Electrochemical water treatment method applied to data center cooling circulating water treatment

Abstract

The invention provides an electrochemical water treatment method applied to data center cooling circulating water treatment, and relates to an electrochemical water treatment technology. The electrochemical water treatment and the RO system are combined, water quality parameters are monitored in real time through multiple sensors, operation parameters of the electrochemical water treatment device are dynamically adjusted through a self-learning control algorithm, calcium and magnesium ions in water are preferentially removed, when the water quality reaches a set value, the RO system is started, and the conductivity is further reduced; accurate regulation and control of water quality and automatic stable operation are realized, and the water treatment efficiency and the intelligent level are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electrochemical water treatment, and in particular to an electrochemical water treatment method applied to cooling circulating water treatment in a data center. Background Art

[0002] As data centers continue to expand in size and computing density increases, water cooling systems are widely adopted as the primary heat dissipation method. In these systems, circulating water quality management is crucial, particularly the control of calcium and magnesium ions, which directly impacts the system's heat dissipation efficiency and equipment lifespan. Currently, a variety of water quality monitoring and control technologies have been developed in the field of industrial circulating water treatment, ranging from traditional chemical treatment to modern intelligent control systems, with continuous technological innovation and improvement.

[0003] However, existing water cooling system control technologies still face the following challenges: First, traditional control systems often rely on a single parameter, such as conductivity or temperature feedback, to adjust the treatment process, which is unable to adapt to the nonlinear variations in water quality. This simple feedback mechanism often struggles to achieve precise control in complex and changing water quality environments, resulting in unstable treatment results. Second, existing systems generally lack the ability to effectively respond to external factors, such as climate change and equipment aging. These factors can significantly affect water treatment effectiveness, but existing technologies struggle to incorporate them into the control logic. Third, existing systems respond poorly during high loads or emergencies, which can easily lead to degraded water quality or waste resources, compromising the system's stable operation and economic viability. Furthermore, while some systems have incorporated adaptive control technologies, these systems can only adjust parameters based on current data and cannot respond to changes in external factors, lacking foresight and comprehensiveness.

[0004] Therefore, there is an urgent need to develop a new type of water cooling system control technology that can comprehensively consider factors such as multi-parameter feedback, external environmental changes, equipment status, etc., and have self-learning and optimization capabilities to achieve more accurate, efficient and stable water quality management. Summary of the Invention

[0005] In order to overcome the shortcomings of the existing technology, the purpose of the present invention is to provide an electrochemical water treatment method for data center cooling circulating water treatment, which combines electrochemical water treatment with RO reverse osmosis system to achieve precise control of water quality and automatic and stable operation, thereby improving water treatment efficiency and system intelligence level.

[0006] To achieve the above object, the present invention provides the following solutions: An electrochemical water treatment method for treating cooling circulating water in a data center, comprising: The cooling circulating water of the data center is transported to the AHU unit through the water supply pump for cooling; The water discharged after cooling is collected in a water pool and treated by an electrochemical water treatment device to precipitate calcium and magnesium ions in the water in the form of scale; Use a filter to filter out the scale precipitated in the pool and discharge the filtered water into the sewage pool; Based on a multi-sensor real-time data acquisition platform, the system dynamically monitors water quality parameters in the pool, acquires and analyzes water quality changes in real time, and optimizes the control steps in the water treatment process through a self-learning control algorithm to ensure that water quality meets set standards. The control steps include adjusting the operating frequency, current density, and reaction time of the electrochemical water treatment device. When the water quality parameters in the pool reach the set dynamic optimization value, the RO reverse osmosis system is started to remove the dissolved salts in the water through RO reverse osmosis to reduce the water conductivity; Drain the wastewater from the RO system into the sewage tank.

[0007] Preferably, it also includes: Adjust the pool water level in real time to ensure that the pool is within the preset optimal working range to meet the water treatment needs under different workloads.

[0008] Preferably, conductivity, temperature, pH value, calcium ion concentration and magnesium ion concentration.

[0009] Preferably, based on a multi-sensor real-time data acquisition platform, the water quality parameters in the pool are dynamically monitored, the water quality changes are acquired and analyzed in real time, and the control steps in the water treatment process are optimized through a self-learning control algorithm to ensure that the water quality meets the set standards, including: Conductivity, temperature, pH value, calcium ion concentration, and magnesium ion concentration in the water pool monitored by the multi-sensor real-time data acquisition platform are sequentially time-aligned, feature-enhanced, and decoupled to obtain anti-interference features; Based on the self-learning control algorithm, automatically identifying the relationship between the anti-interference characteristics and the working state of the electrochemical water treatment device to obtain a dynamic model; The operating frequency, current density and reaction time of the electrochemical water treatment device are adjusted through the dynamic model to ensure that the water quality meets the set standards.

[0010] Preferably, the conductivity, temperature, pH value, calcium ion concentration and magnesium ion concentration in the water pool monitored by the multi-sensor real-time data acquisition platform are sequentially subjected to time alignment, feature enhancement and decoupling operations to obtain anti-interference features, including: Performing dynamic compensation for transmission delay on the multi-source heterogeneous data consisting of the conductivity, the temperature, the pH value, the calcium ion concentration, and the magnesium ion concentration by using a Kalman filter method; Based on a time protocol, the multi-source heterogeneous data are arranged in ascending order with the compensated conductivity as a reference time axis; performing spline interpolation and resampling operations on the multi-source heterogeneous data except the conductivity to obtain resampled data; Performing physical quantity conversion on the resampled data, and eliminating the resampled data whose evaluation index is greater than a preset index threshold; Performing an encoding operation on the multi-source heterogeneous data to obtain encoding features; the encoding operation includes: phase encoding, frequency encoding, and group pulse encoding; Constructing a spiking neural network according to Hebbian rule to competitively decouple the encoding features; Performing scale decomposition and environmental interference suppression on the decoupled coding features to obtain enhanced features; The enhanced feature is thermodynamically reconstructed to generate the anti-interference feature.

[0011] Preferably, the expression of the dynamic model is:

[0012] in, For the moment The control parameters are the operating frequency, current density or reaction time; For the moment Water quality parameters , namely conductivity, pH value, calcium ion concentration, magnesium ion concentration and temperature; is the coefficient calculated by the self-learning control algorithm, which represents the water quality parameter The degree of influence on the control parameters is calculated by the following formula: ; For the moment The amount of change in the control parameter, which represents the difference between the control parameter and the change in water quality; is the number of sample data; is the coefficient of the influence of the anti-interference feature on the control parameter, which represents the effect of the anti-interference feature on the control parameter and is calculated by the following formula: , For the moment The calculated anti-interference characteristics represent the signal characteristics after data preprocessing; For the moment The amount of change in the control parameter represents the difference between the control parameter and the water quality change.

