A refrigeration unit control method and system based on data monitoring

By acquiring and analyzing the operation data of the refrigeration unit in real time, identifying abnormalities and establishing a fault prediction model, the problem of difficulty in time discovering abnormalities and optimizing energy efficiency in the existing technology is solved, and efficient operation and fault prevention of the refrigeration unit are achieved.

CN118623517BActive Publication Date: 2024-11-08SHUZHONG IND EQUIP QIDONG CO LTD
View PDF 9 Cites 0 Cited by

Patent Information

Application Number
CN202411123192.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-11-08
Estimated Expiration
2044-08-15

AI Technical Summary

Technical Problem

In the prior art, it is difficult for refrigeration units to obtain operating data in real time, and they cannot detect abnormal situations in time, resulting in the inability to perform preventive maintenance, which increases production interruptions and maintenance costs caused by failures, and cannot optimize energy efficiency and predict failure trends.

Method used

By obtaining real-time operation data of the refrigeration unit, performing pre-processing and in-depth analysis, identifying abnormal data, evaluating key factors, establishing a fault prediction model, adjusting operating parameters, and building an abnormality monitoring model to monitor prone abnormal points in real time and responding to faults in advance.

Benefits of technology

It realizes timely detection of abnormal situations, preventive maintenance, reduce losses caused by failures, optimize energy efficiency, improve operating efficiency and reliability, and reduce downtime and maintenance costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118623517B_ABST
    Figure CN118623517B_ABST
Patent Text Reader

Abstract

The present invention discloses a refrigeration unit control method and system based on data monitoring, and relates to the technical field of refrigeration units. The refrigeration unit control method based on data monitoring includes the following steps: obtaining refrigeration unit operation characteristic data; identifying abnormal data of the current refrigeration unit operation state; evaluating key factors affecting the refrigeration unit operation state; judging whether it matches the known refrigeration unit operation failure mode, and adjusting the operation parameters of the refrigeration unit. The present invention can timely discover abnormal conditions, thereby performing preventive maintenance, avoiding production interruptions and maintenance costs caused by sudden failures, and establishing a fault prediction model based on key factors, which can predict future fault trends, allowing operators to respond in advance, reducing losses caused by failures, and thereby improving the operation efficiency and reliability of the refrigeration unit.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of refrigeration units, and in particular to a refrigeration unit control method and system based on data monitoring. Background Art

[0002] In the refrigeration industry, the refrigeration unit is the heart of the refrigeration system, and its performance directly affects the efficiency and energy consumption of the entire system. In particular, the two-stage compression refrigeration cycle is widely used in various occasions due to its efficient refrigeration capacity and suitability for large refrigeration sites. In order to further improve the cycle efficiency, many modern refrigeration units have added economizers to the two-stage compression system to optimize the cycle performance by supplementing air in the middle. This design requires the refrigerant to go through two throttling processes during the circulation process, which increases the complexity of the system.

[0003] Since the throttling flow area required under different working conditions and load conditions is different, in order to make the refrigeration unit always operate in the high-efficiency zone, a precise control method is needed to adjust the throttling device. The refrigeration unit control method based on data monitoring came into being. This method dynamically adjusts the flow area of ​​the throttling device by real-time monitoring of the working parameters of the refrigeration unit, such as temperature, pressure, flow, etc., and combining historical data with the preset optimization model. In this way, the system can automatically adjust to the optimal working state according to the changes in actual load and working conditions, thereby achieving efficient and energy-saving operation.

[0004] In the prior art, it is not convenient to obtain the operating data of the refrigeration unit in real time, and it is not possible to detect abnormal situations in time, so it is impossible to carry out preventive maintenance, and it is impossible to avoid production interruptions and repair costs caused by sudden failures. In addition, it is not possible to identify abnormal data during the operation of the refrigeration unit, so it is not convenient to adjust the operating parameters in a targeted manner, optimize the energy efficiency of the refrigeration unit, and reduce energy consumption. At the same time, it is not convenient to predict the future failure trend of the refrigeration unit, and it is not convenient for operators to respond in advance, which increases the losses caused by the failure.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention proposes a refrigeration unit control method and system based on data monitoring, which solves the problems mentioned in the above background technology that the existing system is inconvenient to obtain the operating data of the refrigeration unit in real time, cannot detect abnormal situations in time, and thus cannot perform preventive maintenance, cannot avoid production interruptions and repair costs caused by sudden failures, and cannot identify abnormal data during the operation of the refrigeration unit, thereby making it inconvenient to adjust the operating parameters in a targeted manner, optimize the energy efficiency of the refrigeration unit, and reduce energy consumption. At the same time, it is inconvenient to predict the future failure trend of the refrigeration unit, and it is inconvenient for operators to respond in advance, thereby increasing the losses caused by failures.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] According to one aspect of the present invention, a refrigeration unit control method based on data monitoring is provided, and the refrigeration unit control method based on data monitoring comprises the following steps:

[0009] S1. Acquire the real-time operation data of each part of the refrigeration unit, and pre-process the real-time operation data to obtain the operation characteristic data of the refrigeration unit;

[0010] S2. Use an analysis algorithm to perform in-depth analysis on the obtained refrigeration unit operation characteristic data to identify abnormal data of the current refrigeration unit operation status;

[0011] S3. Based on the evaluation algorithm, compare the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit to evaluate the key factors affecting the operating state of the refrigeration unit;

[0012] S4. Based on key factors, an integrated algorithm is used to establish a fault prediction model that affects the operating status of the refrigeration unit, and to determine whether it matches the known operating fault mode of the refrigeration unit and adjust the operating parameters of the refrigeration unit.

[0013] Further, using an analysis algorithm to deeply analyze the obtained refrigeration unit operation characteristic data, identifying abnormal data of the current refrigeration unit operation state includes the following steps:

[0014] S21, collecting historical operation data of various components of the refrigeration unit;

[0015] S22, calculating the number of abnormal occurrences of each refrigeration unit component within a time period, and dividing the number of abnormal occurrences by the total operating time of each refrigeration unit component to obtain the abnormal frequency distribution of each refrigeration unit component;

[0016] S23, comprehensively analyzing the abnormal frequency to obtain the abnormal occurrence law, and drawing a probability distribution curve, and taking the area with higher abnormal frequency in the probability distribution curve as the abnormal point prone area;

[0017] S24, using the abnormal rate distribution and abnormal occurrence rule of each refrigeration unit component as parameters for constructing an abnormality monitoring model, and deploying the abnormality monitoring model on the refrigeration unit components in the abnormal point prone area;

[0018] S25. If abnormality of refrigeration unit components is detected in the abnormal point area prone to occur, an analysis algorithm is used to detect abnormal values ​​of the refrigeration unit operation characteristic data to identify potential abnormal data of the current refrigeration unit operation status.

