A pump station energy efficiency diagnosis system and diagnosis method based on multi-curve collaborative analysis

Abstract

This invention discloses a pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis, including a data acquisition module, a data preprocessing module, a data analysis module, a data correction and calculation module, and a decision module. The data acquisition module collects data; the data preprocessing module preprocesses the pump station data; the data analysis module identifies outliers; the data correction and calculation module corrects outliers and writes the required calculation values ​​into formula parameters for calculation; and the decision module evaluates and annotates the data, providing solutions and data early warnings. This invention is applied to energy consumption analysis, efficiency assessment, automatic generation of optimal scheduling schemes, and early warning of operational anomalies in large-scale water purification and sewage pump stations, providing decision support for manual control and inspection.

Description

Technical Field

[0001] This invention belongs to the field of pump station energy efficiency diagnosis technology, specifically, it relates to a pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis. Background Technology

[0002] In the operation and management of modern pumping station systems, energy efficiency diagnosis and optimized scheduling are core components to ensure efficient and stable system operation. However, as the scale of pumping stations continues to expand and operating conditions become increasingly complex, traditional pumping station energy efficiency diagnosis methods have gradually revealed many limitations and are unable to meet actual production needs.

[0003] As a core energy-consuming unit in the water industry, the operating efficiency of pumping stations directly affects energy consumption and carbon emissions. Statistics show that pumping station systems account for 40%-60% of the total electricity consumption of water plants (including advanced treatment facilities), but their actual operating efficiency is generally lower than the design value, indicating significant potential for energy conservation.

[0004] The main problems in the current energy efficiency management of pumping stations are as follows:

[0005] 1. Outdated design methods: Traditional pump station design is based solely on the most unfavorable operating conditions (maximum flow rate, highest head), while in actual operation, the unit operates in non-peak conditions for most of the time, causing it to deviate from the high-efficiency zone for a long period of time.

[0006] 2. Inefficient regulation method: When the flow rate changes, most pumping stations rely on valve throttling or bypass backflow regulation, resulting in additional head loss; very few are equipped with frequency converters, but the combined operating conditions are different, and the frequency converter itself increases additional power consumption.

[0007] 3. Lagging energy efficiency assessment: Existing monitoring systems only collect basic parameters such as current and voltage, lacking real-time correlation analysis of pump station system head-flow-efficiency, and cannot determine in real time whether it is operating at the optimal energy efficiency point. This leads to higher average annual electricity consumption per unit.

[0008] 4. Inconsistent operation and maintenance standards: Different operators have significantly different judgments on the operating status, and scheduling decisions rely on subjective experience, making it difficult to achieve refined energy efficiency management.

[0009] The main bottleneck of existing technology lies in:

[0010] 1. Limited data dimensions: Current pumping stations lack a complete database of energy efficiency characteristic curves, making dynamic optimization impossible;

[0011] 2. Insufficient analytical methods: Traditional single-curve analysis methods are inefficient. Even with complete data, it takes a long time to complete a comprehensive energy efficiency assessment, which cannot meet the needs of real-time scheduling.

[0012] 3. Low level of automation: Energy efficiency diagnosis relies on manual calculation and experience judgment, which is difficult to adapt to complex and ever-changing operating conditions. Summary of the Invention

[0013] To address the shortcomings of existing technologies, the present invention aims to provide a pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis.

[0014] To achieve the aforementioned objectives, the technical solution adopted by this invention includes:

[0015] A pump station energy efficiency diagnostic system based on multi-curve collaborative analysis includes:

[0016] The data acquisition module is used to import offline and online pump station data, transmit the pump station data to the data preprocessing module, and simultaneously input the single pump curve and pipeline characteristic curve data of the water pump characteristic curve.

[0017] The data preprocessing module preprocesses and cleans the collected pump station data and transmits the pump station data to the data analysis module.

[0018] The data analysis module identifies abnormal values ​​in the pump station data and highlights them, then transmits the data to the data correction and calculation module.

[0019] The data correction and calculation module corrects outlier data at the pumping station, calculates the pumping station data according to the corresponding formula, and finally transmits the data to the decision module. The outlier data correction and calculation includes:

[0020] By establishing a state determination model based on fuzzy logic, the operating conditions of water pumps and the combination of parallel pumps can be identified.

[0021] A smart reconstruction system for pump characteristic curves based on hybrid modeling generates single-pump curve equations and optimizes data.

[0022] Based on the intelligent generation system for the dynamic characteristics of parallel pump sets, the system generates characteristic curves of parallel pumps and a dynamic calculation model of parallel operators. Combined with coefficient matrix solving and three-dimensional parameterized display, it realizes the rapid calculation and visualization of characteristic curves of parallel pumps.

[0023] The hydraulic system characteristic modeling system based on multi-scale fluid dynamics generates system characteristic curves. Through multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation and adaptive polynomial fitting, the dynamic and accurate construction of pump station system characteristic curves is realized.

[0024] The decision-making module evaluates, analyzes, and annotates the data, displays it in chart form on the page, proposes corresponding solutions and early warning prompts, and feeds back the information to pump station staff and central control personnel for human-computer interaction.

[0025] Furthermore, the data analysis module is used to identify and highlight outliers in the pump station data by setting upper and lower limits.

[0026] Upper and lower limit settings refer to setting upper and lower limits for data such as liquid level, flow rate, pressure, and power consumption that are directly measured by instruments for monitoring the operation of pump stations, combined with design documents and operational experience. Data exceeding the upper limit and falling below the lower limit are displayed in red.

[0027] A pump station energy efficiency diagnosis method based on multi-curve collaborative analysis includes the following diagnostic steps:

[0028] S1. Import offline pump station data and online pump station data, and transfer the pump station data to the data preprocessing module. At the same time, input the single pump curve and pipeline characteristic curve data of the water pump characteristic curve.

[0029] S2. Preprocess the collected pump station data;

[0030] S3. Identify abnormal values ​​in pump station data and highlight them.

[0031] S4. Correct the outlier data of the pumping station and calculate the pumping station data according to the corresponding formula. The correction of outlier data and calculation of pumping station data according to the corresponding formula includes:

[0032] a1. By establishing a state determination model based on fuzzy logic, the operating conditions of water pumps and the combination of parallel pumps can be identified.

[0033] a2. A smart reconstruction system for pump characteristic curves based on hybrid modeling generates single-pump curve equations and optimizes data.

[0034] a3. Based on the intelligent generation system for the dynamic characteristics of parallel pump sets, the characteristic curves of parallel pumps are generated, and the dynamic calculation model of parallel operators is combined with the solution of coefficient matrix and three-dimensional parameterized display to realize the rapid calculation and visualization of the characteristic curves of parallel pumps.

