Drainage pump station energy consumption prediction method and system based on operation data analysis

Abstract

This invention discloses a method and system for predicting the energy consumption of drainage pumping stations based on operational data analysis, belonging to the field of energy consumption prediction technology for drainage pumping stations. The method first decouples the historical operating cycle into start-stop transient and steady-state operating stages based on collected high-frequency operational data. Building upon this, it extracts the transient baseline energy consumption characteristics by averaging the instantaneous power integrals of the transient stage and extracts the steady-state average power characteristics by averaging the power integrals of the steady-state stage. Furthermore, it constructs a transient degradation correction coefficient using an attenuation weighting function. Finally, combining the planned number of start-stop cycles and the steady-state operating duration, it calculates and superimposes the transient and steady-state total energy consumption to obtain the target predicted total energy consumption. This invention avoids errors caused by averaging by independently modeling the high-energy-consumption process during start-stop transients and introducing an equipment degradation correction mechanism. It dynamically quantifies the impact of equipment degradation, accurately assesses the true energy consumption cost of frequent start-stop cycles, and improves the accuracy and reliability of pumping station energy consumption prediction.

Description

Technical Field

[0001] This invention relates to the field of energy consumption prediction technology for drainage pumping stations, and in particular to a method and system for predicting energy consumption of drainage pumping stations based on operational data analysis. Background Technology

[0002] With the acceleration of urbanization, the scale of urban drainage systems is constantly expanding. As a core facility, drainage pumping stations account for a significant proportion of urban public energy consumption. In recent years, the development of the Internet of Things, sensor technology, and intelligent control systems has enabled pumping stations to collect multi-dimensional data such as water level, flow rate, pump status, and power consumption in real time, making energy consumption analysis and optimization possible.

[0003] In existing technologies, energy consumption prediction for drainage pumping stations typically begins with monitoring equipment within the pumping station and SCADA (Supervisory Control and Data Acquisition) systems. The Acquisition (data acquisition and monitoring control system) collects real-time operational data such as pump start-up and shutdown status, operating power, water flow rate, head, water level, and external environmental parameters to form a historical operational database. Secondly, the collected raw data undergoes preprocessing, including outlier removal, missing data completion, time series alignment, and data standardization to improve data quality and usability. Subsequently, after data processing, feature analysis is performed on the operational data to extract key parameters affecting pump station energy consumption, such as flow rate changes, pump load rate, head changes, and start-up / shutdown frequency, establishing a set of energy consumption impact features. Then, an energy consumption prediction model is constructed based on historical operational data. Existing technologies typically employ linear regression, time series analysis, or machine learning algorithms to train the model on the relationship between pump station operating parameters and energy consumption. After model training, real-time operational data is input into the prediction model, outputting energy consumption prediction results for the corresponding time period, and error analysis is used to verify the model's prediction accuracy. Finally, based on the prediction results, the pump station's operating strategy is optimized and adjusted, such as optimizing pump start-up / shutdown sequence, load allocation, and operating periods, to reduce overall energy consumption and improve operational efficiency.

[0004] However, in the process of implementing the inventive technical solution in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems: Existing technologies for energy consumption feature extraction typically treat "start-stop frequency" as a discrete, coarse-grained statistical feature and average the operating power over a time scale (e.g., taking minute-level or hourly averages). However, large drainage pumps generate extremely high inrush currents (up to 5-7 times the rated current) at startup, and the transient nonlinear process of accelerating from a standstill to rated speed exhibits extremely low motor efficiency and extremely high instantaneous energy consumption. Under the specific operating conditions of frequent pump start-stop operations during heavy rain and flood seasons, the averaging process in existing technologies essentially treats start-stop operations as stationary statistical events, masking the transient nonlinear high-energy consumption characteristics during these processes. This leads to a systematic underestimation of errors in model predictions, making it impossible to accurately quantify and assess the true energy load and operational costs of frequent start-stop operations on the pumping station as a whole. Summary of the Invention

[0005] To address the technical problem in existing technologies for predicting the energy consumption of large drainage pumps under frequent start-stop conditions, where time-scale averaging of operating power masks the transient nonlinear high-energy consumption characteristics during start-stop operations, resulting in systematically low predictions and an inability to accurately quantify the actual energy load, this invention provides a method and system for predicting the energy consumption of drainage pumping stations based on operational data analysis. The technical solution is as follows: On the one hand, a method for predicting the energy consumption of drainage pumping stations based on operational data analysis is provided. This method includes: collecting high-frequency operational data of the drainage pumps during historical operating cycles, including instantaneous current, instantaneous voltage, and instantaneous speed; identifying the operating state of the drainage pumps based on the rate of change of instantaneous speed, decoupling the historical operating cycle into a start-stop transient phase and a steady-state operating phase in the time dimension; for the start-stop transient phase within the historical operating cycle, calculating the instantaneous power based on instantaneous current and instantaneous voltage and integrating along the time axis to obtain the actual energy consumption of a single start-stop operation; averaging the actual energy consumption of a single start-stop operation across multiple start-stop transient phases to obtain the transient baseline energy consumption characteristics; for the steady-state operating phase, calculating the average power within this phase to obtain the steady-state average power characteristics; and obtaining the current operating energy consumption of the drainage pumps during the current operating cycle. The ratio of the peak starting inrush current to the rated current during the pre-start-stop transient phase is used to obtain the inrush multiple characteristic, and the time taken to accelerate from standstill to rated speed is used to obtain the acceleration time characteristic. Based on the inrush multiple characteristic and the acceleration time characteristic, a transient degradation correction coefficient is constructed using a preset attenuation weight function. The planned number of start-stop cycles and the planned steady-state running time within the prediction period are obtained. The single predicted transient energy consumption is calculated based on the transient baseline energy consumption characteristic and the transient degradation correction coefficient. The total start-stop transient energy consumption for the prediction period is calculated based on the planned number of start-stop cycles and the single predicted transient energy consumption. The total steady-state running energy consumption for the prediction period is calculated based on the steady-state average power characteristic and the planned steady-state running time. The total start-stop transient energy consumption and the total steady-state running energy consumption are superimposed to obtain the target predicted total energy consumption of the drainage pumping station.