[0019] Furthermore, if abnormalities of refrigeration unit components are detected in the abnormal point area prone to abnormality, the abnormal value detection of the refrigeration unit operation characteristic data is performed using an analysis algorithm, and the identification of potential abnormal data of the current refrigeration unit operation state includes the following steps:

[0020] S251, dividing the refrigeration unit operation characteristic data into a plurality of sub-operation characteristic data, and obtaining a data point of each sub-operation characteristic data;

[0021] S252, calculating the K nearest neighbor distance and local reachability density of the data point of each sub-operation characteristic data, and calculating the local outlier factor value by comparing the local reachability density of the data point of each sub-operation characteristic data with the local reachability density of its K nearest neighbors;

[0022] S253, setting a threshold value for the local outlier factor value. If the local outlier factor value of a data point of a certain sub-operation characteristic data is greater than the preset threshold value, it is regarded as an abnormal value, and potential abnormal data of the current operation status of the refrigeration unit is obtained.

[0023] Further, calculating the K nearest neighbor distance and local reachability density of the data point of each sub-operation characteristic data, and calculating the local outlier factor value by comparing the local reachability density of the data point of each sub-operation characteristic data with the local reachability density of its K nearest neighbors includes the following steps:

[0024] S2521, extracting the features of each sub-operation feature data and constructing a feature matrix;

[0025] S2522, using Euclidean distance to calculate the distance between the data points of each sub-operation feature data to quantify the similarity between the data points of each sub-operation feature data;

[0026] S2523, selecting the best K value, for each data point of the molecular running characteristic data, finding the first K neighbors of the data point, and recording the positions of the K neighbors in the characteristic matrix;

[0027] S2524, for each data point of the sub-operation characteristic data and one of its K nearest neighbors, respectively calculate the reachable distance, the local reachable density and the local outlier factor value;

[0028] S2525. Calculate the density of neighbors around the data point of each sub-operation characteristic data according to the reachable distance, local reachable density and local outlier factor value, evaluate the abnormality of the data point of each sub-operation characteristic data, and sort the data points of each sub-operation characteristic data according to the local outlier factor value.

[0029] Furthermore, for each data point of the molecular running characteristic data and one of its K nearest neighbors, respectively calculating the reachable distance, the local reachable density and the local outlier factor value includes the following steps:

[0030] S25241. For each data point of each sub-operational characteristic data and one K nearest neighbor of the data point of each sub-operational characteristic data, calculate the larger value between the actual distance between the data point of each sub-operational characteristic data and one K nearest neighbor of the data point of each sub-operational characteristic data and the maximum distance between one K nearest neighbor of the data point of each sub-operational characteristic data and all K nearest neighbors of each sub-operational characteristic data, and obtain the reachable distance;

[0031] S25242, using the K value to divide the data point of each sub-operation characteristic data, obtain the sum of the reachable distances of the K nearest neighbors of each sub-operation characteristic data, and obtain the local reachable density;

[0032] S25243. The sum of the local reachable densities of the K nearest neighbors of each data point of the sub-operation characteristic data is divided by the local reachable density of the data point of each sub-operation characteristic data itself and then divided by the K value to obtain the local outlier factor value.

[0033] Furthermore, the calculation formula of the local outlier factor value is:

[0034] ;

[0035] in, The local outlier factor value represented as the data point of each molecular sub-run feature data;

[0036] The data points represented as characteristic data for each sub-run;

[0037] K represents the size of the neighborhood;

[0038] Represents a K nearest neighbor of the data point of each molecular running feature data;

[0039] The local reachable density of a K nearest neighbor of a data point represented as each molecular running feature data;

[0040] The local reachable density of data points representing the characteristic data for each molecule run.

[0041] Further, based on the evaluation algorithm, the identified abnormal data of the current operating state of the refrigeration unit is compared with the known abnormal data of the operating state of the refrigeration unit, and the key factors affecting the operating state of the refrigeration unit are evaluated, including the following steps:

[0042] S31, determining key parameters that affect the operating state of the refrigeration unit and setting an initial search range;

[0043] S32, preset an initial value of the evaluation cycle, and set the number of iterations as the end condition of the evaluation process;

[0044] S33, comparing the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit, and updating the key parameters according to a preset comparison method;

[0045] S34. Evaluate the refrigeration efficiency after each key parameter comparison by statistical methods;

[0046] S35. Calculate the fitness score of each iteration based on the evaluation result of the cooling efficiency;

[0047] S36, comparing the fitness scores of two consecutive iterations, and selecting the parameter update that has the greatest impact on the operating state of the refrigeration unit as the current best solution;

[0048] S37, increasing the initial value of the evaluation cycle, and if the preset upper limit of the number of iterations is reached, terminating the evaluation process;

[0049] S38. Analyze the operating status of the refrigeration unit after the evaluation process is completed, and compare it with the preset refrigeration efficiency standard to determine whether there is a significant impact. If so, make adjustments. If not, use the key parameters of the current operating status of the refrigeration unit as the key factors affecting the operating status of the refrigeration unit.

[0050] Furthermore, by means of statistical methods, the evaluation of the refrigeration efficiency after each key parameter comparison includes the following steps:

[0051] S341, initializing the verification round number to the first round, and dividing the acquired data set of the operating status of the refrigeration unit into several subsets, in each round, selecting one subset as a test set, and the union of all the remaining subsets as a training set;

[0052] S342, training the statistical model using the key parameters adjusted on the training set, and testing the prediction performance of the statistical model on the refrigeration efficiency on the test set;

[0053] S343. Calculate the generalization error of the statistical model based on the actual performance of the key parameters on the test set;

[0054] S344, looping step S341 to step S343 until each subset has been used as a test set;

[0055] S345. Obtain an average value of the generalization errors calculated in all verification rounds as an overall evaluation of the refrigeration efficiency after each key parameter comparison.

[0056] Furthermore, based on the key factors, a fault prediction model affecting the operating state of the refrigeration unit is established through an integrated algorithm, and it is determined whether it matches the known operating fault mode of the refrigeration unit. Adjusting the operating parameters of the refrigeration unit includes the following steps:

[0057] S41, collecting fault characteristic data of refrigeration units under different operating conditions to form an initial sample set;

[0058] S42, classifying the fault characteristic data, and dividing them into different sample subsets according to key factors affecting the operating state of the refrigeration unit;

[0059] S43, performing in-depth analysis on each sample subset to extract a coefficient matrix related to the failure type of the refrigeration unit;

[0060] S44, using the coefficient matrix to transform the original refrigeration unit fault feature data to generate a sample set reflecting new fault features;

[0061] S45, using the converted new sample set to train a fault prediction model that affects the operating state of the refrigeration unit, and obtaining an integrated fault prediction model through iterative optimization;

[0062] S46, collecting operation fault characteristic data of the refrigeration unit in real time during the operation of the refrigeration unit, and inputting the fault characteristic data into an integrated fault prediction model;

[0063] S47. Analyze and predict the real-time fault characteristic data through the integrated fault prediction model to obtain an evaluation of the future operating performance of the refrigeration unit, determine whether it matches the known operating fault mode of the refrigeration unit, and adjust the operating parameters of the refrigeration unit.

