A water supply and drainage pump unit assembly and a digital control system

Abstract

This application relates to the technical field of water pump control, and provides a water supply and drainage pump group assembly and a digital control system. By integrating multi-source data such as pressure, flow rate, and temperature, it realizes adaptive control under complex working conditions, breaks through the dependence on a single parameter in traditional control, and improves the response ability of the control system to water use fluctuations and equipment state changes; Based on the long short-term memory network trained with historical multi-source data, the control system can predict the pressure set value in advance, solve the pressure set value through the particle swarm algorithm, and perform weight fusion according to the predicted pressure set value and the solved pressure set value to generate adjustment instructions, so as to reduce the number of equipment starts and stops and mechanical wear, reduce maintenance costs, and avoid waste of water resources through precise control; At the same time, through redundant design and fault tolerance mechanism, it ensures that the control system can still operate stably in scenarios such as sensor failures and equipment abnormalities.

Description

Technical Field

[0001] This application relates to the technical field of water pump control, and particularly to a water supply and drainage pump group assembly and a digital control system. Background Art

[0002] In the early stage, relays were used for logical control of pump group start and stop, and independent instruments such as pressure gauges and flow meters were used to monitor parameters. Manual adjustment of valve openings and pump group configurations was relied on, resulting in lagged response and low control accuracy. In the stage of automatic control, programmable logic controllers and PID control algorithms were introduced to achieve closed-loop control of pressure and flow, enabling automatic start and stop of pump groups and adjustment of valves according to a single parameter, improving system stability and efficiency. In recent years, with the development of sensor technology, communication technology, and intelligent algorithms, some systems have attempted to introduce intelligent algorithms such as fuzzy control and neural networks to optimize control strategies, or achieve remote data monitoring through Internet of Things technology. However, there are still limitations in multi-parameter fusion, global optimization, and fault self-healing.

[0003] Traditional PID control relies on fixed parameters and is difficult to cope with complex working conditions such as pipeline leakage, peak water consumption fluctuations, and equipment aging, which can lead to problems such as large pressure fluctuations and high energy consumption. For example, constant pressure control is prone to energy waste during low water consumption periods; single reliance on pressure or flow parameters for control without integrating multi-dimensional data such as temperature and equipment status cannot comprehensively evaluate the system operation status; existing systems mostly use single-parameter threshold alarms, lacking dynamic analysis of equipment operation curves and making it difficult to identify early faults in advance. Most do not solve how to achieve the operation control and scheduling of pump groups through multi-source data processing. Summary of the Invention

[0004] In view of the deficiencies of the prior art, this application provides a water supply and drainage pump group assembly and a digital control system.

[0005] In a first aspect, this application provides a water supply and drainage pump group assembly, which includes: a main inlet pipe, a first digital pump, a flexible joint, a digital device, a digital adjustment device, a non-throttling device, a digital monitoring system, and a second digital pump;

[0006] The main inlet pipe includes a first-side main inlet pipe flange, a negative pressure monitoring device, a water inlet pipeline, and a second-side main inlet pipe flange. The main inlet pipe is connected to the first digital pump and the second digital pump through a flexible joint. The flexible joint is used to compensate for the displacement of the main inlet pipe and is connected to an external water source and the water inlet pipeline through the first-side main inlet pipe flange and the second-side main inlet pipe flange. The water flow in the water inlet pipeline enters the first digital pump and the second digital pump after passing through the negative pressure monitoring device in sequence. The negative pressure monitoring device is used to monitor the pressure data in the main inlet pipe.

[0007] As an alternative implementation, the digital device is connected to the first digital pump and the second digital pump, and is used to monitor pressure data, flow data and temperature data. The digital adjustment device is connected to the digital device and is used to adjust the opening and closing state of the valve flap, the valve flap speed and the valve flap execution angle;

[0008] The no-throttling device includes a no-throttling device flange, a pressure monitoring device, a no-throttling main pipeline, an eccentric reducer, a no-throttling pipeline for the second digital pump and an elbow. The pressure monitoring device is installed in the no-throttling main pipeline to monitor the pressure data in the no-throttling main pipeline. The no-throttling pipeline for the second digital pump is connected to the no-throttling main pipeline, and the water flows into the no-throttling main pipeline after being rectified by the elbow. The eccentric reducer is used for pipe diameter transformation in the no-throttling main pipeline. The no-throttling device is connected to the water-using point through the no-throttling device flange;

[0009] The digital monitoring system is used to collect the operation data of the digital pump, the digital device and the digital adjustment device, and adjust the pump group configuration information and the opening and closing state of the valve flap, the valve flap speed and the valve flap execution angle in the digital adjustment device in real time according to the historical operation data.

[0010] In a second aspect, the present application provides a digital control system for a water supply and drainage pump group assembly. The system includes: obtaining pressure data, flow data and temperature data in real time, generating a valve flap action instruction according to the pressure data and the flow data and combining with a fuzzy control algorithm, and predicting a pressure set value through a long short-term memory network based on the temperature data and the pipe network leakage rate;

[0011] Based on the pressure data and the flow data, the hydraulic power of the digital pump is calculated in real time, and the pump efficiency of the digital pump is determined by combining with the motor power monitored in real time. According to the variance of the pump efficiency of the digital pump and the pump efficiency of the current pump group, a start-stop instruction for the standby pump is generated. At the same time, based on the pressure data and the flow data, the pressure set value is solved through a particle swarm algorithm, and the predicted pressure set value and the solved pressure set value are weighted and fused to generate a pressure target value. An adjustment instruction is generated according to the pressure deviation between the pressure target value and the pressure data;

[0012] The operation state of the pump group is monitored in real time, an operation curve is generated by combining the pressure data, the flow data and the temperature data, and the operation curve is matched with the historical normal operation curve through a dynamic time warping algorithm to trigger fault detection, locate the faulty pump and switch to the standby pump, and at the same time trigger a fault tolerance mechanism.

[0013] As an alternative implementation, the generation logic of the start-stop instruction for the standby pump includes:

[0014] Receiving pressure data and flow data, and monitoring the motor operation data of the digital pump in real time to convert it into the motor power of the digital pump;

[0015] Based on the Bernoulli equation, the hydraulic power of the digital pump is calculated in real time based on pressure data and flow data. According to the ratio of the hydraulic power to the motor power, the pump efficiency of the digital pump is determined in real time, and the change trend of the pump efficiency is analyzed for anomaly detection;

[0016] According to the pump efficiency variance of the current pump group, the load balance degree is determined, and an enabling threshold is configured. The load balance degree is compared with the enabling threshold to judge and trigger the start and stop of the standby pump;

[0017] According to the pump efficiency of the digital pump, the total flow rate of the current pump group is determined. Based on the change trend of the pump efficiency, the total flow rate of the current pump group is corrected, and the flow difference between the target flow rate and the corrected total flow rate of the current pump group is calculated. According to the flow difference, the number of standby pumps is determined.

[0018] As an optional implementation manner, the generation logic of the adjustment instruction includes:

[0019] According to the prediction error distribution of the long short-term memory network, the prediction credibility is determined, the weight is dynamically adjusted based on the prediction credibility, and the predicted pressure set value and the solved pressure set value are weighted and fused to generate a pressure target value;

[0020] Calculate the pressure deviation between the pressure target value and the pressure data, and at the same time determine the change rate of the pressure deviation. According to the pressure deviation and the change rate of the pressure deviation, the control area is divided on the two-dimensional plane;

[0021] According to the control area, a hierarchical control mechanism is determined, and an adjustment instruction is generated according to the hierarchical control mechanism;

[0022] Obtain the operating state of the pump group and the predicted deviation between the predicted pressure set value and the pressure data, and compensate and adjust the instruction according to the operating state of the pump group and the predicted deviation;

[0023] Receive the execution feedback data of the digital adjustment device, determine the execution error between the execution feedback data and the adjustment instruction, and judge based on the execution error to optimize the adjustment instruction.

[0024] As an optional implementation manner, the solution sub-logic of the pressure set value includes:

[0025] Determine the fitness function according to system energy consumption, flow stability and pressure deviation control, and use the predicted pressure set value and the pipeline network properties as constraint conditions;

[0026] Initialize the particle swarm, calculate the fitness value of each particle according to the fitness function, update the speed and position of each particle, and continuously iterate until the convergence condition is met to solve the pressure set value;

[0027] Verify the feasibility of the solved pressure set value through the constraint conditions, and output the pressure set value that passes the verification.

[0028] As an alternative implementation, the trigger logic for the fault detection includes:

[0029] Monitor the operating status of the pump group in real time, receive pressure data, flow rate data, and temperature data, and simultaneously identify and eliminate outliers through the Isolation Forest algorithm;

[0030] Generate a pressure curve based on the pressure data, generate a flow rate curve by combining the flow rate data and the pump efficiency of the digital pump, correlate the temperature data and the operating status of the pump group to generate a power curve, form an operating curve, and extract statistical features and morphological features from the operating curve;

[0031] Calculate the similarity between the operating curve and the historical normal operating curve through the Dynamic Time Warping algorithm, configure a similarity threshold, and compare the similarity with the similarity threshold to trigger fault detection.

