space weight extreme value represents the maximum influence degree of pressure regulation in a specific area. The comprehensive standard deviation of all tree nodes is calculated, which reflects the overall parameter fluctuation level and embodies the stability of the scheduling scheme. The time deviation coefficient is obtained by taking the time weight variance of a single leaf node as the dividend, multiplying the time weight variance of the same node, and multiplying the comprehensive standard deviation; at the same time, the ratio of the space weight extreme value to the comprehensive standard deviation is calculated, and the geometric mean of the two is the energy efficiency deviation index of the node. The index quantifies the deviation degree of energy efficiency and the optimal state under different scheduling sequences, and the smaller the index value, the closer the scheduling scheme is to the optimal energy efficiency state.

[0110] When the mapping calibration module fits the optimal energy efficiency reference curve, the control record closest to the current energy efficiency deviation index in the historical database is called first. The historical database stores successful energy-saving control cases in the past, including the energy efficiency state before control, the control measures taken, and the energy efficiency improvement after control. The spatial distribution cosine similarity between the current energy efficiency deviation index and the historical record is calculated as the first mapping reference value. The cosine similarity measures the directional similarity of two vectors in space, and the closer the value is to 1, the more similar the spatial distribution form is.

[0111] The coincidence degree of the current energy efficiency deviation index and the peak point distribution of the historical control record is counted as the second mapping reference value. The peak point distribution coincidence degree focuses on the matching degree of key feature points (such as the maximum power point and the highest efficiency point) in the energy efficiency curve, including the matching degree of the time point and the amplitude size of the peak value. The reciprocal of the coincidence degree is taken as the second mapping reference value, because the higher the coincidence degree, the more similar the current state is to the historical successful case, and the greater the reference value is, so the reciprocal is taken and combined with the first mapping reference value.

[0112] Based on the vector synthesis result of the first mapping reference value and the second mapping reference value, an optimal energy efficiency benchmark curve that best fits the current working condition is matched from a preset energy efficiency curve library. The preset energy efficiency curve library contains theoretical optimal energy efficiency curves under different working conditions, which are fitted by the design parameters of the water pump, the fluid mechanics model and historical operation data. The vector synthesis result comprehensively considers the similarity of the spatial distribution form and the peak value feature, and ensures that the matched benchmark curve can accurately reflect the energy efficiency optimization target under the current working condition.

[0113] When generating the energy efficiency deviation gradient atlas, the current running energy efficiency value is compared with the optimal energy efficiency benchmark curve. For each running mode, the difference between the energy efficiency value and the corresponding point on the benchmark curve is calculated to obtain the energy efficiency deviation value. According to the size and direction of the energy efficiency deviation value, the atlas is divided into a positive adjustment region and a reverse adjustment region. The positive adjustment region indicates that the current energy efficiency can be developed in a more optimal direction through parameter adjustment, for example, reducing the frequency or reducing the valve opening can reduce energy consumption; the reverse adjustment region indicates that the energy efficiency deviation needs to be corrected through compensation measures, for example, increasing the power to maintain the head or increasing the flow.

[0114] In the gradient atlas, the color depth or gray value of each point represents the size of the energy efficiency deviation, and the deeper the color, the greater the deviation and the greater the adjustment amplitude required; the arrow direction represents the adjustment direction, pointing to the direction of energy efficiency improvement. Through this visualization method, the energy efficiency optimization space and adjustment path of each running mode are intuitively displayed, providing clear control targets for the control execution module.

[0115] Embodiment 5:

[0116] This embodiment relates to the specific implementation of the control execution module. The core function of the control execution module is to analyze the energy efficiency deviation gradient atlas, convert it into an executable water pump energy-saving control instruction set, and drive the execution mechanism to complete the adjustment action, thereby realizing real-time optimization of the water pump running state. This process includes atlas region division, adjustment parameter extraction, instruction set generation, execution mechanism driving and closed-loop feedback, etc., and realizes the physical landing of the energy-saving strategy through multi-dimensional data processing and control technology.

