A method and system for dynamically evaluating the health of a refrigeration unit

Abstract

The present application relates to the technical field of equipment management, and particularly relates to a refrigeration unit health condition dynamic evaluation method and system, comprising the following steps: a gradient boosting regression tree algorithm is used to construct a refrigeration capacity and energy efficiency ratio prediction model, historical data is fully utilized to train a high-precision benchmark, and a stable and reliable prediction benchmark and data analysis basis are provided for subsequent dynamic evaluation; a static model is converted into a dynamic model responding to real-time working conditions, real-time evaluation of the performance of the refrigeration unit is realized, and the sensitivity and accuracy of state monitoring are effectively improved in combination with degradation degree cooperative calculation; the real-time score of the refrigeration unit is compared with a preset judgment threshold, and the health state of the refrigeration unit can be quickly and accurately judged; when it is detected that the refrigeration unit is in an abnormal state, the model update period and the sensor sampling frequency can be dynamically adjusted, and more accurate and timely data can be obtained at a critical moment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of equipment management, and in particular to a refrigeration unit health condition dynamic evaluation method and system. BACKGROUND

[0002] Under the background of rapid development and wide application of current refrigeration technology, the field of refrigeration unit management still faces multiple challenges: traditional methods may not accurately evaluate the real-time health status of the refrigeration unit, leading to misjudgment of the unit performance and failure to timely detect potential faults or performance degradation; traditional evaluation methods may not fully consider the performance changes of the refrigeration unit under different working conditions, resulting in evaluation results that do not match the actual operation; traditional methods may lack real-time monitoring of the performance degradation of the refrigeration unit, leading to maintenance only after a fault occurs, increasing maintenance costs and downtime; traditional maintenance strategies may use fixed maintenance cycles and methods, which cannot develop targeted maintenance plans according to the actual operating status and performance degradation of the unit. Therefore, the present application proposes a refrigeration unit health condition dynamic evaluation method and system. SUMMARY

[0003] The purpose of the present application is to solve the problems in the background art and propose a refrigeration unit health condition dynamic evaluation method and system.

[0004] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0005] A refrigeration unit health condition dynamic evaluation method, comprising:

[0006] S1, obtaining an electronic expansion valve opening degree data stream, and using a gradient boosting regression tree algorithm to establish a baseline model M1 for predicting refrigeration capacity and a baseline model M2 for predicting energy efficiency ratio;

[0007] S2, converting the trained baseline model M1 and baseline model M2 into dynamic baseline models responsive to real-time working conditions; inputting real-time electronic expansion valve opening degree into the dynamic baseline models and performing real-time performance index and degradation degree collaborative calculation to obtain real-time scores of the refrigeration unit; wherein the real-time performance index is standard refrigeration capacity and standard energy efficiency ratio;

[0008] S3, according to the real-time score of the refrigeration unit, comparing with the configuration judgment threshold to judge the state of the refrigeration unit, including healthy, sub-healthy, fault, serious fault or complete failure state;

[0009] S4, according to the evaluation result, establishing a dynamic adjustment strategy: if the refrigeration unit is detected to be in a fault, serious fault or complete failure state, dynamically adjusting the update period of the dynamic baseline model and the sensor sampling frequency.

[0010] Further, the process of acquiring an electronic expansion valve opening degree data stream and using a gradient boosting regression tree algorithm to respectively establish a baseline model M1 for predicting refrigeration capacity and a baseline model M2 for predicting energy efficiency ratio includes:

[0011] S11, from the historical operation database of the water chiller or refrigeration system, collect a historical operation data set: electronic expansion valve opening degree , refrigeration capacity Q and energy efficiency ratio EER; clean the collected historical operation data set, process missing values and abnormal values; randomly shuffle the cleaned data set, and divide it into a training set and a test set according to a preset proportion;

[0012] S12, normalize the input features mainly based on the electronic expansion valve opening degree ; at the same time, construct derived features closely related to the physical relationship of refrigeration capacity Q and energy efficiency ratio EER;

[0013] S13, using the preprocessed electronic expansion valve opening degree and derived features as common input feature sets, train model M1 and model M2 in the same gradient boosting regression tree training framework by replacing the output target, the process is as follows: select a gradient boosting regression tree regressor, preset the total number of iterations, and initialize the initial prediction values of the two models; train round by round, a total of rounds, update model M1 and model M2 synchronously in each round; wherein the iterative training process includes calculating the negative gradient, training a new regression tree, calculating the leaf node output value, and updating the overall prediction function;

[0014] S14, use the divided test set to evaluate the performance of baseline model M1 and baseline model M2 respectively.