[0064] According to another aspect of the present invention, a refrigeration unit control system based on data monitoring is also provided, and the refrigeration unit control system based on data monitoring includes:

[0065] The data processing module is used to obtain the real-time operation data of each part of the refrigeration unit and perform preprocessing to obtain the operation characteristic data of the refrigeration unit;

[0066] A data analysis module is used to perform in-depth analysis on the obtained refrigeration unit operation characteristic data using an analysis algorithm to identify abnormal data of the current refrigeration unit operation status;

[0067] A data evaluation module, for comparing the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit based on an evaluation algorithm, and evaluating key factors affecting the operating state of the refrigeration unit;

[0068] The model building module is used to establish a fault prediction model that affects the operating status of the refrigeration unit based on key factors through an integrated algorithm, and to determine whether it matches the known operating fault mode of the refrigeration unit and adjust the operating parameters of the refrigeration unit;

[0069] Among them, the data processing module is connected with the data evaluation module through the data analysis module, and the data evaluation module is connected with the model establishment module.

[0070] The beneficial effects of the present invention are:

[0071] 1. The present invention can timely discover abnormal situations by acquiring the operating data of the refrigeration unit in real time, so as to carry out preventive maintenance and avoid production interruptions and repair costs caused by sudden failures. By deeply analyzing the operating characteristic data of the refrigeration unit and identifying abnormal data, the operating parameters can be adjusted in a targeted manner to optimize the energy efficiency of the refrigeration unit and reduce energy consumption. By using an evaluation algorithm to compare the current abnormal data with the known abnormal data, the problem can be quickly located, intelligent decision support can be provided for operators, and human errors can be reduced. In addition, a fault prediction model is established based on key factors to predict future fault trends, so that operators can respond in advance and reduce losses caused by faults, thereby improving the operating efficiency and reliability of the refrigeration unit.

[0072] 2. The present invention can accurately identify the abnormal point areas prone to occur in each component of the refrigeration unit by collecting historical operation data and calculating the abnormal frequency distribution, thereby realizing targeted monitoring and management. The constructed abnormal monitoring model can monitor the abnormal point areas prone to occur in real time. Once an abnormality is detected, an early warning can be issued immediately to provide operators with timely response time. In addition, by identifying potential abnormal data in advance, maintenance can be performed before a failure occurs, reducing downtime and maintenance costs.

[0073] 3. The present invention can more effectively optimize parameters and improve refrigeration efficiency by determining key parameters that affect the operating state of the refrigeration unit and setting an initial search range. By comparing current abnormal data with known abnormal data and updating key parameters according to a preset comparison method, the refrigeration unit can adapt to different operating conditions and load changes. The refrigeration efficiency after each key parameter comparison is evaluated by statistical methods, so that the impact of parameter changes on refrigeration efficiency can be accurately understood. Based on the evaluation results of refrigeration efficiency, the fitness score of each iteration is calculated, and the parameter update with the greatest impact on the operating state of the refrigeration unit is selected as the current best solution, thereby achieving parameter optimization.

[0074] 4. The present invention constructs a fault prediction model, thereby being able to predict potential fault risks in advance, and classifies and analyzes fault feature data, extracts coefficient matrices related to refrigeration unit fault types, and helps to diagnose and identify fault modes more accurately. The original data is converted using the coefficient matrix to generate a new fault feature sample set, which can better reflect the actual operating status of the refrigeration unit, thereby driving model optimization. During the operation of the refrigeration unit, fault feature data is collected in real time and input into the prediction model, which can achieve real-time monitoring and rapid response to the status of the refrigeration unit. At the same time, according to the evaluation results of the fault prediction model, it is determined whether it matches the known fault mode, and the operating parameters of the refrigeration unit are adjusted accordingly. Then, signs of faults can be discovered in advance through the prediction model, and preventive maintenance can be performed to reduce the losses caused by sudden failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. 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 creative work.

[0076] Figure 1 is a flow chart of a refrigeration unit control method based on data monitoring according to an embodiment of the present invention;

[0077] Figure 2 The invention is a block diagram of a refrigeration unit control system based on data monitoring according to an embodiment of the invention.

[0078] In the figure:

[0079] 1. Data processing module; 2. Data analysis module; 3. Data evaluation module; 4. Model building module. DETAILED DESCRIPTION

[0080] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0081] In the description of the present invention, unless otherwise specified, the meaning of "plurality" is two or more. In addition, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0082] According to an embodiment of the present invention, a refrigeration unit control method and system based on data monitoring are provided.

[0083] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to the refrigeration unit control method based on data monitoring according to an embodiment of the present invention, the refrigeration unit control method based on data monitoring includes the following steps:

[0084] S1. Acquire the real-time operation data of each part of the refrigeration unit, and pre-process the real-time operation data to obtain the operation characteristic data of the refrigeration unit;

[0085] Specifically, real-time operating data of various parts are obtained through various sensors, including temperature sensors, pressure sensors, flow sensors, vibration sensors, etc., which are used to monitor the physical parameters of different parts.

[0086] Specifically, real-time operation data includes:

[0087] Temperature: including the temperature of key components such as evaporator, condenser, expansion valve, etc.

[0088] Pressure: The suction pressure and discharge pressure of the compressor, as well as the pressure on the high-pressure and low-pressure sides of the system.

[0089] Flow rate: The flow rate of refrigerant and may also include the flow rate of cooling water or chilled water.

[0090] Vibration: Vibration data from rotating equipment such as compressors and fans to monitor machinery condition.

[0091] Electrical parameters: such as current, voltage, power, etc., used to monitor the status of the electrical system.

[0092] Other parameters: such as humidity, liquid level, etc. Other monitoring parameters may also be included depending on the specific refrigeration system design.

[0093] Specifically, the various parts refer to the main components of the refrigeration unit, such as the compressor, evaporator, condenser, expansion valve, oil separator, dry filter, safety valve, etc.

