Modular intelligent pump station system

Abstract

The application relates to the field of automation control and discloses a modular intelligent pump station system. The system comprises a standardized water pump module, a distributed cooperative control unit, a pipe network state sensing network, a communication interaction network and a remote algorithm updating platform. The standardized water pump module integrates a local intelligent control terminal and a sensing component, is used for collecting operation state data and dynamically adjusting operation parameters. The distributed cooperative control unit is based on a non-cooperative game theory of multi-target constraint fusion to construct a control logic, so that each standardized water pump module reaches a Nash equilibrium of load distribution through autonomous negotiation. The pipe network state sensing network broadcasts global hydraulic working condition information to each standardized water pump module as a game environmental variable. The communication interaction network supports millisecond-level data exchange between modules. The remote algorithm updating platform is used for pushing optimized control algorithm parameters or game strategy models, so that self-organization and adaptive operation of the pump station system are realized, operation resilience is improved and energy consumption is optimized.

Description

Technical Field

[0001] This invention belongs to the field of automation control technology, specifically relating to a modular intelligent pumping station system. Background Technology

[0002] With the acceleration of urbanization and the in-depth advancement of smart water management, pumping station systems, as core infrastructure for fluid transportation and pressure regulation, play a vital role in municipal water supply, flood control and drainage, and industrial circulation. Modern pumping stations are evolving towards high integration and intelligence, aiming to achieve efficient allocation of water resources through precise flow control and energy management. Under complex and variable pipeline network conditions, the operational stability and response sensitivity of pumping station systems directly affect the safety of end-use water and the energy efficiency of transmission and distribution, requiring the system to possess excellent collaborative regulation capabilities and environmental adaptability.

[0003] Modular intelligent pumping stations, through the parallel combination of multiple standardized pump units, have become a major technical means to improve the flexibility and operational redundancy of water supply systems. These systems typically rely on a centralized control architecture, with a central logic controller uniformly scheduling the speed, start / stop, and valve opening of each pump module. The basic principle is to distribute the load of each module through linear adjustment logic based on a preset pressure target function and real-time feedback data, aiming to reduce overall system energy loss while meeting fluctuating water demand.

[0004] Traditional centralized control strategies have limitations in handling the dynamic coupling effects of multiple pumps operating in parallel. They are highly susceptible to water competition among modules when faced with transient flow surges, leading to frequent system pressure oscillations and increasing the risk of water hammer damage to pipelines. Furthermore, traditional architectures are heavily reliant on the core control unit, resulting in a very high risk of single-point failure. Failure of the central node can paralyze the entire system, lacking effective self-organization and self-healing capabilities. Existing load distribution algorithms often ignore the lifespan differences between individual devices, making it difficult to achieve dynamic balancing of operating wear and tear. This leads to premature failure of some pump sets due to excessive fatigue, shortening the overall lifespan of the system. Summary of the Invention

[0005] The purpose of this invention is to provide a modular intelligent pumping station system that can solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The system includes: multiple standardized water pump modules configured to perform fluid transport tasks; a distributed collaborative control unit embedded in each of the standardized water pump modules; a pipeline status sensing network distributed at key nodes of the main pipeline network; a communication interaction network for data exchange between the standardized water pump modules; and a remote algorithm update platform connected to the standardized water pump modules through a secure encrypted channel; wherein: each standardized water pump module integrates a local intelligent control terminal and integrated sensing components for real-time collection of the operating status data of the standardized water pump module and dynamic adjustment of operating parameters according to collaborative instructions; the distributed collaborative control unit constructs control logic based on non-cooperative game theory with multi-objective constraints, enabling each of the standardized water pump modules to reach a Nash equilibrium of load allocation through autonomous negotiation without central scheduling; the pipeline status sensing network is used to collect global hydraulic condition information and broadcast the data to each of the standardized water pump modules; the communication interaction network adopts an industrial-grade bus or wireless mesh network architecture, supporting millisecond-level data exchange between the standardized water pump modules; and the remote algorithm update platform is used to push optimized distributed control algorithm parameters or game strategy models.

[0007] Compared with the prior art, the present invention has at least the following beneficial effects:

[0008] The modular intelligent pumping station system provided by this invention abandons the traditional centralized control architecture and adopts a decentralized distributed intelligent agent collaborative mechanism. By embedding distributed collaborative control units in each standardized pump module and combining them with the pipeline network status perception network and communication interaction network, a decentralized collaborative control system is constructed. This allows each pump module to autonomously negotiate and achieve Nash equilibrium of load distribution based on a non-cooperative game model that integrates multi-objective constraints without central scheduling. This improves the system's resistance to single-point failures and operational reliability while enabling rapid response to flow changes. Furthermore, it can dynamically and smoothly adjust operating parameters based on global hydraulic condition information, effectively suppressing pressure oscillations and water hammer effects, and improving pipeline network operational stability. It can also achieve balanced load distribution among modules through a balanced scheduling mechanism, delaying equipment wear and aging and extending the overall service life. Relying on a remote algorithm update platform, it enables online optimization and iteration of control strategies, allowing the system to adapt to different operating conditions and possess continuous evolution capabilities. Attached Figure Description

[0009] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0010] Figure 1 This is a schematic diagram of the overall technical solution architecture according to the present invention;

[0011] Figure 2 This is a schematic diagram of the core principle framework of distributed collaborative control based on a biomimetic bee colony behavior model and non-cooperative game theory according to the present invention.

[0012] Figure 3 This is a flowchart illustrating the logical process of the distributed collaborative control unit in this invention performing multiple rounds of strategy iteration based on sudden changes in water demand to converge to an equilibrium solution.

[0013] Figure 4 This is a schematic diagram of the multi-level interaction relationship and data flow between the standardized water pump module, the pipeline network status sensing network and the remote algorithm update platform according to the present invention.

[0014] Figure 5 This is a flowchart illustrating the logical flow framework of the integrated lifetime balancing strategy and dynamic load reconfiguration in the event of module failure, as described in this invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] Example 1: Please refer to the appendix Figure 1 To be continued Figure 5 To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments.