[0015] Further, the process of model iterative training includes:

[0016] A1, for each sample i in the training set, calculate the negative gradient between the prediction value and the true value of the current model;

[0017] A2, using the current entire training set, taking the electronic expansion valve opening degree as the input, and the calculated negative gradient as the output target, train a new regression tree, and divide the data into J leaf nodes through the new regression tree; wherein J represents that the mth tree is divided into J leaf nodes, i.e. the number of leaf nodes; R is a region, which divides the input space into multiple non-overlapping regions; Rj represents the jth leaf node region in the mth tree;

[0018] A3, calculate an output value for each leaf node j of the new tree ,in, This represents the output value of the j-th leaf node in the m-th tree;

[0019] A4. Add the new tree to the ensemble model and update the overall prediction function; fine-tune the hyperparameters shared by models M1 and M2 on the training set through cross-validation, and find the optimal hyperparameter combination that minimizes the validation set error of models M1 and M2; use the optimal hyperparameter combination to retrain on the entire training set to obtain the final baseline model M1 and baseline model M2.

[0020] Furthermore, the trained baseline models M1 and M2 are transformed into dynamic baseline models that respond to real-time operating conditions; the real-time electronic expansion valve opening is input into the dynamic baseline model, and real-time performance indicators and degradation are calculated collaboratively to obtain the real-time score of the refrigeration unit. The process includes:

[0021] The trained baseline models M1 and M2 are packaged into a callable model file package and deployed to the real-time monitoring device of the chiller unit; when the device starts up, all parameters of the baseline models M1 and M2 are automatically loaded.

[0022] Real-time acquisition of current operating condition data of the unit; alignment of real-time operating condition data with the common input feature set of benchmark model M1 and benchmark model M2;

[0023] If there is a difference between the real-time operating conditions and the operating conditions range of the training data for the baseline models M1 and M2, a dynamic calibration mechanism is initiated; the model obtained after calibration becomes the dynamic baseline model. With dynamic benchmark model ;

[0024] Input real-time electronic expansion valve opening and aligned real-time operating condition features to the dynamic baseline model Dynamic benchmark model , by dynamic benchmark model Output the standard cooling capacity under the current operating conditions. , by dynamic benchmark model Output the standard energy efficiency ratio under the current operating conditions. ;

[0025] Real-time cooling capacity is collected synchronously through the unit's sensors. Real-time energy efficiency ratio Based on the calculated standard cooling capacity and standard energy efficiency ratio as benchmarks, the real-time degradation rate of the cooling capacity of the refrigeration unit is calculated respectively. Real-time degradation rate of energy efficiency ratio ;

[0026] The real-time score of the refrigeration unit is calculated by combining the real-time degradation rate of its cooling capacity and the real-time degradation rate of its energy efficiency ratio. .

[0027] Furthermore, the process of determining the status of the refrigeration unit based on its real-time score and the configuration threshold includes:

[0028] If the real-time score of the refrigeration unit If so, the refrigeration unit is in a healthy state; if If so, the refrigeration unit is in a sub-healthy state; if If so, the refrigeration unit is in a fault state; if If so, the refrigeration unit is in a state of serious malfunction; if If this occurs, the refrigeration unit is in a state of complete failure; among which, All of these are judgment thresholds.

[0029] Furthermore, based on the evaluation results, a dynamic adjustment strategy is established: if the refrigeration unit is detected to be in a faulty, severely faulty, or completely failed state, the process of dynamically adjusting the update cycle of the dynamic baseline model and the sensor sampling frequency includes:

[0030] Get the real-time score of the refrigeration unit Real-time degradation rate of cooling capacity Real-time degradation rate of energy efficiency ratio Construct multi-dimensional state quantification indicators:

[0031] Define performance degradation acceleration It reflects the rate of performance degradation and is used to determine whether degradation is accelerating.

[0032] If performance degradation accelerates This indicates accelerated performance degradation; if the performance degradation acceleration is... This indicates that the degradation rate has slowed down or stabilized;

[0033] Introducing fault impact factors Quantify the impact of the current fault on the function of the refrigeration unit;

[0034] Real-time score of integrated refrigeration unit Accelerated performance degradation Fault influencing factors This quantifies the urgency of maintenance tasks, thereby obtaining a maintenance priority index. ;

[0035] Real-time score based on refrigeration unit With performance degradation acceleration For dynamic benchmark models Dynamic benchmark model The update cycle is adaptively adjusted:

[0036] In the formula, For update cycle; The baseline update cycle; The adjustment coefficient for the real-time score of the refrigeration unit; This is the degradation acceleration adjustment coefficient;

[0037] Based on fault impact factors With maintenance priority index The sensor sampling frequency is dynamically adjusted.

[0038] In the formula, The adjusted sampling frequency; The reference sampling frequency; The adjustment coefficient is the factor for the impact of faults. This is the priority adjustment coefficient.

[0039] A dynamic health assessment system for refrigeration units, comprising:

[0040] Benchmark model construction module: Acquire the electronic expansion valve opening data stream, and use the gradient boosting regression tree algorithm to establish benchmark model M1 for predicting cooling capacity and benchmark model M2 for energy efficiency ratio, respectively;

[0041] Dynamic benchmark conversion module: Converts the trained benchmark model M1 and benchmark model M2 into a dynamic benchmark model that responds to real-time operating conditions; inputs the real-time electronic expansion valve opening degree into the dynamic benchmark model, and performs joint calculation of real-time performance indicators and degradation degree to obtain the real-time score of the refrigeration unit; among which, the real-time performance indicators are standard cooling capacity and standard energy efficiency ratio.