[0013] Preferably, when the water quality parameters in the pool reach the set dynamic optimization value, the RO reverse osmosis system is started to remove the dissolved salts in the water through RO reverse osmosis to reduce the conductivity of the water, including: Obtaining preset dynamic optimization values; the dynamic optimization values ​​include the set upper limit of conductivity, upper limits of calcium ion and magnesium ion concentration, and preset ranges of pH value and temperature; When any water quality parameter in the water pool reaches the dynamic optimization value, the self-learning control algorithm outputs a start instruction to start the RO reverse osmosis system, and determines the initial operating parameters of the RO reverse osmosis system based on the difference between the dynamic optimization value and the real-time water quality parameter; the initial operating parameters include working pressure, water inlet flow rate and target water production rate; After the RO reverse osmosis system is started, the water quality parameters of the water pool are monitored in real time. According to the dynamic difference between the conductivity of the water in the water pool and the set target value, the working pressure and water flow rate of the RO reverse osmosis system are adjusted to maintain the stability of the RO reverse osmosis membrane flux. The system recovery rate is adjusted according to the difference between the real-time water production rate and the target water production rate to improve the desalination efficiency and avoid the risk of membrane fouling or concentration polarization; The self-learning control algorithm combines the real-time operating data of the electrochemical water treatment device and the RO reverse osmosis system to dynamically correct the optimized values ​​of the pool water quality parameters and the operating parameters of the RO reverse osmosis system, so that the electrochemical water treatment device and the RO reverse osmosis system form a closed-loop collaborative control relationship, preferentially achieving effective removal of calcium and magnesium ions through electrochemical water treatment, thereby reducing subsequent RO load; When the conductivity monitoring value of the water in the water pool drops to the target range within the dynamic optimization value, the self-learning control algorithm outputs a stop instruction, shuts down the RO reverse osmosis system, and continues to monitor changes in the water quality of the water pool to maintain the water quality parameters stable within the set range for a long time.

[0014] Preferably, the water level in the pool is adjusted in real time to ensure that the pool is within the preset optimal working range to adapt to the water treatment needs under different workloads, including: Based on the multi-sensor real-time data acquisition platform, the water level in the water pool is monitored in real time to obtain current water level data, and the water level data is compared with a preset optimal working water level range; the optimal working water level range is pre-set based on the design processing capacity of the electrochemical water treatment device and the RO reverse osmosis system and the load demand of the data center cooling system; When the water level in the pool is lower than the lower limit of the optimal working water level range, the self-learning control algorithm outputs a control instruction to start the water supply pump based on the water level change trend of the pool and the subsequent water treatment load forecast, and determines the water supply flow rate based on the difference between the current water level in the pool and the lower limit, thereby raising the water level in the pool to within the optimal working water level range. At the same time, the operating parameters of the water supply pump are adjusted in conjunction with the current water inflow and treatment capacity of the electrochemical water treatment device to prevent fluctuations in water quality parameters due to sudden changes in water volume; When the water level in the pool is higher than the upper limit of the optimal working water level range, the self-learning control algorithm outputs a control instruction to start the drainage pump based on the real-time water level data of the pool and the current water quality parameters, the inlet and outlet water volume and the processing capacity of the RO reverse osmosis system, and determines the drainage flow rate according to the difference between the current water level in the pool and the upper limit, thereby lowering the water level in the pool to within the optimal working water level range. At the same time, the inlet and outlet water rhythms of the electrochemical water treatment device and the RO reverse osmosis system are adjusted in a coordinated manner to avoid system load imbalance caused by abnormal water levels; During the water supply and drainage regulation process, the water level changes in the water pool and the operating status of the water supply pump and drainage pump are monitored in real time. Based on the water level change trend of the water pool and the water load of the data center cooling system, the start and stop status and flow control instructions of the water supply pump and drainage pump are dynamically adjusted to ensure that the water level in the water pool is stably maintained within the optimal working water level range; The self-learning control algorithm dynamically corrects the optimal working water level range based on the water level data of the water pool, the real-time operating data of the electrochemical water treatment device and the RO reverse osmosis system, and the actual load changes of the data center cooling system to adapt to changes in water treatment requirements in different seasons, different operating modes or different loads, ensuring that the water level of the water pool is always in an adaptable and optimal working state.

[0015] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects: (1) The present invention dynamically monitors the water quality parameters of the cooling circulating water through a multi-sensor real-time data acquisition platform, and establishes a dynamic optimization model in combination with a self-learning control algorithm. This model can accurately identify the trend of water quality changes and dynamically adjust the operating frequency, current density and reaction time of the electrochemical water treatment device, thereby effectively removing calcium and magnesium ions in the water and improving the stability and automation level of the water treatment process.

[0016] (2) The present invention realizes the linkage control of water quality parameters and the operating status of the electrochemical water treatment device by constructing a dynamic model that includes the water quality change rate, anti-interference characteristics and equipment operating status. It can automatically optimize the control parameters according to the real-time water quality changes, reduce manual intervention, and improve the reliability of water quality compliance and the overall operating efficiency of the water treatment system.

[0017] (3) The present invention intelligently activates the RO reverse osmosis system when the water quality parameters reach the dynamic optimization value, and automatically adjusts the RO system operating parameters (such as working pressure, flow rate and water production rate) according to the changes in water quality, thereby achieving accurate removal of dissolved salts, further reducing the water conductivity, and ensuring that the water quality is stable in the long term to meet the cooling needs of the data center.

[0018] (4) The present invention adjusts the water level of the pool in real time and combines the self-learning control algorithm to dynamically correct the water level setting to ensure that the pool always operates in the optimal water level range, adapting to the water treatment needs under different loads, forming a closed-loop feedback mechanism for coordinated regulation of electrochemical water treatment and RO reverse osmosis system, and improving system stability and energy saving effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 A flow chart of a method provided by an embodiment of the present invention; Figure 2 A schematic diagram of the system structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] The purpose of the present invention is to provide an electrochemical water treatment method for data center cooling circulating water treatment. In view of the defects of existing water treatment systems such as delayed response of water quality control, high dependence on manual intervention, and inability to dynamically adapt to load changes, a dynamic optimization method based on a self-learning control algorithm is proposed. By combining electrochemical water treatment with the RO reverse osmosis system, precise water quality control and automatic stable operation can be achieved, thereby improving water treatment efficiency and the level of system intelligence.

[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0024] Figure 1 A flow chart of the method provided in the embodiment of the present invention is shown in FIG. Figure 1 As shown, the present invention provides an electrochemical water treatment method for data center cooling circulating water treatment, comprising: Step 100: The cooling circulating water of the data center is delivered to the AHU unit through a water supply pump for cooling; Step 200: Collecting the discharged water after cooling into a water pool, and treating the water in the water pool through an electrochemical water treatment device to precipitate calcium ions and magnesium ions in the water in the form of scale; Step 300: Using a filter to filter scale precipitated in the pool, and discharging the filtered water into a sewage pool; Step 400: Based on a multi-sensor real-time data acquisition platform, dynamically monitor water quality parameters in the pool, obtain and analyze water quality changes in real time, and optimize control steps in the water treatment process through a self-learning control algorithm to ensure that water quality meets set standards. The control steps include adjusting the operating frequency, current density, and reaction time of the electrochemical water treatment device. Step 500: When the water quality parameters in the pool reach the set dynamic optimization value, the RO reverse osmosis system is started to remove the dissolved salts in the water through RO reverse osmosis to reduce the conductivity of the water; Step 600: Discharge the wastewater discharged from the RO system into a sewage pool.