[0032] As an alternative implementation, the generation logic for the valve flap action instruction includes:

[0033] Obtain pressure data, flow rate data, and temperature data in real time, calculate the pressure change rate in real time based on the pressure data, configure a target flow rate, and compare the flow rate data with the target flow rate to obtain a flow rate deviation;

[0034] Train the historical operating data through a Convolutional Neural Network to generate a fuzzy rule base, transmit the pressure change rate and the flow rate deviation to the fuzzy rule base for fuzzy inference, and obtain a fuzzy matching result;

[0035] Convert the fuzzy matching result into a valve flap action instruction through the centroid method. The valve flap action instruction includes the opening and closing state of the valve flap, the valve flap speed, and the valve flap execution angle, and feed back the execution result of the valve flap action instruction to the fuzzy rule base.

[0036] As an alternative implementation, the prediction logic for the pressure set value includes:

[0037] Obtain pressure data, flow rate data, temperature data, pipeline leakage rate, and environmental data to form multi-source data, preprocess the multi-source data and align it according to the time series;

[0038] Train a Long Short-Term Memory network with historical multi-source data, and input the multi-source data into the trained Long Short-Term Memory network to output the predicted pressure set value;

[0039] Judge the prediction deviation between the predicted pressure set value and the pressure data to adjust the Long Short-Term Memory network, and dynamically adjust the predicted pressure set value according to the pipeline leakage rate.

[0040] As an alternative implementation, the fault tolerance mechanism includes:

[0041] Close the valve flap of the faulty pump, switch to the standby pump, and dynamically adjust the pipe network connection relationship;

[0042] Send the configuration information of the current pump group to trigger the re - solution of the pressure set value and update the fuzzy rule base;

[0043] Real - time monitor the operation status of the pump group to detect the faults of the current pump group and optimize the configuration information of the current pump group.

[0044] Compared with the prior art, the beneficial effects of this application are as follows: By integrating multi - source data such as pressure, flow rate, and temperature, it realizes adaptive control under complex working conditions, breaks through the dependence on a single parameter in traditional control, and improves the response ability of the control system to water use fluctuations and equipment state changes; Based on the long - short - term memory network trained by historical multi - source data, the control system can predict the pressure set value in advance, solve the pressure set value through the particle swarm algorithm, and perform weight fusion on the predicted pressure set value and the solved pressure set value to generate adjustment instructions, so as to reduce the number of equipment starts and stops and mechanical wear, reduce maintenance costs, and at the same time avoid water resource waste through precise control; At the same time, through redundant design and fault - tolerance mechanism, it ensures that the control system can still operate stably in scenarios such as sensor failures and equipment anomalies.

[0045] The fuzzy rule base trained based on historical operation data enables the valve speed to be dynamically matched with the water flow characteristics, avoids water hammer impact or response delay caused by traditional fixed - speed control, improves the pressure stability of the pipe network, and continuously optimizes the valve action through closed - loop feedback to adapt to long - term operation scenarios such as equipment aging and pipe network resistance changes. The multi - source data - driven long - short - term memory network captures the temporal characteristics and environmental correlations of pressure changes, realizes the trend prediction of the pressure set value, real - time monitoring and self - correction of prediction deviation, enabling the control system to adapt to sudden working conditions and reducing pressure fluctuations caused by prediction errors.

[0046] Evaluate the load balance degree through the pump efficiency variance of the current pump group to generate standby pump start - stop instructions, avoid over - load operation of a single pump, extend the equipment life, dynamically calculate the number of standby pumps according to the flow demand and pump efficiency, ensure sufficient flow supply and reasonable equipment configuration economy, search for the optimal solution through the particle swarm algorithm, avoid the bias of traditional control towards a single target, and at the same time perform weight fusion on the predicted pressure set value and the solved pressure set value to adjust the perception control module.

[0047] The dynamic time warping matching of multi - parameter curves realizes the comparative analysis of the real - time operation curve and the historical normal operation curve, can capture the subtle changes in equipment performance, realizes early fault warning, is more comprehensive and reliable than single - parameter alarm, and the hierarchical warning mechanism ensures that faults of different severity levels are reasonably responded to. Through the rapid isolation of the faulty pump and the switching of the standby pump, and the automatic reconstruction of system parameters after switching, it ensures the control accuracy under the new pump group configuration. Brief Description of the Drawings

[0048] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Among them:

[0049] Figure 1 It is the assembly structure diagram of a water supply and drainage pump group assembly provided by the embodiment of the present application;

[0050] Figure 2 It is the system flow chart of a digital control system for a water supply and drainage pump group assembly provided by the embodiment of the present application;

[0051] Figure 3 It is the generation logic diagram of the start-stop command for the standby pump of a digital control system for a water supply and drainage pump group assembly provided by the embodiment of the present application;

[0052] Figure 4 It is the sub-logic diagram for solving the pressure set value of a digital control system for a water supply and drainage pump group assembly provided by the embodiment of the present application.

[0053] Reference numerals: 1, main inlet pipe; 101, first-side main inlet pipe flange; 102, negative pressure monitoring device; 103, inlet pipeline; 104, second-side main inlet pipe flange; 2, first digital pump; 3, flexible joint; 4, digital device; 5, digital adjustment device; 6, non-throttling device; 601, non-throttling device flange; 602, pressure monitoring device; 603, non-throttling main pipeline; 604, eccentric reducer; 605, non-throttling pipeline for the second digital pump; 606, elbow; 7, digital monitoring system; 8, second digital pump. Detailed Embodiments

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present application more apparent and understandable, the following clearly and completely describes the technical solutions in the embodiments of the present application with reference to the drawings in the specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0055] Embodiment 1:

[0056] As Figure 2 shown, a system flow chart of a digital control system for a water supply and drainage pump group assembly is provided by the embodiment of the present application. The system includes a sensing control module, a pump group scheduling module, and a fault-tolerant switching module.

[0057] The perception control module is used to obtain pressure data, flow data and temperature data in real time, generate a valve flap action instruction according to the pressure data and flow data and combined with the fuzzy control algorithm, and predict the pressure set value based on the temperature data and the pipeline leakage rate through a long short-term memory network.

[0058] Specifically, the generation logic of the valve flap action instruction includes:

[0059] Obtain pressure data, flow data and temperature data in real time, calculate the pressure change rate in real time based on the pressure data, configure the target flow, and compare the flow data with the target flow to obtain the flow deviation;

[0060] Train the historical operation data based on the convolutional neural network to generate a fuzzy rule base, transmit the pressure change rate and the flow deviation to the fuzzy rule base for fuzzy inference, and obtain the fuzzy matching result;

[0061] Convert the fuzzy matching result into a valve flap action instruction through the centroid method. The valve flap action instruction includes the opening and closing state of the valve flap, the valve flap speed and the valve flap execution angle, and feedback the execution result of the valve flap action instruction to the fuzzy rule base.

[0062] The pressure change rate and the flow deviation are the core indicators reflecting the nature of the pipeline network, which need to be calculated and obtained through real-time data, providing input variables for fuzzy control; the pressure sensors are deployed in a distributed and built-in manner, directly embedded inside the flange at the outlet of the pump cavity to ensure real-time acquisition of the pressure data at the pump body outlet, the main pipeline of the pipeline network and the end nodes, and the ultrasonic flowmeter is installed by clamping to obtain the flow data, where it is rigidly connected to the pipeline through the flange to avoid the influence of flow channel disturbance on the measurement accuracy. The temperature sensors cover the digital pump and the main pipeline of the pipeline network. The pressure change rate is calculated by the pressure difference between two adjacent sampling periods, such as an interval of 100 ms. To avoid instantaneous interference, a median filtering method with a sliding window containing 5-cycle data is used to smooth the result.

[0063] The target flow is dynamically issued by the pump group scheduling module according to the current working conditions, where the working conditions include peak or low valley periods. The flow deviation between the flow data and the target flow is expressed in percentage. For example, if the target flow is 1000 m³ / h and the actually obtained flow data is 850 m³ / h, the flow deviation is -15%. At the same time, a mutual verification mechanism for sensor signals is established. The pressure data and the flow data need to meet physical consistency. For example, when the pressure data rises, the flow data should not reverse. If contradictory signals appear, such as the pressure data rising while the flow data drops by more than 5%, trigger the sensor fault diagnosis process and automatically switch to the backup sensor for acquisition; thus dynamically capturing the pressure fluctuation trend and the flow matching degree, providing a quantitative basis for the valve flap speed and avoiding the hysteresis of static threshold control.