[0042] Health status determination module: Based on the real-time score of the refrigeration unit and the configuration determination threshold, the status of the refrigeration unit is determined, including healthy, sub-healthy, faulty, serious faulty or completely failed status.

[0043] Dynamic adjustment strategy module: Based on the evaluation results, a dynamic adjustment strategy is established: if the refrigeration unit is detected to be in a fault, serious fault or complete failure state, the update cycle of the dynamic benchmark model and the sensor sampling frequency are dynamically adjusted.

[0044] Compared with existing technologies, the advantages of this invention, which provides a method and system for dynamically assessing the health status of refrigeration units, are as follows:

[0045] The application can accurately capture the relationship between the electronic expansion valve opening degree and the refrigeration performance by constructing the benchmark model M1 and the benchmark model M2 of the refrigerating capacity and the energy efficiency ratio through the gradient boosting regression tree algorithm; the model is trained by using the historical data, the accuracy and reliability of the prediction are effectively improved, data support is provided for subsequent dynamic evaluation; the real-time evaluation of the refrigerating unit performance is realized by converting the static benchmark model into the dynamic benchmark model responding to the real-time working condition, the model parameters can be adjusted according to the actual operation condition, the accuracy and timeliness of the evaluation result are ensured, and potential problems can be found in time; the state of the refrigerating unit can be accurately judged according to the real-time score of the refrigerating unit and the judgment threshold, the healthy, sub-healthy, fault and other states can be distinguished, maintenance guidance is provided for the operation and maintenance personnel, and the non-planned downtime is reduced; when the refrigerating unit is in the abnormal state such as fault, the model update cycle and the sensor sampling frequency can be dynamically adjusted, the data acquisition and processing efficiency can be optimized, and more accurate data can be obtained at the critical moment. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 A flowchart of a refrigerating unit health condition dynamic evaluation method is provided for the application.

[0047] Figure 2 A module diagram of a refrigerating unit health condition dynamic evaluation system is provided for the application. DETAILED DESCRIPTION

[0048] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the application.

[0049] Please refer to Figure 1 The application provides a refrigerating unit health condition dynamic evaluation method, which comprises the following steps:

[0050] S1, acquiring an electronic expansion valve opening degree data stream, and respectively establishing a benchmark model M1 for predicting refrigerating capacity and a benchmark model M2 for predicting energy efficiency ratio by using a gradient boosting regression tree algorithm;

[0051] S2, converting the trained benchmark model M1 and the benchmark model M2 into dynamic benchmark models responding to real-time working conditions; inputting real-time electronic expansion valve opening degrees into the dynamic benchmark models, and performing real-time performance index and degradation degree collaborative calculation to obtain real-time scores of the refrigerating unit; wherein the real-time performance index is a standard refrigerating capacity and a standard energy efficiency ratio;

[0052] S3, judging the refrigerating unit state including healthy, sub-healthy, failure, serious failure or complete failure state according to the real-time score of the refrigerating unit and comparing with the configuration decision threshold;

[0053] S4, establishing a dynamic adjustment strategy according to the evaluation result: if it is detected that the refrigerating unit is in a failure, serious failure or complete failure state, the update period of the dynamic baseline model and the sensor sampling frequency are dynamically adjusted; wherein the dynamic adjustment strategy can adjust the operation parameters of the model and the sensor in time according to the change of the unit health condition, so as to improve the response speed and accuracy.

[0054] It needs to be further explained that in the specific implementation process, the process of obtaining the electronic expansion valve opening degree data stream and establishing the baseline model M1 for predicting the refrigerating capacity and the baseline model M2 for predicting the energy efficiency ratio by using the gradient boosting regression tree algorithm includes:

[0055] S11, collecting historical operation data sets: electronic expansion valve opening degree , refrigerating capacity Q and energy efficiency ratio EER from the historical operation database of the water chiller or refrigerating system; cleaning the collected historical operation data sets, processing missing values and abnormal values (checking whether there are missing electronic expansion valve opening degree , refrigerating capacity Q and energy efficiency ratio EER records in the data set, directly deleting missing samples or filling them by using interpolation method; identifying and eliminating obviously unreasonable data (such as negative electronic expansion valve opening degree , extremely high or extremely low energy efficiency ratio EER) caused by sensor failure or operation error by using box plot or 3σ rule method; randomly shuffling the cleaned data set, and dividing it into training set and test set according to the preset proportion (such as 8:2), wherein the training set is used to train the model parameters, and the test set is used to finally evaluate the generalization performance of the model;