[0017] This embodiment provides a modular intelligent pumping station system, which includes: multiple standardized water pump modules configured to perform fluid transport tasks; a distributed collaborative control unit embedded in each of the standardized water pump modules; a pipeline status sensing network distributed at key nodes of the main pipeline network; a communication interaction network for data exchange between the standardized water pump modules; and a remote algorithm update platform connected to the standardized water pump modules through a secure encrypted channel.

[0018] Each of the standardized water pump modules integrates a local intelligent control terminal and integrated sensing components, which are used to collect the operating status data of the standardized water pump module in real time and dynamically adjust the operating parameters according to the collaborative instructions;

[0019] The distributed collaborative control unit constructs control logic based on non-cooperative game theory that integrates multi-objective constraints, enabling each of the standardized water pump modules to reach a Nash equilibrium of load distribution through autonomous negotiation without central scheduling.

[0020] The pipeline status sensing network is used to collect global hydraulic condition information and broadcast the data to each of the standardized water pump modules.

[0021] The communication network adopts an industrial-grade bus or wireless mesh network architecture, supporting millisecond-level data exchange between the standardized water pump modules.

[0022] The remote algorithm update platform is used to push optimized distributed control algorithm parameters or game strategy models.

[0023] The standardized water pump modules can be used to carry out specific fluid pressurization and transportation functions, and each standardized water pump module can have independent logical decision-making capabilities; the distributed collaborative control unit can be used to establish a decentralized decision-making mechanism among the standardized water pump modules, and achieve system energy efficiency optimization through the integration of bionics and game theory models; the pipeline network status sensing network can be used to construct a global and local coupled operating condition monitoring system to provide real-time data support for decision-making; the communication interaction network can be used to support high-speed, low-latency data handshakes between various physical entities; and the remote algorithm update platform can be used to realize the dynamic evolution and environmental adaptation of the system control logic.

[0024] Optionally, in some embodiments, the standardized water pump module includes a physical pump body assembly, a drive motor unit, an integrated sensing assembly, and a local intelligent control terminal; the flow-through components of the physical pump body assembly are configured to maintain a preset hydraulic efficiency at different speeds; the drive motor unit is equipped with a frequency converter, which is configured to receive a control signal output by the local intelligent control terminal to achieve continuous adjustment of the motor speed; the integrated sensing assembly includes a high-frequency pressure sensor deployed at the outlet manifold of the standardized water pump module, the sampling frequency of which is configured to a preset frequency threshold capable of capturing transient water hammer signals; the integrated sensing assembly also includes a temperature sensor, a vibration sensor, and an energy metering chip.

[0025] Specifically, the standardized pump module serves as the core physical execution unit of the system, exhibiting a high degree of integration and standardization in its structure. Each standardized pump module may include a physical pump body assembly, a drive motor unit, integrated sensing components, and a local intelligent control terminal. The physical pump body assembly may include an impeller, pump casing, and mechanical seal structure, and its flow-through components can be optimized through computational fluid dynamics to ensure high hydraulic efficiency at different speeds. The drive motor unit may employ a permanent magnet synchronous motor or a high-performance asynchronous motor and may be equipped with a dedicated variable frequency drive. The variable frequency drive can be configured to receive pulse width modulation signals from the local intelligent control terminal to achieve continuous stepless adjustment of the motor speed.

[0026] The integrated sensing component may include a high-frequency pressure sensor and an electromagnetic or ultrasonic flow meter. The high-frequency pressure sensor can be deployed at the outlet manifold of the standardized pump module to collect millisecond-level pressure fluctuation data. Its sampling frequency can be configured to be no less than 1000 Hz to ensure the capture of transient water hammer signals generated by valve action or pump unit switching. The flow meter can be used to monitor the output flow of a single standardized pump module in real time and feed the flow data back to the local intelligent control terminal in real time. The standardized pump module may also integrate a temperature sensor, a vibration sensor, and an energy metering chip to monitor the motor stator winding temperature, bearing vibration intensity, and real-time power consumption.

[0027] The local intelligent control terminal, serving as the control core of the standardized water pump module, can be based on a high-performance microprocessor architecture with large-capacity flash memory and random access memory to run complex distributed computing tasks. The local intelligent control terminal can be pre-installed with a real-time operating system, which can be configured to concurrently process multiple task threads, such as sensor data acquisition, motor control logic calculation, communication protocol stack maintenance, and anomaly diagnosis. At the functional logic layer, the local intelligent control terminal can be configured to have self-sensing capabilities, enabling it to calculate the current operating efficiency curve, estimated remaining design life, and real-time mechanical loss index of the standardized water pump module based on internally integrated multi-dimensional sensor data and a preset transfer function model.

[0028] The physical carrier of the distributed collaborative control unit can be a software logic module embedded in the local intelligent control terminal of each standardized water pump module, and logically it can form a virtual decision-making cluster that transcends physical boundaries. In some examples, the distributed collaborative control unit can construct initial response logic based on a biomimetic bee colony behavior model. Under this logic, each standardized water pump module is endowed with worker bee attributes. When the pipeline network status sensing network detects pressure deviations caused by fluctuations in water demand, each standardized water pump module no longer passively waits for the central controller to schedule it, but can autonomously initiate a strategy search program based on its own perceived local pressure gradient and flow changes to achieve a decentralized response.

[0029] Optionally, in some embodiments, the distributed collaborative control unit guides each standardized water pump module to define a private revenue function based on its operating state in each decision cycle. The revenue function is a combination of multiple weighted terms, specifically including: an energy efficiency gain term, the value of which is positively correlated with the degree to which the standardized water pump module operates in the high-efficiency zone; a pressure deviation penalty term, which is positively correlated with the square of the deviation of the current outlet pressure value from the set target value; and a mechanical loss term, which is determined by the motor's start-stop frequency and the ratio of its operating speed to its rated speed. Under the constraint that the total flow demand equals the sum of the output flow of each standardized water pump module, each standardized water pump module uses a Nash equilibrium search algorithm based on iterative optimal response or gradient adjustment to correct the predicted speed value until all strategy changes are less than a preset threshold, thus achieving a Nash equilibrium state and optimizing the system's total energy consumption.