[0035] a4. The hydraulic system characteristic modeling system based on multi-scale fluid dynamics generates system characteristic curves. Through multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation and adaptive polynomial fitting, the dynamic and accurate construction of pump station system characteristic curves is realized.

[0036] S5. After evaluating, analyzing, and annotating the data, display it on the page in the form of charts, propose corresponding solutions and early warning prompts, and provide feedback to pump station staff and central control personnel for human-computer interaction.

[0037] This invention supports both offline and online data acquisition methods. It can perform offline analysis by importing historical data or perform online monitoring by directly connecting to the production database through an interface. This flexibility enables the system to adapt to the energy efficiency diagnosis needs of pump stations in different scenarios.

[0038] In this invention, the data preprocessing module cleans and processes the collected pump station data to ensure the data quality for subsequent analysis. This step is crucial for improving the accuracy of diagnostic results because it eliminates noise and outliers in the data, making the analysis results more reliable.

[0039] In this invention, the data analysis module can identify outliers in the pump station data and present them to the user intuitively by highlighting them in red. This visualization method allows users to quickly locate the problem, providing strong support for subsequent troubleshooting and energy efficiency optimization.

[0040] In this invention, the data correction and calculation module not only corrects outliers but also calculates pump station data according to corresponding formulas. Based on hybrid modeling, intelligent generation of dynamic characteristics of parallel pump groups, and multi-scale fluid dynamics technologies, this module can intelligently reconstruct pump characteristic curve equations and generate parallel pump curves and hydraulic system characteristic models. This intelligent processing method significantly improves the accuracy and efficiency of diagnosis.

[0041] In this invention, by simultaneously considering multiple curve information such as pump characteristic curves, pipeline characteristic curves, and parallel pump curves, the system can timely and comprehensively evaluate the energy efficiency of the pumping station and provide more accurate diagnostic results and optimization suggestions, thus having higher accuracy and practicality.

[0042] In this invention, energy consumption analysis, efficiency evaluation, automatic generation of optimal scheduling schemes, and early warning of operational anomalies are performed within the pumping station, providing decision support for manual control and manual inspection.

[0043] Furthermore, step a1, by establishing a state determination model based on fuzzy logic, enables the identification of pump operating conditions and parallel pump combinations, including:

[0044] a1.1, Feature parameter normalization processing,

[0045] Define a multidimensional feature vector:

[0046] ;

[0047] in: Multidimensional feature vectors

[0048] , : Normalized runtime value, T opThe operating time of each pump is measured in minutes.

[0049] , Normalized pressure value Pressure at the outlet of each pump, P out -Pressure at the pump station outlet

[0050] , Energy efficiency ratio, P i,act Hourly power consumption per pump, P rated : The rated power of the corresponding water pump, f i,act The operating frequency of the water pump.

[0051] Flow ratio, Q act Q: Hourly water output of each pump rated : The rated flow rate of the corresponding water pump;

[0052] a1.2, State Decision Matrix

[0053] Establish membership function:

[0054]

[0055] Where, μ i Membership degree of the i-th feature parameter - calculated in real time, a i : Membership degree of the i-th feature parameter - historical data statistics and device attributes, b i Critical upper limit of the i-th type of operating condition - historical data statistics and equipment attributes. : The i-th multidimensional feature vector;

[0056] a1.3 Decision Fusion Algorithm

[0057] A weighted voting mechanism will be adopted:

[0058] ;

[0059] Weighting coefficient Optimized using the backpropagation algorithm, if a pumping station is not equipped with the corresponding parameter measurement, such as a flow meter for a single water pump, then... The weight of x4 is 0;

[0060] a1.4 Anomaly Detection Module

[0061] Constructing a residual detection model:

[0062] ;

[0063] when Triggering a Level 3 alarm:

[0064] Yellow Alert

[0065] Orange Alert

[0066] Yellow Alert ,in, Threshold : Residual.

[0067] Step a1: This technology integrates normalized multidimensional features (running time, pressure, energy consumption ratio, flow ratio) through a fuzzy logic model, and uses membership functions and weighted voting mechanisms to accurately identify the pump's operating state (such as the transition state "S"), thus solving the problem of misjudgment caused by noise or fluctuations in the traditional threshold method.

[0068] Furthermore, step a2 generates a single-pump curve equation based on the intelligent reconstruction system for pump characteristic curves using hybrid modeling, and the optimized data includes:

[0069] 2.1 Multi-source data fusion modeling framework,

[0070] Establish a three-dimensional feature space:

[0071] ;

[0072] Where: D is the three-dimensional feature space of each pump in the pumping station, Q is the flow column vector, H is the head column vector, and η is the efficiency column vector;

[0073] a 2.2 Physical constraint model,

[0074] Using regularized constrained least squares method:

[0075] ;

[0076] Where the basis function matrix is:

[0077] ,

[0078] Constraints:

[0079] , ;

[0080] Where, Q: Actual flow rate test value - manufacturer's allowable operating range, H: Actual head value - manufacturer's allowable operating range, a: Quadratic coefficient of the QH characteristic curve - ensures the curve opens downwards, b: Linear coefficient of the QH characteristic curve, c: Constant term of the QH characteristic curve - shut-off head at zero flow. Regularization coefficient - to prevent overfitting; a head-flow quadratic equation is established for each pump using the method in step a.2.2;

[0081] 2.3 Intelligent Modeling:

[0082] Building a deep hybrid network architecture:

[0083] ;

[0084] Where H: Real-time measured head value - core operating condition output parameter, Q: Real-time flow test value - core operating condition output parameter, f VFD : Frequency ratio of inverter operating frequency to power frequency operating frequency - key control variable of variable frequency pump, t: equipment operating timestamp - reflects equipment aging trend, W1: first layer weight evidence - feature space change, b1: first layer bias vector - nonlinear offset compensation, W2: second layer weight evidence - feature weighted aggregation, b2: output layer bias - reference head calibration.

[0085] Model selection metrics

[0086] ;

[0087] Among them, R 2 Prediction accuracy, AIC: Model complexity penalty, Inference Time: Real-time performance metric, α: Fit coefficient, β: Generalization ability coefficient, γ: Efficiency coefficient.

[0088] Step a2 employs a hybrid modeling approach, combining a physical constraint model (using regularized least squares fitting for small samples to ensure the curve conforms to the head-flow rate decreasing trend) and an intelligent deep learning model (using neural networks to fuse flow rate, frequency, and other data for large datasets to improve nonlinear fitting capabilities) to accurately reconstruct the single pump characteristic curve.

