A method and system for evaluating the health status of a variable-operating-condition chiller

Abstract

The present invention discloses a health status assessment method and system for a variable operating condition chiller. The method comprises: obtaining operating data of the chiller and preprocessing the operating data, and obtaining steady-state data in the operating data; extracting and eliminating variable operating condition features of the variable operating condition data of the steady-state data to obtain operating condition-independent data in the operating data; extracting slow return performance feature data from the operating condition-independent data, the slow degradation performance feature data being used to reflect the performance degradation of the chiller; and obtaining a chiller health status assessment result based on the slow degradation performance feature data. The present invention extracts and eliminates feature information related to operating condition changes of the chiller, then extracts features related to equipment degradation performance, constructs a performance health status indicator, and realizes monitoring of health status change trends and health status assessment of the chiller under variable operating conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of fault prediction, and in particular to a health status assessment method and system for a variable-operating-condition chiller. Background Art

[0002] Chillers use a chiller mechanism to cool water, providing constant temperature, flow, and pressure. Chillers are widely used in industries such as plastics, electronics, machinery, food, and pharmaceuticals, as well as in central air conditioning systems in large buildings. Chillers are major energy consumers in public buildings and industrial production. As the core of central air conditioning in public buildings, chillers account for over 30% of the building's total energy consumption. Chillers typically consist of four components: an evaporator, a condenser, a compressor, and an expansion valve. They operate through the interaction of a refrigerant circulation system, a water circulation system, and an automatic control system, achieving cooling and supplying heat. This is a typical electromechanical-hydraulic coupled system. Due to slow performance degradation during operation, such as component wear, scaling of heat exchange tubes, and refrigerant leakage, system energy consumption gradually increases. Severe performance degradation can even lead to damage and failure. Accurately identifying chiller performance effectively monitors equipment health, enabling condition-based maintenance. This plays a crucial role in reducing equipment maintenance costs and enhancing the ability to ensure safe, stable, and efficient operation. Existing chiller health assessment methods can be categorized into three types: model-driven, data-driven, and knowledge-driven. Model-driven methods establish mathematical analytical models based on the equipment's failure mechanisms to assess the equipment's health. Data-driven methods use equipment operating data to construct nonlinear relationships between degradation characteristics and health status. Common data models include neural networks, hidden Markov models, support vector machines, and ensemble learning. Knowledge-driven methods build mappings between degradation characteristics and health status based on expert knowledge through reasoning and analysis. Model-driven methods place high demands on the integrity and accuracy of mathematical analytical models, while knowledge-driven methods require comprehensive, precise, and clear prior knowledge to guide health assessment. Complex equipment often faces challenges such as difficulty measuring key performance variables, lack of precise prior knowledge of physical degradation, and difficulty constructing models of equipment failure mechanisms. These factors make model-driven and knowledge-driven health assessments difficult to apply to complex equipment. Data-driven methods, relying solely on large amounts of historical monitoring data, have garnered increasing attention in recent years.

[0003] Condenser heat exchange tube fouling is a common problem in chiller operation. It reduces the heat exchange efficiency between cooling water and refrigerant, which in turn increases the energy consumption of the equipment. The key performance indicator that directly reflects the degree of condenser heat exchange tube fouling is the heat transfer coefficient of the heat exchange tube. However, this parameter is not measurable, making it difficult to use model-driven and knowledge-driven methods to evaluate the degree of condenser heat exchange tube fouling. At the same time, traditional data-driven methods require multiple complete lifecycle data for data modeling, and most methods are based on a single operating condition or a small number of limited operating conditions. Since the operating conditions of the chiller vary with the outdoor environment and cooling load demand, it is a typical complex variable operating condition system. The degree of influence of the equipment health status on the operating status of the chiller under different operating conditions is also different. How to extract the relevant characteristics of the chiller degradation performance in the complex variable operating process and evaluate the equipment health status remains a huge challenge.

[0004] Therefore, the existing technology needs to be improved. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for evaluating the health status of a variable-operating-condition chiller in response to the above-mentioned defects of the prior art, aiming to solve the problem in the prior art of difficulty in evaluating the health status of equipment during complex variable-operating-condition operation.