[0064] The pipeline network operating conditions are complex, such as pump startup and shutdown, valve flap movement, and water consumption fluctuations. Nonlinear control needs to be achieved through a data-driven fuzzy rule base to replace the traditional fixed logic. By analyzing historical operation data through a convolutional neural network, where the historical operation data includes the pressure change rate, flow deviation, and valve movement records under different operating conditions, features are automatically extracted and the correlation patterns between pressure changes and flow deviations are identified. For example, the convolutional neural network can identify that the combination of "high pressure change rate and large negative flow deviation" corresponds to the operating condition of "rapid valve flap opening", and then a fuzzy rule base is generated. The fuzzy rule base includes 25 basic rules. Through a reinforcement learning algorithm, the rule weights are dynamically adjusted according to the historical error of valve flap movement. For example, when the error rate of a certain fuzzy rule execution is continuously less than 5%, the trigger priority of this fuzzy rule is increased.

[0065] Among them, fuzzy inference includes fuzzifying the pressure change rate and flow deviation through a triangular membership function. The pressure change rate is divided into three levels: "low", "medium", and "high". For example, when the pressure change rate ≤ 0.02 MPa / s is low, the pressure change rate between (0.02 MPa / s, 0.1 MPa / s) is medium, and the pressure change rate ≥ 0.1 MPa / s is high. The flow deviation is divided into three levels: "large negative", "close", and "large positive". For example, when the flow deviation ≤ -10% is large negative, the flow deviation within (-10%, 10%) is close, and the flow deviation ≥ 10% is large positive. There are a total of 9 combinations. For example, if the current pressure change rate is 0.12 MPa / s and the degree of belonging to "high" is 0.8, and the flow deviation is -18% and the degree of belonging to "large negative" is 0.9, then the fuzzy matching result of "high pressure change rate and large negative flow deviation → rapid valve flap opening" is activated. The inference result is defuzzified by the centroid method and converted into specific valve movement parameters. For example, the valve flap opens, the valve flap speed is 50° / s, and the valve flap execution angle is 70°. The valve flap speed and valve flap execution angle are adaptively adjusted to avoid the water hammer effect or response delay caused by a fixed speed. The fuzzy matching result output by fuzzy inference is converted into specific command parameters through defuzzification.

[0066] It is necessary to convert the fuzzy matching result into a physical instruction recognizable by the digital regulating device 5 and continuously optimize the fuzzy rule base through a feedback mechanism. The defuzzified valve flap speed needs to be adapted in combination with the valve flap specifications. For example, the maximum safe opening speed of a certain valve flap is 40° / s. If the defuzzification result is a valve flap speed of 50° / s, it is automatically adjusted to 40° / s and marked as "speed limited". The valve flap movement instruction includes the valve flap opening and closing state, valve flap speed, valve flap execution angle, and priority. For example, the priority of "emergency opening" is higher than that of "normal regulation", and it is transmitted to the digital regulating device 5 to control the valve flap movement of the digital pump.

[0067] After receiving the valve flap action instruction, the digital adjustment device 5 first detects the initial resistance through the built-in torque sensor. If the resistance is greater than 110% of the rated torque, for example, the normal resistance is 100 N·m and the measured value is 115 N·m, it is determined that the valve flap is stuck. After automatically retracting 5°, it retries. If it fails to retry 3 times continuously, it sends a fault signal to the fault-tolerant switching module. During the action process, the encoder real-time feedbacks the valve execution angle and dynamically adjusts the motor voltage to ensure that the deviation between the actual valve flap speed and the valve flap speed in the valve flap action instruction is within ±5%. When the valve flap action is completed, the actual execution time, valve flap execution angle and other data are compared with the parameters of the valve flap action instruction to generate an execution error. If the execution error is greater than the error threshold, for example, the angle error threshold is ±3°, the pressure change rate, flow deviation and error value under this working condition are stored in the historical database as the correction samples for the next training of the fuzzy rule base. The closed-loop control improves the execution accuracy of the valve flap action instruction. At the same time, the precise valve flap control directly affects the stability of the pipe network pressure and provides more reliable front-end execution data for the prediction of the pressure set value.

[0068] Specifically, the prediction logic of the pressure set value includes:

[0069] Obtain pressure data, flow data, temperature data, pipe network leakage rate and environmental data to form multi-source data, preprocess the multi-source data and align it according to the time series;

[0070] Train the long short-term memory network through historical multi-source data, and input the multi-source data into the trained long short-term memory network to output the predicted pressure set value;

[0071] Judge the prediction deviation between the predicted pressure set value and the pressure data to adjust the long short-term memory network, and dynamically adjust the predicted pressure set value according to the pipe network leakage rate.

[0072] The pressure set value is affected by multiple factors, such as temperature, pipeline network leakage, and environment. It is necessary to integrate multi-source data and preprocess it to improve the accuracy of the long short-term memory network; real-time obtain pressure data, flow data, temperature data, pipeline network leakage rate, and environment data to form multi-source data. Among them, the pipeline network leakage rate is measured by the night minimum flow method and can be automatically measured at 2 am every day because the flow is the smallest at this time period, and the leakage ratio is significant. The environment data includes time period labels, including peak, flat peak, and trough time periods, and weather, including rainy days and sunny days. Convert the time period labels into one-hot encoding, and convert the weather into numerical features. For example, rainy day = 1, sunny day = 0. Align the multi-source data in time, and identify outliers in the pressure data and flow data through the isolation forest algorithm. For example, an outlier with a sudden pressure drop of 0.5 MPa and no corresponding flow change, and fill in the interpolation through the data of the previous and subsequent time points; multi-dimensional data coverage can improve the comprehensiveness of the input of the long short-term memory network. The preprocessing of multi-source data can eliminate data noise and dimensional differences. The processed multi-source data is input into the long short-term memory network according to the time series for training and prediction.

[0073] The pressure change has time series dependence and non-linear characteristics. The long short-term memory network can capture long-term dependence relationships. A three-layer LSTM network is adopted. The first layer receives a 5-dimensional feature vector of pressure data, flow data, temperature data, pipeline network leakage rate, and environment data. The second layer captures time series features through the attention mechanism. The third layer outputs the pressure set value. The long short-term memory network is trained through historical multi-source data, where the historical multi-source data includes a training set and a validation set. The preprocessed multi-source data is input into the trained long short-term memory network, and the predicted pressure set value is output. For example, according to the pipeline network leakage rate of 3% and the peak time period, the predicted pressure set value for the peak time period is the current pressure data plus the pressure data to be compensated, which is 0.4 MPa. The predicted pressure set value is smoothed through a moving average filter window of 3 predicted values to avoid misjudgment caused by single-step fluctuations; by calculating the prediction error distribution in the long short-term memory network, a confidence interval for the predicted pressure set value is generated. For example, the 95% confidence interval is ±0.02 MPa. If the current working condition is significantly different from the training set, such as a sudden heavy rain causing a mutation in the water use pattern, the confidence interval automatically expands, prompting the pump group scheduling module to reduce the weight of the predicted pressure set value.

[0074] Actual operating conditions fluctuate, such as unexpected surges in water usage and aging of pipeline equipment, which can lead to prediction deviations. Dynamic adjustment is required to maintain accuracy. The prediction deviation between the pressure set value predicted in real-time and the pressure data is calculated, and different responses are triggered according to the magnitude of the prediction deviation. According to the prediction deviation threshold, the magnitude of the prediction deviation is divided into minor deviation, significant deviation, and severe deviation. When it is a minor deviation, only the prediction deviation is recorded, and the sensing control module compensates by adjusting the valve flap execution angle. When it is a significant deviation, the weights of the input data of the long short-term memory network are dynamically adjusted, such as increasing the weight of the pipe network leakage rate. When it is a severe deviation, the long short-term memory network is reconstructed, and the current abnormal condition data is automatically appended to the training set to trigger incremental training of the long short-term memory network. At the same time, according to the real-time monitored pipe network leakage rate, when the pipe network leakage rate is greater than 5%, the pressure set value is manually or automatically increased to compensate for the pressure loss caused by pipe network leakage. The closed-loop feedback enables the long short-term memory network to have self-adaptability, and dynamic correction can reduce long-term prediction errors, ensuring that the pressure set value is close to the actual operation requirements. Accurate pressure prediction provides more reliable target parameters for the pump group scheduling module, reducing system energy consumption and pressure fluctuations.

[0075] The pump group scheduling module is used to calculate the hydraulic power of the digital pump in real-time based on pressure data and flow data, and determine the pump efficiency of the digital pump by combining the motor power monitored in real-time. Generate start-stop commands for standby pumps according to the variance of the pump efficiency of the digital pump and the current pump group. At the same time, solve the pressure set value through the particle swarm algorithm based on pressure data and flow data, fuse the weights of the predicted pressure set value and the solved pressure set value to generate a pressure target value, and generate adjustment commands according to the pressure deviation between the pressure target value and the pressure data.