[0030] Specifically, the distributed collaborative control unit can employ non-cooperative game theory to resolve load allocation conflicts among standardized pump modules. Within each decision cycle, each standardized pump module can define a private revenue function based on its own operating status (e.g., current speed, efficiency range, accumulated operating time). This revenue function can be configured to include multiple weighted terms: the first term can be an energy efficiency gain term, the value of which is positively correlated with the degree to which the standardized pump module operates in its high-efficiency range; the second term can be a pressure deviation penalty term, the value of which is positively correlated with the square of the deviation of the current outlet pressure value from the set target value, and is treated as a deduction term in the revenue calculation; the third term can be a mechanical loss term, the value of which can be determined by the motor's start-stop frequency and the ratio of its operating speed to its rated speed.

[0031] The distributed cooperative control unit can be configured to guide each standardized pump module into a multi-round iterative game process. During this game, each standardized pump module can broadcast its expected output level and corresponding marginal cost to the communication network, while also receiving similar information from other standardized pump modules. Based on the collected neighborhood information, each standardized pump module can continuously adjust its predicted rotational speed using a Nash equilibrium search algorithm, under the hard constraint that the total flow demand equals the sum of the output flow of all standardized pump modules.

[0032] When the strategy changes of all standardized pump modules are less than a preset small threshold, the system can be considered to have reached a Nash equilibrium state. In the Nash equilibrium state, the load distribution scheme determined collaboratively by each standardized pump module can ensure that the total energy consumption of the system reaches the optimization point while meeting the pipeline pressure requirements, and the loss rate of each standardized pump module can tend to be balanced. This avoids the accelerated fatigue phenomenon caused by some pump sets being under high load for a long time in the traditional control mode.

[0033] For example, if the total system flow demand is fixed during a certain water supply period, each standardized pump module first constructs a corresponding revenue function based on its collected operating status data (including current speed, outlet pressure, start-stop history, and efficiency curve range) at the beginning of the current decision cycle. This function assigns a higher energy efficiency gain weight to the speed range operating in the high-efficiency zone, sets a penalty weight for the squared deviation of the outlet pressure from the target pressure, and calculates mechanical loss costs by combining the start-stop frequency and the ratio of speed to rated speed. Subsequently, each standardized pump module can obtain the current flow allocation information of other modules through a communication network, ensuring the total flow is conserved. Under constraints, each standardized pump module updates its own rotational speed based on an iterative optimal response strategy. This involves finding the optimal rotational speed adjustment for its own benefit function while keeping the strategies of other modules unchanged, or using a gradient adjustment method to correct the predicted rotational speed along the gain direction of the benefit function. As multiple iterations proceed, the rotational speed adjustment range of each standardized pump module gradually decreases. When the strategy changes of all standardized pump modules are less than a preset threshold, the system converges to a Nash equilibrium state. At this point, each standardized pump module achieves adaptive load distribution while satisfying pressure constraints, thereby optimizing overall operating energy consumption and avoiding local overload or frequent start-stop.

[0034] Optionally, in some embodiments, the local intelligent control terminal uses the collected global pressure drop and total flow data to dynamically correct the resistance characteristic coefficient of the pipeline network through a system identification method, and incorporates the corrected pipeline network characteristic parameters as environmental disturbance terms into the revenue function of the distributed collaborative control unit to improve the response accuracy to changes in actual operating conditions.

[0035] Specifically, the pipeline status sensing network can consist of multi-type sensor arrays deployed at key nodes in the pump station inlet pipeline, main outlet pipeline, and distribution network. Redundant pressure transmitters can be configured on the main outlet pipeline to provide highly reliable global pressure feedback. The pipeline status sensing network can also include remote pressure monitoring units deployed in long-distance pipelines, which can use low-power broadband IoT technology to transmit data back. All data streams collected by the pipeline status sensing network can be processed by a digital filter within the local intelligent control terminal to remove measurement noise caused by electromagnetic interference, sensor temperature drift, or pipeline cavitation.

[0036] Furthermore, the pipeline network status sensing network can possess online hydraulic model verification capabilities. The local intelligent control terminal utilizes the collected global pressure drop and total flow data to dynamically correct the pipeline network's resistance characteristic coefficients through a system identification method. The corrected pipeline network characteristic parameters can be incorporated as environmental disturbance terms into the game payoff function of the distributed collaborative control unit, enabling the game process to consider not only the equipment characteristics within the pumping station but also to proactively respond to real-time load changes in the pipeline network.

[0037] For example, when an abnormally high increase in the pipeline resistance coefficient is detected (which may indicate that the valve is mistakenly closed or partially blocked), the game strategy can automatically shift towards increasing the head of a single pump in order to maintain the stability of the end-point water supply pressure.

[0038] Optionally, in some embodiments, the communication network adopts an industrial-grade controller area network bus structure or a wireless network with multipath routing characteristics at the physical layer; when adopting a bus structure, the system applies a priority arbitration mechanism to ensure that emergency alarm signals obtain priority access; the communication network adopts a hybrid mode based on a combination of event-driven and periodic transmission, reducing the message sending frequency when the system is in a steady state, and triggering a high-frequency continuous transmission mechanism when transient disturbances are detected.

[0039] Specifically, the communication network can serve as a communication link connecting the various standardized water pump modules. Its physical layer can adopt an industrial-grade CAN bus structure or a wireless mesh network with multipath routing characteristics. In the implementation using a CAN bus, the system can apply a communication protocol with a priority arbitration mechanism to ensure that emergency alarm signals and synchronization control commands can obtain the preset highest bus access permissions. In the implementation using a wireless mesh network architecture, the communication network can be configured to have dynamic topology discovery and self-healing functions. When a standardized water pump module goes offline due to maintenance or other reasons, the remaining standardized water pump modules can automatically establish a new relay link to maintain the continuous transmission of control messages.

[0040] In practical applications, the communication network can be customized at the transmission protocol layer, and its communication cycle can be controlled between 5 milliseconds and 20 milliseconds to meet the real-time requirements of distributed game algorithms. The transmitted content can include not only conventional operating parameters such as current, voltage, and frequency, but also intermediate variables related to the game, such as the current bidding power of the standardized water pump module, the predicted payoff gradient, and the confidence level of strategy adjustments.