[0025] like Figure 2 As shown, the data center cooling circulating water treatment system of this embodiment includes a water pool, a filter, an electrochemical water treatment device, a multi-sensor real-time data acquisition platform, a self-learning control algorithm module, an RO reverse osmosis system and an RO water tank. The water pool is connected to the filter through a water pump, and the water outlet of the filter is treated by the electrochemical water treatment device and then returned to the water pool. The multi-sensor real-time data acquisition platform is respectively connected to the conductivity sensor, temperature sensor, pH sensor, calcium ion concentration sensor, and magnesium ion concentration sensor arranged in the water pool, and is used to collect water quality parameters in real time and transmit the data to the self-learning control algorithm module; the self-learning control algorithm module is respectively connected to the electrochemical water treatment device and the RO reverse osmosis system, and is used to control their operating status. The water inlet end of the RO reverse osmosis system is connected to the water pool, the water production end is connected to the RO water tank, and the concentrated water is discharged into the sewage pool.

[0026] The self-learning control algorithm module establishes a dynamic water quality optimization model based on the pool water quality parameters collected by the multi-sensor and the operating status of the electrochemical water treatment device, combined with historical water quality data, and outputs control instructions for the operating frequency, current density, and reaction time of the electrochemical water treatment device to effectively remove calcium and magnesium ions in the water and reduce hardness. When the pool water quality parameters reach the dynamic optimization value, the self-learning control algorithm module controls the RO reverse osmosis system to start up and adjust its operating pressure, water flow rate, and water production rate according to changes in water quality to further remove soluble salts in the water and reduce conductivity. The water produced by the RO reverse osmosis system is stored in the RO water tank for standby use, and the concentrated water is discharged into the sewage pool.

[0027] The process flow of the embodiment is: the water in the pool is pumped into the filter by the water pump, and after removing the scale impurities precipitated in the water, it enters the electrochemical water treatment device for treatment, promoting the precipitation of calcium and magnesium ions, and the treated water flows back to the pool. The multi-sensor real-time collects water quality data in the pool and transmits it to the self-learning control algorithm module, which dynamically adjusts the operating parameters of the electrochemical water treatment device according to water quality changes to optimize water quality. When the water quality parameters of the pool reach the dynamic optimization value, the RO reverse osmosis system starts, and the self-learning control algorithm module dynamically adjusts the operating parameters of the RO reverse osmosis system according to water quality changes to accurately remove dissolved salts and reduce electrical conductivity. The RO system produces water that enters the RO water tank, and the concentrated water is discharged into the sewage pool, ultimately achieving continuous optimization of water quality and efficient and stable operation of the system.

[0028] Preferably, it also includes: Real-time adjustment of the water level in the pool ensures that the pool is within the preset optimal working range to meet water treatment needs under different workloads.

[0029] Preferably, electrical conductivity, temperature, pH value, calcium ion concentration, and magnesium ion concentration.

[0030] Preferably, based on the multi-sensor real-time data acquisition platform, the water quality parameters in the pool are dynamically monitored, the water quality changes are obtained and analyzed in real time, and the control steps in the water treatment process are optimized through the self-learning control algorithm to ensure that the water quality meets the set standard, including: The electrical conductivity, temperature, pH value, calcium ion concentration, and magnesium ion concentration in the pool monitored by the multi-sensor real-time data acquisition platform are sequentially time-aligned, feature-enhanced, and decoupled to obtain anti-interference features; Based on the self-learning control algorithm, the relationship between the anti-interference features and the working state of the electrochemical water treatment device is automatically identified to obtain a dynamic model; The working frequency, current density, and reaction time of the electrochemical water treatment device are adjusted through the dynamic model to ensure that the water quality meets the set standard.

[0031] Specifically, this embodiment uses a multi-sensor real-time data acquisition platform to dynamically monitor water quality parameters such as conductivity, temperature, pH, calcium ion concentration, and magnesium ion concentration in the water pool, acquiring continuous time series data in real time. First, the collected multi-source heterogeneous data is time-aligned to eliminate differences in sampling time between different sensors and unify them to a standard time axis based on conductivity. Subsequently, feature enhancement is performed based on physical mechanisms and statistical characteristics. Dynamic delay compensation is achieved through Kalman filtering, and spline interpolation is combined to reconstruct missing points, improve data stability and continuity, further eliminate abnormal data points, and ensure the quality of subsequent modeling data. On this basis, a spiking neural network structure is combined with Hebbian rule to competitively decouple multidimensional encoding features, extracting anti-interference features with weak correlation between physical quantities and strong robustness, and constructing standardized feature vectors that can be used for control. A self-learning control algorithm is used to automatically identify the dynamic relationship between water quality change characteristics and control parameters based on historical water quality change data and operating status data of the electrochemical water treatment device. A dynamic model is established using least squares fitting to clarify the quantitative impact of different water quality characteristics on operating frequency, current density, and reaction time. Ultimately, based on the dynamic model, the control parameters of the electrochemical water treatment device are adjusted in real time to achieve dynamic optimization response as water quality changes, ensuring that water quality is stably maintained within the set target range, and improving the adaptability and control accuracy of the water treatment process. This technical link cannot be omitted in the extraction of anti-interference features and the establishment of a dynamic model. It is directly related to the core innovative effect of the water quality regulation of the present invention. Compared with traditional single data-driven or static rule-based strategies, it can effectively solve the problems of unstable water treatment control and delayed response to water quality fluctuations under environmental disturbances.

[0032] Preferably, the conductivity, temperature, pH value, calcium ion concentration and magnesium ion concentration in the water pool monitored by the multi-sensor real-time data acquisition platform are sequentially subjected to time alignment, feature enhancement and decoupling operations to obtain anti-interference features, including: Performing dynamic compensation for transmission delay on the multi-source heterogeneous data consisting of the conductivity, the temperature, the pH value, the calcium ion concentration, and the magnesium ion concentration by using a Kalman filter method; Based on a time protocol, the multi-source heterogeneous data are arranged in ascending order with the compensated conductivity as a reference time axis; performing spline interpolation and resampling operations on the multi-source heterogeneous data except the conductivity to obtain resampled data; Performing physical quantity conversion on the resampled data, and eliminating the resampled data whose evaluation index is greater than a preset index threshold; Performing an encoding operation on the multi-source heterogeneous data to obtain encoding features; the encoding operation includes: phase encoding, frequency encoding, and group pulse encoding; Constructing a spiking neural network according to Hebbian rule to competitively decouple the encoding features; Performing scale decomposition and environmental interference suppression on the decoupled coding features to obtain enhanced features; The enhanced feature is thermodynamically reconstructed to generate the anti-interference feature.

[0033] Specifically, this embodiment establishes a state-space model in which the state variable is the transmission delay time and the observed variable is the rate of change of the sensor's physical quantity. The current transmission delay is predicted based on the delay value at the previous moment, and the error is corrected by combining the gradient change of the current sensor data. Then, conductivity with good sampling stability is selected as the time reference. The conductivity, after Kalman filter compensation, generates a reference time axis according to the UTC time protocol, dividing time slots into 10-millisecond units. Data from other sensors is inserted into the corresponding positions in chronological order based on the compensated timestamps. If multiple data sources appear in the same time slot, a timestamp-weighted average method is used to generate a unique data point, thus forming a six-dimensional heterogeneous data sequence.