[0006] On the other hand, an energy consumption prediction system for drainage pumping stations based on operational data analysis is provided, including: an operational data acquisition module, an operational phase decoupling module, an energy consumption feature extraction module, a transient correction analysis module, and an energy consumption prediction output module. The operational data acquisition module collects high-frequency operational data of the drainage pumps during historical operational cycles, including instantaneous current, instantaneous voltage, and instantaneous speed. The operational phase decoupling module identifies the operational status of the drainage pumps based on the rate of change of instantaneous speed, decoupling the historical operational cycle into a start-stop transient phase and a steady-state operational phase in the time dimension. The energy consumption feature extraction module calculates instantaneous power based on instantaneous current and instantaneous voltage during the start-stop transient phase within the historical operational cycle and integrates it along the time axis to obtain the true energy consumption for a single start-stop operation. It then averages the true energy consumption of a single start-stop operation across multiple start-stop transient phases to obtain the transient baseline energy consumption characteristics. For the steady-state operational phase, it calculates... The average power during this phase yields the steady-state average power characteristics. The transient correction analysis module obtains the ratio of the peak starting impact current to the rated current during the current start-stop transient phase of the drainage pump, yielding the impact multiple characteristics, and obtains the time taken to accelerate from standstill to rated speed, yielding the acceleration time characteristics. Based on the impact multiple characteristics and acceleration time characteristics, a transient degradation correction coefficient is constructed using a preset attenuation weight function. The energy consumption prediction output module obtains the planned number of start-stop cycles and the planned steady-state running time during the prediction period, calculates the single predicted transient energy consumption based on the transient baseline energy consumption characteristics and the transient degradation correction coefficient, calculates the total start-stop transient energy consumption for the prediction period based on the planned number of start-stop cycles and the single predicted transient energy consumption, calculates the total steady-state running energy consumption for the prediction period based on the steady-state average power characteristics and the planned steady-state running time, and superimposes the total start-stop transient energy consumption and the total steady-state running energy consumption to obtain the target predicted total energy consumption of the drainage pumping station.

[0007] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: This invention effectively overcomes the information loss problem caused by time-scale averaging in existing technologies by using high-frequency data-driven modeling of the operation process of drainage pumping stations and introducing a phased energy consumption analysis mechanism for start-stop transient and steady-state operation. In specific implementation, the operating cycle is decoupled based on the instantaneous speed change rate, separating the start-stop transient stage from the steady-state operation stage. This avoids the mutual masking of energy consumption characteristics under different physical mechanisms, achieving a structured expression of complex operating processes and improving the physical consistency and interpretability of energy consumption modeling.

[0008] Meanwhile, by performing power integration calculations based on instantaneous current and instantaneous voltage during the start-up and shutdown transient phases, the system can accurately capture the true energy consumption characteristics corresponding to the inrush current and low-efficiency acceleration phases during pump startup. Compared with traditional methods that use average power or low-frequency statistical characteristics, this effectively avoids the systematic underestimation problem caused by smoothing out transient high energy consumption, thereby significantly improving the refinement and accuracy of energy consumption assessment.

[0009] Furthermore, by constructing transient baseline energy consumption characteristics and steady-state average power characteristics, this invention achieves the separation and quantification of energy consumption contributions at different operating stages, enabling the model to separately characterize the energy consumption patterns during start-up and shutdown processes and continuous operation processes. This enhances the model's adaptability to complex operating conditions and improves the stability and generalization ability of the prediction results.

[0010] Furthermore, by introducing a transient degradation correction coefficient based on the impact current multiple and acceleration time, adaptive corrections can be made for different operating states and equipment conditions, enabling energy consumption prediction results to reflect the impact of equipment aging, load fluctuations and changes in operating conditions, thereby enhancing the robustness and engineering applicability of the model in long-term operating scenarios.

[0011] Finally, by calculating and superimposing the transient energy consumption during start-up and shutdown and the steady-state energy consumption during the prediction period, the total energy consumption is obtained, thus achieving a complete characterization of the energy consumption of the drainage pumping station throughout its entire operating cycle. This enables the prediction results to truly reflect the overall energy load level under frequent start-up and shutdown conditions, thereby significantly improving the accuracy, reliability, and practical application value of the energy consumption prediction for drainage pumping stations. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A flowchart of a drainage pumping station energy consumption prediction method based on operational data analysis provided in an embodiment of this application; Figure 2 A schematic diagram of the structure of a drainage pumping station energy consumption prediction system based on operational data analysis provided in an embodiment of this application; Figure 3 The overall module logic diagram of the drainage pumping station energy consumption prediction system based on operational data analysis provided in the embodiments of this application is shown. Detailed Implementation

[0014] Embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of the present disclosure are shown in the drawings, it should be understood that embodiments of the present disclosure may be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure.

[0015] It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure. In the description of the embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "this embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects.

[0016] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0017] In urban drainage systems, large-scale drainage pumping stations serve as critical infrastructure for ensuring safe flood control and drainage operations, and their operating energy consumption typically accounts for a significant proportion of the overall drainage system's operating costs. Especially during extreme rainfall or flood season conditions, pumping stations often require frequent start-stop scheduling based on changes in water inflow to meet rapid drainage demands. In these highly dynamic conditions, drainage pumps not only experience long periods of steady-state operation but also numerous transient start-stop processes. These start-stop phases typically exhibit significant electrical and mechanical nonlinear characteristics, such as a sharp increase in starting inrush current, low efficiency during motor acceleration, and drastic instantaneous power fluctuations. These characteristics make the energy consumption during this phase distinctly different from those during steady-state operation.

[0018] However, existing methods for analyzing and predicting energy consumption in drainage pumping stations are generally based on low-frequency sampling data or time-scale average power models. They typically use minute-level or even hour-level average power as the primary energy consumption characterization method, supplemented by discrete statistical indicators such as the number of start-stop cycles for auxiliary description. While this approach can reflect the overall operating trend of the pumping station to some extent, the lack of refined characterization of transient processes leads to the high energy consumption impact during start-stop phases being significantly smoothed out by the time averaging process. Therefore, it fails to accurately reflect the actual energy consumption cost under frequent start-stop conditions.

[0019] In actual operation, especially under complex conditions of frequent rainstorms and frequent dispatching, the proportion of transient energy consumption during start-up and shutdown in the total energy consumption of pumping stations increases significantly, which has an important impact on the overall energy consumption level and operational economy. Ignoring the contribution of this nonlinear transient energy consumption will lead to systematic biases in energy consumption prediction results, thereby affecting the formulation of optimized dispatching strategies for pumping station operation and reducing the engineering guidance value of energy consumption assessment results.

[0020] Therefore, how to effectively decouple the start-up and shutdown transient and steady-state phases of drainage pump operation based on high-frequency operating data, accurately characterize the transient nonlinear energy consumption characteristics, and take into account the impact of equipment operating status changes and operating condition differences on energy consumption has become a key technical issue for improving the accuracy of energy consumption prediction and engineering applicability of drainage pumping stations.

[0021] like Figure 1 The diagram shown is a flowchart of a drainage pumping station energy consumption prediction method based on operational data analysis provided in an embodiment of this application. Figure 1 Based on this, it should be understood that, compared with existing technologies, this invention does not simply use start-stop frequency statistics or operating power over time as the basis for energy consumption analysis. Instead, it uses high-frequency operating data such as instantaneous current, instantaneous voltage, and instantaneous speed to perform refined, phased modeling of the drainage pump's operation process. Existing technologies typically treat start-stop processes as ordinary statistical events and estimate energy consumption using minute-level or hourly average power. This easily masks the physical details of high inrush current, high torque load, and low efficiency present at the moment of large drainage pump startup, thereby reducing the authenticity and accuracy of energy consumption prediction results in frequent start-stop scenarios.