[0089] Its advantages are: small data does not get out of control (physical constraints ensure rationality) and big data is more accurate (deep learning adapts to complex working conditions), which solves the problem that traditional methods have large errors when there is insufficient data or are difficult to model when there is a lot of data, and provides a reliable theoretical curve basis for the optimization of parallel pump systems.

[0090] Furthermore, step a3 generates parallel pump characteristic curves and a parallel operator dynamic calculation model based on the intelligent generation system for the dynamic characteristics of parallel pump groups. Combined with coefficient matrix solving and three-dimensional parameterized display, it realizes the rapid calculation and visualization of parallel pump characteristic curves.

[0091] a 3.1 Parallel Pump Characteristic Calculation Model

[0092] Define parallel operators:

[0093] ;

[0094] in:

[0095] Let i be the characteristic function of the i-th pump.

[0096] The total flow function of a parallel system, the sum of the flow rates of each operating pump at head H.

[0097] This represents the total number of pumps allowed to operate in parallel within the pump system.

[0098] This is a pump status indication function;

[0099] a 3.2 Parallel Pump Characteristic Calculation Model

[0100] Calculating the characteristic curves of parallel pumps requires determining the output of each pump based on its head, then superimposing the flow rates of the pumps participating in the parallel operation, and finally solving using a coefficient matrix.

[0101] ,

[0102] in: basis functions ;

[0103] a 3.3 Parallel Pump Characteristic Calculation Model

[0104] Achieve 3D parametric display:

[0105]

[0106] in, System head; Q: Total flow rate, calculated from step a.3.2. Overall efficiency; MLP: Multilayer Perceptron. The relative head value is used to determine whether the head exceeds the limit. Efficiency standardization value, used to assess energy efficiency health.

[0107] Step a3, through a dynamic calculation model of parallel operators, combined with coefficient matrix solving and three-dimensional parameterized display, achieves rapid and accurate calculation and visualization of the characteristic curves of parallel pumps. Its advantages are:

[0108] 1) Status awareness (determining the operating status of the pump group based on fuzzy logic to ensure the reliability of the computational basis);

[0109] 2) Efficient solution (coefficient matrix decomposition of the characteristic equations of each pump, parallel calculation of the speed increase);

[0110] 3) Intelligent interaction (MLP model 3D visualization supports real-time operating condition simulation and abnormal alarm).

[0111] This solves the problems of complex calculations and inability to dynamically respond to changes in operating conditions in traditional methods, providing real-time theoretical support for the optimized scheduling of pump stations.

[0112] Furthermore, step a4 generates system characteristic curves based on the hydraulic system characteristic modeling system of multi-scale fluid dynamics. Through multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation and adaptive polynomial fitting, the dynamic and accurate construction of the pump station system characteristic curves is realized.

[0113] a4.1 Intelligent head calculation model,

[0114] Establish a three-dimensional spatial head matrix:

[0115] ;

[0116] Wherein: H static : Static range, P dis Water pump outlet pressure, Z dis : Elevation of water pressure gauge / sensor, P suct Water pump inlet pressure, Z suct : Inlet pressure gauge / sensor elevation, This is a time-varying water level correction term. The vertical height difference between the pressure measuring point on the main outlet pipe of the pumping station and the bottom of the suction well;

[0117] a4.2 Hybrid Loss Calculation Engine

[0118] The solution is obtained by combining the Darcy-Weisbach and Navier-Stokes methods:

[0119]

[0120] in: Head loss / dynamic head : , , ;

[0121] a4.3 Adaptive Characteristic Curve Modeling

[0122] Variable-order polynomial model

[0123] ;

[0124] Where Q: flow rate, H(Q): system characteristic curve, c k: Polynomial coefficients, k: Polynomial order;

[0125] Order selection criteria:

[0126] ,

[0127] Among them, K * Optimal order, candidate order values ​​for K-polynomial fitting. Regularization coefficient, Norm penalty term;

[0128] Online parameter recognition system

[0129] Establish a recursive least squares algorithm:

[0130] ,

[0131] : Parameter vector Measured head Kalman gain, : Feature vector;

[0132] in:

[0133] , ,

[0134] Jingyangcheng Linear drag coefficient, Secondary drag coefficient,

[0135] By using multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation, and adaptive polynomial fitting, the dynamic and accurate construction of the characteristic curves of the pumping station system was achieved.

[0136] This solves the problem that traditional methods ignore local resistance and cannot adapt to time-varying operating conditions, providing real-time dynamics for pump station energy efficiency assessment.

[0137] Furthermore, step S5 uses the intelligent diagnostic system for the health status of the pump based on the flow-head characteristics and the historical case matching and optimization system for energy efficiency special diagnosis to feed back the monitored results to the pump station staff and send early warning information to the computer terminal to remind the pump station staff to check and repair abnormal problems in a timely manner.

[0138] Furthermore, the intelligent health status diagnosis system for water pumps based on flow-head characteristics includes the following steps:

[0139] S5.1 Flow Condition Zoning and Wear Feature Mapping Model

[0140] S5.11, Establish the flow-wear characteristic matrix:

[0141]

[0142] S5.12, Employing a time-intensity dual-weighted model:

[0143] ;

[0144] ,

[0145] in:

[0146] The cumulative amount of wear of type k is dimensionless and calculated by real-time integration. The strength coefficient of the i-th type of wear, flow; For the flow range of the kth type of wear, Temperature correction factor (when hour , (Refer to the equipment user manual provided by the supplier for alarm temperature)

[0147] Initial time;

[0148] S5.2 Real-time Health Status Assessment

[0149] S5.21 3D Special Diagnosis Visualization

[0150] Establish the feature space:

[0151] ;

[0152] Color mapping rules:

[0153] green Health status

[0154] Yellow status Pay attention to the state.

[0155] red Dangerous state

[0156] S5.22, Identification of Dominant Wear Factors

[0157] Employing a fuzzy inference system:

[0158]

[0159] in, The cumulative duration of the k-th type of wear. Total effective runtime Type k wear intensity;

[0160] The system outputs the top three wear types and their confidence levels. This technology achieves intelligent diagnosis and prediction of pump health status through a flow-wear feature mapping model and a time-intensity weighted assessment.

[0161] Furthermore, the matching optimization steps of the historical case matching optimization system based on energy efficiency characteristics are as follows:

[0162] S5.3, Construct a four-dimensional energy efficiency characteristic space:

[0163] ;

[0164] in:

[0165] Electricity consumption per unit (kWh / m³) 3 )

[0166] Comprehensive unit power consumption (kWh / km) 3 .MPa)

[0167] Increase unit consumption (kWh / mH2O)

[0168] :Number of runs in history;

[0169] S5.4. Based on the four-dimensional energy efficiency feature space in S5.3, improve the similarity measurement algorithm.

