Adaptive Control Method for Water-cooled Air Conditioners in Machine Rooms Based on Big Data Analysis

Abstract

The present invention discloses an adaptive control method for a water-cooled air conditioner in a computer room based on big data analysis, which relates to the technical field of adaptive regulation and control and is used to improve the problem of too long response time caused by fixed thermal inertia. It includes deploying a distributed optical fiber sensing network, screening environmental data in the water-cooled air conditioner adaptive system using a sliding time window algorithm, suppressing noise and entering a thermal inertia compensation mechanism, constructing an LSTM neural network based on the initial system thermal inertia to predict thermal disturbances in subsequent multiple control cycles, introducing a spatio-temporal alignment compensation matrix to suppress harmonic oscillations in the thermal inertia compensation mechanism, generating a thermal compensation correction amount using a dual closed-loop correction method based on the predicted thermal disturbances and the chilled water flow rate in the air conditioner, performing gain switching on the dual closed-loop correction method, synchronously aligning and correcting asynchronously collected data packets through a timing alignment engine, setting a bending window and performing feature fusion to judge the environmental data collection points and their abnormal collection response rates.

Description

Technical Field

[0001] The present invention relates to the technical field of adaptive control, and more specifically, to an adaptive control method for a water-cooled air conditioner in a computer room based on big data analysis. Background Art

[0002] When the load fluctuation changes significantly in the traditional adaptive control method for a water-cooled air conditioner in a computer room, the initial learning and adaptation time is long, and it cannot respond quickly.

[0003] The existing technology has the following deficiencies:

[0004] The existing system has a large thermal inertia. Even when the control system receives a signal of load change, the cooling capacity and response speed of the system need time to adapt to the new conditions. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the existing technology, an embodiment of the present invention provides an adaptive control method for a water-cooled air conditioner in a computer room based on big data analysis. By introducing an adaptive control algorithm and a thermal inertia compensation mechanism, the thermal inertia of the water-cooled air conditioner system is recorded and screened in real time to solve the problems raised in the above-mentioned background art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An adaptive control method for a water-cooled air conditioner in a computer room based on big data analysis, comprising the following steps:

[0008] Step S1: Deploy a distributed optical fiber sensing network, use a sliding time window algorithm to screen environmental data in the adaptive system of the water-cooled air conditioner, and enable an IIR digital filter to suppress noise and enter the thermal inertia compensation mechanism;

[0009] Step S2: Based on the initial system thermal inertia, construct an LSTM neural network to predict thermal disturbances in subsequent multiple control cycles, and introduce a spatio-temporal alignment compensation matrix to suppress harmonic oscillations in the thermal inertia compensation mechanism;

[0010] Step S3: Synthesize the predicted thermal disturbances and the chilled water flow rate in the air conditioner, use a dual closed-loop correction method to generate a thermal compensation correction amount, detect the environmental load, and set a variable structure sliding mode controller to switch the gain of the dual closed-loop correction method;

[0011] Step S4: Use an improved DTW algorithm through a timing alignment engine to synchronously align and correct asynchronously collected data packets, set a bending window and perform feature fusion to judge the environmental data collection points and their abnormal collection response rates.

[0012] In a preferred embodiment, in step S1, when deploying the distributed optical fiber sensing network, on the surface of the main pipeline and branch pipelines of the chilled water loop, each group of sensor nodes is designed to include 4 FBG sensors, which are arranged in a circular distribution at 120° intervals.

[0013] The environmental data, which is the temperature change rate of the chilled water tank, is acquired at a sampling rate of 5 Hz and transmitted to the central controller.

