Efficient energy-saving integrated control method and system for secondary water supply pipe network

Abstract

The invention discloses a high-efficiency energy-saving integrated control method and system for a secondary water supply pipe network, and relates to the technical field of water supply pipe network control, and the system forms a complete behavior recording-performance quantification-strategy correction closed loop path by introducing an intelligent start-stop execution module and an efficiency evaluation and feedback optimization module. Each start-stop behavior is recorded in a structured mode, joint evaluation of the energy-saving effect and the disturbance response is carried out at the end of the cycle, and a data-driven self-adaptive reference is provided for the next round of strategy decision. According to the learning type scheduling mechanism, the adaptive capacity of the system to uncertain load disturbance is remarkably improved, and the scheduling precision and the long-term energy efficiency performance are improved. Scheduling judgment logic cooperatively driven by an energy-saving scheduling window prediction module and a pressure wave recognition and classification module is constructed, hydraulic disturbance characteristics and the unit time energy consumption trend can be dynamically integrated, and optimal configuration of a start-stop time window is achieved on the premise that system stability is not sacrificed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water supply network control, in particular to an efficient and energy-saving integrated control method and system for secondary water supply network. BACKGROUND

[0002] As the main water supply guarantee means for urban high-rise buildings and areas with insufficient pressure in remote pipe networks, the system structure of secondary water supply usually includes a water storage tank, a water pump set, a frequency converter controller and a distributed end pipe network. In this system, the water pump start-stop scheduling is a key link for water supply energy efficiency regulation and control, and its operation mode directly affects the power consumption, hydraulic stability and equipment fatigue life. Therefore, improving the intelligence and adaptability of its scheduling has become an important direction for the industry.

[0003] In the prior art, water pump start-stop scheduling is mostly based on timing control or simple pressure upper and lower limit feedback strategy. This kind of method does not fully consider the dynamic evolution characteristics of pressure fluctuations and the complex fluid response inside the system, especially when the night load changes frequently or the water use behavior mutates, it is easy to produce short period and high frequency start-stop switching, forming pump set frequency drift or "water hammer excitation". This not only causes the power consumption to rise per unit time, but also causes the pump body parts to fatigue prematurely, the pipe joints to loosen and other hidden dangers.

[0004] The above problems are mostly caused by the neglect of "weak disturbance signals" in the pressure waveform sequence or the too rough processing method. When there are phenomena such as micro-disturbance frequency aggregation and asynchronous reflected wave overlap in the water supply system, if the start-stop operation is just in these interference frequency bands, it is easy to form transient water hammer disturbance and frequency spectrum resonance excitation, causing the system pressure to fluctuate violently. Therefore, it is urgent to build an intelligent identification and scheduling window prediction mechanism based on dynamic analysis of pressure waveform, to realize fine energy-saving control and active steady-state maintenance of the system. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides an efficient and energy-saving integrated control method and system for secondary water supply network, which solves the problems mentioned in the background art.

[0006] To achieve the above purpose, the present application is implemented by the following technical scheme: an efficient and energy-saving integrated control system for secondary water supply network, comprising a dynamic pressure wave acquisition and processing module, a micro-disturbance feature extraction module, a pressure wave identification and classification module, an energy-saving scheduling window prediction module, an intelligent start-stop execution module and an efficiency evaluation and feedback optimization module;

[0007] The dynamic pressure wave acquisition and processing module acquires water flow state data through sensors installed in the water supply network, and obtains a water flow data set SW;

[0008] The perturbation feature extraction module extracts features from the water flow dataset SW, including a spectrum splitting distance Cse, a phase mutation rate Vds, and a perturbation root mean square amplitude Sa, and constructs a perturbation comprehensive index RDZ;

[0009] The pressure wave recognition and classification module combines the acquired perturbation comprehensive index RDZ with the water flow dataset SW, compares it with a preset interference sensitivity threshold Thm and an abnormal drift rate Ysb, and obtains a risk time window set Wr;

[0010] The energy-saving scheduling window prediction module combines the acquired risk time window set Wr with the pump station energy consumption Es and the perturbation change trend Rsp, evaluates all non-disabled time periods, and obtains a high-risk time set Ω;

[0011] The intelligent start-stop execution module sends start-stop control instructions to the water pump controller according to the acquired risk time window set Wr and the high-risk time set Ω, and obtains actual start-stop behavior records Σon / off;

[0012] The performance evaluation and feedback optimization module quantitatively analyzes the energy-saving effect after execution according to the actual start-stop behavior records Σon / off, and obtains a comprehensive energy-saving and stability score Znps.

[0013] Preferably, the dynamic pressure wave acquisition and processing module includes a multi-dimensional perturbation data acquisition unit and a data preprocessing unit.

[0014] The multi-dimensional perturbation data acquisition unit acquires water flow state data through sensors, including an instantaneous pressure value Sp, a number of micro pressure spikes Sc, a minimum frequency Sr of disturbance waveband frequency distribution, a total amount of pressure wave spectrum phase jumps Sz, a perturbation root mean square amplitude Sa, a number of pump group frequency drifts Sb, and a pump station energy consumption Es, and fits them into an original dataset QW.

[0015] The instantaneous pressure value Sp is acquired through a high-frequency pressure sensor.

[0016] The number of micro pressure spikes Sc is acquired through a pressure wave mutation detector.

[0017] The number of micro pressure spikes Sc is acquired through the following formula:

[0018] ;

[0019] In the formula, n represents the total number of time points collected, Spi+1 represents the instantaneous pressure value at the i+1th time point, Spi represents the instantaneous pressure value at the ith time point, Δth represents the pressure mutation threshold, indicates an indicator function, which is 1 when the condition is met.

[0020] The minimum frequency interval Sr of the disturbance waveband frequency distribution is obtained by an embedded frequency spectrum collector;

[0021] The minimum frequency interval Sr of the disturbance waveband frequency distribution is obtained by the following formula:

[0022] ;

[0023] In the formula, fa represents the a-th frequency peak, fa+1 represents the a+1-th frequency peak, Fpeak represents the set of all peak frequencies in the pressure spectrum, and min() represents the minimum value in the formula;

[0024] The total amount of phase jumps Sz of the pressure wave spectrum is obtained by an embedded frequency spectrum demodulator;

[0025] The total amount of phase jumps Sz of the pressure wave spectrum is obtained by the following formula:

[0026] ;

[0027] In the formula, represents the main frequency phase value at the i-th time point, represents the main frequency phase value at the i+1-th time point;

[0028] The disturbance root mean square amplitude Sa is obtained by a pressure signal RMS module;

[0029] The number of frequency drifts Sb of the pump group is obtained by a variable frequency drive electric signal listener;

[0030] The pump station energy consumption Es is obtained by a current and voltage sensor;

[0031] The data preprocessing unit performs cleaning and normalization processing on the collected original data set QW to obtain a water flow data set SW;

[0032] The cleaning includes removing abnormal values in the original data set QW by using a three-sigma method, and filling the removed abnormal values by using a median interpolation processing method;

[0033] The normalization processing is performed on the original data set QW by using a Min-Max normalization method to obtain the water flow data set SW;

[0034] The water flow data set SW is obtained by the following formula:

[0035] ;

[0036] In the formula, SWo represents the oth data in the water flow data set SW, QWo represents the oth data in the original data set QW, maxQWo represents the peak value of the oth data in the original data set QW, and minQWo represents the valley value of the oth data in the original data set QW.