[0056] S12, normalizing the input features mainly with electronic expansion valve opening degree to optimize the convergence speed and performance of the gradient boosting regression tree; at the same time, constructing derived features closely related to the physical relationship of refrigerating capacity Q and energy efficiency ratio EER to form input feature sets suitable for models M1 and M2; wherein the introduction of derived features enhances the model's ability to capture complex physical processes and improves the accuracy of prediction;

[0057] Specifically, in the refrigerating system, the heat exchange amount on the evaporator side is regarded as the chiller refrigerating capacity Q; by measuring the inlet and outlet water temperature and flow of the evaporator, the calculation formula of the chiller refrigerating capacity Q is obtained as follows:

[0058] , wherein, is the specific heat capacity of water at constant pressure, is the mass flow of water on the evaporator e side, Tin as the inlet water temperature of the evaporator e, Tout as the outlet water temperature of the evaporator e; and further constructing a first derived feature, i.e. the refrigeration contribution per unit mass flow rate This feature reflects the refrigeration effect per unit mass of water flowing through the evaporator, and the larger the value, the stronger the refrigeration capacity under the same flow rate;

[0059] It can be understood that, in the feature construction process, in addition to the refrigeration contribution per unit mass flow rate, there are other derived features, such as the heat exchange efficiency on the evaporator side (calculated based on the temperature difference and flow rate of the inlet and outlet water of the evaporator), the rate of change of the electronic expansion valve opening (calculated based on the opening difference at the previous and subsequent time points), etc., to more comprehensively reflect the operating state of the refrigeration system;

[0060] S13, using the pre-processed electronic expansion valve opening and the derived features as the common input feature set, training model M1 (predicting refrigeration capacity Q) and model M2 (predicting energy efficiency ratio EER) in the same gradient boosting regression tree training framework by replacing the output target, the process being: selecting a gradient boosting regression tree regressor, presetting the total number of iterations (T) rounds), and initializing the initial prediction values of the two models (the initial prediction value of model M1 is set to the average of the refrigeration capacity Q in the training set, and the initial prediction value of model M2 is set to the average of the energy efficiency ratio EER in the training set); it can be understood that the gradient boosting regression tree algorithm gradually reduces the prediction error by iteratively adding new regression trees, thereby improving the fitting ability of the model; training round by round, a total of T rounds, updating model M1 and model M2 synchronously round by round; wherein the iterative training process includes calculating the negative gradient (residual error), training a new regression tree, calculating the leaf node output value, and updating the overall prediction function:

[0061] A1, for each sample i in the training set, calculating the negative gradient between the prediction value of the current model (consisting of the first m-1 trees) and the true value:

[0062] In the formula, G is the negative gradient, i is the sample index, and m is the tree or iteration index; is the loss function; represents the negative gradient of the mth iteration or the mth tree, for example, the 5th tree constructed by the algorithm at present, then represents the negative gradient (residual error) calculated for the i-th training sample when constructing the m-th tree; represents the true value of the i-th sample, in model M1, is the true refrigeration capacity of the i-th sample , in model M2, ​The true energy efficiency ratio of the ith sample ; represents the prediction value of the current ensemble model for the ith sample, is a quantity that changes with iterations, when building the mth tree, refers to the prediction value given by the ensemble model composed of the first m-1 trees, that is ; represents the feature input of a sample i;

[0063] A2, using the current entire training set, with the electronic expansion valve opening as input, the calculated negative gradient as output target, train a new regression tree , divide the data into J leaf nodes through the new regression tree ; wherein, J represents that the mth tree is divided into J leaf nodes, that is, the number of leaf nodes, J is an integer, and its size is controlled by hyperparameters such as the maximum depth of the tree; R is the region, which divides the input space into multiple non-overlapping regions; represents the jth leaf node region in the mth tree (j is the leaf node index, and j = 1, 2, …, J), for example, represents the 3rd leaf node on the 5th tree;

[0064] A3, calculate an output value for each leaf node j of the new tree , wherein, represents the output value of the jth leaf node of the mth tree, which is obtained by linear search to minimize the loss function, that is, the average value of the negative gradient of all samples divided into the leaf node ;

[0065] A4, add the new tree to the ensemble model and update the overall prediction function: , wherein, is the learning rate, which is used to control the contribution of each tree, and the two models are kept unified; is an indicator function, which outputs 1 when belongs to the leaf node region , otherwise outputs 0, which is used to locate the split position in the tree; when updating the overall prediction function, determine the adjustment strategy of the learning rate, for example, dynamically adjust the learning rate according to the performance on the validation set, to avoid overfitting or underfitting; through cross-validation, optimize the hyperparameters shared by the model M1 and the model M2 on the training set, to find the optimal combination of hyperparameters that minimize the validation set error of the model M1 and the model M2; wherein the hyperparameters include the number of trees (or the number of iterations), the learning rate (reducing the contribution of each tree), the maximum depth of each decision tree (controlling the model complexity), the subsampling ratio (used to introduce randomness and prevent overfitting), and the minimum number of samples required for a leaf node; using the optimal combination of hyperparameters, reiterate the training on the entire training set to obtain the final baseline model M1 (outputting the refrigeration capacity Q) and the baseline model M2 (outputting the energy efficiency ratio EER);

[0066] S14, using the split test set to evaluate the performance of the baseline model M1 and the baseline model M2 respectively, and recording the mean square error, the mean absolute error and the determination coefficient as the baseline for subsequent model comparison.