[0014] In a preferred embodiment, in step S1, the dynamic data screening based on the environmental data using the sliding time window algorithm is as follows:

[0015] Preprocessing: Initialize the window length and step size, and mark the window length as T_win;

[0016] Generate the parameter sequence: After performing the dimensionless operation on each environmental data, a parameter sequence is generated ;

[0017] It represents the data sequence within the sampling time window, that is, the standard deviation rate is calculated from the measurement data collected within a specific time window T_win , where is the standard deviation of the parameter sequence, is the mean of the parameter sequence; the kurtosis coefficient is ; where kurtosis is used to measure the steepness of the data distribution. Generally speaking, when the kurtosis coefficient is approximately equal to 3, the data follows a normal distribution. When the kurtosis coefficient is greater than 3, the data is concentrated near the mean. When the kurtosis coefficient is less than 3, the data shows a flat distribution;

[0018] In the formula, is each data point in the sequence, is the mean of the sequence, is the standard deviation of the parameter sequence, and n is the total number of data points; the calculation formula for the flow rate change rate is: , where is the flow rate at the current time t, is the flow rate at the previous time , is the sampling time interval.

[0019] In a preferred embodiment, in step S2, an LSTM neural network is constructed based on the initial system thermal inertia to predict the thermal disturbances in subsequent multiple control cycles. The specific steps are as follows:

[0020] Construct an LSTM network with multiple hidden units. The input layer receives the system thermal inertia at multiple time points, which are respectively marked as: ;

[0021] Among them, is the past load prediction data, is the past return water temperature, is the past heat transfer rate change data;

[0022] Hidden layer: Multiple LSTM units are used to learn the thermal inertia characteristics of the system;

[0023] The output layer predicts the thermal disturbance in the next n control cycles, that is, ;

[0024] Set the initial learning rate and define the loss function as:

[0025] ;

[0026] In the formula, MAE is the mean absolute difference between the predicted value and the true value;

[0027] Among them, the calculation formula of the mean squared logarithmic error MSLE is:

[0028] ;

[0029] Among them, MSLE is used to reduce the influence of extreme errors;

[0030] In the formula, is the regularization term, which is used to control the model complexity.

[0031] In a preferred embodiment, in step S2, a compensation coefficient is defined within the spatio-temporal alignment compensation matrix calculation rule:

[0032] When , the time deviation is less than half of the thermal inertia time constant, and the calculation rule of the compensation coefficient is: ;

[0033] When , the time deviation is greater than half of the thermal inertia time constant, and the calculation rule of the compensation coefficient is: ;

[0034] Among them, is the tracking error, and the harmonic oscillation in the compensation process is suppressed through the above calculation rule.

[0035] In a preferred embodiment, in step S3, a double closed-loop correction method is used to generate the thermal compensation correction amount. The specific steps are as follows:

[0036] Define the state vector: , where τ is the thermal inertia time constant, Ceq is the equivalent heat capacity, and Qacc is the cumulative heat deviation; among them, τ reflects the response speed of the system to temperature changes, Ceq measures the heat storage capacity of the system, directly affecting the heat compensation calculation, and Qacc records the heat loss or accumulation of the system to adjust the compensation strategy;

[0037] Adaptive adjustment formula of the observation matrix: ,

[0038] Among them, is when the heat compensation amount changes, 0.8 is a preset fixed weight to stabilize the influence of the observed data, is the time decay factor;

[0039] Set the trigger rule: When the residual covariance satisfies the following inequality, that is, when ;

[0040] The system triggers the parameter update mechanism, and the calculation formula of the gain matrix is as follows:

[0041] , among which, the process noise covariance is set to: , the observation noise covariance is set to: , among which, = , 0.02, 0.01, and 0.05 are respectively the perturbations of the preset thermal inertia time constant, the errors of the equivalent heat capacity, and the corrections of the cumulative heat deviation.

[0042] In a preferred embodiment, in step S4, the data packets collected asynchronously are synchronously aligned and corrected, and a long short-term memory (LSTM) neural network containing 128 hidden units is constructed to predict the thermal inertia dynamic characteristics in the chilled water system.