[0037] Preferably, the perturbation feature extraction module comprises a non-steady spectrum analysis unit and a comprehensive disturbance modeling unit.

[0038] The non-steady spectrum analysis unit is based on fast Fourier transform and continuous wavelet transform on the water flow data set SW, and identifies the frequency spectrum structure cracking area, extracts the frequency spectrum cracking distance Cse, the phase mutation rate Vds and the disturbance root mean square amplitude Sa.

[0039] The frequency spectrum cracking distance Cse is obtained by the following formula:

[0040] ;

[0041] In the formula, σSp represents the standard deviation of the instantaneous pressure value, η represents the non-equilibrium modulation coefficient, μΔf represents the average frequency distance between the main frequencies, Δfmax represents the maximum frequency distance between the main frequencies, and C represents the anti-zero constant.

[0042] The phase mutation rate Vds is obtained by the following formula:

[0043] ;

[0044] In the formula, n represents the total number of time points collected, T represents the total observation time window length, represents the main frequency phase value at the k+1th time point, represents the main frequency phase value at the kth time point, tk+1 represents the k+1th time point, and tk represents the kth time point.

[0045] The comprehensive disturbance modeling unit receives the obtained frequency spectrum cracking distance Cse, the phase mutation rate Vds and the disturbance root mean square amplitude Sa, and constructs a disturbance comprehensive index RDZ using a nonlinear structure.

[0046] The disturbance comprehensive index RDZ is obtained by the following formula:

[0047] ;

[0048] In the formula, ln represents the natural logarithm function, tanh represents the hyperbolic tangent function, C1 represents a positive number, and C2 represents an interference suppression constant.

[0049] Preferably, the pressure wave recognition and classification module comprises a dynamic discriminant index fusion unit and a risk time window screening unit.

[0050] The dynamic discrimination index fusion unit fuses and models the obtained disturbance comprehensive index RDZ and the pump group frequency drift times Sb in the water flow data set SW to construct a discrimination driving index Arisk.

[0051] The discrimination driving index Arisk is obtained by the following formula:

[0052] ;

[0053] In the formula, Thm represents an interference sensitivity threshold value, and Ysb represents an abnormal drift rate.

[0054] The interference sensitivity threshold value Thm is obtained by the following formula:

[0055] ;

[0056] In the formula, μRDZ represents the mean value of the disturbance comprehensive index, σRDZ represents the standard deviation of the disturbance comprehensive index, and α1 represents an adjustment coefficient.

[0057] The abnormal drift rate Ysb is obtained by the following formula:

[0058] ;

[0059] In the formula, μSb represents the mean value of the historical drift times, σSb represents the standard deviation of the drift times, and α2 represents a drift fluctuation tolerance coefficient.

[0060] Preferably, the risk time window screening unit dynamically compares the discrimination driving index Arisk at each time point, screens out a time period exceeding an activation threshold value, and combines the interference sensitivity threshold value Thm and the abnormal drift rate Ysb to form a time index set as a risk time window set Wr, and logically discriminates the disturbance type occurring at the time ti to obtain a disturbance risk mechanism.

[0061] ;

[0062] In the formula, Arisk(ti) represents the discrimination driving index at the time ti, and Task represents a preset risk index activation threshold value.

[0063] The disturbance risk mechanism is obtained by the following formula:

[0064] ;

[0065] In the formula, Type(ti) represents a risk disturbance value at the time ti, RDZ(ti) represents a disturbance comprehensive index at the time ti, and Sb(ti) represents a pump group frequency drift time at the time ti.

[0066] When the risk disturbance value Type (ti) = 1 at time ti, it represents a structural disturbance; the system spectrum / fluctuation is strong, but the pump group control is smooth, which belongs to passive disturbance, such as pipe water hammer, end wave reflection, etc.

[0067] When the risk disturbance value Type (ti) = 2 at time ti, it represents a behavioral disturbance; the pump group control frequently fluctuates, but the disturbance energy is not high, which belongs to active disturbance, such as frequent start-stop, unreasonable frequency conversion strategy;

[0068] When the risk disturbance value Type (ti) = 3 at time ti, it represents a coupling aggravation disturbance; structural disturbance and behavioral disturbance appear at the same time, there may be resonance between scheduling and environment, which causes large-scale instability.

[0069] Preferably, the energy-saving scheduling window prediction module comprises an energy consumption disturbance trend analysis unit and a scheduling window identification unit;

[0070] The energy consumption disturbance trend analysis unit constructs a disturbance evolution function Ψ based on the pump station energy consumption Es and the disturbance change trend Rsp;

[0071] The disturbance evolution function Ψ is obtained by the following formula:

[0072] ;

[0073] In the formula, B1 represents an energy consumption amplification factor, B2 represents a trend amplification factor, d represents an integral symbol, Es(t) represents the pump station energy consumption at time t, and Rsp(t) represents the disturbance change trend at time t, which is obtained by calculating the second derivative of the instantaneous pressure value Sp;

[0074] The scheduling window identification unit dynamically scores all the remaining schedulable time t∉Wr after excluding the risk time window set Wr according to the obtained disturbance evolution function Ψ, and obtains a scheduling optimization index function Φ;

[0075] The scheduling optimization index function Φ is obtained by the following formula:

[0076] ;

[0077] In the formula, Φ(t) represents the scheduling optimization index function at time t, represents infinity;

[0078] A high-risk time set Ω is constructed by the scheduling optimization index function Φ;

[0079] The high-risk time set Ω is obtained by the following formula:

[0080] ;

[0081] In the formula, Φ(ti) represents the scheduling optimization exponential function at time ti, μΦ represents the mean of the scheduling optimization exponential function, σΦ represents the standard deviation of the scheduling optimization exponential function, and λ represents an empirical factor for adjusting the threshold sensitivity.

[0082] Preferably, the intelligent start-stop execution module comprises a start-stop discrimination and command generation unit and a start-stop execution and behavior recording unit.

[0083] The start-stop discrimination and command generation unit performs scheduling execution judgment on each time t, acquires a start-stop state instruction generation function U according to whether it belongs to the risk time window set Wr or the high-risk time set Ω, and executes different control strategies.

[0084] The start-stop state instruction generation function U is acquired through the following formula:

[0085] ;

[0086] The control strategy is matched and acquired through the following method:

[0087] When the start-stop state instruction generation function U = 1, it means that the start-stop is allowed to be executed.

[0088] When the start-stop state instruction generation function U = 0, it means that the control is disabled or delayed execution.

[0089] When the start-stop state instruction generation function U = X, it means other states, maintaining the status quo.

[0090] Preferably, the start-stop execution and behavior recording unit receives the start-stop state instruction generation function U, schedules the corresponding pump group for start-stop control, records the start-stop behavior data, forms a complete operation trajectory, and acquires the actual start-stop behavior record Σon / off.

[0091] The actual start-stop behavior record Σon / off is acquired through the following formula:

[0092] ;

[0093] In the formula, ton(j) represents the jth pump start time, toff(j) represents the jth pump stop time, Δt(j) represents the jth pump operation duration, ID(j) represents the pump group number involved in the jth operation, Eon(j) represents the cumulative energy consumption from the jth operation start to shutdown, and N represents the total number of operations.