[0034] For non-conductive multi-source heterogeneous data, we construct a piecewise cubic polynomial using two adjacent real sampling points as control nodes to ensure that the first-order derivative is continuous at the boundary. We then resample the data at every 50 millisecond reference time point to obtain a 20 Hz uniform sequence.

[0035] Finally, all data are converted into standard physical units. Calculate the transient mutation index of conductivity and eliminate the conductivity with transient mutation index greater than 3.0, where: is the conductivity at the current moment, is the conductivity of the previous sampling period, =0.05s; data with temperature less than 10° or greater than 60° were eliminated; Haar wavelet transform was performed on the conductivity to eliminate noise energy greater than 50dB.

[0036] Furthermore, this embodiment performs feature enhancement and decoupling operations on multi-source heterogeneous data to obtain anti-interference features, including: Encoding operations are performed on multi-source heterogeneous data to obtain encoding features. Specifically, encoding operations include phase encoding, frequency encoding, and group pulse encoding. For conductivity, phase encoding using quadrature phase shift keying is used. The normalized amplitude of the conductivity value is used as the carrier, and its variation gradient is mapped to four phase offsets: 0°, 90°, 180°, and 270°.

[0037] According to Hebb's rule, a spiking neural network is constructed to compete and decouple the encoded features; specifically, the spiking neural network is composed of an input layer, a competition layer and an output layer. The input layer receives the encoded features, wherein each neuron corresponds to a type of encoding operation. The competition layer adopts a spike-timing-dependent plasticity rule, when a temperature pulse and a conductivity pulse arrive at the same time, an inhibitory intermediate neuron is activated to reduce the conductivity channel weight; the output layer adopts a leaky integrate-and-fire (LIF) model for pulse integration, and adjusts the firing threshold dynamically through an adaptive threshold adjustment mechanism to suppress noise, thereby outputting the decoupled pulse sequence.

[0038] The decoupled encoded features are subjected to scale decomposition and environmental interference suppression to obtain enhanced features; specifically, the decoupled pulse sequence is subjected to continuous wavelet transform, and a Mexican hat wavelet basis function is selected to decompose at 0.1s, 0.5s, 1s, 5s and 10s 5 scales, and the modulus maximum points at each scale are extracted as feature key nodes. A 50Hz band-stop filter is used for interference suppression at the 1s scale. The vibration pulse group of the pulse sequence is extracted at the 0.1s scale, and iterative decomposition is performed until the residual energy is <5%. The slowly varying component of the pulse sequence is removed at the 10s scale.

[0039] The enhanced features are subjected to thermodynamic reconstruction to generate anti-interference features. Specifically, the enhanced features F[ f 1, f 2,..., f n ] are mapped to the state of a thermodynamic system, expressed as: ; wherein, is the Boltzmann constant, is the i th component of the enhanced feature vector, is the probability weight of the feature component, is the feature dimension number. The state equation of the thermodynamic system is minimized by the Gibbs free energy, expressed as: ; wherein, is the square of the L2 norm, is the equilibrium factor, which is 0.1 in this embodiment, is the original feature vector before adjustment, is the to-be-solved feature, and the stable feature component with an entropy change rate <0.05 is taken as the anti-interference feature.

[0040] Preferably, the expression of the dynamic model is:

[0041] in, For the moment The control parameters are the operating frequency, current density or reaction time; For the moment Water quality parameters , namely conductivity, pH value, calcium ion concentration, magnesium ion concentration and temperature; is the coefficient calculated by the self-learning control algorithm, which represents the water quality parameter The degree of influence on the control parameters is calculated by the following formula: ; For the moment The amount of change in the control parameter, which represents the difference between the control parameter and the change in water quality; is the number of sample data; is the coefficient of the influence of the anti-interference feature on the control parameter, which represents the effect of the anti-interference feature on the control parameter and is calculated by the following formula: , For the moment The calculated anti-interference characteristics represent the signal characteristics after data preprocessing; For the moment The amount of change in the control parameter represents the difference between the control parameter and the water quality change.

[0042] Specifically, the dynamic model described in the present invention is used to describe the correspondence between pool water quality parameters and the control parameters of the electrochemical water treatment device. The control parameters include operating frequency, current density, and reaction time, and the water quality parameters include conductivity, pH, calcium ion concentration, magnesium ion concentration, and temperature. A self-learning control algorithm, based on collected historical water quality data and corresponding control parameter change data, automatically identifies the degree of influence of different water quality parameters on the control parameters. A corresponding weight coefficient is calculated for each water quality parameter to quantify the specific impact of water quality changes on the operating status of the water treatment device. This creates a dynamic model that automatically updates with water quality changes, enabling a precise response to water quality fluctuations.

[0043] In the specific implementation process, the weight coefficients are determined based on the historical fitting relationship between water quality parameters and control parameter changes. By statistically analyzing the correlation between water quality parameter changes and the adjustment range of control parameters, the actual contribution of each water quality parameter to the control parameter is determined. Based on this, the weight coefficients are established, thus forming a complete dynamic mapping model between water quality and control parameters. This model can clearly distinguish the proportion of influence of different water quality factors on water treatment effects, avoiding the problem that traditional fixed-proportion adjustment methods are unable to adapt to dynamic changes, and significantly improving the system's adaptability and response accuracy to complex water quality changes.

[0044] Furthermore, the present invention also incorporates the anti-interference features extracted in the data preprocessing stage into the dynamic model calculation. The anti-interference features are obtained through operations such as time alignment, feature enhancement and decoupling, and can effectively reflect the real change trend of water quality data after removing noise and outliers. When constructing the model, the degree of correlation between the anti-interference feature and the change of control parameters is comprehensively considered to further determine its influence coefficient in the model. By introducing the anti-interference feature, not only the robustness of the model to data anomalies and fluctuations is improved, but also higher stability and accuracy can be maintained during the water quality control process, forming a dynamic optimization mechanism that links water quality parameters, control parameters and anti-interference features, which is a key technical step that cannot be omitted in the present invention.