[0043] The data packets collected asynchronously include load prediction, return water temperature, supply water temperature, compressor power, condenser temperature difference, evaporator heat transfer efficiency, ambient temperature, cooling tower fan speed, chilled water flow rate, and pump pressure, which are respectively marked as , , , , , , , , , , and all the asynchronously collected data packets are used as ten-dimensional time series data;

[0044] Define the input features, and define the ten-dimensional time series data as the input features;

[0045] Network training and optimization: The initial learning rate is set to 0.001, and the learning rate decay rate is dynamically adjusted to 0.95 every 10 iterations; the loss function uses the mean square error labeled as MSE:

[0046] ;

[0047] The preset amount of training data is 100,000 time steps, and data augmentation is performed according to the window size;

[0048] Cross-validation is performed: K-fold cross-validation is used to determine the environmental data collection points in the preset time steps.

[0049] In a preferred embodiment, in step S4, the bending window is set by the finite difference method and feature fusion is performed to judge the environmental data collection points and their abnormal collection response rates as follows:

[0050] Calculate the thermal inertia change of the chilled water pipe, and the compensation formula is as follows: ; where, is the specific heat capacity, is the fluid mass flow rate, is the pipe wall thermal conductivity, is the heat transfer area, is the temperature difference, is the pipe length microelement;

[0051] Use the electric load field compensation method to calculate the abnormal collection response rate.

[0052] In a preferred embodiment, in step S4, the specific steps of using the electric load field compensation method to calculate the abnormal collection response rate are as follows:

[0053] Predict the transient fluctuation of the compressor power through the support vector regression model, and the compensation formula is as follows: ; where k is the adaptive adjustment factor, is the compensated compressor load power, is the currently measured compressor power, is the change rate of the supply-return water temperature difference, and the change rate of the supply-return water temperature difference is used as the abnormal collection response rate.

[0054] The technical effects and advantages of the adaptive control method for the computer room water-cooled air conditioner based on big data analysis of the present invention:

[0055] The present invention deploys a distributed optical fiber sensing network, screens environmental data in a water-cooled air conditioner adaptive system using a sliding time window algorithm, enables an IIR digital filter to suppress noise and enters a thermal inertia compensation mechanism, constructs an LSTM neural network based on the initial system thermal inertia to predict thermal disturbances in subsequent multiple control cycles, introduces a spatio-temporal alignment compensation matrix to suppress harmonic oscillations in the thermal inertia compensation mechanism, improves data transmission efficiency, generates a thermal compensation correction amount using a dual closed-loop correction method based on the predicted thermal disturbances and the chilled water flow rate in the air conditioner, detects the environmental load, sets a variable structure sliding mode controller to switch the gain of the dual closed-loop correction method, synchronously aligns and corrects asynchronously collected data packets using an improved DTW algorithm through a timing alignment engine, sets a bending window and performs feature fusion to judge the environmental data collection points and their abnormal collection response rates, and realizes the thermal inertia compensation closed loop of the water-cooled air conditioner, achieving the effect of accelerating the response. Description of the Drawings

[0056] Figure 1 It is a schematic diagram of the adaptive control method for the water-cooled air conditioner in the computer room based on big data analysis of the present invention.

[0057] Figure 2 It is a flowchart of the adaptive control method for the water-cooled air conditioner in the computer room based on big data analysis of the present invention. Detailed Embodiments

[0058] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0059] The present invention deploys a distributed optical fiber sensing network, screens environmental data in a water-cooled air conditioner adaptive system using a sliding time window algorithm, enables an IIR digital filter to suppress noise and enters a thermal inertia compensation mechanism, constructs an LSTM neural network based on the initial system thermal inertia to predict thermal disturbances in subsequent multiple control cycles, introduces a spatio-temporal alignment compensation matrix to suppress harmonic oscillations in the thermal inertia compensation mechanism, generates a thermal compensation correction amount using a dual closed-loop correction method based on the predicted thermal disturbances and the chilled water flow rate in the air conditioner, detects the environmental load, sets a variable structure sliding mode controller to switch the gain of the dual closed-loop correction method, synchronously aligns and corrects asynchronously collected data packets using an improved DTW algorithm through a timing alignment engine, sets a bending window and performs feature fusion to judge the environmental data collection points and their abnormal collection response rates, and realizes the thermal inertia compensation closed loop of the water-cooled air conditioner.