[0094] Preferably, the performance evaluation and feedback optimization module comprises an energy-saving effect evaluation unit and a disturbance stability evaluation unit.

[0095] The energy-saving effect evaluation unit calculates the cumulative energy-saving amount achieved in the current scheduling period based on the cumulative energy consumption Eon of the start-to-stop process, and obtains the unit time energy-saving rate η save;

[0096] The unit time energy-saving rate η save is obtained by the following formula:

[0097] ;

[0098] In the formula, Tw represents the total time of the current scheduling period, and Eref represents the reference energy consumption;

[0099] The disturbance stability evaluation unit analyzes the influence of pump start-stop execution on system hydraulic disturbance, constructs a disturbance mean square response index Ψ di based on the disturbance mean square indexes formed before and after the start-stop operation of the instantaneous pressure value Sp;

[0100] The disturbance mean square response index Ψ di is obtained by the following formula:

[0101] ;

[0102] In the formula, Sp(t) represents the instantaneous pressure value at time t, represents the second derivative of the instantaneous pressure value;

[0103] The unit time energy-saving rate η save and the disturbance mean square response index Ψ di are obtained, and the comprehensive energy-saving and stability score Znps is calculated and obtained;

[0104] The comprehensive energy-saving and stability score Znps is obtained by the following formula:

[0105] ;

[0106] In the formula, KS represents an interference penalty factor.

[0107] An efficient energy-saving integrated control method for secondary water supply pipe networks, comprising the following steps:

[0108] Step one: The dynamic pressure wave acquisition and processing module acquires water flow state data through sensors installed in the water supply pipe network, and obtains a water flow data set SW;

[0109] Step two: The micro-perturbation feature extraction module extracts features from the water flow data set SW, including a spectrum splitting distance Cse, a phase mutation rate Vds, and a disturbance mean square root amplitude Sa, and constructs a disturbance comprehensive index RDZ;

[0110] Step three: The pressure wave recognition and classification module combines the obtained disturbance comprehensive index RDZ with the water flow data set SW, compares it with a preset interference sensitivity threshold Thm and an abnormal drift rate Ysb, and obtains a risk time window set Wr;

[0111] Step four: the energy-saving scheduling window prediction module combines the obtained risk time window set Wr with the pump station energy consumption Es and the disturbance change trend Rsp to evaluate all non-disabled time periods and obtain a high-risk time set Ω;

[0112] Step five: the intelligent start-stop execution module sends start-stop control instructions to the water pump controller according to the obtained risk time window set Wr and the high-risk time set Ω to obtain actual start-stop behavior records Σon / off.

[0113] Step six: the performance evaluation and feedback optimization module quantitatively analyzes the energy-saving effect after execution according to the actual start-stop behavior records Σon / off to obtain a comprehensive energy-saving and stability score Znps.

[0114] The application provides an efficient energy-saving integrated control method and system for secondary water supply pipe networks, which has the following beneficial effects:

[0115] (1) When the system is running, the intelligent start-stop execution module and the performance evaluation and feedback optimization module are introduced, and the system forms a complete behavior record-performance quantification-strategy correction closed loop path. Each start-stop behavior is structured and recorded, and the joint evaluation of energy-saving effect and disturbance response is performed at the end of the cycle to provide data-driven adaptive reference for the next round of strategy decision. This learning type scheduling mechanism significantly improves the adaptability of the system to uncertain load disturbance, improves the scheduling accuracy and long-term energy efficiency performance. The application adopts a standard deviation outlier cleaning mechanism and a median interpolation joint processing method to effectively eliminate abnormal data offsets caused by sensor distortion, sudden interference or echo superposition, ensuring the robustness of the original data set QW. Further introduction of the Min-Max normalization method enables data parameters of different dimensions and scales to have a unified dimension, providing a good mathematical basis for subsequent spectral analysis and model fitting.

[0116] (2) The dynamic discriminant index fusion unit first models the disturbance comprehensive index RDZ and the pump group frequency drift number Sb together, and constructs the discriminant driving index Arisk to realize the collaborative discrimination of pressure wave disturbance trend and frequency instability event, effectively making up for the single-dimensional judgment deficiency of the traditional scheduling system which only relies on the pressure transient amplitude. This design not only improves the sensitivity of the system to high-frequency disturbance causes and abnormal start-stop phenomena, but also significantly enhances the comprehensive recognition ability of the system to multiple types of abnormal aggravation signals, avoiding misjudgment and omission.

[0117] The application utilizes historical statistical characteristics to construct two dynamic threshold indicators of interference sensitivity threshold Thm and abnormal drift rate Ysb, and no longer relies on fixed upper and lower limits or artificial set empirical rules. This threshold strategy can automatically adjust the early warning sensitivity according to the long-term change trend of the disturbance mode and the stable interval of the system operation, thereby improving the recognition ability of the system to the atypical behavior under the sudden disturbance and fatigue accumulation state. The mechanism is particularly effective under the background of frequent night load fluctuation or system aging, and can prevent unnecessary frequent start-stop operation of the dispatching strategy due to "rigid discrimination".

[0118] (3) The energy saving rate ηsave per unit time is introduced as a core evaluation index through the energy saving effect evaluation unit, which can accurately measure the energy saving benefit brought by each round of start-stop control behavior. Unlike the traditional system which only roughly counts the running time or total energy consumption, the present scheme refines the energy consumption characteristics according to the start-stop cycle, so that the energy saving effect not only has comparative significance, but also has feedback significance. The mechanism significantly improves the energy control transparency of the system under dynamic working conditions, so that the dispatching strategy can be iterated and self-optimized according to quantitative data, solving the problem of "unobservable energy saving effect and unclear optimization direction" in the traditional scheme.

[0119] (4) Through the efficiency evaluation and feedback optimization module, not only the energy saving effect is quantitatively scored, but also the disturbance influence brought by the start-stop control is analyzed synchronously, and a unified comprehensive performance index is constructed for subsequent strategy correction. This feedback type design enables the system to continuously calibrate parameters and optimize rhythm in practice, realizing the long-term self-adaptive evolution of the strategy. Compared with the traditional system in which the dispatching strategy is fixed and difficult to adapt to the changes of system aging or user water behavior, the present application not only keeps the system running at a low energy consumption, but also ensures the continuous evolution and adaptability of the dispatching control strategy. BRIEF DESCRIPTION OF DRAWINGS

[0120] Figure 1 It is a flow chart of a high-efficiency energy-saving integrated control system for a secondary water supply pipe network of the application;

[0121] Figure 2 It is a flow chart of a high-efficiency energy-saving integrated control method for a secondary water supply pipe network of the application;

[0122] Figure 3 It is a flow chart of the comprehensive energy saving and stability score of the application;

[0123] Figure 4 It is a column chart of the disturbance comprehensive index of the application;

[0124] Figure 5 It is an energy efficiency trend chart of the comprehensive energy saving and stability score of the application;

[0125] Figure 6 A trend chart of the simulation of the perturbation evolution function algorithm of the application. DETAILED DESCRIPTION

[0126] The technical solutions in the embodiments of the application will be apparently and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the application.

[0127] Embodiment 1

[0128] The application provides a high-efficiency energy-saving integrated control system for a secondary water supply pipe network, please refer to Figures 1 to 6 , comprising a dynamic pressure wave acquisition and processing module, a perturbation feature extraction module, a pressure wave identification and classification module, an energy-saving scheduling window prediction module, an intelligent start-stop execution module and an efficiency evaluation and feedback optimization module.