[0076] Specifically, as Figure 3 shown, the generation logic of the start-stop commands for standby pumps includes:

[0077] Receive pressure data and flow data, and monitor the motor operation data of the digital pump in real-time to convert it into the motor power of the digital pump;

[0078] Calculate the hydraulic power of the digital pump in real-time based on pressure data and flow data through the Bernoulli equation, determine the pump efficiency of the digital pump in real-time according to the ratio of hydraulic power to motor power, and analyze the change trend of the pump efficiency for anomaly detection;

[0079] Determine the load balance degree according to the variance of the pump efficiency of the current pump group, configure an enabling threshold, and compare the load balance degree with the enabling threshold to trigger the start-stop of the standby pump;

[0080] Determine the total flow rate of the current pump group according to the pump efficiency of the digital pump, correct the total flow rate of the current pump group based on the change trend of the pump efficiency, calculate the flow difference between the target flow rate and the corrected total flow rate of the current pump group, and determine the number of standby pumps according to the flow difference.

[0081] The operating status and energy consumption of the pump group are closely related to pressure, flow rate, and motor power. Only by obtaining these data can the working conditions of the digital pump be accurately evaluated, and then it can be determined whether to start or stop the standby pump. Continuously obtain pressure data and flow rate data, and at the same time install a power monitoring device on the motor of the digital pump to monitor the motor operating data in real time, such as voltage and current, etc., and convert it into motor power through a data conversion algorithm. Eliminate the high-frequency noise of the motor power through a sliding window filter with a window size of 10 sampling periods, and perform normalization processing on the pressure data and flow rate data to scale it to the [0,1] interval for subsequent unified processing by the algorithm, and establish a data integrity verification mechanism. If the motor power of a certain digital pump is continuously lost for 3 cycles, automatically switch to the standby sensor to obtain the motor power; thus, it is possible to comprehensively and real-time obtain the key data of the pump group operation, providing a basis for accurately evaluating the pump group operation status subsequently, ensuring that the control system understands the pump group operation situation in a timely and accurate manner, providing the necessary data support for calculating the hydraulic power of the digital pump and determining the pump efficiency, and enabling the subsequent steps to perform calculations and analyses based on accurate data.

[0082] The pump efficiency is an important indicator to measure the working performance of the digital pump. By calculating the pump efficiency and analyzing its change trend, it is possible to determine whether the digital pump is working properly and whether there are problems such as performance degradation. For example, when the pump efficiency suddenly drops, it means that there is a fault inside the digital pump or the operating conditions have changed, and it needs to be processed in a timely manner; utilize the principle of Bernoulli's equation, combined with the pressure data and flow rate data obtained in real time, calculate the hydraulic power of the digital pump at different times, and then compare the calculated hydraulic power with the motor power, and obtain the pump efficiency of the digital pump through division operation. At the same time, use data statistics and analysis algorithms to process the historical data of the pump efficiency, draw the change curve of the pump efficiency over time, and analyze the change trend of the pump efficiency by observing the trend and fluctuation of the change curve, and detect whether there are abnormal fluctuations. For example, set an efficiency threshold. When the pump efficiency of a certain digital pump is less than the efficiency threshold for 30 consecutive seconds and the flow rate data is greater than 90% of the target flow rate, it is determined to be overloaded. When the decrease rate of the pump efficiency is greater than the rate threshold, a warning is triggered; thus, it is possible to accurately evaluate the working performance of the digital pump, timely discover potential faults or performance problems, improve the reliability and stability of the control system, avoid situations such as water supply and drainage interruption caused by pump performance problems, provide a key basis for determining the load balance degree and judging whether to start or stop the standby pump, and enable the control system to reasonably adjust the operation configuration of the pump group according to the actual situation of the pump efficiency.

[0083] To ensure the overall efficient and stable operation of the pump group, it is necessary to make the loads of each digital pump as balanced as possible. By calculating the variance of pump efficiency to determine the load balance degree, the load distribution of the current pump group can be quantitatively evaluated. When the load imbalance reaches a certain level, it is necessary to start or stop the standby pump to adjust. Statistical analysis is carried out on the pump efficiency data of each digital pump in the current pump group, and the variance of the pump efficiency of the current pump group is calculated as a measure of the load balance degree. The load balance degree ranges from [0, 1], and a load balance degree of 1 indicates complete balance. According to the design requirements of the control system and actual operation experience, an appropriate activation threshold is pre-configured. The calculated load balance degree is compared with the activation threshold. If the load balance degree is less than the activation threshold, it means that the degree of load imbalance of the current pump group is relatively large, and the control system will trigger the startup procedure of the standby pump. If the load balance degree is greater than the activation threshold, the control system will trigger the shutdown procedure of the standby pump. Thus, the automatic balance adjustment of the load of the current pump group is realized, the overall operation efficiency of the current pump group is improved, the risk of damage to some digital pumps due to excessive load is reduced, the service life of the current pump group is extended, and at the same time, the stable operation of the control system is ensured, and it is clear whether it is necessary to start or stop the standby pump, providing a prerequisite for determining the number of standby pumps according to the flow demand later and making the configuration of the standby pump more reasonable.

[0084] The flow demand of the control system changes with time and water usage conditions. It is necessary to reasonably configure the number of standby pumps according to the actual demand to ensure the satisfaction of the flow supply. By comprehensively considering the target flow, the total flow of the current pump group, and the change trend of pump efficiency, the number of standby pumps can be determined more accurately. The control system obtains real-time flow data from the sensing control module and predicts the target flow at the current time point, that is, the flow demand at the current time point. According to the pump efficiency of each digital pump and the rated flow of the digital pump, the total flow of the current pump group in the current operating state is calculated. At the same time, referring to the change trend of pump efficiency, the total flow of the current pump group is corrected. For example, when the pump efficiency shows a downward trend, the estimated value of the total flow is appropriately reduced. Then, the flow difference between the target flow and the corrected total flow of the current pump group is calculated. According to the magnitude of the flow difference and the rated flow of the standby pump, the number of standby pumps to be started or stopped is determined. Among them, they are sorted from large to small according to the rated flow of the standby pump, and the digital pump with a large flow is preferentially started to reduce the start-stop times of the standby pump.

[0085] Thus, it is possible to accurately adjust the number of standby pumps according to the actual flow demand and the operating status of the pump group, which not only meets the flow supply requirements of the control system but also avoids the problems of energy waste and unstable operation caused by too many or too few standby pumps, improving the operating economy and reliability of the control system. After determining the specific start-stop quantity of the standby pumps, the control system will execute the start-stop operation of the standby pumps according to this result to complete the configuration adjustment of the pump group, enabling the current pump group to operate in a better state. At the same time, it also provides a new basis for the pump group operating parameters for the subsequent solution of the pressure setpoint and the generation of adjustment instructions.

[0086] Furthermore, as Figure 4 shown, the sub-logic for solving the pressure setpoint includes:

[0087] Determine the fitness function based on system energy consumption, flow stability, and pressure deviation control, and use the predicted pressure setpoint and pipeline network properties as constraint conditions;

[0088] Initialize the particle swarm, calculate the fitness value of each particle according to the fitness function, update the velocity and position of each particle, and continuously iterate until the convergence condition is met to solve the pressure setpoint;

[0089] Verify the feasibility of the solved pressure setpoint through the constraint conditions and output the pressure setpoint that passes the verification.

[0090] The solution of the pressure setpoint needs to comprehensively consider multiple factors to achieve the optimal operation of the control system. The energy consumption of the control system is related to the operating cost, the flow stability affects the water supply and drainage effect, and the pressure deviation control is related to the reliability and safety of the control system. By determining the fitness function and constraint conditions, these factors can be quantified, thus finding the optimal pressure setpoint; combining historical operation data and practical experience, determine the weights of system energy consumption, flow stability, and pressure deviation control in the fitness function, and construct the fitness function. Among them, the system energy consumption refers to taking the total motor power of the pump group as the optimization goal, and the lower the power, the higher the fitness. The flow stability refers to taking the sum of the squares of the flow differences as the index, and the smaller the flow difference, the higher the fitness. The pressure deviation control refers to taking the pressure deviation between the pressure setpoint and the pressure data as the index, and the smaller the pressure deviation, the higher the fitness.

[0091] At the same time, take the pressure setpoint predicted by the long short-term memory network before and the pipeline network properties such as the pipe diameter, material, and length of the pipeline network as the constraint conditions for solving the pressure setpoint; thus making the solution process of the pressure setpoint more scientific and reasonable, being able to comprehensively consider multiple key performance indicators of the system, avoiding unreasonable pressure settings caused by a single factor, improving the overall operating performance of the control system, and providing the objective function and constraint conditions for the initialization and subsequent calculation of the particle swarm, enabling the particle swarm to search for the optimal pressure setpoint within a reasonable range.

[0092] The particle swarm optimization algorithm is an optimization search algorithm that searches for the optimal solution by simulating the foraging behavior of a bird flock. By initializing the particle swarm and continuously updating the velocity and position of the particles, it can search in the solution space and gradually approach the optimal pressure setting value. In the control system, the scale of the particle swarm, that is, the number of particles, is set, and the position and velocity of each particle in the solution space are randomly initialized. Then, the position of each particle is used as a candidate solution for the pressure setting value and substituted into the fitness function to calculate the fitness value of each particle. According to the fitness value and the update rules of the particle swarm algorithm, the velocity and position of each particle are updated. This process is repeated and iterated continuously until the pre-set convergence conditions are met, such as the number of iterations reaching the limit or the optimal solution of the particle swarm no longer changing within a certain range. At this time, the optimal particle position obtained is the solved pressure setting value. Utilizing the global search ability of the particle swarm algorithm, it can efficiently find a pressure setting value close to the optimal in a complex solution space, improving the accuracy and efficiency of solving the pressure setting value. The obtained solved pressure setting value provides a data basis for subsequent feasibility verification and fusion.