[0041] Furthermore, to prevent network congestion, the communication network can adopt a hybrid mode combining event-driven and periodic transmission. When the system is in a steady state, the communication network can reduce the message sending frequency, while when transient disturbances are detected, the communication network can immediately trigger a high-frequency continuous transmission mechanism to improve the real-time performance of data synchronization.

[0042] Optionally, in some embodiments, the remote algorithm update platform executes the update process using dual-partition mirroring technology. The update algorithm code is first downloaded to the backup storage area of ​​the local intelligent control terminal and verified. When the verification is successful and the system meets the preset switching conditions, the primary / backup switching logic is executed. The remote algorithm update platform supports a differential upgrade strategy, pushing only the logic segments that have changed in the strategy model.

[0043] Specifically, the architecture of the remote algorithm update platform can be divided into a cloud resource layer, an algorithm repository layer, and an over-the-air (OTA) push management layer. The cloud resource layer can store historical operational big data of the pumping station under different seasons, time periods, and pipeline network conditions, and extract optimal game parameter features under different operating conditions through deep learning models. The algorithm repository layer can store various validated control models, which may include pressure fluctuation suppression models for high-rise water supply, extremely low energy efficiency optimization models for low-load operation, and fault-tolerant compensation models for equipment aging.

[0044] The remote algorithm update platform can connect to the local intelligent control terminal at the pumping station via a secure and encrypted communication link. When the system operating performance deviates from the baseline value, or when a physical change occurs in the pipeline structure (such as the addition of a water supply branch), the remote algorithm update platform can automatically trigger the algorithm update process. The update process can employ dual-partition mirroring technology, meaning the new algorithm code can first be downloaded to the backup storage area of ​​the local intelligent control terminal and verified.

[0045] When the verification is successful and the system is in low-power standby or redundant operation mode, the local intelligent control terminal can execute the master / slave switching logic to ensure that the algorithm upgrade process does not interfere with normal water supply tasks. Furthermore, the remote algorithm update platform can also support a differential upgrade strategy, pushing only the changed logic segments in the algorithm to reduce data transmission volume and lower system resource consumption during the update process.

[0046] Optionally, in some embodiments, the distributed collaborative control unit is embedded with fault reconstruction logic. When any of the standardized water pump modules experiences a hardware failure or communication interruption, the damaged module sends an exit signal to the communication interaction network. The remaining normal standardized water pump modules automatically trigger the reconstruction process after capturing the exit signal. In the reconstruction process, the remaining standardized water pump modules redefine the scale of the game participants and redistribute the load originally borne by the faulty module according to the remaining capacity and individual loss index of each standardized water pump module.

[0047] Specifically, the distributed collaborative control unit may also embed fault reconfiguration logic. When any standardized pump module experiences a hardware failure (such as motor overheating, inverter error, or communication interruption), the faulty standardized pump module can immediately send a termination signal to the communication network. Upon receiving this signal, the remaining normal standardized pump modules can automatically trigger the reconfiguration process.

[0048] During the refactoring process, the remaining standardized pump modules can redefine the scale of participants in the game and redistribute the load borne by the previously faulty standardized pump modules according to the remaining capacity and loss index of each standardized pump module. This decentralized fault response mechanism ensures that the system can achieve service takeover within seconds without relying on a single host command, thereby improving the overall resilience and water supply security of the pumping station system.

[0049] For example, in a system with four standardized pump modules sharing a predetermined total flow requirement, if one of the standardized pump modules cannot continue to participate in the collaboration due to motor failure or communication interruption, the faulty module can send an exit signal to the communication network through a preset heartbeat detection failure or active reporting mechanism. Upon detecting the exit signal, the remaining three normal modules automatically trigger a reconstruction process through their embedded distributed collaborative control units, dynamically adjusting the original four-participant game model to a three-participant model and simultaneously updating the total flow constraints and strategy space. Subsequently, each remaining module weights the flow share originally borne by the faulty module based on its current available margin (such as the flow capacity corresponding to the maximum achievable speed) and individual loss index (characterizing the degree of equipment wear). For example, more load is allocated to modules with sufficient margin and low loss, while setting an allocation cap for high-loss modules. On this basis, each module continues to use the original iterative optimal response or gradient adjustment mechanism to negotiate the updated load allocation in multiple rounds until it converges to a new Nash equilibrium, thereby achieving adaptive reconstruction and stable operation of the system without central intervention.

[0050] Optionally, in some embodiments, the local intelligent control terminal integrates a life balancing strategy module, which establishes a fatigue aging model based on multi-physics coupling. The life balancing strategy module calculates an individual loss index based on the cumulative running time, current impact stress during start-up and shutdown, ambient temperature during operation, and the evolution trend of bearing vibration spectrum. During the game process in the distributed collaborative control unit, the individual loss index is converted into the penalty term weight in the payoff function, so that each standardized water pump module automatically adjusts the load distribution ratio according to the individual loss index during the strategy iteration process, so that the mechanical wear and electrical aging of all the standardized water pump modules in the pumping station tend to be consistent.

[0051] Specifically, the local intelligent control terminal of the standardized water pump module can also integrate a life balancing strategy module. This life balancing strategy module does not simply count operating hours, but rather establishes a fatigue aging model based on multi-physics coupling. The life balancing strategy module can calculate a dimensionless individual loss index based on cumulative operating time, current impact stress during start-up and shutdown, ambient temperature during operation, and the evolution trend of the bearing vibration spectrum.

[0052] In distributed game theory, the individual loss index can be converted into a penalty term weight in the game payoff function. If a standardized pump module has a high loss index, its bid cost in the game can be increased accordingly, thus automatically obtaining a smaller load allocation ratio during strategy iteration. This mechanism ensures that the mechanical wear and electrical aging of all standardized pump modules in the pumping station remain highly synchronized, unifying the overall maintenance cycle of the system and extending the comprehensive service life of the entire station's equipment.