[0067] It needs to be further explained that, in the specific implementation process, the trained baseline model M1 and the baseline model M2 are converted into dynamic baseline models responding to real-time working conditions; the process of inputting real-time electronic expansion valve opening to the dynamic baseline model and performing real-time performance index and degradation degree collaborative calculation to obtain the real-time score of the refrigeration unit includes:

[0068] The trained baseline model M1 and the baseline model M2 (including the optimal hyperparameters, tree structure, leaf node output value and other core parameters) are packaged as a callable model file package and deployed to the real-time monitoring device of the water chiller; when the device starts, it automatically loads all the parameters of the baseline model M1 and the baseline model M2, ensuring that the initial state of the model is consistent with the trained baseline state, providing a basic framework for dynamic calculation;

[0069] Real-time acquisition of current operating condition data of the unit (including real-time electronic expansion valve opening, real-time environmental temperature, chilled water inlet temperature, cooling water inlet temperature and other key working condition variables); align the real-time working condition data with the common input feature set of the baseline model M1 and the baseline model M2 (for example, calculate the opening rate based on the real-time electronic expansion valve opening and the average opening of the previous 10s to obtain the real-time value of the energy efficiency response coefficient), to ensure that the feature dimension and format of the input baseline model M1 and the baseline model M2 are consistent with those during training, and to avoid prediction bias caused by feature mismatch;

[0070] If there is a difference between the real-time operating condition and the working condition range of the training data of the baseline model M1 and the baseline model M2 (for example, the environmental temperature exceeds the ±10% range of the training data, the difference is large), then start the dynamic calibration mechanism: use the real-time electronic expansion valve opening in the last 1 hour , effective data of real-time refrigeration capacity Q and real-time energy efficiency ratio EER (need to meet data quantity ≥ 50 groups, no abnormal value), incremental training is performed on the loaded baseline model M1 and baseline model M2 (fixing the original hyperparameters, only updating 3-5 regression trees of the latest iteration), the prediction deviation of the model to the extreme working condition is corrected; the model obtained after calibration is the dynamic baseline model (real-time refrigeration capacity Q prediction model) and dynamic baseline model (real-time energy efficiency ratio EER prediction model), the parameters of which are dynamically adjusted with real-time working conditions, ensuring that the output standard value fits the current unit actual running state; in the incremental training process, the selection strategy of the newly added regression tree is determined, for example, the optimal number of regression trees is selected based on the validation set performance, and the contribution degree between the newly added tree and the original tree is balanced;

[0071] input real-time electronic expansion valve opening (and aligned real-time working condition characteristics) to dynamic baseline model , dynamic baseline model , dynamic baseline model output standard refrigeration capacity under current working condition , dynamic baseline model output standard energy efficiency ratio under current working condition ;

[0072] synchronously collect real-time refrigeration capacity (for example, calculated by refrigerated water flow and temperature difference), real-time energy efficiency ratio (for example, calculated by refrigeration capacity and input power), wherein , ; based on the calculated standard refrigeration capacity and standard energy efficiency ratio as the baseline, the real-time degradation rate of the refrigeration capacity of the refrigeration unit and the real-time degradation rate of the energy efficiency ratio are calculated: , ;

[0073] comprehensive real-time degradation rate of refrigeration capacity of refrigeration unit and real-time degradation rate of energy efficiency ratio, the real-time score of the refrigeration unit is calculated :

[0074] .

[0075] It needs to be further explained that, in the specific implementation process, according to the real-time score of the refrigeration unit, the process of judging the state of the refrigeration unit by comparing with the configuration judgment threshold includes:

[0076] if the real-time score of the refrigeration unit , the refrigeration unit is in a healthy state, normally running and excellent in performance; if If the refrigeration unit is in a sub-healthy state, it is operating well and performing normally; if If the refrigeration unit is in a fault state, it will exhibit abnormal behavior but can still operate, although its performance will be significantly reduced; if If the refrigeration unit is in a state of serious malfunction, it will be difficult to operate and its performance will be severely degraded; if If this occurs, the refrigeration unit is in a state of complete failure and is essentially unable to operate; among them, All are judgment thresholds, and .