[0093] The solved pressure setting value is theoretically optimal, but it may not be feasible under the actual pipeline network properties and control system operation requirements. Through feasibility verification, it can be ensured that the finally solved pressure setting value meets the actual operation requirements. The solved pressure setting value is substituted into the previously set constraint conditions, including the predicted pressure setting value range and pipeline network property constraints, etc., to check whether this pressure setting value meets all the constraint conditions, such as whether it is within the pressure range that the pipeline network can withstand, whether the deviation from the predicted value is within a reasonable range, etc. If all the constraint conditions are met, then this pressure setting value passes the verification and is output as the final pressure setting value. If not, it returns to readjust the parameters of the particle swarm algorithm or conduct a new search. Thus, the actual feasibility of the pressure setting value is guaranteed, avoiding problems such as pipeline network damage or unstable operation of the control system caused by unreasonable pressure settings, and improving the safety and reliability of the control system operation. The pressure setting value that passes the verification will be fused with the predicted pressure setting value to generate a pressure target value, providing a key parameter for the generation of subsequent adjustment instructions, enabling the control system to adjust the operation of the pump group according to the accurate pressure target value.

[0094] Specifically, the generation logic of the adjustment instruction includes:

[0095] Determine the prediction credibility according to the prediction error distribution of the long short-term memory network, dynamically adjust the weight based on the prediction credibility, and perform weighted fusion on the predicted pressure setting value and the solved pressure setting value to generate a pressure target value.

[0096] Calculate the pressure deviation between the calculated pressure target value and the pressure data, and at the same time determine the change rate of the pressure deviation. Then, divide the control area on the two-dimensional plane according to the pressure deviation and the change rate of the pressure deviation;

[0097] Determine the hierarchical control mechanism according to the control area, and generate an adjustment instruction according to the hierarchical control mechanism;

[0098] Obtain the operating state of the pump group and the predicted deviation between the predicted pressure setting value and the pressure data, and compensate the adjustment instruction according to the operating state of the pump group and the predicted deviation;

[0099] Receive the execution feedback data of the digital regulating device 5, determine the execution error between the execution feedback data and the adjustment instruction, and judge according to the execution error to optimize the adjustment instruction.

[0100] The pressure setting value predicted by the long short-term memory network and the pressure setting value solved by the particle swarm optimization algorithm each have their own advantages and limitations. By considering the prediction error distribution to determine the prediction credibility and dynamically adjusting the weights for fusion, the advantages of both can be integrated to obtain a more accurate and reliable pressure target value; the control system conducts statistical analysis on the prediction errors of the long short-term memory network in historical operations, calculates distribution characteristics such as the mean and variance of the prediction errors, and determines the prediction credibility based on this. According to the level of the prediction credibility, dynamically adjust the weights of the predicted pressure setting value and the solved pressure setting value in the weighted fusion. For example, when the prediction credibility is relatively high, appropriately increase the weight of the predicted pressure setting value; conversely, increase the weight of the solved pressure setting value. Then, perform weighted calculation on the two pressure setting values according to the adjusted weights to obtain the final pressure target value; thus, fully utilize the information of the two pressure setting value acquisition methods, and improve the accuracy and reliability of the pressure target value through reasonable weighted fusion, enabling the control system to more accurately control the operating pressure of the pump group, obtaining an accurate pressure target value, providing a benchmark for subsequent calculation of the pressure deviation and determination of the adjustment instruction, and enabling the adjustment instruction to more specifically adjust the operation of the pump group.

[0101] To effectively control the operating pressure of the pump group, it is necessary to understand the pressure deviation between the pressure target value and the pressure data, as well as the rate of change of the pressure deviation. By dividing the control area on a two-dimensional plane, different control strategies can be adopted according to different pressure deviation and rate of change situations; the control system obtains the current pressure data in real time, compares it with the generated pressure target value, calculates the pressure deviation, and at the same time analyzes the change of the pressure deviation over a certain period of time to determine the rate of change of the pressure deviation. Then, with the pressure deviation as the abscissa and the rate of change of the pressure deviation as the ordinate, different control areas are divided on the two-dimensional plane according to the preset pressure deviation threshold, such as the emergency area, the transition area, and the stable area; the pressure control situation is clearly classified, enabling the control system to adopt corresponding control strategies according to different pressure states, improving the flexibility and effectiveness of pressure control, avoiding the blindness of control strategies, determining the control area to which the current pressure state belongs, providing a basis for generating adjustment instructions according to the hierarchical control mechanism later, and making the adjustment instructions more in line with the actual pressure control requirements.

[0102] Different control areas represent different pressure states and require different intensities and methods of control strategies. Through the hierarchical control mechanism, refined control can be carried out for different pressure situations, improving the response speed and control effect of the control system; the control system determines the corresponding control strategies for each control area according to the preset hierarchical control rules. For example, in the emergency area, it is necessary to give priority to ensuring pressure stability and can quickly start the standby pump. In the transition area, it is necessary to balance pressure control and system energy consumption and finely adjust the operating parameters of the pump group. In the stable area, it is necessary to minimize control actions to reduce the wear of digital pumps. Then, according to the determined control strategies, specific adjustment instructions are generated, such as adjusting the speed of digital pumps and changing the valve flap execution angle, etc.; thus, hierarchical and precise control of the operating pressure of the pump group is achieved, improving the adaptability and stability of the control system under different pressure states, ensuring that the control system can operate stably within the target pressure range, and the generated adjustment instructions will be used to control the operation of the pump group, while providing a basis for optimizing the adjustment instructions according to the pump group operation state and execution feedback data later.

[0103] The actual operating state of the pump group is affected by various factors and may differ from the expected situation. At the same time, there will also be certain errors in the predicted pressure set value. By obtaining this information and compensating the adjustment instructions, the adjustment instructions can be made more in line with the actual operating requirements and the control effect can be improved. The control system obtains the operating state information of the pump group in real time through sensors and monitoring devices, such as the rotation speed, current, and vibration of the digital pump. At the same time, it calculates the predicted deviation between the predicted pressure set value and the pressure data. Then, based on the operating state of the pump group and the predicted deviation, it uses a pre-set compensation algorithm to correct and compensate the adjustment instructions. If the operating state of the pump group shows abnormal vibration, the adjustment range of the rotation speed of the digital pump in the adjustment instructions is appropriately reduced. If the predicted deviation is large, the adjustment instructions will be strengthened or adjusted in direction accordingly. Thus, the adjustment instructions can better adapt to the actual operating conditions of the pump group, make up for the influence brought by the prediction error and the change of the operating state, improve the accuracy and stability of the control system, and ensure the normal operation of the control system. The compensated adjustment instructions will more accurately control the operation of the pump group, and at the same time, the execution effect will be evaluated through the execution feedback data of the digital adjustment device 5, providing a basis for optimizing the adjustment instructions in the future.

[0104] Due to various interference factors in the actual operation process, there will be errors in the execution of the adjustment instructions by the digital adjustment device 5. By analyzing the execution errors, problems can be discovered in a timely manner and the adjustment instructions can be optimized to ensure that the pump group operates according to the expected goals. The control system receives the execution data feedback by the digital adjustment device 5 in real time, such as the valve flap execution angle and the rotation speed of the digital pump. It compares the execution feedback data with the target parameters in the adjustment instructions, calculates the execution errors, and then optimizes the adjustment instructions according to the magnitude and direction of the execution errors. If the execution errors are large, the control system will readjust the parameters of the adjustment instructions. If the execution errors are small and within the acceptable range, the control system will slightly adjust the adjustment instructions to further improve the control accuracy. Thus, a closed-loop control of the execution process of the adjustment instructions is achieved, which can discover and correct the execution deviations in a timely manner, continuously optimize the adjustment instructions, improve the accuracy and reliability of the control system, ensure the stability and accuracy of the pressure control of the control system, and the optimized adjustment instructions will continue to be used to control the operation of the pump group, forming a continuously optimized control loop, enabling the control system to continuously adapt to the changing operating conditions and maintain efficient and stable operation.

[0105] The fault-tolerant switching module is used to monitor the operating state of the pump group in real time, generate an operating curve by combining pressure data, flow data, and temperature data, and match the operating curve with the historical normal operating curve through the dynamic time warping algorithm to trigger fault detection, locate the faulty pump and switch it to the standby pump, and at the same time trigger the fault-tolerant mechanism.