[0170] Define a weighted similarity function:

[0171] ;

[0172] Feature dimensions and weight allocation:

[0173] Electricity consumption similarity (weight) =0.4)

[0174] ;

[0175] Similarity of overall unit power consumption ( =0.3)

[0176] ;

[0177] Improve unit consumption similarity ( =0.2)

[0178] ;

[0179] Reliability of running frequency ( =0.1)

[0180] ;

[0181] Output: The top 3 optimized combinations based on overall ranking.

[0182] The technology achieves intelligent screening of historical high-efficiency operating conditions by constructing a four-dimensional energy efficiency feature space (electricity consumption per unit, comprehensive electricity consumption, improved energy consumption per unit, and historical frequency) and improving the similarity algorithm (weighted multi-index matching).

[0183] Compared with the prior art, the advantages of the present invention include:

[0184] (1) Efficient diagnosis: Develop a dynamic curve fitting algorithm to generate equivalent characteristic curves and pipeline resistance curves of parallel pump groups every hour;

[0185] (2) Precise optimization: Identify the best operating range and recommend the optimal unit combination to greatly improve the annual operating efficiency of the pumping station;

[0186] (3) Intelligent early warning: Real-time monitoring of energy efficiency anomalies, early detection of equipment deterioration trends, and reduction of unplanned downtime.

[0187] (4) The present invention provides a pump station energy efficiency diagnosis system and diagnosis method based on multi-curve collaborative analysis, which can automatically diagnose pump station energy efficiency abnormalities, solve the problem that human experience cannot accurately judge the energy efficiency status in a timely manner, automatically identify inefficient equipment, and provide the optimal scheduling scheme in real time; reduce energy consumption, reduce carbon emissions, improve operating efficiency and extend equipment service life. Attached Figure Description

[0188] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0189] Figure 1 This is a schematic diagram of the system flow of a pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis in this invention.

[0190] Figure 2 This is a schematic diagram of a pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis in this invention;

[0191] Figure 3 This is a diagram showing state S4 of a pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis in this invention.

[0192] Figure 4 This is a diagram showing the second state of step S4 in the pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis in this invention.

[0193] Figure 5 This is a diagram showing the three states of step S4 in the system step S4 of the pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis in this invention.

[0194] Figure 6 This is a diagram showing the four states of step S4 in the system step S4 of the pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis in this invention. Detailed Implementation

[0195] In view of the shortcomings of the prior art, the inventors of this invention, through long-term research and extensive practice, have proposed the technical solution of this invention. The technical solution, its implementation process, and principles will be further explained below with reference to the accompanying drawings and specific implementation examples in the embodiments of this application.

[0196] It should be noted that the embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. The described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, the present invention covers any substitutions, modifications, equivalent methods and solutions made on the spirit, principles and scope of the present invention as defined by the claims. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0197] In the description of this application, the terms "first," "second," "third," and similar words do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, the terms "a" or "one," and similar words, do not indicate a quantity limitation, but rather indicate the presence of at least one. The terms "comprising" or "including," and similar words, mean that the elements or objects preceding "comprising" or "including" encompass the elements or objects listed following "comprising" or "including," and their equivalents, but do not exclude other elements or objects. The terms "connected" or "linked," and similar words, are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect.

[0198] In the description of this application, the terms "center," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used solely for the convenience of describing this application and for simplification, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, when using positional terms such as "both sides," "outer side," and "upper and lower," it should be understood that they are used only for ease of understanding and description, taking into account that the structure may be oriented to other positions.

[0199] In the description of this application, unless otherwise expressly specified and limited, the technical or scientific terms used shall have the ordinary meaning understood by a person with ordinary skills in the art to which this application pertains. Terms such as “installation,” “connection,” and “joining” shall be interpreted broadly, for example, as fixed connection, detachable connection, mating connection, or integral connection. For a person skilled in the art, the specific meaning of the above terms in this application can be understood according to the specific circumstances.

[0200] The present invention aims to introduce and explain the structural composition and the coordination relationship between the components of a pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis. Unless otherwise specified, the dimensions, materials and manufacturing processes of the components in the pump station energy efficiency diagnosis system and method based on multi-curve collaborative analysis in the present invention can be selected according to specific circumstances, and no special limitations or explanations are made here.

[0201] Furthermore, to provide the public with a better understanding of the present invention, certain specific details are described in detail in the following description of the invention. However, those skilled in the art will fully understand the invention even without these detailed descriptions.

[0202] Please see Figures 1-6 A pump station energy efficiency diagnostic system based on multi-curve collaborative analysis includes:

[0203] The data acquisition module is used to import offline and online pump station data and transmit the pump station data to the data preprocessing module. It also records the single pump curve and pipeline characteristic curve data of the water pump characteristic curve. The data acquisition module monitors the pump station data by importing data or accessing the database in real time via an interface.

[0204] The data preprocessing module preprocesses and cleans the collected pump station data, mainly by filling in null values, converting character variables (actually time variables) to time format, converting character variables (actually numeric variables containing strings) to numeric variables, and transmitting the pump station data to the data analysis module.

[0205] The data analysis module identifies and highlights abnormal values ​​in the pump station data. It mainly uses upper and lower limit settings to judge abnormal values ​​in the pump station data and highlight them in red, and then transmits the data to the data correction and calculation module. The upper and lower limit settings refer to the upper and lower limits that can be set for the pump station operation monitoring data such as liquid level, flow rate, pressure, and power directly measured by instruments, combined with design documents and operational experience. Values ​​exceeding the upper limit and falling below the lower limit are displayed in red.

[0206] The data correction and calculation module corrects outlier data at the pumping station, calculates the pumping station data according to the corresponding formula, and finally transmits the data to the decision module. The outlier data correction and calculation includes:

[0207] By establishing a state determination model based on fuzzy logic, the operating conditions of water pumps and the combination of parallel pumps can be identified.

[0208] A smart reconstruction system for pump characteristic curves based on hybrid modeling generates single-pump curve equations and optimizes data.

[0209] Based on the intelligent generation system for the dynamic characteristics of parallel pump sets, the system generates characteristic curves of parallel pumps and a dynamic calculation model of parallel operators. Combined with coefficient matrix solving and three-dimensional parameterized display, it realizes the rapid calculation and visualization of characteristic curves of parallel pumps.

[0210] The hydraulic system characteristic modeling system based on multi-scale fluid dynamics generates system characteristic curves. Through multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation and adaptive polynomial fitting, the dynamic and accurate construction of pump station system characteristic curves is realized.

[0211] The decision-making module evaluates, analyzes, and annotates the data, displays it in chart form on the page, proposes corresponding solutions and early warning prompts, and feeds back the information to pump station staff and central control personnel for human-computer interaction.