[0060] Embodiment, an adaptive control method for water-cooled air conditioners in computer rooms based on big data analysis, as follows Figure 1 and Figure 2 shown, includes the following steps:

[0061] Step S1: Deploy a distributed optical fiber sensing network, use a sliding time window algorithm to screen environmental data in the water-cooled air conditioner adaptive system, and enable an IIR digital filter to suppress noise and enter the thermal inertia compensation mechanism;

[0062] Step S2: Based on the initial system thermal inertia, construct an LSTM neural network to predict thermal disturbances in subsequent multiple control cycles, and introduce a spatio-temporal alignment compensation matrix to suppress harmonic oscillations in the thermal inertia compensation mechanism;

[0063] Step S3: Synthesize the predicted thermal disturbances and the chilled water flow rate in the air conditioner, use a dual closed-loop correction method to generate a thermal compensation correction amount, detect the environmental load, and set a variable structure sliding mode controller to switch the gain of the dual closed-loop correction method;

[0064] Step S4: Use an improved DTW algorithm through a timing alignment engine to perform synchronous alignment correction on asynchronously collected data packets, set a warping window and perform feature fusion to judge the environmental data collection points and their abnormal collection response rates.

[0065] The specific implementation is as follows:

[0066] In step S1, when deploying a distributed optical fiber sensing network, on the surface of the main pipe and branch pipes of the chilled water circuit, select an optical fiber Bragg grating (FBG) array inscribed by ultraviolet laser direct writing method with a spacing of 0.8 mm. This technology has the characteristics of high precision and high stability, and can effectively monitor temperature changes;

[0067] Each group of sensor nodes is designed to contain 4 FBG sensors and is arranged in a 120° circular distribution to ensure that the temperature distribution information of the pipe section can be comprehensively captured. For example, in the chilled water pipe, the water temperature change may be uneven due to the influence of flow rate, pipe wall thermal conductivity and external environmental temperature. The circular arrangement can more accurately obtain these temperature gradient information and provide basic data for subsequent analysis;

[0068] To achieve multiplexing, the sensor network uses wavelength division multiplexing (WDM) technology. The center wavelength is set in the range of 1529.55 nm to 1567.36 nm, and the wavelength interval is set to 0.8 nm. In this way, multiple signals can be transmitted simultaneously in the same optical fiber, greatly improving the efficiency and accuracy of data collection;

[0069] For example, in a large - scale chilled - water pipe network, a control unit may contain hundreds of FBG sensors. If each sensor needs to be independently connected, it will lead to complex wiring and high costs. The WDM technology can enable multiple sensors to share the same optical fiber, significantly reducing the amount of optical fiber used and lowering the system cost.

[0070] Adopt a sampling rate of 5 Hz and a resolution of 0.01 °C to obtain environmental data and transmit it to the central controller. Among them, the environmental data is the temperature change rate of the chilled - water storage tank.

[0071] Perform dynamic data screening based on environmental data using the sliding - time - window algorithm as follows:

[0072] Pre - processing: Initialize the window length and step size, and mark the window length as T_win.

[0073] Generate a parameter sequence: After performing dimensionless operation on each environmental data, generate a parameter sequence ;

[0074] It represents the data sequence within the sampling time window, that is, the measurement data collected within a specific time window T_win.

[0075] Calculate the standard - deviation rate , where is the standard deviation of the parameter sequence, is the mean of the parameter sequence;

[0076] Kurtosis coefficient ;

[0077] Among them, kurtosis is used to measure the steepness of the data distribution. Generally speaking, when the kurtosis coefficient is approximately equal to 3, the data follows a normal distribution. When the kurtosis coefficient is greater than 3, the data is concentrated near the mean. When the kurtosis coefficient is less than 3, the data shows a flat distribution;

[0078] In the formula, is each data point in the sequence, is the mean of the sequence, is the standard deviation of the parameter sequence, and n is the total number of data points;

[0079] For example, effective data screening conditions can be set:

[0080] When SDR < 0.15 and |Kurtosis - 3| < 1, it is determined as stable - operating - condition data for system analysis and decision - making;

[0081] When the flow - rate change rate > 15% / min, shorten T_win to half of the previous time and enable an IIR digital filter for high - frequency noise suppression, with the cut - off frequency set to 0.5 Hz.