[0106] Specifically, the triggering logic of fault detection includes:

[0107] Monitor the operating status of the pump unit in real time, receive pressure data, flow data and temperature data, and identify and eliminate outliers through the Isolation Forest algorithm;

[0108] Generate a pressure curve based on the pressure data, generate a flow curve by combining the flow data and the pump efficiency of the digital pump, correlate the temperature data and the operating status of the pump unit to generate a power curve, form an operating curve, and extract statistical features and morphological features from the operating curve;

[0109] Calculate the similarity between the operating curve and the historical normal operating curve through the Dynamic Time Warping algorithm, configure a similarity threshold, and compare the similarity with the similarity threshold to trigger fault detection.

[0110] During the operation of the pump unit, abnormal values may appear in data such as pressure, flow, and temperature due to factors such as sensor failures and environmental interference. These abnormal values will affect the judgment of the true operating status of the pump unit. Using the Isolation Forest algorithm to eliminate outliers can ensure that subsequent analysis is based on reliable data and accurately detect faults. For example, if the sensor is occasionally affected by electromagnetic interference and generates incorrect data, it will misjudge the pump unit failure if not processed; continuously obtain data and use the Isolation Forest algorithm to construct multiple isolation trees, evaluate whether a data point is an outlier by calculating the path length of each data point in the isolation tree, and eliminate the data points determined to be abnormal; thus effectively improving the data quality, avoiding misjudgment caused by abnormal data, enabling the control system to analyze the operating status of the pump unit more accurately, providing a reliable basis for fault detection, reducing unnecessary shutdown maintenance, improving the operating efficiency of the control system, laying a foundation for generating accurate operating curves, ensuring that the features extracted from the subsequent operating curve truly reflect the operating conditions of the pump unit, and making the fault detection results more reliable.

[0111] Single data is difficult to comprehensively reflect the operating conditions of the pump group. By separately plotting data such as pressure, flow rate, and temperature into curves and combining them with the operating state of the pump group, the operating laws of the pump group can be presented in a visual way, statistical features and morphological features can be extracted, which is convenient for comparison with the historical normal operating curves, timely detection of changes in the operating trend, and detection of potential faults. The control system plots the pressure curve according to the processed pressure data in chronological order, combines the flow rate data with the pump efficiency of the digital pump at the corresponding time point to plot the flow rate curve, and at the same time correlates the temperature data with the operating state of the pump group to generate the power curve, integrates the three curves to form a complete operating curve, and then extracts statistical features such as mean, variance, and slope from the operating curve, as well as morphological features such as the curve fluctuation form, rising or falling trend, etc. Thus, the operating state of the pump group is displayed in an intuitive and comprehensive manner. Through quantitative feature extraction, complex operating data is transformed into information that is convenient for analysis and comparison, improving the accuracy and timeliness of fault detection, being able to detect abnormalities at the initial stage of the fault, and the extracted features are the key basis for matching with the historical normal operating curves, determining the accuracy of fault detection triggering, and providing clues for subsequent location of the faulty pump.

[0112] Due to factors such as equipment aging and working condition changes, the operating curve of the pump group will have a certain difference from the historical normal operating curve, but the normal fluctuations are within the acceptable range. By setting a similarity threshold and making a comparison, normal fluctuations and abnormal changes caused by faults can be distinguished, avoiding false triggering of fault detection. The control system pre-stores the historical normal operating curves of the pump group in the normal operating state, inputs the currently generated operating curve and the historical normal operating curve into the dynamic time warping algorithm. This dynamic time warping algorithm calculates their similarity by finding the optimal time alignment path between the two operating curves. At the same time, according to the operating characteristics of the pump group and historical data, a suitable similarity threshold is configured, and the calculated similarity is compared with the similarity threshold. When the similarity is less than the similarity threshold, it is determined that the current pump group has an abnormality and fault detection is triggered. Thus, it can accurately identify the fault situation in the operation of the pump group, reduce the false alarm rate, trigger detection in a timely manner when a fault occurs, enable the control system to respond quickly, and ensure the stable operation of the control system. Once the fault detection is triggered, the control system will start the process of locating the faulty pump and executing the fault tolerance mechanism, and handle the faulty pump in a timely manner to avoid the expansion of the fault affecting the operation of the control system.

[0113] Specifically, the fault tolerance mechanism includes:

[0114] Close the valve flap of the faulty pump and switch to the standby pump, and at the same time dynamically adjust the pipe network connection relationship;

[0115] Send the configuration information of the current pump group to trigger the re-solving of the pressure set value and update the fuzzy rule base;

[0116] Monitor the operation status of the pump group in real time to detect the faults of the current pump group and optimize the configuration information of the current pump group.

[0117] After a faulty pump is detected, closing its valve flap can prevent the spread of the fault and avoid affecting other devices. At the same time, promptly switching to the standby pump can ensure the uninterrupted operation of the control system. Dynamically adjusting the pipe network connection relationship enables the control system to still maintain a reasonable water flow distribution and maintain the stability of the control system after the pump group configuration changes. After the control system triggers the fault detection, it sends a closing instruction to the digital adjustment device 5 of the faulty pump to close the valve flap through the digital adjustment device 5, and at the same time sends a start instruction to the standby pump to start the standby pump according to the preset start procedure. In addition, the control system recalculates the water flow path and flow distribution based on the current pump group configuration and the pipe network topology structure, and dynamically adjusts the pipe network connection relationship by adjusting the valve flap execution angle of the digital pump and other methods to ensure smooth water flow. Thus, the rapid isolation of the faulty pump and the seamless switching of the standby pump are achieved, ensuring the continuity of the control system, avoiding problems such as water supply interruption caused by pump failures, dynamically adjusting the pipe network connection relationship, optimizing the operation efficiency of the control system, reducing the pressure fluctuations and water flow instability caused by pump group switching, and providing new pump group operation conditions for re-solving the pressure set value and updating the fuzzy rule base. Because after the pump group configuration changes, the pressure and flow requirements will also change, and the control parameters need to be re-optimized.

[0118] After the pump group configuration changes, the original pressure set value and fuzzy control rules are no longer applicable. Re-solving the pressure set value enables the control system to adjust the pressure according to the new operation status to meet the water supply and drainage requirements. Updating the fuzzy rule base can optimize the generation logic of the valve flap action instruction and improve the control accuracy of the control system. The control system re-solves the pressure set value and updates the fuzzy rule base according to the current pump group configuration information, including the number and model of digital pumps and the status of standby pumps. Among them, based on the new pump group configuration, combined with pressure data, flow data, etc., methods such as the particle swarm algorithm are re-used to solve the pressure set value. The update of the fuzzy rule base re-analyzes and trains the historical operation data according to the changes in the pump group operation status and the new pressure set value, and updates the rules in the fuzzy rule base to adapt to the new operation conditions. Thus, it is ensured that after the pump group configuration changes, the control system can quickly adjust the control parameters, maintain pressure stability, improve the accuracy of valve flap control, keep the control system running efficiently all the time, and enhance the stability and reliability of the control system. The updated pressure set value and fuzzy rule base provide new control bases for the perception control module and the pump group scheduling module, enabling each module of the control system to work together and continuously optimize the operation of the pump group.

[0119] Even after the faulty pump is switched and the parameters are adjusted, faults may still occur during the operation of the new pump set. Continuously monitoring the operating status can timely detect potential problems. Optimizing the configuration information of the pump set can further improve the operating efficiency of the pump set, reduce the probability of faults, and ensure the long-term stable operation of the control system. Continuously obtain the operating data of the pump set, such as pressure, flow rate, temperature, vibration, and rotational speed, analyze and process the obtained data to determine whether there are faults in the pump set. At the same time, according to indicators such as the pump efficiency and energy consumption of the current pump set, combined with historical data and operating experience, optimize the configuration information of the current pump set, such as adjusting the start-stop sequence of digital pumps, valve flap speed, and valve flap execution angle. Thus, the real-time monitoring and dynamic optimization of the operating status of the pump set are achieved, potential fault hazards can be detected in advance and processed in a timely manner, the service life of the pump set is extended, the overall operating efficiency of the control system is improved, the operating cost is reduced, and the long-term stable and reliable operation of the control system is ensured. The optimized pump set configuration information provides better operating parameters for each module of the control system, forming a continuously improving closed-loop control system, continuously enhancing the performance of the control system. If a fault occurs again, the control system can quickly respond and process based on a more reasonable configuration.

[0120] Embodiment 2:

[0121] As Figure 1 shown, the present application embodiment provides an assembly structure diagram of a water supply and drainage pump set assembly. The assembly includes: an inlet main pipe 1, a first digital pump 2, a flexible joint 3, a digital device 4, a digital adjustment device 5, a non-throttling device 6, a digital monitoring system 7, and a second digital pump 8.