[0212] A pump station energy efficiency diagnosis method based on multi-curve collaborative analysis includes the following diagnostic steps:

[0213] S1. Import offline pump station data and online pump station data, and transfer the pump station data to the data preprocessing module. At the same time, input the single pump curve and pipeline characteristic curve data of the water pump characteristic curve.

[0214] S2. Preprocess the collected pump station data, mainly by filling in empty values, converting character variables (actually time variables) to time format, and converting character variables (actually numeric variables containing strings) to numeric variables, and then transmit the preprocessed data to the data analysis module.

[0215] S3, Identifying and highlighting abnormal values ​​in pump station data, involves judging abnormal values ​​in pump station data based on upper and lower limit settings and highlighting them in red.

[0216] Upper and lower limit settings refer to setting upper and lower limits for data such as liquid level, flow rate, pressure, and power consumption that are directly measured by instruments for monitoring the operation of pump stations, combined with design documents and operational experience. Data exceeding the upper limit and falling below the lower limit are displayed in red.

[0217] S4. Correct the outlier data of the pumping station and calculate the pumping station data according to the corresponding formula. The correction of outlier data and calculation of pumping station data according to the corresponding formula includes:

[0218] Step a1, by establishing a state determination model based on fuzzy logic, identifies the operating conditions of water pumps and the combination of parallel pumps, including:

[0219] a1.1, Feature parameter normalization processing,

[0220] Define a multidimensional feature vector:

[0221] Define a multidimensional feature vector:

[0222] ;

[0223] in: Multidimensional feature vectors

[0224] , : Normalized runtime value, T op The operating time of each pump is measured in minutes.

[0225] , Normalized pressure value Pressure at the outlet of each pump, P out -Pressure at the pump station outlet

[0226] , Energy efficiency ratio, P i,act Hourly power consumption per pump, P rated : The rated power of the corresponding water pump, f i,act The operating frequency of the water pump.

[0227] Flow ratio, Qact Q: Hourly water output of each pump rated : The rated flow rate of the corresponding water pump;

[0228] a1.2, State Decision Matrix

[0229] Establish membership function:

[0230]

[0231] Where, μ i Membership degree of the i-th feature parameter - calculated in real time, a i : Membership degree of the i-th feature parameter - historical data statistics and device attributes, b i Critical upper limit of the i-th type of operating condition - historical data statistics and equipment attributes. : The i-th multidimensional feature vector;

[0232] a1.3 Decision Fusion Algorithm

[0233] A weighted voting mechanism will be adopted:

[0234] ;

[0235] Weighting coefficient Optimized using the backpropagation algorithm, if a pumping station is not equipped with the corresponding parameter measurement, such as a flow meter for a single water pump, then... The weight of x4 is 0;

[0236] a1.4 Anomaly Detection Module

[0237] Constructing a residual detection model:

[0238] ;

[0239] when Triggering a Level 3 alarm:

[0240] Yellow Alert

[0241] Orange Alert

[0242] Yellow Alert ,in, Threshold : Residual.

[0243] Step a1: This technology integrates normalized multidimensional features (running time, pressure, energy consumption ratio, flow ratio) through a fuzzy logic model, and uses membership functions and weighted voting mechanisms to accurately identify the pump's operating state (such as the transition state "S"), thus solving the problem of misjudgment caused by noise or fluctuations in the traditional threshold method.

[0244] Its residual detection model can dynamically capture anomalies (such as sensor failure or sensor communication anomalies). Its advantages include strong fault tolerance, adaptive weight optimization, cross-pump station versatility, and improved data cleaning efficiency through hierarchical alarms. It is particularly suitable for monitoring the status of pump combinations with large amounts of historical data under complex operating conditions.

[0245] Step a2: The intelligent reconstruction system for pump characteristic curves based on hybrid modeling generates single-pump curve equations, and the optimized data includes:

[0246] 2.1 Multi-source data fusion modeling framework,

[0247] Establish a three-dimensional feature space:

[0248] ;

[0249] Where: D is the three-dimensional feature space of each pump in the pumping station, Q is the flow column vector, H is the head column vector, and η is the efficiency column vector;

[0250] a 2.2 Physical constraint model,

[0251] Using regularized constrained least squares method:

[0252] ;

[0253] Where the basis function matrix is:

[0254] ,

[0255] Constraints:

[0256] , ;

[0257] Where, Q: Actual flow rate test value - manufacturer's allowable operating range, H: Actual head value - manufacturer's allowable operating range, a: Quadratic coefficient of the QH characteristic curve - ensures the curve opens downwards, b: Linear coefficient of the QH characteristic curve, c: Constant term of the QH characteristic curve - shut-off head at zero flow. Regularization coefficient - to prevent overfitting; a head-flow quadratic equation is established for each pump using the method in step a.2.2;

[0258] 2.3 Intelligent Modeling:

[0259] Building a deep hybrid network architecture:

[0260] ;

[0261] Where H: Real-time measured head value - core operating condition output parameter, Q: Real-time flow test value - core operating condition output parameter, f VFD : Frequency ratio of inverter operating frequency to power frequency operating frequency - key control variable of variable frequency pump, t: equipment operating timestamp - reflects equipment aging trend, W1: first layer weight evidence - feature space change, b1: first layer bias vector - nonlinear offset compensation, W2: second layer weight evidence - feature weighted aggregation, b2: output layer bias - reference head calibration.

[0262] Model selection metrics

[0263] ;

[0264] Among them, R 2 Prediction accuracy, AIC: Model complexity penalty, Inference Time: Real-time performance metric, α: Fit coefficient, β: Generalization ability coefficient, γ: Efficiency coefficient.

[0265] Step a2, in which steps 2.2 and 2.3 are parallel. Generally, if the characteristic curves or test data records provided by the manufacturer are less than 50, step 2 is used; if all the pumps in a pumping station are equipped with single-pump outlet flow meters and more than 100 historical data records can be traced, then step 3 can be used for modeling.

[0266] This technology employs a hybrid modeling approach, combining a physical constraint model (using regularized least squares fitting for small samples to ensure the curve conforms to the head-flow rate decreasing trend) and an intelligent deep learning model (using neural networks to fuse flow rate, frequency, and other data to improve nonlinear fitting capabilities for large datasets) to accurately reconstruct the characteristic curve of a single pump. Its advantages are: no loss of control with small data (physical constraints ensure rationality) and greater accuracy with large data (deep learning adapts to complex operating conditions). It solves the problems of large errors with insufficient data or difficulty in modeling with abundant data in traditional methods, providing a reliable theoretical curve foundation for the optimization of parallel pump systems.