[0022] It should also be understood that most existing technologies use fixed parameter models or static empirical coefficients for energy consumption prediction, lacking the ability to dynamically represent changes in equipment operating status and performance degradation factors. Therefore, they are difficult to adapt to energy consumption differences caused by equipment aging, load changes, and operating condition fluctuations during long-term operation. In contrast, this invention extracts the ratio of the peak starting inrush current to the rated current and the time to accelerate to the rated speed to construct a transient degradation correction coefficient, realizing dynamic correction of start-stop energy consumption. This allows the prediction model to adaptively adjust according to the actual operating status of the equipment, exhibiting higher robustness and adaptability to operating conditions. At the same time, existing technologies typically use an overall average energy consumption model to uniformly estimate the energy consumption of pumping stations, failing to distinguish the energy consumption contributions of start-stop losses and steady-state operation losses. This makes it difficult to accurately assess the impact of frequent start-stop strategies on the overall energy load of the pumping station. In contrast, this invention constructs transient baseline energy consumption characteristics and steady-state average power characteristics separately, calculates start-stop transient energy consumption and steady-state operation energy consumption separately, and superimposes them in the prediction stage, achieving a refined quantitative analysis of the energy consumption of the drainage pumping station throughout its entire operating cycle.

[0023] Therefore, this invention can not only improve the accuracy of energy consumption prediction, but also more realistically reflect the actual energy consumption cost under different scheduling strategies, different start-stop frequencies and different operating conditions, providing more reliable data support and decision-making basis for energy-saving optimization operation and scheduling control of pumping stations.

[0024] As the first step in the energy consumption prediction method for drainage pumping stations based on operational data analysis, the specific steps are as follows: collect high-frequency operational data of drainage pumps during historical operating cycles. The high-frequency operational data includes instantaneous current, instantaneous voltage, and instantaneous speed. This step introduces multi-source operational data with high temporal resolution, enabling energy consumption analysis to be based on the real dynamic operation process. This avoids the information loss problem caused by traditional low-frequency data sampling and provides an accurate and complete data foundation for subsequent energy consumption feature extraction and prediction modeling.

[0025] As a further step, after collecting high-frequency operating data (including instantaneous current, instantaneous voltage, and instantaneous speed) of the drainage pump during its historical operating cycle, a preprocessing step is also included for the collected high-frequency operating data, specifically including: S11. Abrupt Dead Value Anomaly Identification and Removal Based on a Two-Dimensional Isolation Forest Algorithm: Due to communication delays or electromagnetic interference at drainage pumping stations, instantaneous numerical jumps (abrupt changes) or signal freezes (dead values) often occur in high-frequency operating data. Simultaneously, large drainage pumps experience a real 5-7 times rated current surge at startup. If anomaly detection is based solely on the absolute amplitude of the values, it is highly likely that genuine physical shocks will be misjudged as anomalies and removed. Therefore, the following two-dimensional detection logic is adopted: First, the first-order difference values ​​of adjacent sampling points in the instantaneous current, instantaneous voltage, and instantaneous speed time series are calculated respectively to obtain the corresponding rate of change series. The absolute values ​​of the original sampling points are combined with their corresponding first-order difference values ​​to construct a six-dimensional feature vector (i.e., absolute value of current, rate of change of current, absolute value of voltage, rate of change of voltage, absolute value of speed, and rate of change of speed).

[0026] Secondly, the Isolation Forest algorithm is used to evaluate anomalies in the six-dimensional feature space. When constructing the isolation tree, feature dimensions and segmentation values ​​are randomly selected. Abrupt dead values ​​caused by electromagnetic interference exhibit extreme statistical isolation in the rate of change dimension, with path lengths significantly shorter than normal data. While the absolute value of the actual starting inrush current is large, it is limited by the motor inductance, resulting in a relatively gentle rate of change, and therefore does not exhibit isolation in the rate of change dimension. Next, anomaly scores are calculated for each data point. Data points with anomaly scores exceeding a preset threshold are identified as anomalous dead values ​​caused by communication delays or electromagnetic interference, and these are removed from the original time series, forming data gaps.

[0027] S12. Data Imputation and Time Series Alignment Based on Cubic Spline Interpolation: After removing outliers and creating missing data points, numerical imputation is required. Since subsequent steps involve calculating instantaneous power based on instantaneous current and voltage and performing integration along the time axis, the smoothness and continuity of the data are extremely important. The specific process is as follows: First, the start and end positions of the missing data breakpoints are located, and several valid sampling points adjacent to both ends of the breakpoints are extracted as interpolation nodes. Then, using cubic spline interpolation, piecewise cubic polynomial functions are constructed between adjacent interpolation nodes. ,in For piecewise indexing; set natural boundary conditions (i.e., the second derivative is zero at both ends of the interpolation interval), solve the linear equation system consisting of the interpolation node function values ​​and the continuity constraints of the first and second derivatives, and determine the coefficients of each piecewise cubic polynomial.

[0028] Next, equally spaced fill timestamps are generated within the interpolation interval according to the original high-frequency sampling frequency. The interpolated values ​​of instantaneous current, instantaneous voltage and instantaneous speed corresponding to each fill timestamp are calculated using a determined piecewise cubic polynomial. This completes the filling of missing data points and alignment of the time series, and outputs smoothed high-frequency operating data.

[0029] Building upon the above, it should also be understood that by constructing a six-dimensional feature space containing absolute values ​​and first-order difference values ​​and combining it with the isolated forest algorithm, the physical difference between the real start-up impact ("large absolute value but slow rate of change") and the interference mutation ("extremely isolated rate of change") is cleverly utilized. While accurately eliminating the mutation dead values ​​caused by communication interference, the real high-impact characteristics of the pump start-up and shutdown transients are well preserved, laying a reliable data foundation for subsequent transient energy consumption calculations.

[0030] Furthermore, subsequent steps require the integration of instantaneous power along the time axis, and this integration is extremely sensitive to the continuity of the data. Using cubic spline interpolation with defined natural boundary conditions not only ensures the continuity of values ​​at the point of discontinuity filling but also enforces the continuity of the first derivative (slope) and second derivative (curvature), restoring the smooth physical nature of the electrical signal transition and avoiding the integration truncation error and numerical oscillation caused by the "hard angles" of simple linear interpolation.

[0031] The second step in the energy consumption prediction method for drainage pumping stations based on operational data analysis is as follows: the operating status of the drainage pump is identified based on the rate of change of instantaneous rotational speed, and the historical operating cycle is decoupled into the start-stop transient stage and the steady-state operation stage in the time dimension. By separating the start-stop transient and steady-state operation, this step can distinguish the characteristics of transient high energy consumption and steady-state energy consumption, avoid the loss of energy consumption information caused by averaging, and thus improve the pertinence and accuracy of energy consumption analysis.

[0032] It should be added that the operating status of the drainage pump is identified based on the rate of change of instantaneous rotational speed, and the historical operating cycle is decoupled into a start-stop transient phase and a steady-state operating phase in the time dimension, specifically including: First, the calculation of the rate of change of rotational speed based on the sliding time window: for the smoothed instantaneous rotational speed sequence. A sliding time window of preset width slides along the time axis in set step lengths; for the position located at the th Instantaneous rotational speed data within a time window, using the rotational speed value at the end of the window. Rotational speed value at the beginning of the time The difference, divided by the window's time span The rate of change of rotational speed at the center of the window was calculated. The calculation formula is as follows: This transforms the original instantaneous speed sequence into a speed change rate sequence.