[0053] For example, in a system with three standardized pump modules, each local intelligent control terminal can calculate individual loss indices (e.g., 0.2 for module A, 0.5 for module B, and 0.8 for module C) based on cumulative running time, the number and amplitude of start-stop current impacts, ambient temperature, and the changing trends of the bearing vibration spectrum's main frequency and energy distribution, using a preset multi-physics coupled fatigue aging model. Subsequently, when constructing the benefit function in the distributed collaborative control unit, the individual loss indices are mapped to the weighting coefficients of the mechanical loss penalty term, so that modules with higher loss indices correspond to higher penalty weights. Under the premise of satisfying the total flow constraint, each module will automatically reduce its target speed or allocate flow due to the decrease in benefit caused by high loss during the iterative optimal response or gradient adjustment process. For example, due to its higher loss index, module C's optimal strategy tends to bear a smaller load, while modules A and B correspondingly increase their load proportions. With multiple rounds of strategy iteration, the system gradually converges to a new Nash equilibrium state, allowing the load to be dynamically redistributed among modules in different health states, thereby achieving convergence of the wear and aging processes of each module in long-term operation.

[0054] In some examples, the data collected by the pipeline status sensing network can pass through an outlier removal module before being connected to the distributed collaborative control unit. This outlier removal module can employ a sliding window averaging algorithm and statistical outlier detection logic to identify and filter false data caused by temporary sensor malfunctions or transient water pressure pulses. The cleaned data can then be broadcast to each standardized pump module as a shared knowledge background during the game process.

[0055] When the system detects that the actual rate of change in the pressure of the main outlet pipe exceeds the preset normal water usage fluctuation range, the outlier removal module can work with the fault diagnosis module to determine whether there is a risk of pipe burst. Once a pipe burst is determined, the game logic can immediately switch from the pressure stabilization water supply mode to the emergency pressure reduction and protection mode. All standardized water pump modules can operate according to the preset safe deceleration curve to prevent large-scale water loss and protect the pump station itself from damage.

[0056] Example 2: Building upon Example 1, this example provides a modular intelligent pumping station system based on edge computing enhancement and a heterogeneous communication architecture. While maintaining the core logic of distributed collaborative control, this system further optimizes the configuration of hardware resources and the hierarchical structure of data processing to adapt to larger-scale and more complex smart water management application scenarios.

[0057] Optionally, in some embodiments, the local intelligent control terminal integrates a processing unit for running a predictive maintenance model. This predictive maintenance model, based on a time-series predictive analysis architecture, uses historical operating data from the standardized pump modules as input to predict water flow demand trends over a certain future time span. These water flow demand trends are input as prior boundary conditions to the distributed collaborative control unit, enabling each standardized pump module to adjust its speed in advance through a pre-game process, thereby reducing control lag and improving regulation stability. The standardized pump module also includes an adaptive hydraulic characteristic mapping module configured to reconstruct the real-time efficiency characteristic surface of the pump using collected speed, power, pressure, and flow data, and to replace the preset static model with the reconstructed real-time efficiency characteristic surface in the game calculation.

[0058] Specifically, the modular intelligent pumping station system in this embodiment may include standardized pump modules, heterogeneous redundant communication networks, multi-dimensional state deep perception matrix, and cloud-edge-device collaborative update platform.

[0059] The standardized water pump modules can possess edge computing capabilities, integrating dedicated neural processing units or high-performance floating-point arithmetic units within the local intelligent control terminal. This enables each standardized water pump module to not only execute basic PID regulation and game logic but also run complex predictive maintenance models. These predictive maintenance models, based on a long short-term memory neural network architecture, use historical operating data from the standardized water pump modules as input to predict water flow demand trends over a preset time span in real time. The prediction results can be used as prior information input into the distributed collaborative control unit. For example, when a water consumption peak is predicted within a preset timeframe, each standardized water pump module can preemptively and gradually increase its speed through pre-game analysis, rather than drastically accelerating after a pressure drop. This can smooth out energy consumption peaks and reduce thermal stress impact on the pipeline network.

[0060] The heterogeneous redundant communication network can adopt a topology combining a fiber optic backbone network and an industrial wireless 5G network. Inside the pumping station, each standardized pump module can be connected to the core switch via a gigabit Ethernet interface; as a backup path, each standardized pump module can be configured with a 5G communication module. In the event of physical damage to the wired link (such as an accidental cable cut), the system can switch to the wireless link to ensure zero interruption of the collaborative control logic. Furthermore, the heterogeneous redundant communication network can also support time-sensitive networking protocols, ensuring millisecond-level synchronization accuracy among the standardized pump modules when performing collaborative control actions through a precise clock synchronization mechanism, thereby eliminating micro-pressure oscillations caused by asynchronous actions.

[0061] The multi-dimensional state depth perception matrix can be built upon Example 1 by incorporating an acoustic emission sensor and an infrared thermal imaging monitoring unit. The acoustic emission sensor can be deployed on the pump body surface to capture weak high-frequency sound waves generated by cavitation within the fluid, and convert the cavitation intensity into real-time constraint parameters added to the game function. When the load assigned to a standardized pump module by the game algorithm leads to entry into the cavitation zone, this constraint parameter can generate a negative payoff penalty, forcing the strategy to revert to a safe range. The infrared thermal imaging monitoring unit can obtain thermal distribution maps by scanning the motor and cable joint areas. If local hotspots are detected, the local intelligent control terminal can automatically reduce the activity level of the standardized pump module in the game.

[0062] The cloud-edge-device collaborative update platform enables vertical integration of the logical architecture. In the cloud, by mining massive amounts of operational data from similar pump stations, various optimized game strategy templates can be generated. These templates are not directly distributed but are first pushed to the edge computing gateway at the pump station. The edge computing gateway, acting as the system's middle layer, is responsible for localizing and refining the cloud-based algorithms. It can combine the pump station's unique pipe network topology (such as the difference between tree-like and ring-like networks) and the actual efficiency degradation characteristics of the pumps, using reinforcement learning algorithms to practice in a simulation environment to find parameter settings suitable for the current physical environment. Only algorithm parameters verified through edge simulation can be distributed to each standardized pump module. This approach utilizes the powerful computing power and big data resources of the cloud while ensuring the accurate implementation and operational security of the control logic.

[0063] The distributed collaborative control unit can also incorporate a decision-making and evidence storage mechanism based on blockchain technology. The load allocation scheme reached in each round of the game, the original bidding data of each standardized water pump module, and the final execution result can be encapsulated into timestamped data blocks and stored in the local storage of each standardized water pump module in the form of a lightweight hash chain. This decentralized ledger mechanism ensures the transparency and immutability of the system's decision-making process, providing authoritative data for subsequent accident tracing, energy efficiency audits, and equipment maintenance. For example, when an unexpected system downtime occurs, by retrieving the consensus ledger of each standardized water pump module, the decision chain that led to system instability can be reconstructed, providing data support for improving the game model.