[0129] The dynamic pressure wave acquisition and processing module acquires water flow state data through sensors installed in the water supply pipe network, and obtains a water flow dataset SW.

[0130] The perturbation feature extraction module extracts features from the water flow dataset SW, including a spectrum splitting distance Cse, a phase mutation rate Vds and a perturbation root mean square amplitude Sa, and constructs a perturbation comprehensive index RDZ.

[0131] The pressure wave identification and classification module combines the obtained perturbation comprehensive index RDZ with the water flow dataset SW, compares them with a preset interference sensitivity threshold Thm and an abnormal drift rate Ysb, and obtains a risk time window set Wr.

[0132] The energy-saving scheduling window prediction module combines the obtained risk time window set Wr with a pump station energy consumption Es and a perturbation change trend Rsp, evaluates all non-disabled time periods, and obtains a high-risk time set Ω.

[0133] The intelligent start-stop execution module sends start-stop control instructions to a water pump controller according to the obtained risk time window set Wr and the high-risk time set Ω, and obtains actual start-stop behavior records Σon / off.

[0134] The efficiency evaluation and feedback optimization module quantitatively analyzes the energy-saving effect after execution according to the actual start-stop behavior records Σon / off, and obtains a comprehensive energy-saving and stability score Znps.

[0135] In this embodiment, a dynamic pressure wave acquisition and processing module and a perturbation feature extraction module are provided, a continuous perception and high-resolution extraction mechanism for micro-perturbation patterns is introduced for the first time, and fine identification of internal non-steady-state fluctuations of the water supply system is realized. This mechanism significantly enhances the pre-warning capability of the system to potential water hammer disturbances or frequency drift causes, effectively avoids the pump set start-stop operation falling into the high-risk disturbance window, and reduces the instability probability and scheduling impact of the hydraulic system.

[0136] The system constructs a scheduling judgment logic driven by an energy-saving scheduling window prediction module and a pressure wave identification and classification module, which can dynamically integrate hydraulic disturbance features and unit time energy consumption trends, realize the optimal configuration of the start-stop time window without sacrificing system stability. Compared with the traditional start-stop strategy relying on fixed upper and lower threshold values, the present application can significantly reduce the frequency of pump body idle operation and repeated start-stop actions, thereby ensuring smooth water pressure waveform while achieving greater energy saving.

[0137] By introducing an intelligent start-stop execution module and an efficiency evaluation and feedback optimization module, the system forms a complete behavior record-performance quantification-strategy correction closed loop path. Each start-stop behavior is structured and recorded, and the joint evaluation of energy-saving effect and disturbance response is performed at the end of the cycle to provide data-driven adaptive reference for the next round of strategy decision. This learning-type scheduling mechanism significantly improves the system's adaptability to uncertain load disturbances, improves scheduling accuracy and long-term energy efficiency performance.

[0138] Embodiment 2

[0139] This embodiment is an explanation and description in embodiment 1, please refer to Figure 1 and Figure 4 , in particular: the dynamic pressure wave acquisition and processing module includes a multi-dimensional disturbance data acquisition unit and a data preprocessing unit;

[0140] The multi-dimensional disturbance data acquisition unit acquires water flow state data through a sensor, including instantaneous pressure value Sp, micro-pressure peak number Sc, minimum frequency distance Sr of disturbance waveband frequency distribution, total amount of pressure wave spectrum phase jump Sz, root mean square amplitude of disturbance Sa, number of pump set frequency drift Sb and pump station energy consumption Es, and is fitted into an original data set QW;

[0141] The instantaneous pressure value Sp is acquired by a high-frequency pressure sensor;

[0142] The micro-pressure peak number Sc is acquired by a pressure wave mutation detector;

[0143] The minimum frequency distance Sr of disturbance waveband frequency distribution is acquired by an embedded spectrum collector;

[0144] The total amount of phase jumps of the pressure wave spectrum Sz is acquired by an embedded spectrum demodulator;

[0145] The root mean square amplitude of the disturbance Sa is acquired by a pressure signal RMS module;

[0146] The number of frequency drifts of the pump group Sb is acquired by a variable frequency drive electric signal listener;

[0147] The energy consumption Es of the pump station is acquired by a current and voltage sensor;

[0148] The data preprocessing unit cleans and normalizes the acquired original data set QW to obtain a water flow data set SW;

[0149] The cleaning includes an outlier processing, which removes outliers in the original data set QW by using a three-sigma method, and fills the removed outliers by a median interpolation processing method;

[0150] The normalization processing processes the original data set QW by using a Min-Max normalization method to obtain the water flow data set SW;

[0151] The water flow data set SW is obtained by the following formula:

[0152] ;

[0153] In the formula, SWo represents the oth data in the water flow data set SW, QWo represents the oth data in the original data set QW, maxQWo represents the peak value of the oth data in the original data set QW, and minQWo represents the valley value of the oth data in the original data set QW.

[0154] The perturbation feature extraction module includes a non-stationary spectrum analysis unit and a comprehensive disturbance modeling unit;

[0155] The non-stationary spectrum analysis unit is based on fast Fourier transform and continuous wavelet transform of the water flow data set SW, and identifies a spectrum structure cracking region, extracts a spectrum cracking distance Cse, a phase mutation rate Vds, and a root mean square amplitude of the disturbance Sa;

[0156] The spectrum cracking distance Cse is obtained by the following formula:

[0157] ;

[0158] In the formula, σSp represents the standard deviation of the instantaneous pressure value, η represents a non-equilibrium modulation coefficient, μΔf represents an average frequency distance between main frequencies, Δfmax represents a maximum frequency distance between main frequencies, and C represents a zero prevention constant;

[0159] The phase mutation rate Vds is obtained by the following formula:

[0160] ;

[0161] In the formula, n represents the total number of time points collected, T represents the total observation time window length, represents the dominant frequency phase value of the k+1th time point, represents the dominant frequency phase value of the kth time point, tk+1 represents the k+1th time point, and tk represents the kth time point.

[0162] The comprehensive disturbance modeling unit receives the acquired spectral splitting distance Cse, the phase mutation rate Vds and the disturbance root mean square amplitude Sa, and constructs a disturbance comprehensive index RDZ using a nonlinear structure.

[0163] The disturbance comprehensive index RDZ is obtained by the following formula:

[0164] ;

[0165] In the formula, ln represents the natural logarithm function, tanh represents the hyperbolic tangent function, C1 represents a positive number, and C2 represents an interference suppression constant.

[0166] Specific examples:

[0167] Table 1: Disturbance comprehensive index calculation table

[0168] In this embodiment, a plurality of non-traditional data parameters are introduced by the multi-dimensional disturbance data acquisition unit, including the instantaneous pressure value Sp, the number of micro pressure spikes Sc, the minimum frequency distance Sr of the disturbance waveband frequency distribution, the total amount of pressure wave spectrum phase jumps Sz, the disturbance root mean square amplitude Sa, the number of pump group frequency drifts Sb and the pump station energy consumption Es, which far exceeds the mode of collecting only flow and pressure values in traditional systems, significantly improving the system's ability to identify potential water sources, pulsating disturbance patterns and high-frequency abnormal signals. At the same time, all parameter acquisition is completed through professional sub-modules, ensuring high-precision coverage of data in space and time dimensions. This multi-dimensional perception capability lays a solid foundation for the system to build a stable scheduling strategy, effectively breaking through the dilemma of "scheduling lag and disturbance response disconnection" of existing systems.