[0094] Specifically, obtaining the real-time operation data of each part of the refrigeration unit and preprocessing the real-time operation data to obtain the operation characteristic data of the refrigeration unit includes the following steps:

[0095] S11, obtaining duplicate data, missing values ​​and abnormal values ​​of real-time operation data of each part of the refrigeration unit, and performing denoising, filtering and smoothing on the duplicate data, missing values ​​and abnormal values;

[0096] S12, connecting the unprocessed operation data in various parts of the refrigeration unit to generate a new data table, and associating different data tables through external key values ​​to generate a complete data table, and obtain an accurate data set;

[0097] S13, extracting features from the accurate data set, and obtaining corresponding refrigeration unit operation feature data.

[0098] S2. Use an analysis algorithm to perform in-depth analysis on the obtained refrigeration unit operation characteristic data to identify abnormal data of the current refrigeration unit operation status;

[0099] Specifically, the abnormal data of the refrigeration unit operation status includes:

[0100] Abnormal temperature: If the monitored temperature value is outside the normal operating range, it may be due to insufficient refrigerant flow, system blockage, decreased heat exchange efficiency or other reasons.

[0101] Pressure anomalies: Unusual changes in system pressure may indicate leaks, overloads, or other fluid dynamics problems.

[0102] Abnormal flow: Abnormal flow of refrigerant or cooling water may indicate pipe blockage, reduced pump efficiency, or system design problems.

[0103] Abnormal vibration: Vibration of rotating equipment outside normal ranges may indicate a mechanical problem such as worn, unbalanced, or loose bearings.

[0104] Abnormal electrical parameters: Unusual fluctuations in current, voltage, or power may indicate problems with the electrical system, such as a short circuit, overload, or unstable power supply.

[0105] Abnormal performance indicators: such as a sudden drop in COP (energy efficiency ratio) or cooling capacity, may indicate a decrease in system efficiency.

[0106] Control system feedback: If the automatic control system reports an error code or protective shutdown, this is also considered abnormal data.

[0107] User-reported issues: Direct reports from operators, such as abnormal noise, leaks, unusual odors, etc., are also an important source of abnormal data.

[0108] S3. Based on the evaluation algorithm, compare the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit to evaluate the key factors affecting the operating state of the refrigeration unit;

[0109] Specifically, the known abnormal data of the refrigeration unit operation status refers to the abnormal data set that has been identified and classified through historical data collection and analysis. These data come from past failure events, maintenance records, system performance tests or other related research.

[0110] Specifically, the key factors affecting the operating status of the refrigeration unit refer to those variables or conditions that can significantly affect the performance and stability of the refrigeration unit, including:

[0111] Environmental conditions: such as external temperature, humidity, altitude, etc.

[0112] Operating parameters: such as set temperature, pressure, flow rate, compression ratio, etc.

[0113] Hardware status: such as the degree of wear of the compressor, the cleanliness of the heat exchanger surface, the leakage of the pipeline, etc.

[0114] Control system: such as control logic, sensor accuracy, actuator response time, etc.

[0115] Maintenance history: such as the time of the most recent maintenance, the quality and frequency of maintenance, etc.

[0116] Energy supply: such as voltage and frequency stability, energy prices, etc.

[0117] S4. Based on key factors, an integrated algorithm is used to establish a fault prediction model that affects the operating status of the refrigeration unit, and to determine whether it matches the known operating fault mode of the refrigeration unit and adjust the operating parameters of the refrigeration unit.

[0118] Specifically, known refrigeration unit operation failure modes refer to typical failure modes of refrigeration units that have been identified and classified based on historical data, professional knowledge, and technical literature. These failure modes include the characteristics of the failure, the cause of the failure, the scope of the impact, and possible manifestations. For example, compressor overheating, refrigerant leakage, electrical system short circuit, control system failure, etc. are all common failure modes.

[0119] Specifically, adjusting the operating parameters of the refrigeration unit includes:

[0120] Change set point: such as adjusting the set point of temperature, pressure or flow.

[0121] Changing the operating mode: such as switching from single-stage compression to multi-stage compression, or adjusting the operating frequency of the compressor.

[0122] Implement maintenance measures: such as cleaning the heat exchanger, replacing the filter, adding refrigerant, etc.

[0123] Optimize control systems: such as updating control logic, adjusting sensor sensitivity, etc.

[0124] Specifically, the operating parameters of the refrigeration unit include:

[0125] Temperature setting: including evaporator temperature, condenser temperature, etc.

[0126] Pressure Setting: The highest and lowest pressure limits of the system.

[0127] Flow control: such as refrigerant flow, cooling water flow, etc.

[0128] Compressor operation mode: such as single-stage or multi-stage compression, variable frequency control, etc.

[0129] Energy-saving mode: such as nighttime energy-saving mode, seasonal adjustment, etc.

[0130] Safety protection parameters: such as overload protection, low voltage protection, etc.

[0131] Preferably, using an analysis algorithm to perform in-depth analysis on the obtained refrigeration unit operation characteristic data to identify abnormal data of the current refrigeration unit operation state includes the following steps:

[0132] S21, collecting historical operation data of various components of the refrigeration unit;

[0133] S22, calculating the number of abnormal occurrences of each refrigeration unit component within a time period, and dividing the number of abnormal occurrences by the total operating time of each refrigeration unit component to obtain the abnormal frequency distribution of each refrigeration unit component;

[0134] S23, comprehensively analyzing the abnormal frequency to obtain the abnormal occurrence law, and drawing a probability distribution curve, and taking the area with higher abnormal frequency in the probability distribution curve as the abnormal point prone area;

[0135] S24, using the abnormal rate distribution and abnormal occurrence rule of each refrigeration unit component as parameters for constructing an abnormality monitoring model, and deploying the abnormality monitoring model on the refrigeration unit components in the abnormal point prone area;

[0136] S25. If abnormality of refrigeration unit components is detected in the abnormal point area prone to occur, an analysis algorithm is used to detect abnormal values ​​of the refrigeration unit operation characteristic data to identify potential abnormal data of the current refrigeration unit operation status.

[0137] Specifically, the analysis algorithm is a fault detection algorithm based on K-means clustering and local outlier factor algorithm. First, the K-means clustering algorithm is used to cluster multimodal industrial data to separate the data of each mode. Then, the local outlier factor algorithm is used to establish a separate model in each mode and determine the local outlier factor control limit in each mode. During detection, the sample is first determined to which category it belongs, and then the local outlier factor value is obtained in the corresponding category and compared with the control limit in this category to determine whether it is faulty data.

[0138] Specifically, the Local Outlier Factor (LOF) algorithm is a density-based anomaly detection algorithm. Its main idea is to evaluate whether a data point is an outlier by comparing the density of the neighbors around the data point.