[0045] Preferably, when the water quality parameters in the pool reach the set dynamic optimization value, the RO reverse osmosis system is started to remove the dissolved salts in the water through RO reverse osmosis to reduce the conductivity of the water, including: Obtaining preset dynamic optimization values; the dynamic optimization values ​​include the set upper limit of conductivity, upper limits of calcium ion and magnesium ion concentration, and preset ranges of pH value and temperature; When any water quality parameter in the water pool reaches the dynamic optimization value, the self-learning control algorithm outputs a start instruction to start the RO reverse osmosis system, and determines the initial operating parameters of the RO reverse osmosis system based on the difference between the dynamic optimization value and the real-time water quality parameter; the initial operating parameters include working pressure, water inlet flow rate and target water production rate; After the RO reverse osmosis system is started, the water quality parameters of the water pool are monitored in real time. According to the dynamic difference between the conductivity of the water in the water pool and the set target value, the working pressure and water flow rate of the RO reverse osmosis system are adjusted to maintain the stability of the RO reverse osmosis membrane flux. The system recovery rate is adjusted according to the difference between the real-time water production rate and the target water production rate to improve the desalination efficiency and avoid the risk of membrane fouling or concentration polarization; The self-learning control algorithm combines the real-time operating data of the electrochemical water treatment device and the RO reverse osmosis system to dynamically correct the optimized values ​​of the pool water quality parameters and the operating parameters of the RO reverse osmosis system, so that the electrochemical water treatment device and the RO reverse osmosis system form a closed-loop collaborative control relationship, preferentially achieving effective removal of calcium and magnesium ions through electrochemical water treatment, thereby reducing subsequent RO load; When the conductivity monitoring value of the water in the water pool drops to the target range within the dynamic optimization value, the self-learning control algorithm outputs a stop instruction, shuts down the RO reverse osmosis system, and continues to monitor changes in the water quality of the water pool to maintain the water quality parameters stable within the set range for a long time.

[0046] Specifically, when the water quality parameters in the pool reach the dynamic optimization value, this embodiment performs a control operation through a self-learning control algorithm to obtain and compare the real-time monitored pool water quality data with the preset dynamic optimization value. The dynamic optimization value is determined based on the historical operation data of the pool and the long-term treatment effect of the electrochemical water treatment device, specifically including the set upper limit of conductivity, the upper limit of calcium ion concentration and magnesium ion concentration, and the reasonable range of pH value and temperature. By dynamically analyzing the difference between the real-time water quality of the pool and the above-mentioned optimization value, this embodiment determines that after the startup conditions are met, it outputs a startup instruction to control the RO reverse osmosis system to start operation, and determines the initial operating parameters based on the deviation between the real-time water quality and the dynamic optimization value, and clearly sets key parameters such as working pressure, water flow rate and target water production rate to ensure that the RO reverse osmosis device can achieve stable operation during the startup phase, avoid the risk of membrane contamination or abnormal fluctuations in energy consumption, and improve the safety of operation and the accuracy of water quality control.

[0047] During the operation of the RO reverse osmosis system, this embodiment monitors changes in pool water quality data in real time, focusing on conductivity trends. Based on the dynamic difference between the real-time conductivity and the target set value, the operating pressure and inlet flow rate of the RO reverse osmosis device are adjusted in real time to ensure that the membrane flux remains within a stable and reasonable range and prevent membrane performance degradation due to fluctuations in inlet water quality. At the same time, based on the difference between the water production rate and the set target water production rate, this embodiment further adjusts the RO system recovery rate parameters to improve desalination efficiency and prevent risks such as concentration polarization and membrane fouling. Through a dynamic parameter optimization mechanism, the RO reverse osmosis process achieves real-time and accurate response to water quality changes, ensuring water quality stability and system safety during long-term operation, and ensuring that water quality continues to converge to the target range.

[0048] Furthermore, this embodiment continuously acquires real-time operating data from the electrochemical water treatment device and the RO reverse osmosis device through a self-learning control algorithm. Based on changes in the pool water quality and the operating status of the water treatment equipment, it dynamically adjusts the water quality optimization value and RO system operating parameters. This forms a coordinated closed-loop control mechanism that prioritizes the removal of calcium and magnesium ions by electrochemical water treatment, followed by precise desalination through RO reverse osmosis. This reduces the load and operational risks of the RO device and improves the overall efficiency and stability of the system. When the pool water conductivity drops to within the target set range, this embodiment automatically outputs a stop command, shuts down the RO reverse osmosis device, and enters a continuous monitoring and dynamic maintenance phase, ensuring that water quality parameters remain within the set standard range over the long term, achieving intelligent, dynamic, and precise management of the entire water treatment process.

[0049] Preferably, the water level in the pool is adjusted in real time to ensure that the pool is within the preset optimal working range to adapt to the water treatment needs under different workloads, including: Based on the multi-sensor real-time data acquisition platform, the water level in the water pool is monitored in real time to obtain current water level data, and the water level data is compared with a preset optimal working water level range; the optimal working water level range is pre-set based on the design processing capacity of the electrochemical water treatment device and the RO reverse osmosis system and the load demand of the data center cooling system; When the water level in the pool is lower than the lower limit of the optimal working water level range, the self-learning control algorithm outputs a control instruction to start the water supply pump based on the water level change trend of the pool and the subsequent water treatment load forecast, and determines the water supply flow rate based on the difference between the current water level in the pool and the lower limit, thereby raising the water level in the pool to within the optimal working water level range. At the same time, the operating parameters of the water supply pump are adjusted in conjunction with the current water inflow and treatment capacity of the electrochemical water treatment device to prevent fluctuations in water quality parameters due to sudden changes in water volume; When the water level in the pool is higher than the upper limit of the optimal working water level range, the self-learning control algorithm outputs a control instruction to start the drainage pump based on the real-time water level data of the pool and the current water quality parameters, the inlet and outlet water volume and the processing capacity of the RO reverse osmosis system, and determines the drainage flow rate according to the difference between the current water level in the pool and the upper limit, thereby lowering the water level in the pool to within the optimal working water level range. At the same time, the inlet and outlet water rhythms of the electrochemical water treatment device and the RO reverse osmosis system are adjusted in a coordinated manner to avoid system load imbalance caused by abnormal water levels; During the water supply and drainage regulation process, the water level changes in the water pool and the operating status of the water supply pump and drainage pump are monitored in real time. Based on the water level change trend of the water pool and the water load of the data center cooling system, the start and stop status and flow control instructions of the water supply pump and drainage pump are dynamically adjusted to ensure that the water level in the water pool is stably maintained within the optimal working water level range; The self-learning control algorithm dynamically corrects the optimal working water level range based on the water level data of the water pool, the real-time operating data of the electrochemical water treatment device and the RO reverse osmosis system, and the actual load changes of the data center cooling system to adapt to changes in water treatment requirements in different seasons, different operating modes or different loads, ensuring that the water level of the water pool is always in an adaptable and optimal working state.

[0050] This embodiment uses a multi-sensor real-time data acquisition platform to acquire real-time pool water level data. This water level data is dynamically compared with a preset optimal operating water level range, which is determined based on the designed processing capacity of the electrochemical water treatment device and the RO reverse osmosis device, as well as the load requirements of the data center cooling system. When the pool water level falls below the lower limit of the optimal operating water level, this embodiment uses a self-learning control algorithm, combined with the pool water level change trend and future water treatment load forecasts, to output water supply pump start and stop and water supply flow adjustment instructions. The water supply flow is adjusted based on the difference between the water level and the lower limit, raising the pool water level to within the optimal range. At the same time, the water supply rhythm is adjusted in conjunction with the real-time water inflow and operating capacity of the electrochemical water treatment device to ensure stable water quality and avoid abnormal water quality fluctuations caused by sudden changes in water volume.