[0064] Furthermore, the standardized pump module may also include an adaptive hydraulic characteristic mapping module. Since the actual characteristic curve of a pump deviates from the factory-designed standard curve due to wear, scaling, or changes in fluid viscosity during operation, the adaptive hydraulic characteristic mapping module can utilize the collected four-dimensional data of speed, power, pressure, and flow rate to reconstruct the real-time efficiency characteristic surface of the pump. This reconstructed surface model replaces the preset static model in game theory calculations, enabling load allocation to be based on the actual health status and energy efficiency performance of the equipment, thus improving the overall energy-saving effect of the system throughout its entire lifecycle.

[0065] Optionally, in some embodiments, the distributed collaborative control unit further includes a virtual coupling stiffness adjustment module, which is used to suppress pressure oscillations caused by multi-module coupling and improve system operational stability.

[0066] To address the water competition phenomenon when multiple pumps are connected in parallel, a virtual coupling stiffness adjustment term can be added to the distributed collaborative control unit. This virtual coupling stiffness adjustment term can logically adjust the response sensitivity of each standardized pump module to pressure fluctuations by simulating the stiffness characteristics of a mechanical coupling. When a resonance is detected in the system, each standardized pump module can collaboratively adjust the virtual coupling stiffness adjustment term, changing the phase relationship of each module's control loop to generate a damping torque opposite to the resonance direction. This allows for proactive softening control of pressure vibrations within the pipeline network without changing the total output flow rate.

[0067] Example 3: This example provides a modular intelligent pumping station system implementation scheme for ultra-large-scale urban water supply boosting operations. In this scenario, the number of pumping station modules may reach dozens or even hundreds, and traditional point-to-point game theory would lead to a geometric increase in communication overhead. Therefore, this example introduces a hierarchical clustering game theory mechanism and a multi-objective dynamic evolution algorithm.

[0068] The modular intelligent pumping station system in this embodiment includes a large-scale standardized pump group cluster, a hierarchical distributed collaborative center, a distributed optical fiber sensing network, and an adaptive scenario simulation platform.

[0069] Specifically, the standardized water pump modules can be logically divided into multiple functional subgroups. Each functional subgroup can contain 5 to 8 standardized water pump modules, which are geographically close and share the same physical water collector. Within a subgroup, non-cooperative game logic as described in Example 1 can be executed to achieve load balancing. Between subgroups, a leader module automatically elected by each subgroup can represent the subgroup in higher-level global games. Through this recursive game architecture, the hierarchical distributed collaborative center can decompose the optimization problem of large-scale water pumps into multiple parallel low-dimensional optimization problems, significantly reducing the computational load and ensuring that the system maintains a preset response speed even during large-scale expansion.

[0070] The distributed optical fiber sensing network can acquire pressure fluctuation spectra of the entire water supply main in real time using vibration sensing optical fibers laid along the pipe wall. These pressure fluctuation spectra can be converted into a three-dimensional visualization matrix and input into each standardized pump module. The distributed collaborative control unit can then identify the propagation direction and attenuation pattern of pressure waves in the pipeline network and add a spatial compensation term to the game payoff function. For example, standardized pump modules located at the far end of the pipeline network can be given a higher pressure compensation weight in the game, spontaneously enhancing their support for the pressure at the pipeline's end without requiring central commands.

[0071] The adaptive scenario simulation platform can be deployed at the pump station's monitoring center or in the cloud, receiving real-time operational data from the field by establishing a digital twin model of the pump station. The platform can be configured to simulate various extreme scenarios, such as sudden fire-induced water supply demands, the pump restart process after a large-scale power outage, and emergency dispatching after external damage to the pipeline network. The optimal response strategy generated during the simulation can be converted into a strategy correction factor and pushed in real-time to the strategy library of each standardized pump module through the remote algorithm update platform. This enables the system not only to cope with routine load fluctuations but also to have intelligent decision-making reserves for specific events.

[0072] In some embodiments, a parameter optimizer based on a quantum genetic algorithm can be introduced into the local control logic of the standardized water pump module. During the intervals between each round of the game, the parameter optimizer can utilize the idle computing power of the standardized water pump module to fine-tune the weighting factors in the game payoff function by simulating quantum superposition and collapse processes. This self-evolving capability enables the pumping station system to automatically adjust its operating parameters according to seasonal changes, changes in water source temperature, and equipment aging, in order to maintain optimal operating conditions.

[0073] The hierarchical distributed collaborative center also possesses dynamic reorganization capabilities. When multiple standardized water pump modules within a functional subgroup fail simultaneously, causing the subgroup to be unable to meet its allocated water supply share, the navigation modules of adjacent subgroups can negotiate and take over the coverage area of ​​the failed subgroup through a communication network. Each subgroup can dynamically adjust the game boundary based on the real-time hydraulic connectivity matrix, enabling the system to exhibit self-healing characteristics and improving the safety margin of the ultra-large-scale water supply system.

[0074] Furthermore, the communication network in this embodiment can employ deterministic communication technology based on time-slice multiplexing. When sending game data packets, the standardized water pump module can complete the transmission within a specified microsecond time slot according to a high-precision clock synchronization signal. This deterministic communication environment can provide stable latency guarantees for complex distributed algorithms, thereby improving the closed-loop control bandwidth of the system and more effectively suppressing high-frequency pressure pulsations.

[0075] Optionally, in some embodiments, the remote algorithm update platform also integrates a bypass evaluation function. Before the control strategy to be updated takes effect, it is first tested in the isolated memory area of ​​the local intelligent control terminal using real-time input data. The system compares the predicted performance of the control strategy to be updated with the actual performance of the current strategy. When the control strategy to be updated continues to dominate in the preset indicator evaluation and does not trigger security restrictions, the physical control right of the control strategy to be updated is officially granted.