[0139] Specifically, the K-means clustering algorithm is an iterative algorithm that aims to divide a data set into K clusters, where K is a user-specified number. Each cluster is composed of the closest points, and the definition of "closest" is the Euclidean distance (the straight-line distance between two points in multidimensional space). The goal of the algorithm is to group similar data together to achieve data classification.

[0140] Preferably, if abnormality of refrigeration unit components is detected in an abnormal point area prone to abnormality, abnormal value detection is performed on the refrigeration unit operation characteristic data using an analysis algorithm, and identifying potential abnormal data of the current refrigeration unit operation state includes the following steps:

[0141] S251, dividing the refrigeration unit operation characteristic data into a plurality of sub-operation characteristic data, and obtaining a data point of each sub-operation characteristic data;

[0142] S252, calculating the K nearest neighbor distance and local reachability density of the data point of each sub-operation characteristic data, and calculating the local outlier factor value by comparing the local reachability density of the data point of each sub-operation characteristic data with the local reachability density of its K nearest neighbors;

[0143] Specifically, local reachable density is a density concept defined based on local properties and is often used to identify outliers in a data set. The definition of local reachable density is based on the concept of "reachable distance", which is the maximum distance from a point to its k nearest neighbors.

[0144] S253, setting a threshold value for the local outlier factor value. If the local outlier factor value of a data point of a certain sub-operation characteristic data is greater than the preset threshold value, it is regarded as an abnormal value, and potential abnormal data of the current operation status of the refrigeration unit is obtained.

[0145] Preferably, calculating the K nearest neighbor distance and the local reachable density of the data point of each sub-operation characteristic data, and calculating the local outlier factor value by comparing the local reachable density of the data point of each sub-operation characteristic data with the local reachable density of its K nearest neighbors comprises the following steps:

[0146] S2521, extracting the features of each sub-operation feature data and constructing a feature matrix;

[0147] S2522, using Euclidean distance to calculate the distance between the data points of each sub-operation feature data to quantify the similarity between the data points of each sub-operation feature data;

[0148] Specifically, Euclidean distance, also known as Euclidean distance, is a distance measurement method in mathematics, which is used to calculate the straight-line distance between two points in Euclidean space (such as two-dimensional or three-dimensional space).

[0149] S2523, selecting the best K value, for each data point of the molecular running characteristic data, finding the first K neighbors of the data point, and recording the positions of the K neighbors in the characteristic matrix;

[0150] S2524, for each data point of the sub-operation characteristic data and one of its K nearest neighbors, respectively calculate the reachable distance, the local reachable density and the local outlier factor value;

[0151] S2525. Calculate the density of neighbors around the data point of each sub-operation characteristic data according to the reachable distance, local reachable density and local outlier factor value, evaluate the abnormality of the data point of each sub-operation characteristic data, and sort the data points of each sub-operation characteristic data according to the local outlier factor value.

[0152] Preferably, for each data point of the molecular running characteristic data and one of its K nearest neighbors, respectively calculating the reachable distance, the local reachable density and the local outlier factor value comprises the following steps:

[0153] S25241. For each data point of each sub-operational characteristic data and one K nearest neighbor of the data point of each sub-operational characteristic data, calculate the larger value between the actual distance between the data point of each sub-operational characteristic data and one K nearest neighbor of the data point of each sub-operational characteristic data and the maximum distance between one K nearest neighbor of the data point of each sub-operational characteristic data and all K nearest neighbors of each sub-operational characteristic data, and obtain the reachable distance;

[0154] S25242, using the K value to divide the data point of each sub-operation characteristic data, obtain the sum of the reachable distances of the K nearest neighbors of each sub-operation characteristic data, and obtain the local reachable density;

[0155] S25243. The sum of the local reachable densities of the K nearest neighbors of each data point of the sub-operation characteristic data is divided by the local reachable density of the data point of each sub-operation characteristic data itself and then divided by the K value to obtain the local outlier factor value.

[0156] Preferably, the calculation formula of the local outlier factor value is:

[0157] ;

[0158] in, The local outlier factor value represented as the data point of each molecular sub-run feature data;

[0159] The data points represented as characteristic data for each sub-run;

[0160] K represents the size of the neighborhood;

[0161] Represents a K nearest neighbor of the data point of each molecular running feature data;

[0162] The local reachable density of a K nearest neighbor of a data point represented as each molecular running feature data;

[0163] The local reachable density of data points representing the characteristic data for each molecule run.

[0164] Preferably, based on the evaluation algorithm, the identified abnormal data of the current operating state of the refrigeration unit is compared with the known abnormal data of the operating state of the refrigeration unit, and the evaluation of the key factors affecting the operating state of the refrigeration unit includes the following steps:

[0165] S31, determining key parameters that affect the operating state of the refrigeration unit and setting an initial search range;

[0166] S32, preset an initial value (usually 0 or 1) of the evaluation cycle (i.e., preset a time counter), and set the number of iterations as the end condition of the evaluation process;

[0167] Specifically, a time counter is a counting variable used to track the iteration process in an algorithm or program. In optimization and computational tasks, a time counter is used to monitor the number of iterations that an algorithm has executed and decide when to stop the iteration process.

[0168] S33, comparing the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit, and updating the key parameters according to a preset comparison method;

[0169] Specifically, the comparison method refers to a set of predefined rules or algorithms for evaluating and analyzing the similarities or differences between the abnormal data of the current refrigeration unit operating status and the abnormal data of the known (historical) refrigeration unit operating status.

[0170] Specifically, the comparison methods include:

[0171] Similarity metric: Define a mathematical way to quantify the similarity between the current anomaly data and the known anomaly data. Common metrics include Euclidean distance, Manhattan distance, cosine similarity, etc.

[0172] Feature selection: Select the key features that best represent the operating status of the refrigeration unit from the data. These features should be able to reflect the abnormal operation of the refrigeration unit.

[0173] Anomaly identification rules: define under what conditions the current data is considered abnormal. This can be based on statistical analysis, such as how many standard deviations a data point falls outside the normal operating data distribution.

[0174] Update mechanism: If the current anomaly data is significantly different from the known anomalies, it may be necessary to update or adjust the key reference data set, including the library of anomaly data, the selection of key features, etc., to ensure the accuracy and adaptability of the prediction model.

[0175] S34, evaluate the cooling efficiency after each key parameter comparison by statistical methods (i.e., k-fold cross validation);

[0176] Specifically, k-fold cross validation is a machine learning model evaluation method, which refers to dividing the collected set of key parameters that affect the operating status of the refrigeration unit into k subsets, one of which is used as a validation set, and the remaining k-1 subsets are used as training sets.