[0169] The present application adopts a standard deviation outlier cleaning mechanism combined with a median interpolation processing method, effectively eliminates abnormal data offsets caused by sensor distortion, sudden interference or echo superposition, and ensures the robustness of the original data set QW. Further introducing the Min-Max normalization method makes the data parameters of different dimensions and different scales have a unified dimension, providing a good mathematical foundation for subsequent spectral analysis and model fitting.

[0170] The continuous wavelet transform and Fourier transform fusion strategy is introduced by the non-steady spectrum analysis unit, the system can identify the asymmetric spectrum splitting structure in the hydraulic system, so as to capture the intermediate frequency discontinuity, pulsation group aggregation, phase jump and other phenomena which are difficult to find by traditional frequency domain method. The non-linear disturbance index RDZ is constructed on the basis of the comprehensive disturbance modeling unit, realizing the multi-factor fusion evaluation mode. Compared with the traditional disturbance judgment method using single index such as mean square deviation or maximum peak, RDZ has stronger classification and discrimination ability, and provides a scientific basis for subsequent risk time window analysis and intelligent start-stop scheduling.

[0171] Embodiment 3

[0172] This embodiment is an explanation and description in embodiment 2, please refer to Figure 1 and Figure 3 , specifically: the pressure wave recognition and classification module includes a dynamic discrimination index fusion unit and a risk time window screening unit;

[0173] The dynamic discrimination index fusion unit fuses and models the disturbance comprehensive index RDZ and the pump group frequency drift times Sb in the water flow data set SW, and constructs the discrimination driving index Arisk.

[0174] The discrimination driving index Arisk is obtained by the following formula:

[0175] ;

[0176] In the formula, Thm represents the interference sensitivity threshold, and Ysb represents the abnormal drift rate.

[0177] The interference sensitivity threshold Thm is obtained by the following formula:

[0178] ;

[0179] In the formula, μRDZ represents the mean of the disturbance comprehensive index, σRDZ represents the standard deviation of the disturbance comprehensive index, and α1 represents the adjustment coefficient.

[0180] The abnormal drift rate Ysb is obtained by the following formula:

[0181] ;

[0182] In the formula, μSb represents the mean of the historical drift times, σSb represents the standard deviation of the drift times, and α2 represents the drift fluctuation tolerance coefficient.

[0183] The risk time window screening unit dynamically compares the discriminant driving index Arisk at each time point, screens out the time period exceeding the activation threshold, and combines the interference sensitivity threshold Thm and the abnormal drift rate Ysb to form a time index set as a risk time window set Wr, and logically discriminates the disturbance type occurring at the time ti to obtain the disturbance risk mechanism;

[0184] ;

[0185] In the formula, Arisk (ti) represents the discriminant driving index at time ti, and Task represents the preset risk index activation threshold;

[0186] The disturbance risk mechanism is obtained by the following formula:

[0187] ;

[0188] In the formula, Type (ti) represents the risk disturbance value at time ti, RDZ (ti) represents the disturbance comprehensive index at time ti, and Sb (ti) represents the pump group frequency drift frequency at time ti;

[0189] When the risk disturbance value Type (ti) at time ti is 1, it represents a structural disturbance;

[0190] When the risk disturbance value Type (ti) at time ti is 2, it represents a behavioral disturbance;

[0191] When the risk disturbance value Type (ti) at time ti is 3, it represents a coupling aggravation disturbance.

[0192] The embodiment firstly combines the disturbance comprehensive index RDZ and the pump group frequency drift frequency Sb through the dynamic discriminant index fusion unit, models the discriminant driving index Arisk, realizes the collaborative discrimination of the pressure wave disturbance trend and the frequency instability event, and effectively makes up for the single-dimensional judgment deficiency of the traditional scheduling system that only relies on the pressure transient amplitude. The design not only improves the sensitivity of the system to high-frequency disturbance causes and abnormal start-stop phenomena, but also significantly enhances the comprehensive recognition ability of the system to multiple types of abnormal aggravation signals, avoiding the occurrence of misjudgment and omission.

[0193] The present application utilizes historical statistical characteristics to construct two dynamic threshold indicators, the interference sensitivity threshold Thm and the abnormal drift rate Ysb, and no longer relies on fixed upper and lower limits or artificial set experience rules. This threshold strategy can automatically adjust the early warning sensitivity according to the long-term change trend of the disturbance mode and the stable interval of the system operation, thereby improving the recognition ability of the system to non-typical behaviors under sudden disturbance and fatigue accumulation. The mechanism is particularly effective under the background of frequent night load fluctuations or system aging, and can prevent unnecessary frequent start-stop operations of the scheduling strategy due to "rigid discrimination".

[0194] The logical discrimination function introduced by the risk time window screening unit further subdivides the current system disturbance into three typical modes: structural, behavioral, and coupling aggravation, effectively establishing a bridge from pressure wave recognition to dispatch execution. This disturbance risk mechanism not only improves the accuracy of the system's early warning of potential high-risk time periods, but also supports the subsequent module in adopting differentiated response strategies for different types of disturbances, solving the unnecessary loss and instability risk caused by the "one-size-fits-all dispatch" of existing systems.

[0195] Embodiment 4

[0196] This embodiment is an explanation and illustration in Embodiment 3, please refer to Figure 1 and Figure 6 Specifically, the energy-saving dispatch window prediction module includes an energy consumption disturbance trend analysis unit and a dispatch window identification unit.

[0197] The energy consumption disturbance trend analysis unit constructs a disturbance evolution function based on the pump station energy consumption Es and the disturbance change trend Rsp.

[0198] The disturbance evolution function is obtained by the following formula:

[0199] ;

[0200] In the formula, B1 represents the energy consumption amplification factor, B2 represents the trend amplification factor, d represents the integral symbol, Es(t) represents the pump station energy consumption at time t, and Rsp(t) represents the disturbance change trend at time t, which is obtained by calculating the second derivative of the instantaneous pressure value Sp.

[0201] The dispatch window identification unit dynamically scores all remaining dispatchable time t∉Wr after excluding the risk time window set Wr based on the obtained disturbance evolution function, and obtains a dispatch optimization index function.

[0202] The dispatch optimization index function is obtained by the following formula:

[0203] ;

[0204] In the formula, Φ(t) represents the dispatch optimization index function at time t, represents infinity.

[0205] A high-risk time set is constructed by the dispatch optimization index function.

[0206] The high-risk time set is obtained by the following formula:

[0207] ;

[0208] In the formula, Φ(ti) represents the scheduling optimization exponential function at time ti, μΦ represents the mean of the scheduling optimization exponential function, σΦ represents the standard deviation of the scheduling optimization exponential function, and λ represents an empirical factor for adjusting the threshold sensitivity.

[0209] The intelligent start-stop execution module comprises a start-stop discrimination and command generation unit and a start-stop execution and behavior recording unit.

[0210] The start-stop discrimination and command generation unit performs scheduling execution judgment on each time t, acquires a start-stop state instruction generation function U according to whether it belongs to the risk time window set Wr or the high-risk time set Ω, and executes different control strategies.