[0122] The inlet main pipe 1 includes a first-side inlet main pipe flange 101, a negative pressure monitoring device 102, a water inlet pipeline 103, and a second-side inlet main pipe flange 104. The inlet main pipe 1 is connected to the first digital pump 2 and the second digital pump 8 through the flexible joint 3. The flexible joint 3 is used to compensate for the displacement of the inlet main pipe 1 and is connected to the external water source and the water inlet pipeline 103 through the first-side inlet main pipe flange 101 and the second-side inlet main pipe flange 104. The water flow in the water inlet pipeline 103 enters the first digital pump 2 and the second digital pump 8 after passing through the negative pressure monitoring device 102 in sequence. The negative pressure monitoring device 102 is used to monitor the pressure data in the inlet main pipe 1.

[0123] The inlet main pipe 1 serves as a channel for the external water source to enter the pump set. It is composed of a first-side inlet main pipe flange 101, a negative pressure monitoring device 102, a water inlet pipeline 103, and a second-side inlet main pipe flange 104. The function of the inlet main pipe 1 is to connect the external water source and guide the water flow into the pump set. The current pump set includes a first digital pump 2 and a second digital pump 8, and the negative pressure monitoring device 102 is used to monitor the inlet water pressure data in the inlet main pipe 1.

[0124] The first digital pump 2 and the second digital pump 8 serve as the power core equipment for water supply and drainage. Taking the first digital pump 2 and the second digital pump 8 as examples here, the number of digital pumps can be increased or decreased as needed. Driven by electric energy, the water flow introduced by the water inlet main pipe 1 is pressurized and transported to the non-throttling device 6. Among them, the flexible joint 3 is installed between the water inlet main pipe 1 and the digital pump, with the characteristics of elasticity and deformability, and is used to compensate for the displacement of the water inlet main pipe 1 caused by factors such as thermal expansion and contraction and installation errors, and reduce the vibration transmission during the operation of the control system.

[0125] The digital device 4 is connected to the first digital pump 2 and the second digital pump 8, and is used to monitor pressure data, flow data and temperature data. The digital adjustment device 5 is connected to the digital device 4, and is used to adjust the opening and closing state, valve flap speed and valve flap execution angle of the valve flap.

[0126] The digital device 4 is equipped with a variety of sensing elements such as a pressure sensor, a supersonic flowmeter and a temperature sensor, and can monitor the pressure data, flow data and temperature data during the operation of the first digital pump 2 and the second digital pump 8 in real time, providing key parameters of the operation state of the pump group for the control system, being an important source for the perception control module to obtain data, and providing a basis for subsequent valve action control, pump group scheduling and fault-tolerant switching. The digital adjustment device 5 is connected to the digital device 4, receives the valve action instruction to adjust the opening and closing state, valve flap speed and valve flap execution angle of the valve flap, and can accurately adjust the opening and closing state, valve flap speed and valve flap execution angle according to the received valve action instruction to control the size of the water flow channel, and then adjust the flow rate of the control system to achieve precise adjustment of the flow rate.

[0127] The non-throttling device 6 includes a non-throttling device flange 601, a pressure monitoring device 602, a non-throttling main pipeline 603, an eccentric reducer 604, a non-throttling pipeline 605 for the second digital pump and an elbow 606. The pressure monitoring device 602 is installed in the non-throttling main pipeline 603 to monitor the pressure data in the non-throttling main pipeline 603. The non-throttling pipeline 605 for the second digital pump is connected to the non-throttling main pipeline 603. The water flow enters the non-throttling main pipeline 603 after being rectified by the elbow 606. The eccentric reducer 604 is used for pipe diameter transformation in the non-throttling main pipeline 603. The non-throttling device 6 is connected to the water use point through the non-throttling device flange 601.

[0128] The non - throttling device 6 is the passage for the water flow output by the digital pump. It is composed of the non - throttling device flange 601, the pressure monitoring device 602, the non - throttling main pipeline 603, the eccentric reducer 604, the second digital pump non - throttling pipeline 605 and the elbow 606. It is used to transport the water flow pressurized by the digital pump to the water - using point. At the same time, the pressure monitoring device 602 monitors the pressure data of the outlet main pipeline. The elbow 606 rectifies the water flow to reduce hydraulic losses. The eccentric reducer 604 realizes the pipe diameter transformation to adapt to different flow and pressure requirements. The flanges on both sides are used to connect the water - using point pipeline, thus optimizing the water flow transportation.

[0129] The digital monitoring system 7 is used to collect the operation data of the digital pump, the digital device 4 and the digital regulating device 5, and adjust the pump group configuration information and the opening and closing state, valve flap speed and valve flap execution angle of the valve flap in the digital regulating device 5 in real time according to the historical operation data.

[0130] The digital monitoring system 7 collects the operation data of the digital pump, the digital device 4 and the digital regulating device 5, stores the historical operation data, analyzes the operation status of the pump group according to the historical operation data and the real - time operation data, adjusts the pump group configuration information in real time, optimizes the control parameters of the valve flap in the digital regulating device 5, ensures the efficient and stable operation of the pump group, and realizes the intelligent control of the pump group.

[0131] Among them, the first - side inlet main pipeline flange 101 and the second - side inlet main pipeline flange 104 provide connection interfaces for the inlet main pipeline 1, realizing the sealed connection between the inlet main pipeline 1 and the external water source and the inlet pipeline 103. The negative - pressure monitoring device 102 is used to monitor the pressure data in the inlet main pipeline 1 in real time and give timely feedback when the pressure drops abnormally, ensuring the stable inlet pressure of the control system and avoiding problems such as cavitation caused by negative pressure from affecting the operation of the pump group. The inlet pipeline 103 serves as the water flow transmission channel, guiding the water source to flow smoothly into the digital pump; the first digital pump 2 and the second digital pump 8 provide power for the transportation of the water flow, pressurize the water flow by means of impeller rotation, etc., to meet the pressure and flow requirements of the control system. The flexible joint 3 is used for shock absorption and noise reduction, reducing the impact of the vibration and noise generated during the operation of the digital pump on the control system, and at the same time compensating for the displacement of the inlet main pipeline 1 to adapt to the expansion or position change of the pipeline due to environmental factors; the digital device 4 is used to accurately measure and real - time feedback the key parameters during the operation of the digital pump, providing a data basis for the evaluation and control of the operation status of the control system.

[0132] The digital adjustment device 5 can, according to the requirements of the control system, adjust the water flow rate in the pipeline by changing the valve flap movement, achieving precise control of the flow rate; the non-throttling device flange 601 in the non-throttling device 6 is used to connect to subsequent water usage points or other pipelines, the pressure monitoring device 602 is used to monitor the pressure data in the non-throttling main pipeline 603 in real time and feedback it to the control system to maintain the stability of the outlet water pressure, the non-throttling main pipeline 603 is used to transport the water flow output by the digital pump to the water usage point, the eccentric reducer 604 changes the pipeline diameter according to actual requirements to optimize the water flow velocity and pressure distribution, the second digital pump non-throttling pipeline 605 is used to guide the water flow output by the second digital pump 8 to the non-throttling main pipeline 603, and the elbow 606 rectifies the water flow to reduce the hydraulic loss caused by water flow disorder and optimize the water flow pattern; the digital monitoring system 7 is used to collect the operation data of the digital pump, digital device 4 and digital adjustment device 5, store, analyze and calculate, and adjust the pump group configuration information based on historical operation data and real-time operation data to achieve the intelligent and efficient operation of the water supply and drainage pump group.

[0133] The digital device 4 transmits the real-time monitored pressure data, flow rate data and temperature data to the perception control module. Based on these data, the perception control module calculates the pressure change rate in combination with the pressure data, compares it with the configured target flow rate to obtain the flow rate deviation. The perception control module generates a valve flap movement instruction through a fuzzy control algorithm according to the pressure change rate and flow rate deviation and sends it to the digital adjustment device 5 for execution. After the digital adjustment device 5 executes the valve movement instruction, it feeds back the execution result of the valve flap movement instruction to the perception control module for the optimization of the fuzzy rule base. The perception control module predicts the pressure set value through a long short-term memory network based on the temperature data and the pipeline network leakage rate, etc. This pressure set value will affect the adjustment instruction of the digital monitoring system 7 for the pump group operation parameters.

[0134] The digital device 4 and the digital monitoring system 7 transmit the pressure data, flow rate data, etc. to the pump group scheduling module. The pump group scheduling module calculates the hydraulic power of the digital pump in real time based on these data, and determines the pump efficiency of the digital pump in combination with the motor power monitored in real time. According to the variance of the pump efficiency of the digital pump and the pump efficiency of the current pump group, the pump group scheduling module generates a start / stop instruction for the standby pump. If it is necessary to start / stop the standby pump, it controls the start / stop operation of the digital pump through the digital monitoring system 7. The pump group scheduling module solves the pressure set value through a particle swarm algorithm based on the pressure data and flow rate data, and performs weighted fusion with the pressure set value predicted by the perception control module to generate a pressure target value. The digital monitoring system 7 controls the digital adjustment device 5 to generate an adjustment instruction according to the pressure deviation between the pressure target value and the actual pressure data to adjust the operation state of the control system.