[0077] It should be further explained that, in the specific implementation process, a dynamic adjustment strategy is established based on the evaluation results: if the refrigeration unit is detected to be in a faulty, severely faulty, or completely failed state, the process of dynamically adjusting the update cycle of the dynamic benchmark model and the sensor sampling frequency includes:

[0078] Get the real-time score of the refrigeration unit Real-time degradation rate of cooling capacity Real-time degradation rate of energy efficiency ratio Construct multi-dimensional state quantification indicators:

[0079] Define performance degradation acceleration It reflects the rate of performance degradation and is used to determine whether degradation is accelerating. In the formula, This represents the change in the average degradation rate. , Let be the average degradation rate at time t. (Cooling capacity and energy efficiency ratio are weighted equally, and adjusted according to the unit's application, such as in scenarios where cooling is prioritized.) ); For time intervals;

[0080] If performance degradation accelerates This indicates accelerated performance degradation; if the performance degradation acceleration is... This indicates that the degradation rate has slowed down or stabilized;

[0081] Introducing fault impact factors Quantify the impact of the current fault on the core functions of the unit: In the formula, The number of currently detected fault types (e.g., both condenser blockage and abnormal fan speed exist simultaneously). ); Index of detected fault types; For the first Weighting coefficients for different types of faults (preset based on the degree of fault impact); For the first State coefficient of the class fault (when confirming the fault , when suspecting the fault , when no fault ); is the degradation rate contribution value corresponding to the first class fault; wherein the value range of the fault influence factor is [0, 1], and the greater the value, the more serious the fault influence;

[0082] The real-time score of the comprehensive refrigeration unit , the performance degradation acceleration , the fault influence factor quantify the urgency of the maintenance task:

[0083] , wherein, is the maintenance priority index, and the value range is [0, 1]; the preset priority threshold is that is high priority, is medium priority, is low priority; all are preset weight coefficients; is the real-time score of the refrigeration unit Reverse indicator, the greater the value, the higher the priority; is the positive degradation acceleration, which only promotes the priority when accelerating the degradation;

[0084] Based on the real-time score of the refrigeration unit and the performance degradation acceleration , the update period of the dynamic benchmark model , the dynamic benchmark model is adaptively adjusted:

[0085] , wherein, is the update period; is the benchmark update period (default 24h, initial period under healthy state); is the adjustment coefficient of the real-time score of the refrigeration unit; is the degradation acceleration adjustment coefficient;

[0086] Based on the fault influence factor and the maintenance priority index , the sensor sampling frequency is dynamically adjusted:

[0087] , wherein, is the adjusted sampling frequency; is the benchmark sampling frequency (default setting is 1 / minute, sampling frequency under healthy state); is the fault influence adjustment coefficient; Priority adjustment coefficient.

[0088] Please refer to Figure 2 The application provides a refrigeration unit health condition dynamic evaluation system, comprising:

[0089] A reference model construction module: acquires an electronic expansion valve opening degree data stream, and uses a gradient boosting regression tree algorithm to respectively establish a reference model M1 for predicting refrigeration capacity and a reference model M2 for predicting energy efficiency ratio;

[0090] A dynamic reference conversion module: converts the trained reference model M1 and the reference model M2 into a dynamic reference model that responds to real-time working conditions; inputs real-time electronic expansion valve opening degree into the dynamic reference model, and performs real-time performance index and degradation degree collaborative calculation, thereby obtaining real-time scores of the refrigeration unit; wherein the real-time performance index is standard refrigeration capacity and standard energy efficiency ratio;

[0091] A health condition determination module: according to the real-time scores of the refrigeration unit, compares with the configured determination threshold, determines the state of the refrigeration unit, including healthy, sub-healthy, failure, serious failure or complete failure state;

[0092] A dynamic adjustment strategy module: according to the evaluation result, establishes a dynamic adjustment strategy: if it is detected that the refrigeration unit is in a failure, serious failure or complete failure state, the update period of the dynamic reference model and the sensor sampling frequency are dynamically adjusted; wherein the dynamic adjustment strategy can adjust the operation parameters of the model and the sensor in time according to the change of the unit health condition, improve the response speed and accuracy.

[0093] In addition, the formulas involved in the above are calculated by removing the dimension and taking the numerical value, which is obtained by collecting a large amount of data to simulate the closest real situation by software. The proportion coefficient in the formula and each preset threshold in the analysis process are set by the person skilled in the art according to the actual situation or obtained by a large amount of data simulation. The size of the proportion coefficient is to obtain a specific numerical value of each parameter for subsequent comparison. The size of the proportion coefficient depends on the number of sample data and the corresponding processing coefficient preliminarily set by the person skilled in the art for each group of sample data. As long as it does not affect the proportional relationship between the parameters and the quantized numerical value.

[0094] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment mainly describes the difference from other embodiments. Especially for the device embodiment, since it basically relies on the method embodiment, the description is relatively simple, and the related parts can be referred to the part of the method embodiment.

[0095] For convenience sake, the above apparatus is described in various units by function for separate description. Of course, the functions of the units can be implemented in one or more software and / or hardware in implementing the present application.

[0096] Those skilled in the art should understand that embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer usable program code.

[0097] The present application is described with reference to the flowcharts and / or block diagrams of the methods, apparatus (systems) and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and a combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus generate a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that implements the functions specified in the flowcharts and / or block diagrams.