[0211] The start-stop state instruction generation function U is acquired through the following formula:

[0212]

[0213] The control strategy is matched and acquired through the following mode:

[0214] When the start-stop state instruction generation function U = 1, it indicates that the start-stop is allowed to be executed.

[0215] When the start-stop state instruction generation function U = 0, it indicates that the control is disabled or the execution is delayed.

[0216] When the start-stop state instruction generation function U = X, it indicates other states, and the status quo is maintained.

[0217] The start-stop execution and behavior recording unit receives the start-stop state instruction generation function U, schedules the corresponding pump group for start-stop control, records the start-stop behavior data, forms a complete operation trajectory, and acquires actual start-stop behavior records Σon / off.

[0218] The actual start-stop behavior records Σon / off are acquired through the following formula:

[0219]

[0220] In the formula, ton(j) represents the jth pump start time, toff(j) represents the jth pump stop time, Δt(j) represents the jth pump operation duration, ID(j) represents the pump group number involved in the jth operation, Eon(j) represents the cumulative energy consumption of the jth operation from start to stop, and N represents the total number of operations.

[0221] ​​The embodiment introduces an energy consumption disturbance trend analysis unit, establishes a dynamic disturbance evolution function Ψ jointly constructed by "pump station unit time energy consumption" and "pressure disturbance change rate", and can accurately capture the "high energy consumption interval induced by load fluctuation" and "superimposed interval of sudden disturbance" in the pump station operation, thereby significantly improving the identification ability of the system to the high-risk period. Compared with the traditional scheduling mode which only relies on the fixed start-stop cycle or the rough upper and lower limit judgment, the scheme provides a dynamic response, multi-factor driven sensing framework, and fundamentally enhances the quantitative analysis ability of the scheduling system to the coupling relationship between the pressure waveform change and the energy consumption pulsation.

[0222] The scheduling window identification unit designed in the embodiment gives a dynamic score to the remaining time region after dynamic time elimination according to the disturbance evolution result and the identified risk time window, and extracts the most suitable start-stop window. This method overcomes the problem of being unable to avoid the mis-triggering of the water hammer excitation window in the traditional "forced execution strategy", and realizes the adaptive optimization of the scheduling behavior. The finally generated "scheduling optimization index" is not only used to identify the start-stop priority, but also supports the system to compress the high energy consumption running period on demand, realizes the synchronous satisfaction of the dynamic balance of supply and demand and the energy saving goal, and greatly improves the matching degree of the "stability-energy saving" double indicators of the scheduling behavior.

[0223] The embodiment forms a logical closed loop between the "start-stop discrimination and command generation unit" and the "start-stop execution and behavior recording unit", and for the first time proposes a state function mechanism of start-stop control, which outputs standardized instructions such as "allow execution", "forcibly prohibit" and "maintain the original state". This module not only significantly reduces the risk of frequent start-stop caused by false scheduling of the system, but also realizes the whole process recording of the execution action, including start-stop time, pump group number, running time and actual energy consumption. The complete scheduling behavior database constructed by the system provides a structured basis for subsequent strategy learning, adaptive adjustment and abnormal backtracking, breaking through the technical limitations of the "black box" of the execution behavior in the existing system.

[0224] Embodiment 5

[0225] The embodiment is an explanation and description in embodiment 4. Please refer to Figure 3 and Figure 5 , specifically: the performance evaluation and feedback optimization module includes an energy saving effect evaluation unit and a disturbance stability evaluation unit;

[0226] The energy saving effect evaluation unit calculates the cumulative energy saving amount realized in the current scheduling period based on the cumulative energy consumption Eon of the start-to-stop process, and obtains the unit time energy saving rate η save.

[0227] The unit time energy saving rate η save is obtained by the following formula:

[0228] ;

[0229] In the formula, Tw represents the total time of the current scheduling period, and Eref represents the reference energy consumption.

[0230] The disturbance stability evaluation unit analyzes the influence of pump start-stop execution on the hydraulic disturbance of the system, and constructs a disturbance mean square response index Ψdi based on the disturbance mean square index formed by the instantaneous pressure value Sp before and after the start-stop operation.

[0231] The disturbance mean square response index Ψdi is obtained by the following formula:

[0232]

[0233] In the formula, Sp(t) represents the instantaneous pressure value at time t, represents the second derivative of the instantaneous pressure value;

[0234] The comprehensive energy saving and stability score Znps is calculated by the obtained unit time energy saving rate ηsave and the mean square response index Ψdi.

[0235] The comprehensive energy saving and stability score Znps is obtained by the following formula:

[0236]

[0237] In the formula, KS represents the interference penalty factor.

[0238] The energy saving effect evaluation unit introduces the unit time energy saving rate ηsave as the core evaluation index, which can accurately measure the energy saving benefit brought by each round of start-stop control behavior. Unlike the traditional system which only roughly counts the running time or total energy consumption, the present scheme refines the energy consumption characteristics according to the start-stop period, so that the energy saving effect not only has comparative significance, but also has feedback significance. This mechanism significantly improves the energy control transparency of the system under dynamic working conditions, so that the scheduling strategy can be iterated and self-optimized according to quantitative data, solving the problem of "unobservable energy saving effect and unclear optimization direction" in the traditional scheme.

[0239] The system sets up a disturbance stability evaluation unit, which constructs a quantitative index of disturbance response by monitoring the second-order change of the instantaneous pressure value before and after the start-stop. This design first takes the system disturbance caused by the start-stop strategy as part of the strategy evaluation, guiding the system to balance the stability of the hydraulic system while pursuing energy saving. This "double-index evaluation" mode greatly overcomes the risks of "water hammer rebound" and "system oscillation" caused by excessive control frequency in traditional energy saving optimization methods, and truly realizes the coordinated optimization of "energy saving-safety-stability".

[0240] ​​The embodiment fuses the energy saving rate and the disturbance response index by function to form a comprehensive energy saving and stability score Znps, which provides a quantifiable, orderable and comparable evaluation basis for the whole system scheduling strategy. The score mechanism is not only used for result display, but also participates in the strategy modification of the next scheduling period as a feedback variable, realizing the self-learning and closed-loop evolution of the scheduling model. This mechanism enables the system to adapt to complex factors such as environmental changes and load structure fluctuations in long-term operation, avoiding performance degradation caused by strategy rigidity and fixed configuration.

[0241] Embodiment 6

[0242] An efficient energy-saving integrated control method for secondary water supply pipe network, please refer to Figure 2 , specifically: comprising the following steps:

[0243] Step one: the dynamic pressure wave acquisition and processing module acquires water flow state data through the sensor installed in the water supply pipe network, and obtains a water flow data set SW;

[0244] Step two: the micro-perturbation feature extraction module extracts features from the water flow data set SW, including the frequency spectrum splitting distance Cse, the phase mutation rate Vds and the disturbance root mean square amplitude Sa, and constructs a disturbance comprehensive index RDZ;

[0245] Step three: the pressure wave recognition and classification module combines the obtained disturbance comprehensive index RDZ with the water flow data set SW, compares it with the preset interference sensitivity threshold Thm and abnormal drift rate Ysb, and obtains a risk time window set Wr;

[0246] Step four: the energy-saving scheduling window prediction module combines the obtained risk time window set Wr with the pump station energy consumption Es and the disturbance change trend Rsp to evaluate all non-disabled time periods, and obtains a high-risk time set Ω;

[0247] Step five: the intelligent start-stop execution module sends start-stop control instructions to the water pump controller according to the obtained risk time window set Wr and the high-risk time set Ω, and obtains actual start-stop behavior records Σon / off;

[0248] Step six: the efficiency evaluation and feedback optimization module quantitatively analyzes the energy-saving effect after execution according to the actual start-stop behavior records Σon / off, and obtains a comprehensive energy saving and stability score Znps.