[0076] In other words, the remote algorithm update platform in this embodiment can also integrate a sandbox operation function. Before the new control strategy officially takes effect, it can be used to perform trial calculations in the background isolated memory area of ​​the local intelligent control terminal using real-time input data. The background isolated memory area can receive real input data and calculate the output results, but the output results do not directly affect the drive motor unit.

[0077] The system can compare the predicted performance of the new control strategy with the actual performance of the current control strategy in the isolated memory area. Only when the new control strategy consistently outperforms other strategies in a comprehensive evaluation of energy efficiency, stability, and other indicators, and does not trigger any safety restrictions, can the local intelligent control terminal officially grant the new control strategy physical control. This mechanism provides a risk-free testing ground for the continuous iteration of algorithms.

[0078] In summary, this embodiment, through a hierarchical game theory architecture and deep environmental awareness, can solve the problems of collaborative complexity, real-time response, and operational reliability in large-scale pumping station systems.

[0079] Example 4: This example provides a specific implementation of a modular intelligent pumping station system in a low-power operation and mobile emergency water supply scenario. In this application context, the system enhances its sensitivity to energy efficiency.

[0080] The modular intelligent pumping station system of this embodiment includes a standardized water pump module, a self-organizing network mobile communication link, an environment-adaptive game engine, and a portable algorithm configuration terminal.

[0081] The standardized water pump module can adopt a lightweight design and be equipped with a high-power-density silicon carbide frequency converter to reduce size and improve power conversion efficiency. Its integrated local intelligent control terminal can use an ultra-low-power heterogeneous processor, consuming only microamps of current in standby mode. The drive motor can employ a special magnetic circuit design to support stable operation at extremely low speeds, thereby broadening the effective adjustment range of the distributed game.

[0082] The self-organizing mobile communication link can employ enhanced low-power wide-area network (LPWAN) technology. In mobile emergency water supply scenarios, the self-organizing mobile communication link can automatically detect available communication base stations or satellite links in the vicinity and establish a multi-hop peer-to-peer network among the standardized water pump modules. Even in the event of a failure of public communication infrastructure, the standardized water pump modules can still maintain synchronization of game messages using the built-in long-distance radio frequency link.

[0083] The environmentally adaptive game engine can introduce resource finiteness constraints into the payoff function. In emergency situations, the game objective can shift from maximizing energy efficiency to maximizing water supply output with limited energy reserves. The engine can monitor voltage fluctuations and remaining power in the power supply system in real time and incorporate these as rigid constraints into each round of Nash equilibrium calculation. If the power level is detected to be below a preset threshold, the game logic can automatically shift towards reducing the operation of high-energy-consuming pumps and increasing the load on high-efficiency pumps, and trigger an energy-saving mode to reduce outlet pressure as needed.

[0084] The portable algorithm configuration terminal allows maintenance personnel to define the game roles of each standardized water pump module on-site via wireless near-field communication technology. Maintenance personnel can set operating priorities such as pressure priority, lifespan priority, or quiet operation priority. These preference settings can be instantly translated into weighting factors in the game function and synchronized to the entire distributed collaborative control unit.

[0085] Furthermore, the standardized pump module in this embodiment also possesses self-learning load characteristic recognition capabilities. When connecting to an unknown pipeline network, the standardized pump module can derive the equivalent topology parameters of the current pipeline network through a single pulse test (i.e., changing the rotational speed and observing pressure feedback) and using an online system identification algorithm. These equivalent topology parameters are then shared with all game participants, enabling the system to complete the transition from physical deployment to intelligent collaboration within minutes. This plug-and-play characteristic has extremely important practical value for tasks such as flood control and disaster relief, and temporary water supply scheduling.

[0086] For example, after a set of standardized water pump modules are temporarily connected to an unknown pipeline (such as an emergency water supply pipeline), each module enters a short-term identification mode during the initialization phase: the local intelligent control terminal controls the motor to perform small-amplitude, multi-frequency speed pulse disturbances (such as step or sinusoidal frequency sweep) within a safe range, and simultaneously collects the corresponding outlet pressure and flow response data; the embedded online system identification algorithm (such as recursive least squares or extended Kalman filter) fits the dynamic relationship between the input speed and the output pressure / flow, and equivalently solves for the set of parameters characterizing the pipeline network characteristics (such as equivalent resistance). The system incorporates force coefficients, inertia coefficients, and potential delay characteristics to form a simplified equivalent model of the pipeline network. Each standardized pump module then shares its identified parameters through a communication network and performs consistent fusion (such as weighted averaging or confidence screening). This ensures that all game participants evaluate payoff functions and update strategies based on a consistent pipeline network model in subsequent collaborative control. On this basis, the distributed collaborative control unit can quickly enter a stable game iteration process, enabling the system to transition from an "unknown environment" to "predictable collaborative control" within minutes, achieving plug-and-play operation.

[0087] In some embodiments, the environmental adaptive game engine may further include a global search strategy based on simulated annealing. When the system detects a step-like jump in water consumption, the environmental adaptive game engine may temporarily increase the randomness of the strategy search, accepting a temporarily poor load scheme with a preset probability, thereby escaping the original inefficient equilibrium point and finding the optimal Nash equilibrium point adapted to the new operating conditions.

[0088] The system in this embodiment not only achieves rapid physical deployment, but also ensures, through the flexibility and adaptability at the algorithm level, that the pumping station system can still provide stable and efficient fluid transport services through the autonomous collaboration of each standardized water pump module in extremely variable environments.

[0089] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention. Without departing from the concept of the present invention, those skilled in the art can make several variations and improvements, all of which fall within the scope of protection of the present invention. Contents not described in detail in this specification are prior art known to those skilled in the art.

Claims

1. A modular intelligent pumping station system, the system comprising: Multiple standardized pump modules are configured to perform fluid transport tasks; A distributed collaborative control unit is embedded in each of the standardized water pump modules; The pipeline status awareness network is distributed across key nodes of the main pipeline network. A communication network is used for data exchange between the standardized water pump modules; Its features include: each of the standardized water pump modules integrates a local intelligent control terminal and integrated sensing components, which are used to collect the operating status data of the standardized water pump module in real time and dynamically adjust the operating parameters according to the collaborative instructions; The distributed collaborative control unit constructs control logic based on non-cooperative game theory that integrates multi-objective constraints, enabling each of the standardized water pump modules to reach a Nash equilibrium of load distribution through autonomous negotiation without central scheduling. The pipeline status sensing network is used to collect global hydraulic condition information and broadcast the data to each of the standardized water pump modules. The communication network adopts an industrial-grade bus or wireless mesh network architecture, supporting millisecond-level data exchange between the standardized water pump modules.