[0098] These computer program instructions can also be stored in a computer-readable memory that can direct the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction apparatus, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that implements the functions specified in the flowcharts and / or block diagrams.

[0099] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that implements the functions specified in the flowcharts and / or block diagrams.

[0100] Secondly, the present application discloses embodiments in the drawings, only involves the structure related to the embodiments of the present disclosure, other structures can refer to the general design, in the case of no conflict, the same embodiments and different embodiments of the present application can be combined with each other;

[0101] Finally: the above is only the preferred embodiment of the present application, but the scope of protection of the present application is not limited to this, any skilled in the technical field within the scope of the present application disclosed by the technical range, according to the technical solution of the present application and its inventive concept to make equivalent replacement or change, should be covered within the scope of protection of the present application.

Claims

1. A method for dynamically evaluating the health status of a refrigeration unit, characterized in that: S1, obtaining an electronic expansion valve opening degree data stream, and using a gradient boosting regression tree algorithm to respectively establish a baseline model M1 for predicting refrigeration capacity and a baseline model M2 for predicting energy efficiency ratio; S2, converting the trained baseline models M1 and M2 into dynamic baseline models responsive to real-time operating conditions; inputting real-time electronic expansion valve opening degree into the dynamic baseline models, and performing real-time performance index and degradation degree collaborative calculation to obtain real-time score of the refrigeration unit; wherein the real-time performance index is standard refrigeration capacity and standard energy efficiency ratio; the process of converting the trained baseline models M1 and M2 into dynamic baseline models responsive to real-time operating conditions; inputting real-time electronic expansion valve opening degree into the dynamic baseline models, and performing real-time performance index and degradation degree collaborative calculation to obtain real-time score of the refrigeration unit comprises: encapsulating the trained baseline models M1 and M2 into callable model file packages, and deploying them to a real-time monitoring device of the water chiller; the device automatically loads all parameters of the baseline models M1 and M2 when starting; real-time acquisition of current operating condition data of the unit; alignment of real-time operating condition data with the common input feature set of the baseline models M1 and M2; S3, according to the real-time score of the refrigeration unit, comparing with the configured determination threshold to judge the state of the refrigeration unit, including healthy, sub-healthy, failure, serious failure or complete failure state; S4, according to the evaluation result, establishing a dynamic adjustment strategy: if it is detected that the refrigeration unit is in a failure, serious failure or complete failure state, dynamically adjusting the update period of the dynamic baseline model and the sensor sampling frequency. The process of obtaining an electronic expansion valve opening degree data stream, and using a gradient boosting regression tree algorithm to respectively establish a baseline model M1 for predicting refrigeration capacity and a baseline model M2 for predicting energy efficiency ratio comprises: S14, using the divided test set to respectively evaluate the performance of the baseline models M1 and M2. The process of model iterative training comprises: A4, adding a new tree to the integrated model, updating the overall prediction function; optimizing the hyperparameters commonly used by the models M1 and M2 on the training set through cross-validation to find the optimal hyperparameter combination that minimizes the validation set error of the models M1 and M2; re-iterating training on the entire training set using the optimal hyperparameter combination to obtain the final baseline models M1 and M2. The process of comparing the real-time score of the refrigeration unit with the configured determination threshold to judge the state of the refrigeration unit comprises: S31, if the real-time score of the refrigeration unit is greater than the configured determination threshold, determining that the refrigeration unit is in a healthy state; if the real-time score of the refrigeration unit is less than the configured determination threshold, determining that the refrigeration unit is in a failure state. The process of establishing a dynamic adjustment strategy according to the evaluation result: if it is detected that the refrigeration unit is in a failure, serious failure or complete failure state, dynamically adjusting the update period of the dynamic baseline model and the sensor sampling frequency comprises: S41, if the refrigeration unit is in a failure state, increasing the update period of the dynamic baseline model and the sensor sampling frequency; if the refrigeration unit is in a serious failure or complete failure state, decreasing the update period of the dynamic baseline model and the sensor sampling frequency. A system for dynamically evaluating the health status of a refrigeration unit according to any one of claims 1-5, comprising: a baseline model construction module: obtaining an electronic expansion valve opening degree data stream, and using a gradient boosting regression tree algorithm to respectively establish a baseline model M1 for predicting refrigeration capacity and a baseline model M2 for predicting energy efficiency ratio; If there is a difference between the real-time running condition and the condition range of the training data of the benchmark model M1 and the benchmark model M2, a dynamic calibration mechanism is started; the model obtained after calibration is the dynamic benchmark model and the dynamic benchmark model ​ input real-time electronic expansion valve opening and aligning the real-time operating condition characteristics to the dynamic reference model , the dynamic reference model , the dynamic reference model output standard refrigeration capacity under current operating condition , the dynamic reference model output standard energy efficiency ratio under current operating condition ; Synchronize through unit sensor collection real-time refrigeration capacity , real-time energy efficiency ratio ; based on the calculated standard refrigeration capacity and standard energy efficiency ratio as a benchmark, respectively calculate the real-time degradation rate of refrigeration capacity of refrigeration unit and real-time degradation rate of energy efficiency ratio ; The real-time degradation rate of the refrigerating capacity of the comprehensive refrigerating unit and the real-time degradation rate of the energy efficiency ratio are comprehensively considered to calculate the real-time score of the refrigerating unit ; ​ ​ 2. The method of claim 1, wherein: ​ S11、from the historical operation database of the water chiller or refrigeration system, collect a historical operation dataset: electronic expansion valve opening degree , refrigerating capacity Q and energy efficiency ratio EER; clean the collected historical operation dataset, process missing values and abnormal values; randomly shuffle the cleaned dataset, and divide it into a training set and a test set according to a preset proportion; S12, the electronic expansion valve opening degree normalization processing is performed on the input features mainly composed of the refrigerating capacity Q and the energy efficiency ratio EER, and derivative features closely related to the physical relationship between the refrigerating capacity Q and the energy efficiency ratio EER are constructed. S13, with the pre-processed electronic expansion valve opening degree and the derived features are common input feature sets, and the model M1 and the model M2 are trained in the same gradient boosting regression tree training framework by replacing the output target, and the process is as follows: selecting a gradient boosting regression tree regressor, presetting the total number of iterations, and initializing the initial prediction values of the two models; training by iteration round by round, a total of synchronously updating the model M1 and the model M2 in each round; wherein the iteration training process includes calculating the negative gradient, training a new regression tree, calculating the leaf node output value, and updating the overall prediction function; ​ 3. The method of claim 2, wherein: ​ A1. For each sample i in the training set, compute the negative gradient between the predicted value of the current model and the true value ; A2, using the current entire training set, to the electronic expansion valve opening degree As input, the calculated negative gradient As output target, train a new regression tree, and divide the data into J leaf nodes through the new regression tree ; wherein, J represents that the mth tree is divided into J leaf nodes, that is, the number of leaf nodes; R is a region, which divides the input space into multiple non-overlapping regions; represents the jth leaf node region in the mth tree. A3. Compute an output value for each leaf node j of the new tree wherein, represents the output value of the jth leaf node of the mth tree. ​ 4. The method of claim 1, wherein: ​ If the real-time score of the refrigerating unit is , the refrigerating unit is in a healthy state; if , the refrigerating unit is in a sub-healthy state; if , the refrigerating unit is in a fault state; if , the refrigerating unit is in a serious fault state; if , the refrigerating unit is in a complete failure state; wherein, are all determination thresholds.