[0033] Secondly, obtain the positive threshold. Ratio of startup completion Among them, the positive threshold The method for obtaining this information is as follows: based on the rated speed parameters on the pump's nameplate. Compared with standard startup time Calculate standard mean acceleration And set the positive threshold to a set percentage of the standard average acceleration (e.g.) To distinguish between active acceleration and speed fluctuations; start-up completion ratio The method for obtaining the value is as follows: Statistically analyze the instantaneous speed corresponding to the current drop from the peak value to the steady-state value during multiple historical startup processes, and calculate the average value of the ratio of this instantaneous speed to the rated speed (typically within the range of 90%-98%); traverse the speed change rate sequence, and when the speed change rate starts at a certain moment... Continuously greater than the positive threshold When the speed change rate is recorded, that moment is taken as the starting point; Falling below the positive threshold And at this time, the instantaneous speed reaches the starting completion ratio of the rated speed. When the above occurs, record that moment as the start-up endpoint; extract the time interval between the start-up start point and the start-up endpoint and mark it as the start-up transient phase.

[0034] Furthermore, obtain the negative threshold. Ratio of shutdown to zero Among them, the negative threshold The method for obtaining the data is based on the typical coasting time when the water pump stops due to power failure. With rated speed Calculate the standard mean deceleration And set the negative threshold to a set percentage of the standard average deceleration (e.g.) ); Stop-to-zero ratio The method for obtaining the value is as follows: It is statistically derived from historical shutdown data at the critical point where the speed no longer significantly decreases (usually taken as 2%-5% of the rated speed); the speed change rate sequence is traversed, and the speed change rate at a certain moment is... Continuously less than the negative threshold When the speed change rate is recorded, that moment is taken as the start of the shutdown process; Rebound to above the negative threshold And at this moment, the instantaneous speed drops to the shutdown zero ratio. The following time is recorded as the shutdown end point; the time interval between the shutdown start point and the shutdown end point is extracted and marked as the shutdown transient phase.

[0035] Furthermore, obtain the steady-state threshold. and the preset judgment period Among them, steady-state threshold The method for obtaining this is: based on the allowable deviation rate of the pump speed during steady-state operation under rated conditions (e.g., Divide by the width of the sliding time window Calculated; Judgment period The method for obtaining the value is as follows: Based on the historical data, the maximum duration of speed oscillation decay during the transition from start-stop transient to steady state is set to ensure that non-steady-state transition processes are filtered out; the speed change rate sequence is traversed, and when the absolute value of the speed change rate is... Less than the steady-state threshold When a point is reached, it is marked as a steady-state candidate point; the interval formed by consecutive steady-state candidate points is taken as a steady-state candidate interval, and the duration of the steady-state candidate interval is calculated. If the duration exceeds a preset judgment period, the point is considered a steady-state candidate. If the duration does not exceed [a certain threshold], then the candidate interval is truncated and marked as the steady-state operation phase; if the duration does not exceed [a certain threshold], then [the relevant parameters will be considered]. If it is not considered, it is regarded as a short-term transition interval between start-up / shutdown and steady state, and is not considered as an effective steady-state operation phase.

[0036] After the above identification and extraction, the historical operating cycle is completely decoupled into the startup transient phase, the shutdown transient phase, and the steady-state operating phase in the time dimension.

[0037] It should also be noted that the startup process involves active acceleration driven by the motor (with a large and positive rate of change), while the shutdown process typically involves deceleration due to resistance after power failure (with a negative rate of change and the absolute value may be less than that during startup). The dynamic mechanisms of the two are different. By setting positive and negative thresholds for independent discrimination, it is possible to accurately distinguish between the "startup transient" and the "shutdown transient," avoiding misjudging the slow deceleration at the end of the shutdown phase as a steady state, and achieving refined extraction of different transient physical processes.

[0038] Meanwhile, in the identification of the steady-state phase, not only is it required that the absolute value of the rate of change of rotational speed be less than the steady-state threshold (stable rotational speed), but it is also mandatory that the duration of this state exceeds the preset judgment period. This constraint in the time dimension effectively eliminates the risk that transient intervals caused by brief "surges" or rotational speed fluctuations during start-up and shutdown are mistakenly identified as steady states, ensuring that the extracted steady-state operating phase has pure physical meaning, thereby guaranteeing the accuracy of subsequent steady-state average power feature extraction.

[0039] The third step in the energy consumption prediction method for drainage pumping stations based on operational data analysis is as follows: For the start-up and shutdown transient phases within the historical operating cycle, the instantaneous power is calculated based on the instantaneous current and instantaneous voltage and integrated along the time axis to obtain the true energy consumption of a single start-up and shutdown. The average of the true energy consumption of a single start-up and shutdown in multiple start-up and shutdown transient phases is taken to obtain the transient baseline energy consumption characteristics. For the steady-state operation phase, the average power within this phase is calculated to obtain the steady-state average power characteristics. This step can realistically depict the high-impact and low-efficiency energy consumption characteristics during the start-up and shutdown process, avoid the smoothing masking of transient high energy consumption by traditional time averaging methods, and simultaneously achieve the separation and quantification of start-up and shutdown energy consumption and steady-state energy consumption, improving the authenticity and stability of the energy consumption characteristic expression.

[0040] In the third step, it should be noted that, for the start-stop transient phase within the historical operating cycle, the instantaneous power is calculated based on the instantaneous current and instantaneous voltage, and integrated along the time axis to obtain the actual energy consumption for a single start-stop cycle. Specifically, this includes: S31. Frequency Domain Decomposition and Harmonic Parameter Extraction: Due to magnetic circuit saturation and power electronic device chopping during the start-up and shutdown transient phases of large drainage pumps, the instantaneous current and voltage waveforms are severely distorted, containing a large number of high-order harmonics. To eliminate the interference of cross-frequency terms and DC bias on the calculation of active power consumption, the instantaneous voltage sequence within the identified single start-up and shutdown transient phases is analyzed. With instantaneous current sequence Windowing is applied (using a Hanning window to suppress spectral leakage), and a Discrete Fourier Transform (DFT) is performed based on this truncated interval. The time-domain signal is then converted to the frequency domain using the DFT, and the fundamental angular frequency is extracted. And the amplitude and phase angle of the fundamental component and each significant harmonic component; let the th The voltage amplitude of the second harmonic (including the fundamental frequency n=1) is Phase angle is The current amplitude is Phase angle is .

[0041] S32. Instantaneous Active Power Sequence Reconstruction Based on Frequency Domain Parameters: Based on the amplitude and phase angle of the extracted fundamental component and each harmonic component, only the voltage and current components of the same frequency are retained for instantaneous active power reconstruction, and cross-product terms of different frequencies that do not generate actual active power consumption are filtered out; for the nth harmonic, its instantaneous active power component in the time domain... The calculation formula is: In the formula, the first term The first term represents the constant active power generated by this harmonic, and the second term represents the active power of the second harmonic oscillation.

[0042] The instantaneous active power components of each harmonic are superimposed in the time domain to generate the total instantaneous active power sequence. The calculation formula is as follows: ;in The highest harmonic order is extracted. This yields an instantaneous active power sequence that reflects only the actual active power consumption.