[0267] Step a3: Based on the intelligent generation system for the dynamic characteristics of parallel pump groups, the characteristic curves of parallel pumps are generated, and the dynamic calculation model of parallel operators is combined with the solution of coefficient matrix and three-dimensional parameterized display to realize the rapid calculation and visualization of the characteristic curves of parallel pumps.

[0268] a 3.1 Parallel Pump Characteristic Calculation Model

[0269] Define parallel operators:

[0270] ;

[0271] in:

[0272] Let i be the characteristic function of the i-th pump.

[0273] The total flow function of a parallel system, the sum of the flow rates of each operating pump at head H.

[0274] This represents the total number of pumps allowed to operate in parallel within the pump system.

[0275] This is a pump status indication function;

[0276] a 3.2 Parallel Pump Characteristic Calculation Model

[0277] Calculating the characteristic curves of parallel pumps requires determining the output of each pump based on its head, then superimposing the flow rates of the pumps participating in the parallel operation, and finally solving using a coefficient matrix.

[0278] ,

[0279] in: basis functions ;

[0280] a 3.3 Parallel Pump Characteristic Calculation Model

[0281] Achieve 3D parametric display:

[0282]

[0283] in, System head; Q: Total flow rate, calculated from step a.3.2. Overall efficiency, MLP: Multilayer Perceptron, Hnorm: Relative head value, used to determine if the head exceeds the limit, Qnorm: Standardized efficiency value, used to assess energy efficiency health.

[0284] Step a3, through a dynamic calculation model of parallel operators, combined with coefficient matrix solving and three-dimensional parameterized display, achieves rapid and accurate calculation and visualization of the characteristic curves of parallel pumps. Its advantages are:

[0285] 1) Status awareness (determining the operating status of the pump group based on fuzzy logic to ensure the reliability of the computational basis);

[0286] 2) Efficient solution (coefficient matrix decomposition of the characteristic equations of each pump, parallel calculation of the speed increase);

[0287] 3) Intelligent interaction (3D parametric display visualization supports real-time operating condition simulation and abnormal alarm).

[0288] Furthermore, step a4 generates system characteristic curves based on the hydraulic system characteristic modeling system of multi-scale fluid dynamics. Through multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation and adaptive polynomial fitting, the dynamic and accurate construction of the pump station system characteristic curves is realized.

[0289] a4.1 Intelligent head calculation model,

[0290] Establish a three-dimensional spatial head matrix:

[0291] ;

[0292] Wherein: H static : Static range, P dis Water pump outlet pressure, Z dis : Elevation of water pressure gauge / sensor, P suct Water pump inlet pressure, Z suct : Inlet pressure gauge / sensor elevation, This is a time-varying water level correction term. The vertical height difference between the pressure measuring point on the main outlet pipe of the pumping station and the bottom of the suction well;

[0293] a4.2 Hybrid Loss Calculation Engine

[0294] The solution is obtained by combining the Darcy-Weisbach and Navier-Stokes methods:

[0295]

[0296] in: Head loss / dynamic head; : , , ;

[0297] a4.3 Adaptive characteristic curve modeling,

[0298] Variable-order polynomial model

[0299] ;

[0300] Where Q: flow rate, H(Q): system characteristic curve, c k : Polynomial coefficients, k: Polynomial order;

[0301] Order selection criteria:

[0302] ,

[0303] Among them, K *Optimal order, candidate order values ​​for K-polynomial fitting. Regularization coefficient, Norm penalty term;

[0304] Online parameter recognition system

[0305] Establish a recursive least squares algorithm:

[0306] ,

[0307] : Parameter vector Measured head Kalman gain, : Feature vector;

[0308] in:

[0309] , ,

[0310] Jingyangcheng Linear drag coefficient, Secondary drag coefficient,

[0311] This technology achieves dynamic and accurate construction of the characteristic curves of pump station systems through multi-scale fluid dynamics modeling, combined with intelligent head calculation (3D matrix + time-varying correction), hybrid loss calculation (joint solution of Darcy-Weisbach and Navier-Stokes equations), and adaptive polynomial fitting (variable order constraint optimization + online parameter identification). Its advantages include:

[0312] 1) Full-element modeling (static head, time-varying water level);

[0313] 2) Adaptive optimization (AIC criterion automatically selects the polynomial order to avoid overfitting);

[0314] 3) Real-time updates (recursive least squares online parameter identification, dynamically tracking changes in operating conditions).

[0315] It solves the problems of traditional methods ignoring local resistance and being unable to adapt to time-varying operating conditions, and provides a real-time dynamic pipeline characteristic curve set for pump station energy efficiency assessment.

[0316] S5. After evaluating, analyzing, and annotating the data, display it on the page in the form of charts, propose corresponding solutions and early warning prompts, and provide feedback to pump station staff and central control personnel for human-computer interaction.

[0317] Furthermore, step S5 uses the intelligent diagnostic system for the health status of the pump based on the flow-head characteristics and the historical case matching and optimization system for energy efficiency special diagnosis to feed back the monitored results to the pump station staff and send early warning information to the computer terminal to remind the pump station staff to check and repair abnormal problems in a timely manner.

[0318] The intelligent health status diagnosis system for water pumps based on flow-head characteristics includes the following steps:

[0319] S5.1, Flow Condition Zoning and Wear Feature Mapping Model

[0320] S5.11 Establish the flow-wear characteristic matrix:

[0321]

[0322] S5.12, Employing a time-intensity dual-weighted model:

[0323] ;

[0324] ,

[0325] in:

[0326] The cumulative amount of wear of type k is dimensionless and calculated by real-time integration. The strength coefficient of the i-th type of wear, flow; For the flow range of the kth type of wear, Temperature correction factor (when hour , (Refer to the equipment user manual provided by the supplier for alarm temperature)

[0327] Initial time;

[0328] S5.2 Real-time Health Status Assessment

[0329] S5.21 3D Special Diagnosis Visualization

[0330] Establish the feature space:

[0331] ;

[0332] Color mapping rules:

[0333] green( Health status

[0334] Yellow status ( ): Pay attention to the state

[0335] red( Dangerous state

[0336] S5.22, Identification of Dominant Wear Factors

[0337] Employing a fuzzy inference system:

[0338]

[0339] in, The cumulative duration of the k-th type of wear. Total effective runtime The kth type of wear intensity,

[0340] The system outputs the top three wear types and their confidence levels. This technology achieves intelligent diagnosis and early warning of pump health status through a flow-wear feature mapping model and a time-intensity dual-weighted evaluation.