[0177] S35. Calculate the fitness score of each iteration based on the evaluation result of the cooling efficiency;

[0178] S36, comparing the fitness scores of two consecutive iterations, and selecting the parameter update that has the greatest impact on the operating state of the refrigeration unit as the current best solution;

[0179] S37, increasing the initial value of the evaluation cycle, and if the preset upper limit of the number of iterations is reached, terminating the evaluation process;

[0180] S38. Analyze the operating status of the refrigeration unit after the evaluation process is completed, and compare it with the preset refrigeration efficiency standard to determine whether there is a significant impact. If so, make adjustments. If not, use the key parameters of the current operating status of the refrigeration unit as the key factors affecting the operating status of the refrigeration unit.

[0181] Preferably, evaluating the refrigeration efficiency after each key parameter comparison by statistical methods comprises the following steps:

[0182] S341, initializing the verification round number to the first round, and dividing the acquired data set of the operating status of the refrigeration unit into several subsets, in each round, selecting one subset as a test set, and the union of all the remaining subsets as a training set;

[0183] S342, training a statistical model using the adjusted key parameters on the training set, and testing the prediction performance of the statistical model (i.e., the SVM model) on the refrigeration efficiency on the test set;

[0184] S343. Calculate the generalization error of the statistical model (i.e., SVM model) based on the actual performance of the key parameters on the test set;

[0185] Specifically, generalization error is an important indicator for evaluating the generalization ability of machine learning models. It reflects the prediction error of the model on new data that has not been seen.

[0186] Specifically, calculating the generalization error of the statistical model (ie, the SVM model) refers to calculating the prediction error of the statistical model (ie, the SVM model) on the test set.

[0187] S344, looping step S341 to step S343 until each subset has been used as a test set;

[0188] S345. Obtain an average value of the generalization errors calculated in all verification rounds as an overall evaluation of the refrigeration efficiency after each key parameter comparison.

[0189] Specifically, the evaluation algorithm is an SVM parameter selection method based on an improved bird flock optimization algorithm, which is a strategy for applying nature-inspired algorithms to machine learning parameter optimization. In this method, an improved bird flock optimization algorithm is used to select or optimize SVM parameters to improve the performance of the model.

[0190] Specifically, SVM is a supervised learning algorithm based on the principle of structural risk minimization. It uses expected risk minimization as the evaluation criterion and aims to obtain a decision function with high generalization ability through finite sample training, that is, to find an optimal hyperplane to completely separate the samples under the condition of linear separability.

[0191] Specifically, the improved bird flock optimization algorithm refers to the improvement and variant of the traditional particle swarm optimization algorithm (Particle Swarm Optimization, PSO). The particle swarm optimization algorithm is an optimization technology based on swarm intelligence, which simulates the social behavior of bird flocks or fish schools to solve optimization problems.

[0192] Specifically, the improved bird flock optimization algorithm includes the following improvements:

[0193] Improvements in speed and position update rules: In standard PSO, the movement of particles is based on their best historical position and the best position of the entire group. Improved algorithms may use different rules to adjust the speed and position of particles to improve convergence speed and avoid falling into local optimality.

[0194] Parameter adjustment and control: Improved algorithms may include dynamic adjustment mechanisms for PSO parameters such as inertia weight, individual best and group best learning factors to balance exploration (global search) and exploitation (local search) during the search process.

[0195] Customization to adapt to specific problems: For specific types of problems (such as nonlinear problems, multi-objective problems, etc.), the improved bird flock optimization algorithm may be specially designed to solve these problems more effectively.

[0196] Preferably, based on key factors, a fault prediction model affecting the operating state of the refrigeration unit is established through an integrated algorithm, and it is determined whether it matches a known refrigeration unit operating fault mode, and adjusting the operating parameters of the refrigeration unit includes the following steps:

[0197] S41, collecting fault characteristic data of refrigeration units under different operating conditions to form an initial sample set;

[0198] S42, classifying the fault characteristic data, and dividing them into different sample subsets according to key factors affecting the operating state of the refrigeration unit;

[0199] S43, performing in-depth analysis (principal component analysis) on each sample subset, and extracting a coefficient matrix (principal component coefficient matrix) related to the failure type of the refrigeration unit;

[0200] Specifically, in-depth analysis of each sample subset and extraction of coefficient matrices related to refrigeration unit fault types include the following steps:

[0201] S431. Re-sampling the sample subset of key factors affecting the operating state of the refrigeration unit by using a stratified sampling method to generate a new sample subset of key factors affecting the operating state of the refrigeration unit;

[0202] Specifically, stratified sampling is used to resample the sample subsets of key factors that affect the operating status of the refrigeration unit. This means that when extracting samples, the data is divided into different layers or subsets, and each layer represents a specific subgroup in the data. The purpose of this method is to ensure that the sample is more representative statistically, because it can ensure that data from each important subgroup in the data set are included in the sample.

[0203] S432, performing principal component analysis on a new sample subset of key factors affecting the operating state of the refrigeration unit, and obtaining a principal component coefficient vector related to the fault type of the refrigeration unit;

[0204] S433, repeatedly executing step S431 to step S432, processing different sample subsets of key factors affecting the operating state of the refrigeration unit, obtaining several groups of principal component coefficient vectors, and merging the principal component coefficient vectors into a principal component coefficient matrix.

[0205] Specifically, in-depth analysis of each sample subset and extraction of coefficient matrices related to refrigeration unit fault types include the following steps:

[0206] S44, using the coefficient matrix (principal component coefficient matrix) to transform the original refrigeration unit fault feature data to generate a sample set reflecting new fault features;

[0207] S45, using the converted new sample set to train a fault prediction model that affects the operating state of the refrigeration unit, and obtaining an integrated fault prediction model through iterative optimization;

[0208] S46, collecting operation fault characteristic data of the refrigeration unit in real time during the operation of the refrigeration unit, and inputting the fault characteristic data into an integrated fault prediction model;

[0209] S47. Analyze and predict the real-time fault characteristic data through the integrated fault prediction model to obtain an evaluation of the future operating performance of the refrigeration unit, determine whether it matches the known operating fault mode of the refrigeration unit, and adjust the operating parameters of the refrigeration unit.

[0210] Specifically, the integration algorithm is a heterogeneous multi-classifier integration algorithm. First, the rotation forest is used to transform and divide the original sample set to obtain a new sample set. Then, a support vector machine with high classification accuracy or a kernel matching pursuit with fast classification speed is selected as the basic integrated individual classifier through a specific proportion, and the new sample set is classified to obtain its prediction label. Finally, the prediction labels under the two models are combined. By combining two different classifier models, the algorithm achieves the complementarity of accuracy and speed, and improves the generalization error of the integrated system after the mixed integration of the two.