[0117] The control execution module first analyzes the energy efficiency deviation gradient map, and divides the map into positive adjustment region and reverse adjustment region according to the energy efficiency deviation angle of each mode. The energy efficiency deviation angle refers to the angle between the current energy efficiency state vector and the optimal energy efficiency direction vector. When the angle is less than 90 degrees, it is a positive deviation, indicating that the energy efficiency can be improved through conventional adjustment. When the angle is greater than 90 degrees, it is a reverse deviation, indicating that compensation measures need to be taken to correct the energy efficiency decline. For example, when the water pump is running in the low efficiency area but no fault occurs, the energy efficiency deviation angle is usually positive, and the energy efficiency can be optimized by reducing the speed or adjusting the valve opening. When the water pump energy efficiency decreases due to impeller wear, power input may need to be increased to maintain the head, and the energy efficiency deviation angle is reverse.

[0118] The adjustment step of the energy efficiency deviation is extracted from the positive adjustment region. The adjustment step represents the change amount of each parameter adjustment. The determination of the adjustment step needs to consider the system response speed and stability. For a fast responding system (such as a centrifugal pump variable frequency speed regulation system), the step can be appropriately increased to speed up the adjustment; for a large inertia system (such as a volumetric pump or a long pipeline system), the step needs to be reduced to avoid excessive adjustment leading to system oscillation. For example, in the centrifugal pump variable frequency speed regulation scene, the initial adjustment step can be set to 1%-2% of the rated frequency, and the subsequent dynamic adjustment is based on the system response.

[0119] The compensation amplitude of the energy efficiency deviation is extracted from the reverse adjustment region. The compensation amplitude represents the additional energy or pressure input required to correct the energy efficiency deviation. The calculation of the compensation amplitude is based on the reverse deviation degree in the energy efficiency deviation gradient map. The greater the deviation degree, the greater the compensation amplitude. For example, when the head decreases due to pipeline blockage, the flow needs to be maintained by increasing the motor power or adjusting the bypass valve opening, and the compensation amplitude is the additional power or pressure increment required to restore the target flow.

[0120] The extracted adjustment step and compensation amplitude need to be dynamically matched with the operating mode working condition weight. The working condition weight reflects the importance of the mode in the current operating state, which can be determined according to the running time proportion, load size, safety level, etc. For example, in a continuous water supply system, the operating mode weight of the main pump is higher than that of the standby pump; in a chemical process, the working condition weight of the key process section of the water pump is greater than that of the auxiliary process section. Through weighted calculation, the core execution parameters of the water pump energy saving control instruction set are generated, including variable frequency regulator frequency adjustment value, valve opening adjustment value, impeller speed adjustment value, etc.

[0121] After the generation of the instruction set, it is transmitted to the actuator of the water pump control system through an industrial bus or a field bus. Industrial buses (such as Modbus and Profibus) are suitable for long-distance communication and have high anti-interference ability; field buses (such as CANopen and DeviceNet) are suitable for short-distance device connection and have fast response speed. Data encryption and verification are required during transmission to ensure the integrity and accuracy of the instructions, such as using CRC verification or digital signature technology to verify whether errors or tampering have occurred during data transmission.

[0122] After receiving the instructions, the actuator adjusts the parameters according to the instructions. For frequency converters, the motor speed is adjusted by changing the output frequency to achieve flow and head regulation; for electric valves, the valve opening and closing are controlled by the motor to adjust the pipeline resistance; for adjustable speed impellers, the efficiency curve of the pump is optimized by changing the impeller speed or blade angle through a servo motor. The action of the actuator must follow the pre-set safety constraints, such as the motor frequency must not be lower than the minimum allowed frequency (to prevent motor overheating) and the valve opening must not exceed the maximum safe opening (to avoid pipeline overpressure).

[0123] During the adjustment process, the control execution module continuously monitors key operating parameters to form a closed-loop feedback mechanism. The monitoring parameters include flow, head, power, temperature, etc., which are compared with the energy efficiency benchmark curve to evaluate the adjustment effect. If there is a deviation between the actual operating parameters and the expected target, the system automatically calculates the deviation value and generates a correction instruction to adjust the actuator parameters again until the expected energy efficiency target is reached or the system enters a stable operating state. For example, during frequency control, if the flow adjustment does not reach the expected value, the system will adjust the frequency adjustment step according to the flow deviation value for secondary adjustment.