[0341] The flow condition zoning and wear feature mapping model technology achieves intelligent diagnosis and early warning of pump health status through a flow-wear feature mapping model and time-intensity dual-weighted evaluation, and color mapping. Its advantages include:

[0342] 1) Accurate fault prediction (based on international standards to divide flow condition ranges and match typical wear types);

[0343] 2) Dynamic health assessment (combining multiple parameters such as temperature and operating time to calculate cumulative wear);

[0344] 3) Visualized decision support (the device status is displayed intuitively in a three-dimensional feature space, and fuzzy reasoning is used to identify the main fault types).

[0345] It solves the problems of slow progress and strong reliance on experience in traditional manual inspections. Through real-time data-driven analysis, it can detect potential problems such as cavitation and bearing wear in advance, significantly reducing the risk of unplanned downtime.

[0346] The matching optimization steps of the historical case matching optimization system based on energy efficiency characteristics are as follows:

[0347] S5.3, Construct a four-dimensional energy efficiency characteristic space:

[0348] ;

[0349] in:

[0350] Electricity consumption per unit (kWh / m³) 3 )

[0351] Comprehensive unit power consumption (kWh / km) 3.MPa)

[0352] Increase unit consumption (kWh / mH2O)

[0353] :Number of runs in history;

[0354] S5.4. Based on the four-dimensional energy efficiency feature space in S5.3, improve the similarity measurement algorithm.

[0355] Define a weighted similarity function:

[0356] ;

[0357] Feature dimensions and weight allocation:

[0358] Electricity consumption similarity (weight) 0.4)

[0359] ;

[0360] Overall unit power consumption similarity (weight) =0.3)

[0361] ;

[0362] Improve unit cost similarity (weight) 0.2)

[0363] ;

[0364] Frequency of operation credibility (weight) 0.1)

[0365] ;

[0366] Case selection method:

[0367] Initial screening criteria: Current flow rate requirement ±100m 3 Historical combinations within the range of / h (can be freely defined)

[0368] Sorting rules:

[0369] First priority: Top 30% in similarity score.

[0370] Second priority: Top 30% of runs.

[0371] Output: The top 3 optimized combinations based on overall ranking.

[0372] This technology achieves intelligent screening of historically high-efficiency operating conditions by constructing a four-dimensional energy efficiency feature space (electricity consumption per unit, total electricity consumption, improved energy consumption per unit, and historical frequency) and an improved similarity algorithm (weighted multi-index matching). Its advantages include:

[0373] 1) Multidimensional energy efficiency assessment (quantifying the advantages and disadvantages of working conditions from multiple perspectives, such as energy conversion efficiency and pipeline transmission and distribution efficiency);

[0374] 2) Dynamic weight optimization (emphasizing direct energy consumption costs);

[0375] 3) Dual screening mechanism (first screening by traffic range, then sorting by similarity and frequency of operation).

[0376] Compared to traditional methods, it solves the problems of one-sidedness in single-indicator decision-making and strong reliance on experience, and can quickly output the top 3 optimized combinations that have been verified by history.

[0377] It should be understood that the above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. It should not be considered that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, several simple deductions or substitutions can be made without departing from the concept of the present invention. All equivalent changes or modifications made in accordance with the spirit and essence of the present invention should be covered within the protection scope of the present invention.

Claims

1. A pump station energy efficiency diagnosis system based on multi-curve collaborative analysis, characterized in that, include: The data acquisition module is used to import offline and online pump station data, transmit the pump station data to the data preprocessing module, and simultaneously input the single pump curve and pipeline characteristic curve data of the water pump characteristic curve. The data preprocessing module preprocesses and cleans the collected pump station data and transmits the pump station data to the data analysis module. The data analysis module identifies abnormal values ​​in the pump station data and highlights them, then transmits the data to the data correction and calculation module. The data correction and calculation module corrects outlier data at the pumping station, calculates the pumping station data according to the corresponding formula, and finally transmits the data to the decision module. The outlier data correction and calculation includes: By establishing a state determination model based on fuzzy logic, the operating conditions of water pumps and the combination of parallel pumps can be identified. A smart reconstruction system for pump characteristic curves based on hybrid modeling generates single-pump curve equations and optimizes data. Based on the intelligent generation system for the dynamic characteristics of parallel pump sets, the system generates characteristic curves of parallel pumps and a dynamic calculation model of parallel operators. Combined with coefficient matrix solving and three-dimensional parameterized display, it realizes the rapid calculation and visualization of characteristic curves of parallel pumps. The hydraulic system characteristic modeling system based on multi-scale fluid dynamics generates system characteristic curves. Through multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation and adaptive polynomial fitting, the dynamic and accurate construction of pump station system characteristic curves is realized. The decision-making module evaluates, analyzes, and annotates the data, displays it in chart form on the page, proposes corresponding solutions and early warning prompts, and feeds back the information to pump station staff and central control personnel for human-computer interaction. a1. By establishing a state determination model based on fuzzy logic, the operating conditions of water pumps and the combination of parallel pumps are identified. Step a1 includes: a1.1, Feature parameter normalization processing, Define a multidimensional feature vector: ； in: Multidimensional feature vectors , : Normalized runtime value, T op The operating time of each pump is measured in minutes. , Normalized pressure value Pressure at the outlet of each pump, P out The pressure at the pump station outlet. , Energy efficiency ratio, P i,act Hourly power consumption per pump, P rated : The rated power of the corresponding water pump, f i,act The operating frequency of the water pump. Flow ratio, Q act Q: Hourly water output of each pump rated : The rated flow rate of the corresponding water pump; a1.2, State Decision Matrix Establish membership function: Where, μ i : Calculate the membership degree of the i-th feature parameter in real time, a i : Membership degree of the i-th feature parameter of historical data statistics and equipment attributes, b i The critical upper limit of the i-th type of working condition. : The i-th multidimensional feature vector; a1.3 Decision Fusion Algorithm A weighted voting mechanism will be adopted: ； Weighting coefficient Optimized using the backpropagation algorithm; a1.4 Anomaly Detection Module Constructing a residual detection model: ； when Triggering a Level 3 alarm: Yellow Alert Orange Alert Yellow Alert , in, Threshold : residual; a2. A smart reconstruction system for pump characteristic curves based on hybrid modeling generates single-pump curve equations and optimizes data. Step a2 includes: 2.1 Multi-source data fusion modeling framework, Establish a three-dimensional feature space: ； Where: D is the three-dimensional feature space of each pump in the pumping station, Q is the flow column vector, H is the head column vector, and η is the efficiency column vector; a 2.2 Physical constraint model, Using regularized constrained least squares method: ； Where the basis function matrix is: , Constraints: , ； Where, Q: actual flow rate measured value, H: actual head value, a: coefficient of the quadratic term of the QH characteristic curve, b: coefficient of the linear term of the QH characteristic curve, c: constant term of the QH characteristic curve. Regularization coefficient; Each pump establishes a head using the method in step a2.2; 2.3 Intelligent Modeling: Building a deep hybrid network architecture: Where H: real-time measured head value, Q: real-time flow rate test value, f VFD : Ratio of inverter operating frequency to power frequency, t: Equipment operating timestamp, W1: First layer weighted evidence, b1: First layer bias vector, W2: Second layer weighted evidence, b2: Output layer bias; Model selection metrics ； Among them, R 2 Prediction accuracy, AIC: Model complexity penalty, Inference Time: Real-time performance metric, α: Fit coefficient, β: Generalization ability coefficient, γ: Efficiency coefficient; S3. Identify abnormal values ​​in pump station data and highlight them. a4. The hydraulic system characteristic modeling system based on multi-scale fluid dynamics generates system characteristic curves. Through multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation and adaptive polynomial fitting, the dynamic and accurate construction of pump station system characteristic curves is realized. Step a4 includes: a4.1 Intelligent head calculation model, Establish a three-dimensional spatial head matrix: ； Wherein: H static : Static range, P dis Water pump outlet pressure, Z dis : Elevation of water pressure gauge / sensor, P suct Water pump inlet pressure, Z suct : Inlet pressure gauge / sensor elevation, This is a time-varying water level correction term. The vertical height difference between the pressure measuring point on the main outlet pipe of the pumping station and the bottom of the suction well; a4.2 Hybrid Loss Calculation Engine The solution is obtained by combining the Darcy-Weisbach and Navier-Stokes methods: ， ： ， ， ； in: Head loss / dynamic head; a4.3 Adaptive characteristic curve modeling, Variable-order polynomial model ； Where Q: flow rate, H(Q): system characteristic curve, c k : Polynomial coefficients, k: Polynomial order; Order selection criteria: ， Among them, K * Optimal order, candidate order values ​​for K-polynomial fitting. Regularization coefficient, Norm penalty term; Online parameter recognition system Establish a recursive least squares algorithm: ， : Parameter vector Measured head Kalman gain, : Feature vector; in: , ， Jingyangcheng Linear drag coefficient, Secondary drag coefficient, By using multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation, and adaptive polynomial fitting, the characteristic curves of the pump station system were dynamically constructed.