[0043] S33. Calculation of Real Energy Consumption During a Single Start-Stop Operation Based on Numerical Integration: A numerical integration algorithm is used to calculate the total instantaneous active power sequence. Time interval during the start-stop transient phase Discrete accumulation integral calculation is performed; specifically, the trapezoidal quadrature formula is used, based on the high-frequency sampling period. The calculation formula is: In the formula, For the first The instantaneous active power at each sampling time. For the first The instantaneous active power at each sampling moment is used to accurately obtain the actual active power consumed during the transient phase of a single start-stop operation, thus representing the true energy consumption of that single start-stop operation. .

[0044] For the multiple start-stop transient phases within the historical operating cycle, repeat the above steps to obtain the actual energy consumption of each start-stop. Calculate the arithmetic mean of the actual energy consumption of multiple start-stop cycles to obtain the transient baseline energy consumption characteristics. For the steady-state operation phase, calculate the instantaneous power at all sampling points within this phase. The arithmetic mean of the values ​​is used to obtain the steady-state average power characteristics.

[0045] It should be understood that the harmonic content is extremely high during the start-up and shutdown of large water pumps. Traditional methods of directly multiplying and averaging in the time domain will include reactive power and cross-frequency terms in the statistics, leading to distortion in the calculation of active power consumption. This invention extracts the fundamental and harmonic parameters through discrete Fourier transform and reconstructs the instantaneous active power strictly according to the same frequency components. It physically removes the reactive and interference components that do not produce thermal effects or mechanical work, ensuring that the extracted transient high energy consumption characteristics are the true active power consumption driving the water pump startup.

[0046] Meanwhile, the power during the start-up and shutdown process exhibits highly nonlinear and rapid changes. This invention employs a trapezoidal numerical integration algorithm to accumulate the instantaneous active power along the time axis, mathematically closely matching the transient fluctuation details of the power curve. This achieves accurate quantification of the nonlinear high-energy-consumption accumulation process and effectively overcomes the problem of systematically low prediction results caused by the averaging of time scales in existing technologies.

[0047] To address the two distinctly different physical characteristics of the two stages, this invention employs a differentiated strategy of "precise transient calculation and simplified steady-state calculation." For the start-up and shutdown stages, high-precision calculation using "frequency domain decomposition + integration" ensures that the dominant transient impact energy consumption is not lost. For the steady-state stages, arithmetic average calculation is used, which not only conforms to the physical law of stable steady-state power but also significantly reduces the computational load of the system, thereby improving the model's computational efficiency while maintaining overall prediction accuracy.

[0048] The fourth step in the energy consumption prediction method for drainage pumping stations based on operational data analysis is as follows: First, obtain the ratio of the peak starting impact current to the rated current during the current start-stop transient phase of the drainage pump, thus obtaining the impact multiple characteristic. Second, obtain the time taken to accelerate from a standstill to the rated speed, thus obtaining the acceleration time characteristic. Third, based on the impact multiple characteristic and the acceleration time characteristic, construct a transient degradation correction coefficient using a preset attenuation weight function. This step can dynamically reflect the impact of pump equipment aging, mechanical wear, or performance degradation on start-stop energy consumption, making transient energy consumption prediction closer to the actual operating state and improving the accuracy and long-term applicability of the prediction.

[0049] In the fourth step, based on the impact factor characteristics and acceleration time characteristics, a transient degradation correction coefficient is constructed using a preset attenuation weighting function, specifically including: S41. Construction of the degradation feature sample set: Obtain the actual energy consumption of a single start-stop cycle of the drainage pump under rated start-stop conditions during the initial stage of operation, and use it as the factory calibration benchmark energy consumption. The rated start-stop condition is the start-up process under rated voltage, with the outlet valve fully closed or at a specified standard opening. During the historical service life of the drainage pump, multiple historical start-stop transient phase data were collected under the same or equivalent conditions as the rated start-stop condition. The historical impact multiple characteristics corresponding to each historical start-stop were extracted. and historical acceleration time characteristics And calculate the corresponding historical actual energy consumption of a single start-stop cycle. Calculate the ratio of the actual energy consumption during a single historical start-stop cycle to the factory-calibrated baseline energy consumption to obtain the actual energy consumption degradation rate. Due to increased resistance caused by mechanical wear of equipment, the actual energy consumption degradation rate... Greater than 1 and showing an increasing trend with the service life; based on multiple sets of data... This constitutes a sample set of degradation features.

[0050] S42. Fitting the decay weight function based on multivariate nonlinear regression: using historical impact factor characteristics. and historical acceleration time characteristics The input independent variable is the actual energy degradation rate. To output the dependent variable, a multivariate nonlinear regression algorithm was used for fitting and training. Considering that the wear of the pump rotor leads to an increase in resistance torque, which not only directly prolongs the acceleration time but also forces the motor to generate a larger electromagnetic torque, manifested as a change in the inrush current characteristics, and that there is a coupling superposition effect between the increased mechanical resistance and the electromagnetic transient response, a quadratic polynomial nonlinear mapping model containing linear, quadratic, and cross terms was constructed, the expression of which is: In the formula, For the initial degradation bias term, and The linear influence coefficient is... and The nonlinear acceleration effect coefficient is... The electromechanical coupling influence coefficient is used; the degradation feature sample set is substituted into the above model, and the least squares method is used to... to The solution involves constructing an objective function that minimizes the sum of squared residuals between the model's predicted degradation rate and the actual energy consumption degradation rate, thereby obtaining definite values ​​for each coefficient. The mapping relationship model with these definite coefficients is then used as a pre-defined attenuation weight function. .

[0051] S43. Calculation of the current transient degradation correction coefficient: Obtain the impact multiple characteristics of the drainage pump during the most recent historical start-stop transient phase before the start of the prediction period. With acceleration time characteristics (Define this phase as the current start-stop transient phase), and input it into the preset decay weight function. In the process, the transient degradation correction coefficients under the current device state are calculated and output. Since equipment degradation is a slow evolution process on a macroscopic timescale, it is assumed that the equipment degradation state is constant within a relatively short prediction period. Therefore, the transient degradation correction coefficient calculated based on the most recent historical data is used as the unified degradation correction coefficient for all planned start-up and shutdown events within the prediction period and substituted into subsequent calculations.

[0052] Traditional static prediction models assume that the performance of equipment remains unchanged throughout its entire lifecycle. However, this invention introduces an "energy consumption degradation rate" extracted from real operating data to construct a transient degradation correction coefficient. This coefficient can dynamically increase as the service life of the equipment extends and mechanical wear intensifies, fundamentally compensating for the deficiency of static models in quantifying the increased energy consumption due to equipment degradation. This enables the prediction model to adaptively correct itself over time.