[0082] Among them, the calculation formula for the flow rate change rate is:

[0083] ;

[0084] In the formula, is the flow rate at the current moment t, is the flow rate at the previous moment ; is the sampling time interval;

[0085] It should be noted that the IIR digital filter is a common signal processing tool used to remove high-frequency noise and smooth data. Its characteristics include using recursive calculation, where the output value can depend on the current input value and also on past output values;

[0086] In this system, the IIR digital filter is enabled for high-frequency noise suppression, and the cut-off frequency is set to 0.5 Hz, which means the system will filter out fluctuations above 0.5 Hz, thereby reducing short-term noise interference and making the flow rate data more stable, which is helpful for subsequent analysis and control;

[0087] In step S2, an LSTM neural network is constructed based on the initial system thermal inertia to predict thermal disturbances in subsequent multiple control cycles. The specific steps are as follows:

[0088] Construct an LSTM network with multiple hidden units. The input layer receives the system thermal inertia at multiple time points, which are respectively marked as: ;

[0089] Among them, is the load prediction data for the past 10 seconds, is the return water temperature for the past 5 seconds, is the heat transfer rate change data for the past 3 seconds;

[0090] Hidden layer: Adopt multiple LSTM units to learn the thermal inertia characteristics of the system;

[0091] The number of LSTM units in the specific hidden layer is not limited, but is set by the experimenter according to the specific implementation method and will not be elaborated here;

[0092] The output layer predicts thermal disturbances in the next n control cycles, that is ;

[0093] The training uses the Nadam optimizer, which combines the advantages of the NAG accelerated gradient and the Adam algorithm, improving the training convergence speed and stability;

[0094] Set the initial learning rate to 0.001, the batch size to 256, and the loss function is defined as:

[0095] ;

[0096] Wherein, MAE is the mean absolute error for measuring the difference between the predicted value and the true value;

[0097] Wherein, the calculation formula of the mean squared logarithmic error MSLE is:

[0098] ;

[0099] Wherein, MSLE is applicable to the case of a large prediction range and can reduce the influence of extreme errors;

[0100] In the formula, is the regularization term, which is used to control the model complexity and prevent overfitting. Specifically, the regularization coefficient λ = 0.01;

[0101] Define the calculation rule of the compensation coefficient in the spatio-temporal alignment compensation matrix:

[0102] When the time deviation is less than half of the thermal inertia time constant, the calculation rule of the compensation coefficient is: ;

[0103] When the time deviation is greater than half of the thermal inertia time constant, the calculation rule of the compensation coefficient is: ;

[0104] Wherein, is the tracking error, and the harmonic oscillation in the compensation process is suppressed through the above calculation rule;

[0105] It should be noted that in data flow analysis, if the SDR of temperature data is less than 0.15 and |Kurtosis - 3| < 1 within a certain period of time, it is determined that the data belongs to stable operating condition data, otherwise abnormal data needs to be further screened or filtered;

[0106] For example, during the operation of a chilled water pipe network, if there is a sudden increase in flow when the pump starts, the system can automatically shorten the time window to capture key data and suppress high-frequency noise, improving the monitoring accuracy;

[0107] In step S3, the specific steps for generating the thermal compensation correction amount using the double closed-loop correction method are as follows:

[0108] Definition of the state vector: ;

[0109] In the formula, τ is the thermal inertia time constant, Ceq is the equivalent heat capacity, and Qacc is the cumulative thermal deviation;

[0110] Among them, τ reflects the response speed of the system to temperature changes, Ceq measures the heat storage capacity of the system, which directly affects the heat compensation calculation, and Qacc records the heat loss or accumulation of the system to adjust the compensation strategy;

[0111] For example, in practical applications, if the change in the equivalent heat capacity Ceq of the chilled water tank exceeds the normal range (such as increasing by more than 10%), it may mean that there is scaling inside the tank or the heat transfer efficiency has decreased, and maintenance or cleaning is required.