[0124] The control execution module also has the ability to handle exceptions and recover from faults. When abnormal parameters (such as motor overload and high bearing temperature) are detected, the system immediately triggers the protection mechanism, suspends energy-saving adjustment and takes emergency measures such as reducing load, starting backup devices or sending alarm signals. After troubleshooting, the system automatically restores the energy-saving control mode and generates a new instruction set based on the current working conditions to ensure that the water pump operates safely and efficiently.

[0125] For multiple pump parallel or series systems, the control execution module needs to coordinate the operating state of each pump. According to the system total demand and single pump performance curve, the optimal pump combination scheme and load distribution strategy are calculated to maximize the overall energy efficiency. For example, in a multiple pump parallel system, the number of operating pumps and load distribution are dynamically adjusted according to the flow demand, high-efficiency pumps are preferentially enabled, and pumps are prevented from operating in the low-efficiency zone.

[0126] The adjustment period of the control execution module is dynamically adjusted according to the system characteristics and the frequency of working condition changes. For systems with frequent working condition changes (such as municipal water supply systems), the adjustment period is short (such as 1-5 minutes) to quickly respond to demand changes; for systems with stable working conditions (such as industrial circulating water systems), the adjustment period can be appropriately extended (such as 10-30 minutes) to reduce the number of adjustments and reduce equipment wear. The dynamic adjustment of the adjustment period is achieved by analyzing the trend and frequency characteristics of historical working condition data, ensuring that the system can maintain efficient operation under different working conditions.

[0127] The control execution module converts energy efficiency optimization strategies into actual device control actions through graph analysis, parameter extraction, instruction generation and execution, closed-loop feedback, and exception handling. The design of this module fully considers the real-time, reliability, and safety of industrial control systems, ensuring that the water pump energy-saving control system can operate stably under complex working conditions and achieve continuous energy efficiency improvement.

[0128] It should be noted that, in this text, 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 actual such relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed or inherent to such a process, method, article or device.

[0129] Although embodiments of the present application have been shown and described, it will be understood by those of ordinary skill 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 present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A water pump energy-saving control system based on an intelligent data model, characterized in that: include: The data sensing module is used to collect flow, head and power data during the operation of the water pump and set a dynamic monitoring window that matches the operating conditions of the water pump; The modal analysis module is used to divide the dynamic monitoring window into multiple operating modes, perform mode decoupling on the power data of each operating mode, and generate the energy efficiency characteristic matrix corresponding to the operating mode; The rule building module is used to extract the core energy efficiency factors from the energy efficiency feature matrix, build the energy-saving control strategy associated with the operating mode, and extract the equipment adjustment parameters corresponding to the strategy; The priority scheduling module is used to analyze the control priority matrix in the equipment adjustment parameters, hierarchically schedule the core energy efficiency factors according to the priority matrix, and evaluate the energy efficiency deviation index of each operating mode under different scheduling sequences; The mapping calibration module is used to fit the optimal energy efficiency benchmark curve according to the energy efficiency deviation index, and generate an energy efficiency deviation gradient map by matching the current operating energy efficiency value with the optimal energy efficiency benchmark curve; The control execution module is used to analyze the energy efficiency deviation gradient map and convert the energy efficiency deviation gradient map into a water pump energy-saving control instruction set based on the energy efficiency deviation direction of the operating mode.

2. A water pump energy-saving control system based on an intelligent data model according to claim 1, characterized in that: The implementation of the modal analysis module includes: establishing an operating condition feature library corresponding to the operating mode, the operating condition feature library including an operating condition parameter matrix mapped with flow rate, head and power data; The operating condition parameter matrix is ​​matched with similar operating conditions, and the operating condition parameter matrix is ​​divided into operating condition clusters according to the matching results; the distribution feature points of the power data are extracted from the operating condition clusters, and the distribution feature points are set as the operating mode.

3. The water pump energy-saving control system based on the intelligent data model according to claim 2 is characterized in that: The operating condition clusters that divide the operating condition parameter matrix also include: According to the equipment model and operation stage in the operating parameter matrix, pipeline pressure, liquid density and environmental vibration parameters are extracted, and the operating condition feature identifier is generated based on the above parameters; The working condition feature identifiers are associated with the working condition parameter matrix. By calculating the working condition correlation between the feature identifiers, the working condition parameter matrix with a correlation exceeding a preset working condition threshold is screened to form a working condition cluster.