[0053] The fifth step in the energy consumption prediction method for drainage pumping stations based on operational data analysis is as follows: First, obtain the planned number of start-ups and shutdowns and the planned steady-state operating time within the prediction period. Then, calculate the single-time predicted transient energy consumption based on transient baseline energy consumption characteristics and transient degradation correction coefficients. Next, calculate the total start-up and shutdown transient energy consumption for the prediction period based on the planned number of start-ups and shutdowns and the single-time predicted transient energy consumption. Finally, calculate the total steady-state operating energy consumption for the prediction period based on steady-state average power characteristics and the planned steady-state operating time. The total start-up and shutdown transient energy consumption and the total steady-state operating energy consumption are then superimposed to obtain the target predicted total energy consumption of the drainage pumping station. This step considers both start-up and shutdown behavior and the energy consumption contribution of steady-state operation, enabling refined prediction of the overall energy consumption of the pumping station and accurately quantifying the impact of frequent start-ups and shutdowns on the total energy load, providing a scientific basis for pumping station scheduling optimization and energy-saving management.

[0054] In the fifth step, it is necessary to specifically explain S51, the start-stop plan simulation based on meteorological and hydrological coupling: obtain meteorological precipitation forecast data for the forecast period, and calculate the predicted inflow process curve of the drainage pumping station forebay by combining the area of ​​the drainage pumping station's catchment area and the comprehensive runoff coefficient. Based on the cross-sectional area and capacity of the pump station forebay The discrete time step iterative method is used to simulate the rise and fall of the liquid level in the forebay; the time step is set. The real-time monitoring liquid level of the forebay at the start of the prediction period is obtained and used as the initial liquid level for iterative simulation. During the iteration, the current pool level Touching or exceeding the preset pump start level At that time, the pumps are put into operation according to the preset pump combination rules to determine the current drainage flow rate. When the liquid level The water level drops to touch or exceed the pump stop level. At that time, the water pump is shut down and replaced according to the preset rules. Based on the law of conservation of mass, the first The forebay level is calculated iteratively using the following formula: By traversing the entire prediction period, the system extracts each time the liquid level touches a target during the simulation. and At specific time points, the number of planned starts and stops within the predicted time period is statistically generated. And the planned steady-state runtime between two adjacent start and stop nodes. .

[0055] S52. Transient energy consumption correction for pipeline fluid coupling: Based on the planned start-up and shutdown simulation results, identify whether multiple pumps start concurrently; if so, obtain the number of pumps starting concurrently. and fluid resistance parameters of the main outlet pipe of the drainage pumping station Based on the fluid dynamics characteristics of pipeline resistance, using the formula Estimate the additional transient water hammer pressure generated in the outlet pipe network due to the surge in flow rate from multiple pumps, among which... The rated flow rate of a single pump is used; the transient water hammer pressure increment is used as the load compensation factor to positively amplify and compensate for the impact factor and acceleration time characteristics of the input. The update formula is as follows: , ,in The rated pressure of the water pump; the updated and Input a preset decay weight function, recalculate the degradation correction coefficient and single-time predicted transient energy consumption under concurrent startup conditions, and then sum them to obtain the total transient energy consumption during startup and shutdown. .

[0056] S53. Dynamic calculation of steady-state power based on surface regression: extracting historical operating head during historical steady-state operation phases. Historical operating flow With the corresponding historical steady-state power Using historical operating head and historical operating flow rate as two-dimensional input variables, and historical steady-state power as the output variable, a bivariate polynomial fitting algorithm is used for surface regression to construct the steady-state algebraic relationship equation between head, flow rate, and power: In the formula, to The coefficients are those determined by fitting using the least squares method. For running traffic, To determine the operating head; obtain the predicted dynamic head and planned flow rate for the predicted period; the predicted dynamic head is calculated in real time based on the difference between the forebay level and the outlet level derived in step S51; the predicted dynamic head is then... With planned water flow Input the above steady-state algebraic equation, calculate and output the dynamically adjusted power that varies with hydraulic conditions. ; Dynamically converted power Instead of static steady-state average power characteristics, it incorporates the planned steady-state operating time. By performing cumulative calculations, the total steady-state energy consumption for the predicted period is obtained. In the formula, For the first The dynamic equivalent power corresponding to each steady-state operation stage For the first The planned steady-state operating time is calculated; the total energy consumption during start-up and shutdown transients is superimposed with the total energy consumption during steady-state operation to obtain the target predicted total energy consumption of the drainage pumping station. .

[0057] S54. Start-up and shutdown strategy optimization with energy minimization as the objective: Generate multiple candidate start-up and shutdown timing strategies for the predicted time period. The candidate strategies are generated through parameterized combination, including setting different pump start-up liquid level values. Different pump stop liquid level values And different combinations of pump activation priorities; substitute each candidate start-stop timing strategy into the prediction calculation process of steps S51-S53 above to obtain the candidate predicted total energy consumption set corresponding to each strategy; take minimizing total energy consumption as the optimization objective function. The constraints are that the liquid level in the front pool does not exceed the overflow alarm level and does not fall below the minimum safe level during the predicted period. The optimal start-stop timing strategy that satisfies the constraints and minimizes the objective function value is selected from the candidate predicted total energy consumption set, and the optimal start-stop timing strategy is converted into a control command and sent to the local controller of the pump station for execution.

[0058] As described above, this step overcomes the limitations of traditional methods that rely on manual input of start-up and shutdown plans. Through iterative derivation of the rainfall-runoff model and the differential equation of mass conservation in the forebay, macroscopic meteorological data is microscopically transformed into the specific liquid level rise and fall process of the pumping station. This automatically generates planned start-up and shutdown times and operating durations that fit future water inflow scenarios, providing a realistic and dynamic time input carrier for energy consumption prediction. Simultaneously, traditional methods typically ignore the impact of water hammer resistance in the outlet pipe network when multiple pumps start concurrently. This step, by introducing fluid impedance to calculate the additional water hammer pressure and using it as a load compensation factor to amplify the impact factor and acceleration time characteristics, truly reflects the negative feedback effect of pipe network fluid resistance on the electromagnetic transients of motor startup, avoiding a severe underestimation of transient energy consumption under multi-pump concurrent operation.

[0059] like Figure 2 The diagram shown is a structural schematic of the drainage pumping station energy consumption prediction system based on operational data analysis provided in an embodiment of this application. (Refer to...) Figure 2 The system includes: a running data acquisition module, a running phase decoupling module, an energy consumption feature extraction module, a transient correction analysis module, and an energy consumption prediction output module.

[0060] Specifically, the operation data acquisition module is used to collect high-frequency operating data of the drainage pump during its historical operating cycle. The high-frequency operating data includes instantaneous current, instantaneous voltage, and instantaneous speed.

[0061] The operation phase decoupling module is used to identify the operating status of the drainage pump based on the instantaneous speed change rate, and decouple the historical operating cycle into the start-stop transient phase and the steady-state operation phase in the time dimension.

[0062] The energy consumption feature extraction module is used to calculate the instantaneous power based on the instantaneous current and instantaneous voltage during the start-stop transient phase within the historical operating cycle, and integrate it along the time axis to obtain the true energy consumption of a single start-stop. The average value of the true energy consumption of a single start-stop during multiple start-stop transient phases is taken to obtain the transient baseline energy consumption feature. For the steady-state operation phase, the average power within this phase is calculated to obtain the steady-state average power feature.

[0063] The transient correction analysis module is used to obtain the ratio of the peak value of the starting impact current to the rated current during the current start-stop transient phase of the drainage pump, thus obtaining the impact multiple characteristic, and to obtain the time from the current acceleration from standstill to the rated speed, thus obtaining the acceleration time characteristic. Based on the impact multiple characteristic and the acceleration time characteristic, a transient degradation correction coefficient is constructed through a preset attenuation weight function.