[0112] Adaptive adjustment formula for the observation matrix:

[0113] ;

[0114] In the formula, When the change in the heat compensation amount is large, the system reduces the compensation intensity to avoid over-adjustment. 0.8 is a preset fixed weight to stabilize the influence of the observation data. is the time decay factor, which ensures that the influence of new observation data on the compensation amount gradually decreases and prevents short-term fluctuations from disturbing the system stability;

[0115] Set the trigger rule: When the residual covariance satisfies the following inequality, that is, when ;

[0116] The system triggers the parameter update mechanism, and the calculation formula of the gain matrix is as follows:

[0117] ;

[0118] Among them, the process noise covariance is set to: , and the observation noise covariance is set to: ;

[0119] In the formula, 0.02 corresponds to the perturbation of the thermal inertia time constant, 0.01 corresponds to the error of the equivalent heat capacity, and 0.05 corresponds to the correction of the cumulative heat deviation;

[0120] Among them, the observation noise covariance is used to limit the influence of measurement errors and ensure the system stability;

[0121] It should be noted that the core of the double closed-loop correction method lies in real-time adjustment of the compensation parameters to optimize the system response, thereby ensuring the efficient and stable operation of the chilled water system, which will not be elaborated here;

[0122] In step S4, the data packets collected asynchronously are synchronously aligned and corrected, and a long short-term memory (LSTM) neural network containing 128 hidden units is constructed to predict the thermal inertia dynamic characteristics in the chilled water system.

[0123] The data packets collected asynchronously include load prediction, return water temperature, supply water temperature, compressor power, condenser temperature difference, evaporator heat exchange efficiency, ambient temperature, cooling tower fan speed, chilled water flow rate, and pump pressure, which are respectively labeled as 、 、 、 、 、 、 、 、 、 , and all the asynchronously collected data packets are regarded as ten-dimensional time series data;

[0124] Define the input features and define the ten-dimensional time series data as the input features;

[0125] For example, in a chilled water system, the ambient temperature and the dynamic change of the cooling tower fan speed will directly affect the heat exchange efficiency of the condenser , and thus affect the overall refrigeration performance. Therefore, the dynamic correlation of these parameters must be incorporated into the input layer of the LSTM neural network to achieve accurate modeling.

[0126] Network training and optimization: The Adam optimization algorithm is adopted, the initial learning rate is set to 0.001, and the learning rate decay rate is dynamically adjusted to 0.95 every 10 rounds of iteration; The loss function uses the mean squared error (MSE):

[0127] ;

[0128] The training data volume is set to 100,000 time steps, and the sliding window method (window size = 50s) is used for data augmentation;

[0129] Perform cross-validation: Adopt K-fold cross-validation (K = 5) to determine the environmental data collection points in the preset time steps;

[0130] For example, during the experiment, if the error of the LSTM model on the test data continuously exceeds 2.5%, it may indicate that there is a lag effect in the input data, and it is necessary to adjust the time window size or increase the data normalization preprocessing to improve the prediction accuracy;

[0131] Among them, the normalization processing can use min-max normalization to scale the input data to the interval [0,1] to reduce the learning bias caused by numerical differences;

[0132] The dynamic thermal inertia of the chilled water system is mainly affected by the coupling of multiple physical fields such as fluid mechanics, heat transfer, and electrical load characteristics. To improve the compensation accuracy, a multi-physical field coupling compensation mechanism is adopted, and a multi-field model based on the finite element method (FEM) is designed, including:

[0133] Thermal field compensation calculates the change in thermal inertia of the chilled water pipeline through the finite difference method (FDM), and the compensation formula is as follows:

[0134] ;

[0135] In the formula, is the specific heat capacity, is the fluid mass flow rate, is the heat transfer coefficient of the pipe wall, is the heat transfer area, is the temperature difference, is the differential element of the pipe length;

[0136] For example: when the chilled water flow rate decreases (e.g., from 0.2 m / s to 0.1 m / s), the heat transfer efficiency of the pipeline decreases, and the compensation parameters need to be automatically adjusted to avoid overcooling or overheating.