2. The pump station energy efficiency diagnosis system based on multi-curve collaborative analysis according to claim 1, characterized in that: The data analysis module is used to identify and highlight outliers in the pump station data by setting upper and lower limits. The aforementioned upper and lower limit settings refer to setting upper and lower limits for data such as liquid level, flow rate, pressure, and power consumption that are directly measured by instruments for monitoring the operation of pump stations, combined with design documents and operational experience. Data exceeding the upper limit and falling below the lower limit are displayed in red.

3. The diagnostic method for a pump station energy efficiency diagnostic system based on multi-curve collaborative analysis according to any one of claims 1-2, characterized in that: The diagnostic steps include the following: S1. Import offline pump station data and online pump station data, and transmit the pump station data to the data preprocessing module. At the same time, input the single pump curve and pipeline characteristic curve data of the water pump characteristic curve. S2. Preprocess the collected pump station data; S3. Identify abnormal values ​​in pump station data and highlight them. S4. Correct the outlier data of the pumping station and calculate the pumping station data according to the corresponding formula. The correction of outlier data and calculation of pumping station data according to the corresponding formula includes: a1. By establishing a state determination model based on fuzzy logic, the operating conditions of water pumps and the combination of parallel pumps can be identified. a2. A smart reconstruction system for pump characteristic curves based on hybrid modeling generates single-pump curve equations and optimizes data. a3. Based on the intelligent generation system for the dynamic characteristics of parallel pump sets, the characteristic curves of parallel pumps are generated, and the dynamic calculation model of parallel operators is combined with the solution of coefficient matrix and three-dimensional parameterized display to realize the rapid calculation and visualization of the characteristic curves of parallel pumps. a4. The hydraulic system characteristic modeling system based on multi-scale fluid dynamics generates system characteristic curves. Through multi-scale fluid dynamics modeling, combined with intelligent head calculation, mixed loss calculation and adaptive polynomial fitting, the dynamic and accurate construction of pump station system characteristic curves is realized. S5. After evaluating, analyzing, and annotating the data, display it on the page in the form of charts, propose corresponding solutions and early warning prompts, and provide feedback to pump station staff and central control personnel for human-computer interaction.

4. The pump station energy efficiency diagnosis method based on multi-curve collaborative analysis according to claim 3, characterized in that: Step S5 is based on the intelligent diagnostic system for the health status of the water pump based on the flow-head characteristics and the historical case matching optimization system based on energy efficiency characteristics. The monitored results are fed back to the pump station staff, and the early warning information is sent to the computer terminal to remind the pump station staff to check and repair abnormal problems in a timely manner.

5. The pump station energy efficiency diagnosis method based on multi-curve collaborative analysis according to claim 4, characterized in that: The diagnostic process of the intelligent health status diagnosis system for water pumps based on flow-head characteristics includes the following steps: S5.1 Flow Condition Zoning and Wear Feature Mapping Model S5.11 Establish the flow-wear characteristic matrix: S5.12, Employing a time-intensity dual-weighted model: ， in: The cumulative amount of wear of type k is dimensionless and calculated by real-time integration. The strength coefficient of the i-th type of wear, flow; For the flow range of the kth type of wear, This is a temperature correction factor; Initial time; S5.2 Real-time Health Status Assessment S5.21 3D Special Diagnosis Visualization Establish the feature space: ;x ,y ,z: , Color mapping rules: green Health status Yellow status Pay attention to the state. red Dangerous state S5.22, Identification of Dominant Wear Factors Employing a fuzzy inference system: ,in in, The cumulative duration of the k-th type of wear. Total effective runtime The kth type of wear intensity, Output the wear type and confidence level of the first three values.

6. The pump station energy efficiency diagnosis method based on multi-curve collaborative analysis according to claim 4, characterized in that: The matching and optimization steps of the historical case matching and optimization system based on energy efficiency characteristics are as follows: S5.3, Construct a four-dimensional energy efficiency characteristic space: ； in: Electricity consumption per unit Comprehensive unit power consumption Increase unit consumption. :Number of runs in history; S5.4 Improve the similarity measurement algorithm based on the four-dimensional energy efficiency feature space in S5.

3. Define a weighted similarity function: Feature dimensions and weight allocation: Electricity consumption similarity, weight 0.4, ； Based on the similarity of unit power consumption, the weights 0.3 ； Improve unit cost similarity, weight =0.2 ； Frequency of operation credibility, weight =0.1 ； Output: The top 3 optimized combinations based on overall ranking.

Images