[0135] The digital monitoring system 7 collects the operation data of digital pumps, digital devices 4, etc. in real time and transmits it to the fault-tolerant switching module. The fault-tolerant switching module monitors the operation status of the pump group in real time through this data, generates an operation curve by combining pressure data, flow data, and temperature data. When a fault in the pump group is detected through algorithms such as the dynamic time warping algorithm, the fault-tolerant switching module locates the faulty pump, controls the digital monitoring system 7 to close the valve flap of the faulty pump, switches to the standby pump, and at the same time triggers the re-solving of the pressure set value and updates the fuzzy rule base, etc., to ensure the stable operation of the control system.

Claims

1. A digital control system for a water supply and drainage pump unit assembly, characterized in that, Including: Obtain pressure data, flow data, and temperature data in real time, generate valve flap action instructions according to the pressure data and flow data combined with the fuzzy control algorithm, and predict the pressure set value based on the temperature data and the pipeline network leakage rate through the long short-term memory network; Calculate the hydraulic power of the digital pump in real time based on the pressure data and flow data, determine the pump efficiency of the digital pump in combination with the motor power monitored in real time, generate standby pump start / stop instructions according to the variance of the pump efficiency of the digital pump and the pump efficiency of the current pump group, and solve the pressure set value through the particle swarm algorithm based on the pressure data and flow data, perform weighted fusion on the predicted pressure set value and the solved pressure set value to generate a pressure target value, and generate an adjustment instruction according to the pressure deviation between the pressure target value and the pressure data; Monitor the operation status of the pump group in real time, generate an operation curve in combination with the pressure data, flow data, and temperature data, match the operation curve with the historical normal operation curve through the dynamic time warping algorithm to trigger fault detection, locate the faulty pump and switch to the standby pump, and trigger the fault tolerance mechanism at the same time; The generation logic of the standby pump start / stop instruction includes: Receive pressure data and flow data, and monitor the motor operation data of the digital pump in real time to convert it into the motor power of the digital pump; Calculate the hydraulic power of the digital pump in real time based on the pressure data and flow data through the Bernoulli equation, determine the pump efficiency of the digital pump in real time according to the ratio of the hydraulic power to the motor power, and analyze the change trend of the pump efficiency for anomaly detection; Determine the load balance degree according to the variance of the pump efficiency of the current pump group, configure the enable threshold, and compare the load balance degree with the enable threshold to trigger the start / stop of the standby pump; Determine the total flow of the current pump group according to the pump efficiency of the digital pump, correct the total flow of the current pump group based on the change trend of the pump efficiency, calculate the flow difference between the target flow and the corrected total flow of the current pump group, and determine the number of standby pumps according to the flow difference; The generation logic of the adjustment instruction includes: Determine the prediction credibility according to the prediction error distribution of the long short-term memory network, dynamically adjust the weight based on the prediction credibility, perform weighted fusion on the predicted pressure set value and the solved pressure set value to generate a pressure target value; Calculate the pressure deviation between the pressure target value and the pressure data, determine the change rate of the pressure deviation at the same time, and divide the control area on the two-dimensional plane according to the pressure deviation and the change rate of the pressure deviation; Determine the hierarchical control mechanism according to the control area, and generate an adjustment instruction according to the hierarchical control mechanism; Obtain the operation status of the pump group and the prediction deviation between the predicted pressure set value and the pressure data, and compensate and adjust the instruction according to the operation status of the pump group and the prediction deviation; Receive the execution feedback data of the digital regulating device, determine the execution error between the execution feedback data and the adjustment instruction, and optimize the adjustment instruction according to the execution error judgment.

2. The digital control system of a water supply and drainage pump unit assembly according to claim 1, characterized in that, The solution sub-logic of the pressure set value includes: Determine the fitness function according to the system energy consumption, flow stability, and pressure deviation control, and use the predicted pressure set value and the pipeline network properties as constraints; Initialize the particle swarm, calculate the fitness value of each particle according to the fitness function, update the velocity and position of each particle, and iterate continuously until the convergence condition is met to solve the pressure set value; Verify the feasibility of the solved pressure set value through the constraint conditions, and output the pressure set value that passes the verification.

3. The digital control system of a water supply and drainage pump unit assembly according to claim 2, characterized in that, The trigger logic of the fault detection includes: Real-time monitor the operating status of the pump group, receive pressure data, flow data and temperature data, and at the same time identify and eliminate outliers through the Isolation Forest algorithm; Generate a pressure curve based on the pressure data, generate a flow curve by combining the flow data and the pump efficiency of the digital pump, correlate the temperature data and the operating status of the pump group to generate a power curve, form an operating curve, and extract statistical features and morphological features from the operating curve; Calculate the similarity between the operating curve and the historical normal operating curve through the Dynamic Time Warping algorithm, configure the similarity threshold, and compare the similarity with the similarity threshold to trigger fault detection.

4. The digital control system of a water supply and drainage pump unit assembly according to claim 3, characterized in that, The generation logic of the valve flap action instruction includes: Real-time obtain pressure data, flow data and temperature data, calculate the pressure change rate in real time based on the pressure data, configure the target flow, and compare the flow data with the target flow to obtain the flow deviation; Train the historical operation data based on the Convolutional Neural Network to generate a fuzzy rule base, transmit the pressure change rate and the flow deviation to the fuzzy rule base for fuzzy inference, and obtain the fuzzy matching result; Convert the fuzzy matching result into a valve flap action instruction through the centroid method. The valve flap action instruction includes the opening and closing state of the valve flap, the valve flap speed and the valve flap execution angle, and feedback the execution result of the valve flap action instruction to the fuzzy rule base.

5. The digital control system of a water supply and drainage pump unit assembly according to claim 4, characterized in that, The prediction logic of the pressure set value includes: Obtain pressure data, flow data, temperature data, pipeline leakage rate and environmental data to form multi-source data, preprocess the multi-source data and align it according to the time series; Train the Long Short-Term Memory network with historical multi-source data, and input the multi-source data into the trained Long Short-Term Memory network to output the predicted pressure set value; Judge the prediction deviation between the predicted pressure set value and the pressure data to adjust the Long Short-Term Memory network, and dynamically adjust the predicted pressure set value according to the pipeline leakage rate.

6. The digital control system of a water supply and drainage pump unit assembly according to claim 5, characterized in that, The fault tolerance mechanism includes: Close the valve flap of the faulty pump, switch to the standby pump, and dynamically adjust the pipeline connection relationship at the same time; Send the configuration information of the current pump group to trigger the re-solving of the pressure set value and update the fuzzy rule base; Real-time monitor the operating status of the pump group to detect the faults of the current pump group and optimize the configuration information of the current pump group.

7. A water supply and drainage pump unit assembly for implementing a digital control system of a water supply and drainage pump unit assembly according to any one of claims 1-6, characterized in that, Include: Inlet main pipe (1), first digital pump (2), flexible joint (3), digital device (4), digital regulating device (5), non-throttling device (6), digital monitoring system (7) and second digital pump (8); The inlet main pipe (1) includes a first-side inlet main pipe flange (101), a negative pressure monitoring device (102), a water inlet pipeline (103), and a second-side inlet main pipe flange (104). The inlet main pipe (1) is connected to the first digital pump (2) and the second digital pump (8) through a flexible joint (3). The flexible joint (3) is used to compensate for the displacement of the inlet main pipe (1), and is connected to an external water source and the water inlet pipeline (103) through the first-side inlet main pipe flange (101) and the second-side inlet main pipe flange (104). The water flow in the water inlet pipeline (103) enters the first digital pump (2) and the second digital pump (8) after passing through the negative pressure monitoring device (102) in sequence. The negative pressure monitoring device (102) is used to monitor the pressure data in the inlet main pipe (1).

8. The assembly of a water supply and drainage pump unit according to claim 7, characterized in that, The digital device (4) is connected to the first digital pump (2) and the second digital pump (8), and is used to monitor pressure data, flow data, and temperature data. The digital regulating device (5) is connected to the digital device (4), and is used to regulate the opening and closing state, the valve flap speed, and the valve flap execution angle of the valve flap; The non-throttling device (6) includes a non-throttling device flange (601), a pressure monitoring device (602), a non-throttling main pipeline (603), an eccentric reducer (604), a non-throttling pipeline for the second digital pump (605), and an elbow (606). The pressure monitoring device (602) is installed in the non-throttling main pipeline (603) to monitor the pressure data in the non-throttling main pipeline (603). The non-throttling pipeline for the second digital pump (605) is connected to the non-throttling main pipeline (603). The water flow enters the non-throttling main pipeline (603) after being rectified by the elbow (606). The eccentric reducer (604) is used for pipe diameter transformation in the non-throttling main pipeline (603). The non-throttling device (6) is connected to the water usage point through the non-throttling device flange (601); The digital monitoring system (7) is used to collect the operation data of the digital pumps, the digital device (4), and the digital regulating device (5), and adjust the pump group configuration information and the opening and closing state, the valve flap speed, and the valve flap execution angle of the valve flap in the digital regulating device (5) in real time according to the historical operation data.

Images