[0051] When the pool water level is higher than the upper limit of the optimal operating water level, this embodiment uses a self-learning control algorithm to combine real-time pool water level data with water quality parameters, pool inlet and outlet water flow, and the operating capacity of the RO reverse osmosis device to output drainage pump start and stop and drainage flow adjustment instructions. The drainage flow is precisely adjusted based on the difference between the water level and the upper limit to restore the pool water level to within the set range. At the same time, the inlet and outlet rhythm of the electrochemical water treatment device and the RO reverse osmosis device are adjusted in conjunction to avoid abnormal water levels causing sudden load changes or unstable operation of the water treatment device. The above steps ensure that the pool water level remains within a reasonable range as the water treatment and cooling loads change dynamically, and are the core mechanism of this embodiment to ensure stable water quality and continuous and efficient operation of the device.

[0052] During the water supply and drainage regulation process, this embodiment monitors the changes in the pool water level and the operating status of the water supply and drainage pumps in real time. Based on the pool water level trends and the water load changes in the data center cooling system, it dynamically adjusts the start and stop status and flow instructions of the water supply and drainage pumps to ensure that the pool water level is maintained within the optimal range. Furthermore, this embodiment combines pool water level data, real-time operating data from the electrochemical water treatment device and RO reverse osmosis device, and cooling load changes to continuously and dynamically adjust the optimal operating water level setpoint to adapt to water treatment needs in different seasons, operating modes, or load variations, thereby improving the overall stability, cost-effectiveness, and operational safety of the water treatment system.

[0053] This embodiment is used for water quality treatment of cooling circulating water in data centers. Through the coordinated operation of the electrochemical water treatment device and the RO reverse osmosis system, calcium and magnesium ions and soluble salts in the water are continuously removed, the stability of the circulating water quality is improved, and scaling, bacterial growth or algae pollution of the heat exchanger and pipes of the air-conditioning unit are prevented. In the specific process, the water supply pump delivers water to the AHU unit for cooling, and the water discharged from the unit is returned to the pool. When the water level in the pool is lower than the set liquid level, municipal tap water is added through the water replenishment device to maintain the water volume in the pool. The water in the pool continuously passes through the electrochemical water treatment device, which causes the calcium and magnesium ions in the water to precipitate in the form of scale. Subsequently, the water is intercepted by the high-efficiency filter and discharged into the sewage pool together with the sewage, ensuring that the hardness of the water continues to decrease. At the same time, it has a sterilization and algae-killing effect, ensuring that the water quality of the system meets the requirements of cooling circulation.

[0054] In this embodiment, when the conductivity of the water in the pool reaches a preset value, the RO reverse osmosis system is started through a self-learning control algorithm, and part of the circulating water is sent to the RO water tank. The soluble salts in the water are separated and removed by the RO reverse osmosis membrane, effectively reducing the conductivity of the water body. The RO produced water is reused in the circulation system, and the concentrated water is discharged into the sewage pool. The water quality of the wastewater discharged from the sewage pool meets the "Urban Wastewater Recycling Green Space Irrigation Water Quality" standard and can be used for park greening irrigation or external transportation for treatment. This embodiment uses this process to increase the concentration multiple of cooling circulating water and reduce the amount of sewage discharged by more than 70%, achieving the dual effects of water saving and energy reduction and water quality stabilization, meeting the stringent requirements of the data center cooling system for long-term operation water quality.

[0055] In this embodiment, the data center circulating water system includes: an electrochemical water treatment device arranged on the raw water side, an RO reverse osmosis system arranged on the deep treatment side, a water pool (also called an equalization pool) connected thereto and its liquid level, flow and water quality online monitoring unit, and a water supply pump and a drainage pump as actuators. For ease of explanation, the following definitions are used: The dynamic optimization value Θ is the threshold vector for adaptive convergence between the target water quality and the operational safety boundary. It includes at least the upper limit of conductivity, the upper limit of calcium and magnesium ions, and the allowable range of pH and temperature, and is updated online with operating conditions. The optimal operating water level range is the upper and lower limits of the water level calculated in real time by the self-learning algorithm under given load and residence time constraints. It is used to ensure the efficiency of electrochemical descaling and the stability of subsequent RO flux. The hysteresis band refers to the convergence margin within the target range set to suppress frequent starts and stops. The linkage beat refers to the execution rhythm of the duty cycle or frequency / valve position of each device in coordination with the master control command within a unified control cycle. The self-learning control algorithm is a comprehensive algorithm that uses the pool water level / quality, inlet and outlet water volume, RO operation measurements, and cooling load as inputs, estimates system disturbances and capacity boundaries online, and outputs closed-loop control and parameter update logic for pump speed, valve position, and RO start and stop. The system operates with a fixed control cycle, preferably in seconds. In each cycle, the algorithm first determines whether the water level and water quality have crossed the limit, then calculates the supply and drainage flow instructions and the RO start and stop and operating status. Finally, it performs a self-learning update of Θ and the optimal operating water level range, forming a coordinated control relationship of "electrochemical desalination first, RO precise desalination as a backup."

[0056] In the closed-loop control of water level and water quality, the target flow of the supply / drainage pump is driven by the difference and generated with amplitude and speed limits. The RO shutdown is constrained by hysteresis and hold time to avoid jitter. This can be achieved as follows: ; And on the execution side Slope constraints are applied to limit the instantaneous rate of change; RO start and stop adopt the following hysteresis-hold logic: when the conductivity monitoring value meets and maintain for not less than When the shutdown command is output, RO is turned off; when If either the calcium or magnesium ion indicator exceeds the upper limit of the dynamic optimization value for a duration of at least the startup delay, a startup command is issued to start the RO system and the inlet flow rate and transmembrane pressure differential are adjusted in conjunction with the system, while ensuring that the combined treatment capacity of the electrochemical device and RO system does not exceed the designed capacity. To prevent water quality fluctuations caused by sudden changes in water volume, the algorithm synchronously adjusts the inlet and outlet frequency / valve position of the electrochemical device and RO system according to the linkage rhythm to match the volume exchange rate of the water tank with the real-time load. The algorithm also performs online verification of the pump motor temperature rise, valve position travel, and sensor validity. In the event of an over-limit or failure, the system enters a degraded mode (freezing the learning step size, expanding the hysteresis band, and increasing the conductivity target margin) to ensure safe availability during abnormal conditions.