[0249] In this embodiment, the method takes dynamic pressure wave acquisition and processing as the starting point, and gradually progresses to energy-saving control instruction generation and performance evaluation, building a six-step functional chain that effectively overcomes the problems of data segmentation processing, control instruction disconnection, and inability to continuously feedback in traditional methods. The logical relationship between its modules is close, ensuring that from the early detection of hydraulic disturbance, impact quantification, to the development and landing of scheduling strategies, the entire process has logical traceability and execution closed-loop, achieving steady-state scheduling optimization under multi-module collaboration from the system design.

[0250] This embodiment emphasizes the identification of scheduling windows based on disturbance behavior characteristics, that is, not passively responding to fluctuations caused by pump start-stop, but actively avoiding high-sensitivity disturbance periods. Through the feature extraction module, a disturbance comprehensive index RDZ is constructed, and combined with pump group operation drift behavior and historical risk windows, potential hydraulic abnormal area is accurately identified. This mechanism ensures that pump start-stop behavior avoids the "water hammer excitation area" or "pressure echo coupling window", effectively reducing system energy consumption while reducing equipment fatigue load, avoiding unnecessary energy consumption and safety hazards caused by traditional fixed-time start-stop schemes.

[0251] Through the efficiency evaluation and feedback optimization module, the method not only quantifies the energy-saving effect, but also analyzes the disturbance impact caused by start-stop control, and builds a unified comprehensive performance index for subsequent strategy correction. This feedback design enables the system to continuously calibrate parameters and optimize rhythm in practice, achieving long-term adaptive evolution of the strategy. Compared with the traditional system where scheduling strategies remain unchanged and are difficult to adapt to system aging or changes in user water behavior, the present invention ensures the continuous evolution and adaptability of the scheduling control strategy while maintaining low energy consumption of the system.

[0252] Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A high efficiency energy saving integrated control system for secondary water supply network, characterized in that: The dynamic pressure wave acquisition and processing module, the perturbation feature extraction module, the pressure wave identification and classification module, the energy-saving scheduling window prediction module, the intelligent start-stop execution module and the performance evaluation and feedback optimization module are included. The dynamic pressure wave acquisition and processing module acquires water flow state data through sensors installed in the water supply network to obtain a water flow dataset SW. The perturbation feature extraction module extracts features from the water flow dataset SW, including a spectral splitting distance Cse, a phase mutation rate Vds and a disturbance root mean square amplitude Sa, and constructs a disturbance comprehensive index RDZ. The pressure wave identification and classification module combines the acquired disturbance comprehensive index RDZ with the water flow dataset SW, compares them with a preset interference sensitivity threshold Thm and an abnormal drift rate Ysb, and obtains a risk time window set Wr. The energy-saving scheduling window prediction module combines the acquired risk time window set Wr with a pump station energy consumption Es and a disturbance change trend Rsp, evaluates all non-disabled time periods, and obtains a high-risk time set Ω. The intelligent start-stop execution module sends start-stop control instructions to a water pump controller according to the acquired risk time window set Wr and the high-risk time set Ω, and obtains actual start-stop behavior records Σon / off. The performance evaluation and feedback optimization module quantitatively analyzes the energy-saving effect after execution according to the actual start-stop behavior records Σon / off, and obtains a comprehensive energy-saving and stability score Znps.

2. The energy-efficient integrated control system for secondary water supply network according to claim 1, characterized in that: The dynamic pressure wave acquisition and processing module includes a multi-dimensional disturbance data acquisition unit and a data preprocessing unit. The multi-dimensional disturbance data acquisition unit acquires water flow state data through sensors, including an instantaneous pressure value Sp, a micro pressure peak number Sc, a minimum frequency distance Sr of disturbance waveband frequency distribution, a total amount of pressure wave spectrum phase jumps Sz, a disturbance root mean square amplitude Sa, a pump group frequency drift number Sb and a pump station energy consumption Es, and fits them into an original dataset QW. The instantaneous pressure value Sp is acquired by a high-frequency pressure sensor. The micro pressure peak number Sc is acquired by a pressure wave mutation detector. The minimum frequency distance Sr of disturbance waveband frequency distribution is acquired by an embedded spectrum collector. The total amount of pressure wave spectrum phase jumps Sz is acquired by an embedded spectrum demodulator. The disturbance root mean square amplitude Sa is acquired by a pressure signal RMS module. The pump group frequency drift number Sb is acquired by a variable frequency drive electric signal listener. The pump station energy consumption Es is acquired by a current and voltage sensor. The data preprocessing unit cleans and normalizes the acquired original dataset QW to obtain a water flow dataset SW. The cleaning includes outlier processing, which removes outliers in the original dataset QW using a three-sigma method and fills in the removed outliers using a median interpolation method. The normalization processing processes the original dataset QW using a Min-Max normalization method to obtain the water flow dataset SW. The water flow dataset SW is obtained by the following formula: ； In the formula, SWo represents the oth data in the water flow data set SW, QWo represents the oth data in the original data set QW, maxQWo represents the peak value of the oth data in the original data set QW, and minQWo represents the valley value of the oth data in the original data set QW.

3. The energy-efficient integrated control system for secondary water supply network according to claim 2, characterized in that: The perturbation feature extraction module comprises a non-steady-state spectrum analysis unit and a comprehensive disturbance modeling unit. The non-steady-state spectrum analysis unit is based on fast Fourier transform and continuous wavelet transform on the water flow data set SW, and identifies the frequency spectrum structure cracking area, extracts the frequency spectrum cracking distance Cse, the phase mutation rate Vds and the disturbance root mean square amplitude Sa. The frequency spectrum cracking distance Cse is obtained by the following formula: ； In the formula, σSp represents the standard deviation of the instantaneous pressure value, η represents the non-equilibrium modulation coefficient, μΔf represents the average frequency distance between the main frequencies, Δfmax represents the maximum frequency distance between the main frequencies, and C represents the anti-zero constant. The phase mutation rate Vds is obtained by the following formula: ； In the formula, n represents the total number of time points collected, T represents the total observation time window length, represents the dominant frequency phase value at the k+1th time point, represents the dominant frequency phase value at the kth time point, tk+1represents the k+1th time point, and tkrepresents the kth time point. The comprehensive disturbance modeling unit receives the obtained frequency spectrum cracking distance Cse, the phase mutation rate Vds and the disturbance root mean square amplitude Sa, and constructs a disturbance comprehensive index RDZ using a nonlinear structure. The disturbance comprehensive index RDZ is obtained by the following formula: ； In the formula, ln represents the natural logarithm function, tanh represents the hyperbolic tangent function, C1 represents a positive number, and C2 represents an interference suppression constant.