2. The modular intelligent pumping station system according to claim 1, characterized in that: The standardized water pump module includes a physical pump body assembly, a drive motor unit, an integrated sensing assembly, and a local intelligent control terminal. The flow-through components of the physical pump assembly are configured to maintain a preset hydraulic efficiency at different speeds; The drive motor unit is equipped with a frequency converter driver, which is configured to receive control signals output by the local intelligent control terminal to achieve continuous adjustment of the motor speed. The integrated sensing component includes a high-frequency pressure sensor deployed at the outlet manifold of the standardized water pump module, and the sampling frequency of the high-frequency pressure sensor is configured to a preset frequency threshold capable of capturing transient water hammer signals. The integrated sensing component also includes a temperature sensor, a vibration sensor, and an energy metering chip.

3. The modular intelligent pumping station system according to claim 2, characterized in that: In each decision cycle, the distributed collaborative control unit guides each standardized water pump module to define a private revenue function based on the operating status of the standardized water pump module. The benefit function is a combination function of multiple weighted terms, specifically including: an energy efficiency gain term, the value of which is positively correlated with the degree to which the standardized water pump module operates in the high-efficiency zone; a pressure deviation penalty term, which is positively correlated with the square value of the deviation of the current outlet pressure value from the set target value; and a mechanical loss term, which is determined by the motor's start-stop frequency and the ratio of its operating speed to its rated speed. Under the constraint that the total flow demand is equal to the sum of the output flow of each standardized water pump module, the predicted rotational speed is corrected by a Nash equilibrium search algorithm based on iterative optimal response or gradient adjustment until all strategy changes are less than a preset threshold, thus achieving a Nash equilibrium state and optimizing the total energy consumption of the system.

4. The modular intelligent pumping station system according to claim 3, characterized in that: The local intelligent control terminal integrates a life balancing strategy module, which establishes a fatigue aging model based on multi-physics field coupling. The life balancing strategy module calculates the individual loss index based on the cumulative running time, the current impact stress during start-up and shutdown, the ambient temperature during operation, and the evolution trend of the bearing vibration spectrum. During the game played by the distributed collaborative control unit, the individual loss index is converted into the penalty term weight in the payoff function, so that each standardized water pump module automatically adjusts its load distribution ratio according to the individual loss index during the strategy iteration process, thereby making the mechanical wear and electrical aging of all the standardized water pump modules in the pumping station tend to be consistent.

5. The modular intelligent pumping station system according to claim 1, characterized in that: The local intelligent control terminal uses the collected global pressure drop and total flow data to dynamically correct the resistance characteristic coefficient of the pipeline network through a system identification method. The corrected pipeline characteristic parameters are then incorporated into the revenue function of the distributed collaborative control unit as an environmental disturbance term to improve the response accuracy to changes in actual operating conditions.

6. The modular intelligent pumping station system according to claim 1, characterized in that: The communication network adopts an industrial-grade controller local area network bus structure or a wireless network with multipath routing characteristics at the physical layer. When using a bus architecture, the system applies a priority arbitration mechanism to ensure that emergency alarm signals receive priority access. The communication network adopts a hybrid mode that combines event-driven and periodic transmission. When the system is in a steady state, the message sending frequency is reduced, and a high-frequency continuous transmission mechanism is triggered when transient disturbances are detected.

7. The modular intelligent pumping station system according to claim 1, characterized in that: The system also includes a remote algorithm update platform, which is connected to the standardized water pump module via a secure encrypted channel; The remote algorithm update platform is used to push optimized distributed control algorithm parameters or game strategy models; The remote algorithm update platform executes the update process through dual-partition mirroring technology. The update algorithm code is first downloaded to the backup storage area of ​​the local intelligent control terminal and verified. When the verification is correct and the system meets the preset switching conditions, the master-slave switching logic is executed. The remote algorithm update platform supports a differential upgrade strategy, which only pushes the logic segments that have changed in the strategy model.

8. The modular intelligent pumping station system according to claim 1, characterized in that: The distributed collaborative control unit is embedded with fault reconstruction logic. When any of the standardized water pump modules experiences a hardware failure or communication interruption, the damaged module sends an exit signal to the communication interaction network. The remaining normal standardized water pump modules automatically trigger the reconstruction process after capturing the exit signal. In the reconstruction process, the remaining standardized pump modules redefine the size of the game participants and redistribute the load borne by the original faulty module according to the remaining capacity and individual loss index of each standardized pump module.

9. The modular intelligent pumping station system according to claim 7, characterized in that: The remote algorithm update platform also integrates a bypass evaluation function. Before the control strategy to be updated takes effect, it is first tested in the isolated memory area of ​​the local intelligent control terminal using real-time input data. The system compares the predicted performance of the control strategy to be updated with the actual performance of the current strategy. When the control strategy to be updated continues to be superior in the preset indicator evaluation and does not trigger security restrictions, the physical control right of the control strategy to be updated is officially granted. The distributed collaborative control unit also includes a virtual coupling stiffness adjustment module, which is used to suppress pressure oscillations caused by multi-module coupling and improve system operational stability.

10. The modular intelligent pumping station system according to any one of claims 1 to 9, characterized in that: The local intelligent control terminal integrates a processing unit for running predictive maintenance models; The predictive maintenance model is based on a time series predictive analysis architecture. It takes the historical operating data of the standardized water pump module as input and predicts the water flow demand trend in real time over a certain time span in the future. The water flow demand trend is input into the distributed collaborative control unit as a priori boundary condition, so that each standardized water pump module can adjust its speed in advance through a pre-game process to reduce control lag and improve regulation stability. The standardized water pump module also includes an adaptive hydraulic characteristic mapping module, which is configured to reconstruct the real-time efficiency characteristic surface of the water pump in real time using the collected speed, power, pressure and flow data, and use the reconstructed real-time efficiency characteristic surface to replace the preset static model in game calculation.

Images