[0211] According to another embodiment of the present invention, Figure 2 As shown, a refrigeration unit control system based on data monitoring is also provided, and the refrigeration unit control system based on data monitoring includes:

[0212] Data processing module 1 is used to obtain real-time operating data of various parts of the refrigeration unit and perform preprocessing to obtain operating characteristic data of the refrigeration unit;

[0213] Data analysis module 2, used to perform in-depth analysis on the obtained refrigeration unit operation characteristic data using an analysis algorithm to identify abnormal data of the current refrigeration unit operation status;

[0214] The data evaluation module 3 is used to compare the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit based on the evaluation algorithm, and evaluate the key factors affecting the operating state of the refrigeration unit;

[0215] Model building module 4 is used to establish a fault prediction model that affects the operating state of the refrigeration unit based on key factors through an integrated algorithm, and to determine whether it matches a known refrigeration unit operating fault mode, and adjust the operating parameters of the refrigeration unit;

[0216] The data processing module 1 is connected to the data evaluation module 3 via the data analysis module 2 , and the data evaluation module 3 is connected to the model building module 4 .

[0217] In summary, with the help of the above technical solution of the present invention, the present invention can accurately identify the abnormal point areas prone to occur in each component of the refrigeration unit by collecting historical operation data and calculating the abnormal frequency distribution, thereby realizing targeted monitoring and management. The abnormal monitoring model constructed can monitor the abnormal point areas prone to occur in real time. Once an abnormality is detected, an early warning can be issued immediately to provide timely response time for operators, and by identifying potential abnormal data in advance, maintenance can be performed before a failure occurs, reducing downtime and maintenance costs. The present invention can more effectively optimize parameters and improve refrigeration efficiency by determining the key parameters that affect the operating state of the refrigeration unit and setting the initial search range. By comparing the current abnormal data with the known abnormal data and updating the key parameters according to the preset comparison method, the refrigeration unit can be adaptive to different operating conditions and load changes. The refrigeration efficiency after each key parameter comparison is evaluated by a statistical method, so that the influence of parameter changes on refrigeration efficiency can be accurately understood, and based on the evaluation result of refrigeration efficiency, the fitness score of each iteration is calculated, and the parameter update with the greatest influence on the operating state of the refrigeration unit is selected as the current optimal solution, thereby realizing parameter optimization. The present invention constructs a fault prediction model, thereby being able to predict potential fault risks in advance, and classifies and analyzes fault feature data, extracts a coefficient matrix related to the fault type of the refrigeration unit, and helps to diagnose and identify fault modes more accurately. The original data is converted using the coefficient matrix to generate a new fault feature sample set, which can better reflect the actual operating status of the refrigeration unit, thereby driving model optimization. During the operation of the refrigeration unit, fault feature data is collected in real time and input into the prediction model, which can achieve real-time monitoring and rapid response to the status of the refrigeration unit. At the same time, according to the evaluation results of the fault prediction model, it is determined whether the known fault mode is matched, and the operating parameters of the refrigeration unit are adjusted accordingly. Furthermore, the prediction model can be used to detect signs of faults in advance, and preventive maintenance can be performed to reduce the losses caused by sudden failures.

[0218] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A refrigeration unit control method based on data monitoring, characterized in that: The refrigeration unit control method based on data monitoring comprises the following steps: S1. Acquire the real-time operation data of each part of the refrigeration unit, and pre-process the real-time operation data to obtain the operation characteristic data of the refrigeration unit; S2. Use an analysis algorithm to perform in-depth analysis on the obtained refrigeration unit operation characteristic data to identify abnormal data of the current refrigeration unit operation status; S3. Based on the evaluation algorithm, compare the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit to evaluate the key factors affecting the operating state of the refrigeration unit; S4. Based on the key factors, a fault prediction model that affects the operating status of the refrigeration unit is established through an integrated algorithm, and whether it matches the known operating fault mode of the refrigeration unit is determined, and the operating parameters of the refrigeration unit are adjusted; The method of using an analysis algorithm to deeply analyze the obtained refrigeration unit operation characteristic data to identify abnormal data of the current refrigeration unit operation state includes the following steps: S21, collecting historical operation data of various components of the refrigeration unit; S22, calculating the number of abnormal occurrences of each refrigeration unit component within a time period, and dividing the number of abnormal occurrences by the total operating time of each refrigeration unit component to obtain the abnormal frequency distribution of each refrigeration unit component; S23, comprehensively analyzing the abnormal frequency to obtain the abnormal occurrence law, and drawing a probability distribution curve, and taking the area with higher abnormal frequency in the probability distribution curve as the abnormal point prone area; S24, using the abnormal rate distribution and abnormal occurrence rule of each refrigeration unit component as parameters for constructing an abnormality monitoring model, and deploying the abnormality monitoring model on the refrigeration unit components in the abnormal point prone area; S25. If abnormality of refrigeration unit components is detected in the abnormal point area prone to occur, abnormal value detection is performed on the refrigeration unit operation characteristic data using an analysis algorithm to identify potential abnormal data of the current refrigeration unit operation status; The method of comparing the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit based on the evaluation algorithm to evaluate the key factors affecting the operating state of the refrigeration unit includes the following steps: S31, determining key parameters that affect the operating state of the refrigeration unit and setting an initial search range; S32, preset an initial value of the evaluation cycle, and set the number of iterations as the end condition of the evaluation process; S33, comparing the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit, and updating the key parameters according to a preset comparison method; S34. Evaluate the refrigeration efficiency after each key parameter comparison by statistical methods; S35. Calculate the fitness score of each iteration based on the evaluation result of the cooling efficiency; S36, comparing the fitness scores of two consecutive iterations, and selecting the parameter update that has the greatest impact on the operating state of the refrigeration unit as the current best solution; S37, increasing the initial value of the evaluation cycle, and if the preset upper limit of the number of iterations is reached, terminating the evaluation process; S38, analyzing the operating state of the refrigeration unit after the evaluation process is completed, and comparing it with the preset refrigeration efficiency standard to determine whether there is a significant impact, if so, make adjustments, if not, take the key parameters of the current operating state of the refrigeration unit as the key factors affecting the operating state of the refrigeration unit; The method of establishing a fault prediction model that affects the operating state of the refrigeration unit based on key factors through an integrated algorithm, and determining whether the known operating fault mode of the refrigeration unit is matched, and adjusting the operating parameters of the refrigeration unit includes the following steps: S41, collecting fault characteristic data of refrigeration units under different operating conditions to form an initial sample set; S42, classifying the fault characteristic data, and dividing them into different sample subsets according to key factors affecting the operating state of the refrigeration unit; S43, performing in-depth analysis on each sample subset to extract a coefficient matrix related to the failure type of the refrigeration unit; S44, using the coefficient matrix to transform the original refrigeration unit fault feature data to generate a sample set reflecting new fault features; S45, using the converted new sample set to train a fault prediction model that affects the operating state of the refrigeration unit, and obtaining an integrated fault prediction model through iterative optimization; S46, collecting operation fault characteristic data of the refrigeration unit in real time during the operation of the refrigeration unit, and inputting the fault characteristic data into an integrated fault prediction model; S47. Analyze and predict the real-time fault characteristic data through the integrated fault prediction model to obtain an evaluation of the future operating performance of the refrigeration unit, determine whether it matches the known operating fault mode of the refrigeration unit, and adjust the operating parameters of the refrigeration unit.