[0057] To adapt to seasonal, operational mode and cooling load variations, the self-learning algorithm jointly updates the conductivity target and the optimal working water level range at each control cycle, preferably using an update-projection integrated rule with feasible region projection and data-driven step size:

[0058] Wherein the parameters are as follows (combined into one paragraph): Wherein, is the instantaneous water level of the pool; and are the lower and upper limits of the optimal working water level range, respectively; and are the target volume flow rates of the water supply pump and the drainage pump, respectively; is a saturation function for limiting the instruction within the physically allowed range; and are the proportional coefficients automatically given by system calibration for mapping the water level deviation to the target flow rate; and are the maximum target flow rates allowed by the pump under the current working condition and are jointly limited by the equipment nameplate and online diagnosis; is the online monitoring value of the water conductivity; is the conductivity target value and is subject to the dynamic optimization value constraint; is the shutdown hysteresis band and is adaptively fine-tuned according to the disturbance intensity; is the RO shutdown holding time to avoid frequent start-stop; is the sliding statistical quantity of the conductivity for reflecting the recent steady-state level; is the downward floating amount for the membrane pollution safety margin to ensure the stable recovery of water quality after shutdown; is the process upper limit of the conductivity for clipping the target value; is a projection operator for mapping the candidate back to a solution that satisfies the feasible region , wherein is defined by the minimum residence time constraint, the minimum hysteresis width, and the pool structure boundary; is the short-term load prediction value of the data center cooling system, which can be estimated online from historical load and weather data; is the minimum water body residence time required to ensure the efficiency of electrochemical softening; is the current water body residence time calculated according to the real-time volume and net exchange flow rate; and are the data-driven step sizes automatically given by the normalized error variance and load prediction uncertainty to reduce manual tuning; is the dynamic optimization value vector, at least including the conductivity, calcium / magnesium upper limit, and pH temperature interval; The minimum water level hysteresis width is used to suppress high-frequency switching; the "linked beat" is the rhythm of the coordinated change of the duty cycle or frequency / valve position of each device with the master control instruction within the control cycle, which is used to maintain the smoothness of volume exchange and the stability of water quality during large load fluctuations.

[0059] Furthermore, this embodiment operates with a fixed control cycle, preferably 1-5 seconds; the sampling / filtering cycle of conductivity and water level is consistent with the control cycle or an integer multiple thereof, and the sliding statistical window of the conductivity target is preferably 10-30 minutes to suppress short-term fluctuations; the RO shutdown hold time is preferably 3-10 minutes, the startup delay is preferably 30-120 seconds, and the minimum startup time is set to 5-20 minutes and the minimum shutdown time is 5-15 minutes to avoid frequent startup and shutdown; the conductivity shutdown hysteresis band is set to the target value. , water level hysteresis zone (i.e. The minimum distance between the effective water depth Or the minimum amount of a project (such as the water level change corresponding to a single effective replenishment and drainage volume), whichever is greater; the slope limit of the water supply / drainage instruction ( ) It is preferred that the flow rate should not exceed the maximum flow rate currently allowed by the corresponding pump. The target maximum flow rate of the pump is preferably limited to the nameplate rating. To reserve linkage and disturbance response capabilities; when the electrochemical device and RO are running at the same time, the linkage beat ensures that the total processing capacity of the two does not exceed the design capacity Small synchronous fine-tuning is performed in steps of 1 to 3 control cycles to suppress transient water quality deviations caused by sudden changes in volume exchange.

[0060] Furthermore, the minimum residence time Determined by bench test or field test, the method is to measure the current density under representative raw water conditions and nominal electrochemical working point (such as plate spacing and current density within the recommended range of the equipment) The fractional removal rate meets the dynamic optimization value The average water residence time required for The safety margin of Minutes; working water level range The candidate update value must be projected back to satisfy , the hysteresis zone is not less than the aforementioned minimum spacing and does not cross the feasible region of the pool structure boundary ; Load forecasting The time domain window is preferably 15-60 minutes and is generated based on the combined characteristics of weather and historical loads. When the forecast uncertainty is large, the self-learning step size is automatically reduced; the step size of the conductivity target and the working water level range in the self-learning The preferred value is 0.02-0.15, and the "freeze-expand" protection is triggered for continuous over-limit or abnormal sensor quality status (such as single point loss for more than 10 seconds or consistency check failure): the freeze parameter updates a statistical window cycle, and expands the conductivity hysteresis band and water level hysteresis band by the original setting value. , and gradually unfreeze according to the original step size after the quality status is restored and the limit is eliminated; the above magnitude can be calibrated once by the project debugging records and equipment supplier data, and subsequently fine-tuned within the above range by the self-learning mechanism.

[0061] The beneficial effects of the present invention are as follows: 1. The present invention dynamically analyzes the changing trends of pool water quality through a self-learning control algorithm. Combined with real-time data from multiple sensors, it accurately identifies the impact of water quality changes on the electrochemical water treatment device and adjusts the operating frequency, current density, and reaction time in real time to ensure the effective removal of calcium and magnesium ions. This significantly improves the intelligent level and adaptive ability of water quality control and reduces the need for manual intervention.

[0062] 2. When the water quality parameters reach the dynamically optimized values, the present invention automatically starts the RO reverse osmosis device and dynamically adjusts the working pressure, water flow rate and recovery rate according to the deviation between the water quality and the target value, thereby ensuring the stability of the membrane flux and the desalination efficiency, avoiding the risks of membrane pollution and concentration polarization, improving the stability and accuracy of the RO reverse osmosis system operation, and ensuring that the water quality continues to converge to the target value.

[0063] 3. The present invention adopts a closed-loop control mechanism that coordinates electrochemical water treatment and RO reverse osmosis to preferentially remove calcium and magnesium ions, reduce the load of the RO system, and dynamically correct the water quality optimization value and operating parameters, forming a dynamic optimization process with fast response and precise control, thereby improving the stability and operating efficiency of the overall water treatment system and adapting to different load changes and water quality disturbance scenarios.

[0064] 4. The present invention dynamically adjusts the water level in the pool throughout the entire process to ensure that the water level is always in the optimal adaptability range. It continuously optimizes the water supply and drainage strategy based on water quality and load changes, ensures the stable operation of electrochemical water treatment and RO reverse osmosis, improves the energy-saving effect and economy of the system, and realizes efficient, stable and intelligent management and control of the entire process of cooling water treatment in the data center.

[0065] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0066] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. An electrochemical water treatment method for cooling circulating water treatment in a data center, characterized in that: include: The cooling circulating water of the data center is transported to the AHU unit through the water supply pump for cooling; The water discharged after cooling is collected in a pool and treated by an electrochemical water treatment device to precipitate calcium and magnesium ions in the water in the form of scale; Use a filter to filter out the scale precipitated in the pool and discharge the filtered water into the sewage pool; Based on a multi-sensor real-time data acquisition platform, the system dynamically monitors water quality parameters in the pool, acquires and analyzes water quality changes in real time, and optimizes the control steps in the water treatment process through a self-learning control algorithm to ensure that water quality meets set standards. The control steps include adjusting the operating frequency, current density, and reaction time of the electrochemical water treatment device. When the water quality parameters in the pool reach the set dynamic optimization value, the RO reverse osmosis system is started to remove the dissolved salts in the water through RO reverse osmosis to reduce the water conductivity; Drain the wastewater from the RO system into the sewage tank.

2. The electrochemical water treatment method for cooling circulating water treatment in a data center according to claim 1 is characterized in that: Also includes: Adjust the pool water level in real time to ensure that the pool is within the preset optimal working range to meet the water treatment needs under different workloads.

3. The electrochemical water treatment method for data center cooling circulating water treatment according to claim 1 is characterized in that: Conductivity, temperature, pH, calcium ion concentration and magnesium ion concentration.