4. The energy-efficient integrated control system for secondary water supply network according to claim 3, characterized in that: The pressure wave identification and classification module comprises a dynamic discriminant index fusion unit and a risk time window screening unit. The dynamic discriminant index fusion unit fuses and models the obtained disturbance comprehensive index RDZ and the pump group frequency drift times Sb in the water flow data set SW to construct a discriminant driving index Arisk. The discriminant driving index Arisk is obtained by the following formula: ； In the formula, Thm represents the interference sensitivity threshold, and Ysb represents the abnormal drift rate. The interference sensitivity threshold Thm is obtained by the following formula: ； In the formula, μRDZ represents the mean of the disturbance comprehensive index, σRDZ represents the standard deviation of the disturbance comprehensive index, and α1 represents an adjustment coefficient. The abnormal drift rate Ysb is obtained by the following formula: ； In the formula, μSb represents the mean of the historical drift times, σSb represents the standard deviation of the drift times, and α2 represents a drift fluctuation tolerance coefficient.

5. The energy efficient integrated control system for secondary water supply network as claimed in claim 4 wherein: The risk time window screening unit dynamically compares the discriminant driving index Arisk at each time point, screens out the time period exceeding the activation threshold, and combines it with the interference sensitivity threshold Thm and the abnormal drift rate Ysb to form a time index set as a risk time window set Wr, and logically discriminates the disturbance type occurring at time ti to obtain a disturbance risk mechanism. ； In the formula, Arisk(ti) represents the discriminant driving index at time ti, and Task represents a preset risk index activation threshold. The disturbance risk mechanism is obtained by the following formula: ； In the formula, Type(ti) represents the risk disturbance value at time ti, RDZ(ti) represents the disturbance comprehensive index at time ti, and Sb(ti) represents the pump group frequency drift times at time ti. When the risk disturbance value Type(ti) at time ti is 1, it represents a structural disturbance. When the risk disturbance value Type (ti) = 2 at time ti, it represents the behavior type disturbance; When the risk disturbance value Type (ti) = 3 at time ti, it represents the coupling intensification type disturbance.

6. The energy-efficient integrated control system for secondary water supply network according to claim 5, characterized in that: The energy-saving scheduling window prediction module includes an energy consumption disturbance trend analysis unit and a scheduling window identification unit; The energy consumption disturbance trend analysis unit constructs a disturbance evolution function based on the pump station energy consumption Es and the disturbance change trend Rsp; The disturbance evolution function is obtained by the following formula: ； In the formula, B1 represents an energy consumption amplification factor, B2 represents a trend amplification factor, d represents an integral symbol, Es(t) represents the pump station energy consumption at time t, and Rsp(t) represents the disturbance change trend at time t, which is obtained by calculating the second derivative of the instantaneous pressure value Sp; The scheduling window identification unit dynamically scores all remaining schedulable time t not in the risk window set Wr based on the obtained disturbance evolution function after excluding the risk window set Wr, and obtains a scheduling optimization index function; The scheduling optimization index function is obtained by the following formula: ； In the formula, Φ(t) represents a scheduling optimization exponential function at time t, represents infinity. A high-risk time set is constructed by the scheduling optimization index function; The high-risk time set is obtained by the following formula: ； In the formula, Φ(ti) represents the scheduling optimization index function at time ti, μΦ represents the mean of the scheduling optimization index function, σΦ represents the standard deviation of the scheduling optimization index function, and λ represents an empirical factor adjusting the threshold sensitivity.

7. The energy-efficient integrated control system for secondary water supply network according to claim 6, characterized in that: The intelligent start-stop execution module includes a start-stop discrimination and command generation unit and a start-stop execution and behavior recording unit; The start-stop discrimination and command generation unit judges the scheduling execution of each time t, obtains a start-stop state instruction generation function U according to whether it belongs to the risk window set Wr or the high-risk time set, and executes different control strategies; The start-stop state instruction generation function U is obtained by the following formula: ； The control strategy is matched and obtained by the following method: When the start-stop state instruction generation function U = 1, it means that the start-stop is allowed to be executed; When the start-stop state instruction generation function U = 0, it means that the control is disabled or delayed execution; When the start-stop state instruction generation function U = X, it means other states, maintaining the status quo.

8. The energy efficient integrated control system for secondary water supply network as claimed in claim 7 wherein: The start-stop execution and behavior recording unit receives the start-stop state instruction generation function U, schedules the corresponding pump group for start-stop control, records the start-stop behavior data, forms a complete operation trajectory, and obtains the actual start-stop behavior record Σon / off; The actual start-stop behavior record Σon / off is obtained by the following formula: ； In the formula, ton(j) represents the jth start pump time, toff(j) represents the jth stop pump time, Δt(j) represents the jth pump operation duration, ID(j) represents the pump group number involved in the jth operation, Eon(j) represents the cumulative energy consumption from the jth operation start to stop, and N represents the total number of operations.

9. The energy efficient integrated control system for secondary water supply network as claimed in claim 8 wherein: The performance evaluation and feedback optimization module includes an energy-saving effect evaluation unit and a disturbance stability evaluation unit; The energy-saving effect evaluation unit calculates the cumulative energy saving amount achieved in the current scheduling period based on the cumulative energy consumption Eon from start to stop, and obtains the unit time energy saving rate ηsave; The unit time energy saving rate ηsave is obtained by the following formula: ； In the formula, Tw represents the total time of the current scheduling period, and Eref represents the reference energy consumption. The disturbance stationarity evaluation unit analyzes the influence of pump start-stop execution on the hydraulic disturbance of the system, and forms a disturbance mean square index based on the instantaneous pressure value Sp before and after the start-stop operation, to construct a disturbance mean square response index Ψdi. The disturbance mean square response index Ψdi is obtained through the following formula: ； where Sp(t) represents the instantaneous pressure value at time t, represents the second derivative of the instantaneous pressure value; Through the obtained unit time energy saving rate ηsave and the mean square response index Ψdi, the comprehensive energy saving and stationarity score Znps is calculated and obtained. The comprehensive energy saving and stationarity score Znps is obtained through the following formula: ； In the formula, KS represents an interference penalty factor.

10. An energy-efficient integrated control method for a secondary water supply network, applied to the energy-efficient integrated control system for a secondary water supply network according to any one of claims 1-9, characterized in that: The method comprises the following steps: Step one: the dynamic pressure wave acquisition and processing module acquires water flow state data through sensors installed in the water supply network, to obtain a water flow data set SW; Step two: the micro-perturbation feature extraction module extracts features from the water flow data set SW, including a spectrum splitting distance Cse, a phase mutation rate Vds, and a disturbance root mean square amplitude Sa, and constructs a disturbance comprehensive index RDZ; Step three: the pressure wave recognition and classification module combines the obtained disturbance comprehensive index RDZ with the water flow data set SW, compares them with a preset interference sensitivity threshold Thm and an abnormal drift rate Ysb, to obtain a risk time window set Wr; Step four: the energy-saving scheduling window prediction module combines the obtained risk time window set Wr with a pump station energy consumption Es and a disturbance change trend Rsp, to evaluate all non-forbidden time periods, and obtain a high-risk time set Ω; Step five: the intelligent start-stop execution module sends start-stop control instructions to a water pump controller according to the obtained risk time window set Wr and the high-risk time set Ω, to obtain actual start-stop behavior records Σon / off; Step six: the efficiency evaluation and feedback optimization module quantitatively analyzes the energy-saving effect after execution according to the actual start-stop behavior records Σon / off, to obtain a comprehensive energy saving and stationarity score Znps.

Images