[0137] Electric power load field compensation: predicts the transient fluctuation of the compressor power through a support vector regression model, and the compensation formula is as follows:

[0138] ;

[0139] Among them, k is an adaptive adjustment factor, which is updated in real time with the load change, is the compensated compressor power, is the currently measured compressor power, is the change rate of the temperature difference between supply and return water;

[0140] For example: when there are large fluctuations in the power grid of the chilled water plant (e.g., a 5% voltage drop), the system automatically adjusts the operating curve of the cooling tower fan to ensure stable power.

[0141] It should be noted that the support vector regression (SVR) for electric power load field compensation is only applicable to the prediction of short-term power fluctuations that are linear or weakly non-linear. If the power disturbance is large or the degree of non-linearity is high, the LSTM or Transformer structure may be more optimal. This solution uses SVR for prediction, but for complex power grid disturbance scenarios, more complex deep learning methods can be used for optimization, which will not be elaborated here.

[0142] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0143] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and the inventive constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0144] In addition, the functional modules in each embodiment of this application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0145] As mentioned above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0146] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. The adaptive control method of water-cooled air conditioner in computer room based on big data analysis is characterized in that: The following steps are included: Step S1: Deploy the distributed optical fiber sensor network, use the sliding time window algorithm to filter the environmental data in the water-cooled air conditioning adaptive system, enable the IIR digital filter to suppress the noise and enter the thermal inertia compensation mechanism; Step S2: Based on the initial system thermal inertia, an LSTM neural network is constructed to predict the thermal disturbances of subsequent multiple control cycles, and a spatiotemporal alignment compensation matrix is ​​introduced to suppress harmonic oscillations in the thermal inertia compensation mechanism; Step S3: using a double closed-loop correction method to generate a thermal compensation correction value based on the predicted thermal disturbance and the chilled water flow in the air conditioner, detecting the environmental load, and setting a variable structure sliding film controller to perform gain switching on the double closed-loop correction method; Step S4: Using the improved DTW algorithm through the timing alignment engine to perform synchronous alignment correction on the asynchronously collected data packets, set the bending window and perform feature fusion to determine the environmental data collection points and their abnormal collection response rates; In step S1, dynamic data screening is performed based on environmental data using a sliding time window algorithm as follows: Preprocessing: Initialize the window length and step size, and mark the window length as T_win; Generate parameter sequence: Generate parameter sequence after de-dimensionalizing each environmental data ; It represents the data sequence within the sampling time window, that is, the standard deviation rate of the measurement data collected within a specific time window T_win ,in, is the standard deviation of the parameter sequence, is the parameter sequence mean; the kurtosis coefficient is ; Among them, kurtosis is used to measure the steepness of data distribution. When the kurtosis coefficient is equal to 3, the data follows a normal distribution. When the kurtosis coefficient is greater than 3, the data is concentrated near the mean. When the kurtosis coefficient is less than 3, the data is flatly distributed. In the formula, For each data point in the sequence, is the mean of the series, is the standard deviation of the parameter sequence, n is the total number of data points; the calculation formula for the flow rate change rate is: ,in, is the flow rate at the current time t, For the previous moment of traffic, Sampling time interval; In step S2, an LSTM neural network is constructed based on the initial system thermal inertia to predict the thermal disturbances of subsequent multiple control cycles. The specific steps are as follows: Construct an LSTM network with multiple hidden units, and the input layer receives the system thermal inertia at multiple time points, which are marked as: ; in, is the historical load forecast data, is the historical return water temperature, It is the historical heat exchange rate change data; Hidden layer: multiple LSTM units are used to learn the thermal inertia characteristics of the system; The output layer predicts the thermal disturbance of the next n control cycles, that is, ; Set the initial learning rate and define the loss function as: ； In the formula, MAE is the average absolute difference between the predicted value and the true value; Among them, the calculation formula of MSLE mean square logarithmic error is: ； Among them, MSLE is used to reduce the impact of extreme errors; In the formula, is a regularization term used to control the complexity of the model; In step S2, the compensation coefficients are defined in the spatiotemporal alignment compensation matrix Calculation rules: when When the time deviation is less than half of the thermal inertia time constant, the calculation rule of the compensation coefficient is: ; when When the time deviation is greater than half of the thermal inertia time constant, the calculation rule of the compensation coefficient is: ; in, To track the error, the above calculation rules are used to suppress the harmonic oscillation during the compensation process; In step S3, the double closed-loop correction method is used to generate the thermal compensation correction value. The specific steps are as follows: Define the state vector: , where τ is the thermal inertia time constant, Ceq is the equivalent heat capacity, and Qacc is the accumulated thermal deviation; where τ reflects the system's response speed to temperature changes, Ceq measures the system's heat storage capacity and directly affects the thermal compensation calculation, and Qacc records the system's heat loss or accumulation to adjust the compensation strategy; Adaptive adjustment formula of the observation matrix: , in, When the thermal compensation amount changes, 0.8 is the preset fixed weight to stabilize the influence of the observed data. is the time decay factor; Set the trigger rule: If the residual covariance satisfies the following inequality, that is, when hour; The system triggers the parameter update mechanism, the gain matrix The calculation formula is as follows: , where the process noise covariance is set as: , the observation noise covariance is set as: ,in, = , 0.02, 0.01, and 0.05 are the disturbance of the preset thermal inertia time constant, the error of the equivalent heat capacity, and the correction of the accumulated thermal deviation, respectively; In step S4, the asynchronously collected data packets are synchronized and corrected, and a long short-term memory (LSTM) neural network with 128 hidden units is constructed to predict the dynamic characteristics of thermal inertia in the chilled water system; The asynchronously collected data packets include load forecast, return water temperature, supply water temperature, compressor power, condenser temperature difference, evaporator heat exchange efficiency, ambient temperature, cooling tower fan speed, chilled water flow and pump pressure, which are marked as , , , , , , , , , , all asynchronously collected data packets are regarded as ten-dimensional time series data; Define input features and define the ten-dimensional time series data as input features; Network training and optimization: The initial learning rate is set to 0.001, and the learning rate decay rate is dynamically adjusted to 0.95 every 10 iterations; the loss function uses the mean square error marked as MSE: ； The amount of training data is preset to 100,000 time steps, and data augmentation is performed according to the window size; Perform cross-validation: Use K-fold cross-validation to determine the environmental data collection points in the preset time step.