4. The water pump energy-saving control system based on the intelligent data model according to claim 1 is characterized in that: The implementation methods for generating the energy efficiency characteristic matrix corresponding to the operating mode include: For each operating mode, according to the time series distribution of the mode in the dynamic monitoring window, the flow deviation data of the mode within the set period is obtained, and the flow deviation coefficient of the mode is calculated; When the flow deviation coefficient exceeds the first deviation threshold, the mode is marked as an abnormal mode, and its power data is extracted to construct an energy efficiency feature matrix; when the flow deviation coefficient is lower than the first deviation threshold, the mode is marked as a stable mode, and the power data of the adjacent modes of the mode are dimensionally aggregated, and the aggregated data are reorganized into an energy efficiency feature matrix.

5. The water pump energy-saving control system based on the intelligent data model according to claim 1 is characterized in that: The implementation of the rule building module includes: Separate the mechanical loss ratio, ineffective cycle ratio, and pressure fluctuation parameters from the energy efficiency characteristic matrix, and generate an energy-saving control strategy for the operating mode based on these parameters; If the number of operating modes covered by the current energy-saving control strategy is less than the preset mode threshold, the energy efficiency feature matrix of the adjacent operating modes is traversed, and the energy efficiency factors not included in the strategies of the adjacent modes are added to the current strategy.

6. The water pump energy-saving control system based on the intelligent data model according to claim 1 is characterized in that: The implementation of the priority scheduling module includes: obtaining the time weight of frequency regulation and the space weight of pressure regulation in the control priority matrix; A scheduling decision tree associated with time weight and space weight is constructed, and the energy efficiency deviation index under different scheduling sequences is determined according to the path score of each branch in the decision tree.

7. The water pump energy-saving control system based on the intelligent data model according to claim 6 is characterized in that: Building a scheduling decision tree also includes: Identify the periodic characteristics of the time weight. If the current periodic characteristics completely match the preset operating cycle, then set the time weight as the root node of the scheduling decision tree. Calculate the coupling degree between time weight and space weight, and generate the intermediate nodes and leaf nodes of the scheduling decision tree in descending order of coupling degree; The path of the leaf node is backtracked. When the coupling degree of the leaf node is lower than the preset coupling threshold, it is marked as the preferred path of the scheduling decision tree.

8. The water pump energy-saving control system based on the intelligent data model according to claim 7 is characterized in that: The implementation methods for evaluating the energy efficiency deviation index include: Count the time weight variance and spatial weight extreme value of each leaf node in the scheduling decision tree, and calculate the comprehensive standard deviation of all tree nodes; The time deviation coefficient is obtained by dividing the time weight variance of a single leaf node by the time weight variance of the same-level node and multiplying it by the comprehensive standard deviation. At the same time, the ratio of the spatial weight extreme value to the comprehensive standard deviation is calculated, and the geometric mean of the two is taken as the energy efficiency deviation index of the node.

9. The water pump energy-saving control system based on intelligent data model according to claim 1, characterized in that: The implementation methods of fitting the optimal energy efficiency benchmark curve include: Call the control record in the historical database that is closest to the current energy efficiency deviation index, calculate the cosine similarity of the two in spatial distribution, and use it as the first mapping reference value; Count the overlap between the current energy efficiency deviation index and the historical control records on the peak point distribution, and use the inverse of the overlap as the second mapping reference value; Based on the vector synthesis result of the first mapping reference value and the second mapping reference value, the optimal energy efficiency reference curve in the preset energy efficiency curve library is matched.

10. The water pump energy-saving control system based on intelligent data model according to claim 1, characterized in that: The implementation method of the control execution module includes: dividing the positive regulation area and the negative regulation area according to the energy efficiency deviation angle of each mode in the energy efficiency deviation gradient map; The adjustment step of the energy efficiency deviation in the positive adjustment area and the compensation amplitude of the energy efficiency deviation in the reverse adjustment area are extracted, and the two are dynamically matched according to the working condition weight of the operating mode to generate the core execution parameters of the water pump energy-saving control instruction set.

Images