[0064] The energy consumption prediction output module is used to obtain the planned number of start-stop cycles and the planned steady-state running time within the prediction period. Based on the transient baseline energy consumption characteristics and transient degradation correction coefficient, it calculates the single predicted transient energy consumption. Based on the planned number of start-stop cycles and the single predicted transient energy consumption, it calculates the total start-stop transient energy consumption for the prediction period. Based on the steady-state average power characteristics and the planned steady-state running time, it calculates the total steady-state running energy consumption for the prediction period. The total start-stop transient energy consumption and the total steady-state running energy consumption are superimposed to obtain the target predicted total energy consumption of the drainage pumping station.

[0065] Based on the above, it should also be understood that, such as Figure 3 The diagram shown is an overall module logic diagram of the drainage pumping station energy consumption prediction system based on operational data analysis provided in this application embodiment. Figure 3It can be seen that the system of the present invention constructs a complete closed loop from bottom-level data perception to top-level energy consumption prediction through the close collaboration of the data acquisition module, the operation phase decoupling module, the energy consumption feature extraction module, the transient correction analysis module, and the energy consumption prediction output module: the high-frequency data provided by the acquisition module is accurately divided into transient steady-state intervals with distinctly different physical characteristics in the time dimension based on the rotational speed change rate by the decoupling module, and the extraction module executes a differentiated algorithm accordingly (transient high-precision integration and steady-state mean calculation). The collaboration of the two completely avoids the full-cycle averaging error caused by the high energy consumption of the start-up impact being diluted by the long steady-state period; furthermore, the historical transient benchmark energy consumption output by the extraction module and the correction module The module superimposes a dynamic degradation coefficient based on the current impact multiple and acceleration time, achieving a leap from "static historical assessment" to "dynamic adaptive calibration," giving the prediction model the ability to automatically correct itself as the equipment's mechanical wear decreases. Finally, the prediction output module deeply integrates the micro-electromechanical degradation characteristics and steady-state power characteristics obtained from the front end with the planned start-stop times and runtime of macro-schedules. This not only accurately quantifies the micro-energy consumption of a single equipment state but also maps the total energy consumption of the pumping station under the macro-schedule strategy. It directly transforms the value of the underlying data into a quantitative basis to support macro-energy-saving scheduling decisions, significantly improving the long-term accuracy and engineering practicality of drainage pumping station energy consumption prediction.

[0066] In summary, this invention overcomes the limitations of traditional coarse-grained statistics by collecting high-frequency operating data and decoupling states based on the rate of change of rotational speed. It extracts differentiated features for the start-stop transient and steady-state operation phases and innovatively introduces a transient degradation correction coefficient, fundamentally solving the problem that existing technologies mask the nonlinear high-energy consumption characteristics of start-stop transients due to time-scale averaging. This effectively eliminates the systematic underestimation error in prediction results under frequent start-stop conditions, achieving accurate quantification of the actual energy consumption load of pumping stations. This invention not only significantly improves the energy consumption prediction accuracy of large drainage pumping stations under complex scheduling conditions such as rainstorms and flood seasons, but also makes the high energy consumption cost of start-stop transients explicit, providing a reliable basis for assessing the actual operating costs brought about by frequent start-stop operations. It has significant engineering application value and significant economic benefits for optimizing pumping station scheduling strategies, extending equipment life, and achieving energy conservation and consumption reduction in pumping station systems.

[0067] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the above functions can be divided into different functional modules to complete all or part of the functions described above.

[0068] In the embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0069] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units, located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0070] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0071] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to the solution, or all or part of the technical solution, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0072] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for predicting the energy consumption of drainage pumping stations based on operational data analysis, characterized in that, Includes the following steps: Collect high-frequency operating data of the drainage pump during its historical operating cycle. The high-frequency operating data includes instantaneous current, instantaneous voltage, and instantaneous speed. The operating status of the drainage pump is identified based on the instantaneous rate of change of rotational speed, and the historical operating cycle is decoupled into the start-stop transient stage and the steady-state operation stage in the time dimension. For the start-stop transient phase within the historical operating cycle, the instantaneous power is calculated based on the instantaneous current and instantaneous voltage and integrated along the time axis to obtain the actual energy consumption of a single start-stop. The average of the actual energy consumption of a single start-stop during multiple start-stop transient phases is taken to obtain the transient reference energy consumption characteristics. For the steady-state operation phase, the average power during this phase is calculated to obtain the steady-state average power characteristics. The ratio of the peak value of the starting impact current to the rated current during the current start-stop transient phase of the drainage pump is obtained to obtain the impact multiple characteristic, and the time taken to accelerate from standstill to rated speed is obtained to obtain the acceleration time characteristic. Based on the impact factor characteristics and acceleration time characteristics, a transient degradation correction coefficient is constructed through a preset attenuation weight function; The planned number of start-stop cycles and the planned steady-state runtime are obtained within the forecast period. The transient energy consumption of a single forecast is calculated based on the transient baseline energy consumption characteristics and the transient degradation correction coefficient. The total transient energy consumption during the start-stop period is calculated based on the planned number of start-stop cycles and the transient energy consumption of a single predicted cycle. The total steady-state operating energy consumption during the predicted period is calculated based on the steady-state average power characteristics and the planned steady-state operating time. The total transient energy consumption during start-stop cycles and the total steady-state operating energy consumption are then superimposed to obtain the target predicted total energy consumption of the drainage pumping station.

2. The method for predicting the energy consumption of drainage pumping stations based on operational data analysis as described in claim 1, characterized in that, The process involves collecting high-frequency operating data of the drainage pump during its historical operating cycle, followed by a high-frequency data preprocessing step. The isolated forest algorithm is used to identify and remove abrupt dead value outliers caused by communication delays or electromagnetic interference in the instantaneous current, instantaneous voltage, and instantaneous speed sequences, forming data missing breakpoints; The cubic spline interpolation method is used to fill in missing data points and align the time series, outputting smoothed high-frequency running data.

3. The method for predicting the energy consumption of drainage pumping stations based on operational data analysis as described in claim 1 or 2, characterized in that, The method of identifying the operating status of the drainage pump based on the rate of change of instantaneous rotational speed decouples the historical operating cycle into a start-stop transient phase and a steady-state operating phase in the time dimension, specifically including: The instantaneous speed sequence in the smoothed high-frequency operating data is differentiated using a preset sliding time window to obtain the speed change rate sequence; When the rate of change of rotational speed within the sliding time window is greater than a preset positive threshold, the corresponding time interval is extracted and marked as the start-up transient phase; when the rate of change of rotational speed is less than a preset negative threshold, the corresponding time interval is extracted and marked as the shutdown transient phase. When the absolute value of the speed change rate is less than the preset steady-state threshold and the duration exceeds the preset judgment period, the corresponding time interval is extracted and marked as the steady-state operation stage.