2. The method for adaptive control of water-cooled air conditioners in a computer room based on big data analysis according to claim 1 is characterized in that: In step S1, when the distributed optical fiber sensor is deployed in the network, on the surface of the main pipe and branch pipe of the chilled water loop, each group of sensor nodes is designed to include 4 FBG sensors, and are arranged in a 120° ring distribution; The environmental data is acquired at a sampling rate of 5 Hz and transmitted to the central controller, where the environmental data is the temperature change rate of the cold storage water tank.

3. The method for adaptive control of water-cooled air conditioners in a computer room based on big data analysis according to claim 1 is characterized in that: In step S4, the curved window is set by the finite difference method and the characteristic fusion is performed to determine the environmental data collection point and its abnormal collection response rate as follows: Calculate the thermal inertia change of the chilled water pipeline and the compensation formula is as follows: ;in, is the specific heat capacity, is the fluid mass flow rate, is the thermal conductivity of the tube wall, is the heat transfer area, is the temperature difference, is the infinitesimal length of the pipeline; The abnormal acquisition response rate is calculated using the power load field compensation method.

4. The method for adaptive control of water-cooled air conditioners in a computer room based on big data analysis according to claim 3 is characterized in that: In step S4, the specific steps of calculating the abnormal acquisition response rate using the power load field compensation method are as follows: The transient fluctuation of compressor electric power is predicted by support vector regression model, and the compensation formula is as follows: ; where k is the adaptive adjustment factor, is the load power of the compressor after compensation, is the currently measured compressor power, is the change rate of the supply and return water temperature difference, and the change rate of the supply and return water temperature difference is used as the abnormal collection response rate.

Images