2. A refrigeration unit control method based on data monitoring according to claim 1, characterized in that: If it is detected that an abnormality occurs in a refrigeration unit component in an abnormal point area, an abnormal value detection is performed on the refrigeration unit operation characteristic data using an analysis algorithm to identify potential abnormal data of the current refrigeration unit operation state, including the following steps: S251, dividing the refrigeration unit operation characteristic data into a plurality of sub-operation characteristic data, and obtaining a data point of each sub-operation characteristic data; S252, calculating the K nearest neighbor distance and local reachability density of the data point of each sub-operation characteristic data, and calculating the local outlier factor value by comparing the local reachability density of the data point of each sub-operation characteristic data with the local reachability density of its K nearest neighbors; S253, setting a threshold value for the local outlier factor value. If the local outlier factor value of a data point of a certain sub-operation characteristic data is greater than the preset threshold value, it is regarded as an abnormal value, and potential abnormal data of the current operation status of the refrigeration unit is obtained.

3. A refrigeration unit control method based on data monitoring according to claim 2, characterized in that: The method of calculating the K nearest neighbor distance and the local reachable density of the data point of each sub-operation characteristic data, and calculating the local outlier factor value by comparing the local reachable density of the data point of each sub-operation characteristic data with the local reachable density of its K nearest neighbors, comprises the following steps: S2521, extracting the features of each sub-operation feature data and constructing a feature matrix; S2522, using Euclidean distance to calculate the distance between the data points of each sub-operation feature data to quantify the similarity between the data points of each sub-operation feature data; S2523, selecting the best K value, for each data point of the molecular running characteristic data, finding the first K neighbors of the data point, and recording the positions of the K neighbors in the characteristic matrix; S2524, for each data point of the sub-operation characteristic data and one of its K nearest neighbors, respectively calculate the reachable distance, the local reachable density and the local outlier factor value; S2525. Calculate the density of neighbors around the data point of each sub-operation characteristic data according to the reachable distance, local reachable density and local outlier factor value, evaluate the abnormality of the data point of each sub-operation characteristic data, and sort the data points of each sub-operation characteristic data according to the local outlier factor value.

4. A refrigeration unit control method based on data monitoring according to claim 3, characterized in that: The calculation of the reachable distance, the local reachable density and the local outlier factor value for each data point of the molecular running characteristic data and one of its K nearest neighbors comprises the following steps: S25241. For each data point of each sub-operational characteristic data and one K nearest neighbor of the data point of each sub-operational characteristic data, calculate the larger value between the actual distance between the data point of each sub-operational characteristic data and one K nearest neighbor of the data point of each sub-operational characteristic data and the maximum distance between one K nearest neighbor of the data point of each sub-operational characteristic data and all K nearest neighbors of each sub-operational characteristic data, and obtain the reachable distance; S25242, using the K value to divide the data point of each sub-operation characteristic data, obtain the sum of the reachable distances of the K nearest neighbors of each sub-operation characteristic data, and obtain the local reachable density; S25243. The sum of the local reachable densities of the K nearest neighbors of each data point of the sub-operation characteristic data is divided by the local reachable density of the data point of each sub-operation characteristic data itself and then divided by the K value to obtain the local outlier factor value.

5. A refrigeration unit control method based on data monitoring according to claim 4, characterized in that: The calculation formula of the local outlier factor value is: ; in, The local outlier factor value represented as the data point of each molecular sub-run feature data; The data points represented as characteristic data for each sub-run; K represents the size of the neighborhood; Represents a K nearest neighbor of the data point of each molecular running feature data; The local reachable density of a K nearest neighbor of a data point represented as each molecular running feature data; The local reachable density of data points representing the characteristic data for each molecule run.

6. A refrigeration unit control method based on data monitoring according to claim 1, characterized in that: The method of evaluating the refrigeration efficiency after each comparison of key parameters by statistical methods includes the following steps: S341, initializing the verification round number to the first round, and dividing the acquired data set of the operating status of the refrigeration unit into several subsets, in each round, selecting one subset as a test set, and the union of all the remaining subsets as a training set; S342, training the statistical model using the key parameters adjusted on the training set, and testing the prediction performance of the statistical model on the refrigeration efficiency on the test set; S343. Calculate the generalization error of the statistical model based on the actual performance of the key parameters on the test set; S344, looping step S341 to step S343 until each subset has been used as a test set; S345. Obtain an average value of the generalization errors calculated in all verification rounds as an overall evaluation of the refrigeration efficiency after each key parameter comparison.

7. A refrigeration unit control system based on data monitoring, used to implement the refrigeration unit control method based on data monitoring according to any one of claims 1 to 6, characterized in that: The refrigeration unit control system based on data monitoring includes: The data processing module is used to obtain the real-time operation data of each part of the refrigeration unit and perform preprocessing to obtain the operation characteristic data of the refrigeration unit; A data analysis module is used to perform in-depth analysis on the obtained refrigeration unit operation characteristic data using an analysis algorithm to identify abnormal data of the current refrigeration unit operation status; A data evaluation module, for comparing the identified abnormal data of the current operating state of the refrigeration unit with the known abnormal data of the operating state of the refrigeration unit based on an evaluation algorithm, and evaluating key factors affecting the operating state of the refrigeration unit; The model building module is used to establish a fault prediction model that affects the operating status of the refrigeration unit based on key factors through an integrated algorithm, and to determine whether it matches the known operating fault mode of the refrigeration unit and adjust the operating parameters of the refrigeration unit; Among them, the data processing module is connected with the data evaluation module through the data analysis module, and the data evaluation module is connected with the model establishment module.

Citation Information

Patent Citations

  • Health state monitoring system for spatial environment simulation test equipment

    CN106610345A

  • The application of condition monitoring and fault diagnosis technology in the operation management of refrigeration unit

    CN109446482A

  • Water chilling unit fault diagnosis method and system based on Bayesian optimization LightGBM, and medium

    CN113792762A

  • Semi-supervised water chilling unit fault diagnosis method and system based on graph neural network

    CN115127192A

  • Intelligent air conditioner monitoring method and system based on multiple dimensions

    CN115899964A