4. The method for predicting the energy consumption of drainage pumping stations based on operational data analysis as described in claim 1, characterized in that, The process of calculating instantaneous power based on instantaneous current and instantaneous voltage and integrating it along the time axis to obtain the actual energy consumption for a single start-stop cycle specifically includes: Discrete Fourier transform is performed on the instantaneous voltage and current sequences during the start-stop transient phase to extract the amplitude and phase angle of the fundamental component and each harmonic component; Based on the amplitude and phase angle of the fundamental component and each harmonic component, the instantaneous active power at each sampling time is calculated, and an instantaneous active power sequence is generated. A numerical integration algorithm is used to accumulate and integrate the instantaneous active power sequence on the time axis to obtain the actual energy consumption of a single start-stop cycle.

5. The method for predicting the energy consumption of drainage pumping stations based on operational data analysis as described in claim 1, characterized in that, The construction of transient degradation correction coefficients through a preset decay weight function specifically includes: A degradation feature sample set of the drainage pump's historical service cycle is collected. The degradation feature sample set includes multiple sets of historical impact multiple features, historical acceleration time features, and corresponding actual energy consumption degradation rate. The actual energy consumption degradation rate is the ratio of the actual energy consumption of a single historical start-stop cycle to the factory-calibrated baseline energy consumption. Using historical impact multiples and historical acceleration time characteristics as input independent variables and actual energy consumption degradation rate as output dependent variable, a multivariate nonlinear regression algorithm is used for fitting and training to obtain a mapping relationship model between the characteristic variables and the degradation rate, and the mapping relationship model is used as a preset decay weight function. The acquired impact factor and acceleration time characteristics are input into a preset attenuation weight function to calculate and output the transient degradation correction coefficient under the corresponding state.

6. The method for predicting the energy consumption of drainage pumping stations based on operational data analysis as described in claim 1, characterized in that, The specific deduction process for obtaining the planned start-stop count and planned steady-state runtime within the prediction period includes: Obtain meteorological precipitation forecast data for the forecast period, and obtain the predicted inflow process curve of the drainage pumping station forebay based on the meteorological precipitation forecast data; Based on the cross-sectional area and capacity of the pump station forebay, a dynamic mass conservation deduction is performed on the predicted inflow process curve and the preset start-up and stop-up liquid levels to simulate the rise and fall of the forebay liquid level over time. Extract the time points when the liquid level touches the start-up liquid level and the stop-up liquid level during the simulation, and count the number of planned start-ups and stop-ups within the prediction period, as well as the planned steady-state running time between two adjacent start-up and stop-ups.

7. The method for predicting the energy consumption of drainage pumping stations based on operational data analysis as described in claim 6, characterized in that, The calculation of the total transient energy consumption during the prediction period based on the planned number of start-ups and shutdowns and the single predicted transient energy consumption also includes a pipeline fluid coupling correction process: Based on the number of concurrently started drainage pumps and the fluid impedance parameters of the main outlet pipe of the drainage pumping station, calculate the additional value of transient water hammer pressure generated by the outlet pipe network. Using the transient water hammer pressure as a load compensation factor, the impact factor and acceleration time characteristics of the compensation input are positively amplified, the degradation correction coefficient under concurrent startup state is updated, and the single predicted transient energy consumption and the target predicted total energy consumption are recalculated.

8. The method for predicting the energy consumption of drainage pumping stations based on operational data analysis as described in claim 6, characterized in that, The step of calculating the total steady-state operating energy consumption for the predicted period based on steady-state average power characteristics and planned steady-state operating duration also includes a step of dynamic hydraulic conversion of steady-state average power characteristics: Extract the historical operating head, historical operating flow rate, and corresponding historical steady-state power during the historical steady-state operation phase; Using historical operating head and historical operating flow rate as two-dimensional input variables and historical steady-state power as output variable, a polynomial fitting algorithm is used to perform surface regression and construct a steady-state algebraic relationship equation between head, flow rate and power. Obtain the predicted dynamic head and planned flow rate for the forecast period, input the predicted dynamic head and planned flow rate into the steady-state algebraic relationship equation, and calculate the output dynamic converted power. The total energy consumption for steady-state operation during the predicted period is calculated by replacing the steady-state average power characteristic with dynamically converted power and combining it with the planned steady-state operating time.

9. The method for predicting the energy consumption of drainage pumping stations based on operational data analysis as described in claim 1, characterized in that, The process of obtaining the target predicted total energy consumption of the drainage pumping station further includes: Multiple candidate start-up and shutdown timing strategies with different combinations of drainage pumps and operating water level ranges are generated for the predicted period. Substitute the planned start-stop times and planned steady-state runtime corresponding to each candidate start-stop timing strategy into the prediction calculation to obtain a candidate predicted total energy consumption set containing multiple prediction results; With the goal of minimizing total energy consumption, and with the constraint that the liquid level in the forepool does not exceed the overflow alarm level and does not fall below the minimum safe level during the predicted period, the optimal start-stop timing strategy is selected from the candidate predicted total energy consumption set, and the optimal start-stop timing strategy is converted into control commands and sent to the local controller of the pump station.

10. A drainage pumping station energy consumption prediction system based on operational data analysis, employing the drainage pumping station energy consumption prediction method based on operational data analysis as described in any one of claims 1-9, characterized in that, include: The system includes a data acquisition module, a decoupling module for the operational phase, an energy consumption feature extraction module, a transient correction and analysis module, and an energy consumption prediction and output module. The operation data acquisition module is used to collect high-frequency operation data of the drainage pump during its historical operation cycle. The high-frequency operation data includes instantaneous current, instantaneous voltage, and instantaneous speed. The operation phase decoupling module is used to identify the operating status of the drainage pump based on the instantaneous speed change rate, and decouple the historical operating cycle into the start-stop transient phase and the steady-state operating phase in the time dimension. The energy consumption feature extraction module is used to calculate the instantaneous power based on the instantaneous current and instantaneous voltage during the start-stop transient phase within the historical operating cycle, and integrate it along the time axis to obtain the true energy consumption of a single start-stop. The average value of the true energy consumption of a single start-stop during multiple start-stop transient phases is taken to obtain the transient reference energy consumption feature. For the steady-state operation phase, the average power during this phase is calculated to obtain the steady-state average power characteristics. The transient correction analysis module is used to obtain the ratio of the peak value of the starting impact current to the rated current during the current start-stop transient phase of the drainage pump, to obtain the impact multiple characteristic, and to obtain the time from the current acceleration from standstill to the rated speed, to obtain the acceleration time characteristic. Based on the impact factor characteristics and acceleration time characteristics, a transient degradation correction coefficient is constructed through a preset attenuation weight function; The energy consumption prediction output module is used to obtain the planned start-stop times and planned steady-state running time within the prediction period, and calculate the single predicted transient energy consumption based on transient baseline energy consumption characteristics and transient degradation correction coefficient. The total transient energy consumption during the start-stop period is calculated based on the planned number of start-stop cycles and the transient energy consumption of a single predicted cycle. The total steady-state operating energy consumption during the predicted period is calculated based on the steady-state average power characteristics and the planned steady-state operating time. The total transient energy consumption during start-stop cycles and the total steady-state operating energy consumption are then superimposed to obtain the target predicted total energy consumption of the drainage pumping station.

Images