4. The electrochemical water treatment method for data center cooling circulating water treatment according to claim 3 is characterized in that: Based on a multi-sensor real-time data acquisition platform, the water quality parameters in the pool are dynamically monitored, water quality changes are acquired and analyzed in real time, and the control steps in the water treatment process are optimized through self-learning control algorithms to ensure that the water quality meets the set standards, including: Conductivity, temperature, pH value, calcium ion concentration, and magnesium ion concentration in the water pool monitored by the multi-sensor real-time data acquisition platform are sequentially time-aligned, feature-enhanced, and decoupled to obtain anti-interference features; Based on the self-learning control algorithm, automatically identifying the relationship between the anti-interference characteristics and the working state of the electrochemical water treatment device to obtain a dynamic model; The operating frequency, current density and reaction time of the electrochemical water treatment device are adjusted through the dynamic model to ensure that the water quality meets the set standards.

5. The electrochemical water treatment method for data center cooling circulating water treatment according to claim 4 is characterized in that: Conductivity, temperature, pH value, calcium ion concentration, and magnesium ion concentration in the water pool monitored by the multi-sensor real-time data acquisition platform are sequentially subjected to time alignment, feature enhancement, and decoupling operations to obtain anti-interference features, including: Performing dynamic compensation for transmission delay on the multi-source heterogeneous data consisting of the conductivity, the temperature, the pH value, the calcium ion concentration, and the magnesium ion concentration by using a Kalman filter method; Based on a time protocol, the multi-source heterogeneous data are arranged in ascending order with the compensated conductivity as a reference time axis; performing spline interpolation and resampling operations on the multi-source heterogeneous data except the conductivity to obtain resampled data; Performing physical quantity conversion on the resampled data, and eliminating the resampled data whose evaluation index is greater than a preset index threshold; Performing an encoding operation on the multi-source heterogeneous data to obtain encoding features; the encoding operation includes: phase encoding, frequency encoding, and group pulse encoding; Constructing a spiking neural network according to Hebbian rule to competitively decouple the encoding features; Performing scale decomposition and environmental interference suppression on the decoupled coding features to obtain enhanced features; The enhanced feature is thermodynamically reconstructed to generate the anti-interference feature.

6. The electrochemical water treatment method for data center cooling circulating water treatment according to claim 4 is characterized in that: The expression of the dynamic model is: ； in, For the moment The control parameters are the operating frequency, current density or reaction time; For the moment Water quality parameters , namely conductivity, pH value, calcium ion concentration, magnesium ion concentration and temperature; is the coefficient calculated by the self-learning control algorithm, which represents the water quality parameter The degree of influence on the control parameters is calculated by the following formula: ; For the moment The amount of change in the control parameter, which represents the difference between the control parameter and the change in water quality; is the number of sample data; is the coefficient of the influence of the anti-interference feature on the control parameter, which represents the effect of the anti-interference feature on the control parameter and is calculated by the following formula: , For the moment The calculated anti-interference characteristics represent the signal characteristics after data preprocessing; For the moment The amount of change in the control parameter represents the difference between the control parameter and the water quality change.

7. The electrochemical water treatment method for data center cooling circulating water treatment according to claim 1, characterized in that: When the water quality parameters in the pool reach the set dynamic optimization value, the RO reverse osmosis system is started to remove the dissolved salts in the water through RO reverse osmosis to reduce the water conductivity, including: Obtaining preset dynamic optimization values; the dynamic optimization values ​​include the set upper limit of conductivity, upper limits of calcium ion and magnesium ion concentration, and preset ranges of pH value and temperature; When any water quality parameter in the water pool reaches the dynamic optimization value, the self-learning control algorithm outputs a start instruction to start the RO reverse osmosis system, and determines the initial operating parameters of the RO reverse osmosis system based on the difference between the dynamic optimization value and the real-time water quality parameter; the initial operating parameters include working pressure, water inlet flow rate and target water production rate; After the RO reverse osmosis system is started, the water quality parameters of the water pool are monitored in real time. According to the dynamic difference between the conductivity of the water in the water pool and the set target value, the working pressure and water flow rate of the RO reverse osmosis system are adjusted to maintain the stability of the RO reverse osmosis membrane flux. The system recovery rate is adjusted according to the difference between the real-time water production rate and the target water production rate to improve the desalination efficiency and avoid the risk of membrane fouling or concentration polarization; The self-learning control algorithm combines the real-time operating data of the electrochemical water treatment device and the RO reverse osmosis system to dynamically correct the optimized values ​​of the pool water quality parameters and the operating parameters of the RO reverse osmosis system, so that the electrochemical water treatment device and the RO reverse osmosis system form a closed-loop collaborative control relationship, preferentially achieving effective removal of calcium and magnesium ions through electrochemical water treatment, thereby reducing subsequent RO load; When the conductivity monitoring value of the water in the water pool drops to the target range within the dynamic optimization value, the self-learning control algorithm outputs a stop instruction, shuts down the RO reverse osmosis system, and continues to monitor changes in the water quality of the water pool to maintain the water quality parameters stable within the set range for a long time.

8. The electrochemical water treatment method for data center cooling circulating water treatment according to claim 2, characterized in that: Adjust the pool water level in real time to ensure that the pool is within the preset optimal working range to meet the water treatment needs under different workloads, including: Based on the multi-sensor real-time data acquisition platform, the water level in the water pool is monitored in real time to obtain current water level data, and the water level data is compared with a preset optimal working water level range; the optimal working water level range is pre-set based on the design processing capacity of the electrochemical water treatment device and the RO reverse osmosis system and the load demand of the data center cooling system; When the water level in the pool is lower than the lower limit of the optimal working water level range, the self-learning control algorithm outputs a control instruction to start the water supply pump based on the water level change trend of the pool and the subsequent water treatment load forecast, and determines the water supply flow rate based on the difference between the current water level in the pool and the lower limit, thereby raising the water level in the pool to within the optimal working water level range. At the same time, the operating parameters of the water supply pump are adjusted in conjunction with the current water inflow and treatment capacity of the electrochemical water treatment device to prevent fluctuations in water quality parameters due to sudden changes in water volume; When the water level in the pool is higher than the upper limit of the optimal working water level range, the self-learning control algorithm outputs a control instruction to start the drainage pump based on the real-time water level data of the pool and the current water quality parameters, the inlet and outlet water volume and the processing capacity of the RO reverse osmosis system, and determines the drainage flow rate according to the difference between the current water level in the pool and the upper limit, thereby lowering the water level in the pool to within the optimal working water level range. At the same time, the inlet and outlet water rhythms of the electrochemical water treatment device and the RO reverse osmosis system are adjusted in a coordinated manner to avoid system load imbalance caused by abnormal water levels; During the water supply and drainage regulation process, the water level changes in the water pool and the operating status of the water supply pump and drainage pump are monitored in real time. Based on the water level change trend of the water pool and the water load of the data center cooling system, the start and stop status and flow control instructions of the water supply pump and drainage pump are dynamically adjusted to ensure that the water level in the water pool is stably maintained within the optimal working water level range; The self-learning control algorithm dynamically corrects the optimal working water level range based on the water level data of the water pool, the real-time operating data of the electrochemical water treatment device and the RO reverse osmosis system, and the actual load changes of the data center cooling system to adapt to changes in water treatment requirements in different seasons, different operating modes or different loads, ensuring that the water level of the water pool is always in an adaptable and optimal working state.

Images