5. The method of claim 1, wherein: ​ Obtaining real-time scores of refrigerating units Real-time degradation rates of refrigerating capacity Real-time degradation rates of energy efficiency ratios Constructing multi-dimensional state quantification indexes: Definition of performance degradation acceleration , reflecting the rate of change of the trend of performance degradation, for determining whether the degradation is accelerated; if the performance degradation acceleration then the performance degradation acceleration is increasing; if the performance degradation acceleration then the degradation rate is slowing or stable; Introducing a failure impact factor quantifies the current failure's impact on the chiller unit's functionality; Real-time score of a comprehensive refrigeration unit , performance degradation acceleration , failure influence factor , quantifying the urgency of maintenance tasks, thereby obtaining a maintenance priority index ; Real-time score based on refrigeration unit With performance degradation acceleration , dynamic benchmark model , dynamic benchmark model Adaptive adjustment of update period ， In the formula, is an update period; is a reference update period; is an adjustment coefficient of the real-time score of the refrigeration unit; is a degradation acceleration adjustment coefficient; Based on failure impact factor With maintenance priority index Dynamic adjustment of sensor sampling frequency: wherein is the adjusted sampling frequency; is the reference sampling frequency; is the fault impact adjustment coefficient; is the priority adjustment coefficient; is low priority.

6. A refrigeration unit health dynamic assessment system, comprising: ​ ​ The dynamic benchmark conversion module converts the trained benchmark model M1 and the benchmark model M2 into a dynamic benchmark model responding to real-time working conditions; inputs the real-time electronic expansion valve opening degree to the dynamic benchmark model, and performs real-time performance index and degradation degree collaborative calculation, thereby obtaining the real-time score of the refrigeration unit; wherein the real-time performance index is the standard refrigerating capacity and the standard energy efficiency ratio; The health state judgment module judges the state of the refrigeration unit according to the real-time score of the refrigeration unit, and compares the configuration judgment threshold to judge the state of the refrigeration unit, including the healthy, sub-healthy, fault, serious fault or complete failure state; The dynamic adjustment strategy module establishes a dynamic adjustment strategy according to the evaluation result: if it is detected that the refrigeration unit is in the fault, serious fault or complete failure state, the update period of the dynamic benchmark model and the sensor sampling frequency are dynamically adjusted.

Images