[0006] The technical solutions adopted by the present invention to solve the problem are as follows:

[0007] In a first aspect, an embodiment of the present invention provides a method for evaluating the health status of a variable-operating-condition chiller, wherein the method comprises:

[0008] Acquiring operating data of the chiller, preprocessing the operating data, and acquiring steady-state data from the operating data;

[0009] Extracting and eliminating the variable operating condition features of the steady-state data to obtain operating condition-independent data in the operating data;

[0010] Extracting slowly degraded performance characteristic data from the working condition-independent data, wherein the slowly degraded performance characteristic data is used to reflect the performance degradation of the chiller;

[0011] A chiller health status assessment result is obtained based on the slowly degrading performance characteristic data.

[0012] In one implementation, the obtaining of operating data of the chiller, preprocessing the operating data, and obtaining steady-state data from the operating data includes:

[0013] Acquire operating data of the chiller and determine a data type of the operating data, wherein the data type includes condition variable data and state variable data;

[0014] Abnormal data is eliminated from the operating data according to the data type of the operating data, and steady-state data in the operating data is obtained, where the steady-state data includes conditional variable steady-state data and state variable steady-state data.

[0015] In one implementation, removing abnormal data from the operating data according to the data type of the operating data to obtain steady-state data from the operating data includes:

[0016] A geometric weighted mean and variance of the operating data are calculated, and steady-state filtering is performed on the operating data to obtain steady-state data in the operating data.

[0017] In one implementation, extracting and eliminating the variable operating condition features of the steady-state data to obtain operating condition-independent data in the operating data includes:

[0018] Extracting the variable operating condition features of the steady-state data based on a deep variational information bottleneck method to obtain variable operating condition feature data of the steady-state data;

[0019] The variable operating condition characteristic data of the steady-state data is eliminated to obtain the operating condition-independent data in the steady-state data.

[0020] In one implementation, extracting the variable operating condition features of the steady-state data based on the deep variational information bottleneck method to obtain variable operating condition feature data of the steady-state data includes:

[0021] The conditional variable steady-state data in the steady-state data is input into a deep variational information bottleneck model, and the variable operating condition features in the conditional variable steady-state data are extracted to obtain the variable operating condition feature data of the steady-state data.

[0022] In one implementation, removing the variable operating condition characteristic data of the steady-state data to obtain the operating condition-independent data in the steady-state data includes:

[0023] A residual is made for the variable operating condition characteristic data of the steady-state data and the steady-state data of the state variables in the steady-state data, and the residual value is obtained as the operating condition-independent data in the steady-state data.

[0024] In one implementation, extracting the slowly degrading performance characteristic data from the operating condition-independent data includes:

[0025] The working condition-independent data is input into a deep variational information bottleneck encoder model to extract slowly degrading performance characteristic data from the working condition-independent data to obtain slowly degrading performance characteristic data.

[0026] In one implementation, obtaining a chiller health status assessment result based on the slowly degrading performance characteristic data includes:

[0027] Setting a monitoring time window, and determining a reference baseline value based on the slow degradation performance characteristic data within the monitoring time window;

[0028] The distance between the mean value of the slowly degrading performance characteristic data in the current time window and the reference benchmark value is calculated, and the health status assessment result of the chiller is obtained according to the distance, where the distance is a health status indicator reflecting the degradation performance of the chiller.

[0029] In one implementation, the method further includes:

[0030] If the health status indicator exceeds the set threshold, an alarm maintenance message will be issued to notify the operation and maintenance personnel to clean the equipment in time.

[0031] In a second aspect, an embodiment of the present invention further provides a variable operating condition chiller health status assessment system, wherein the system comprises:

[0032] A data acquisition module, configured to acquire operating data of the chiller, pre-process the operating data, and acquire steady-state data from the operating data;

[0033] A first feature extraction module is used to extract and eliminate the variable operating condition features of the steady-state data to obtain operating condition-irrelevant data in the operating data;

[0034] a second feature extraction module, configured to extract slowly degrading performance feature data from the operating condition-independent data, wherein the slowly degrading performance feature data is used to reflect the performance degradation of the chiller;

[0035] An evaluation result acquisition module is used to obtain a health status evaluation result of the chiller based on the slow degradation performance characteristic data.

[0036] In one implementation, the data acquisition module includes:

[0037] a type determination unit, configured to obtain operating data of the chiller and determine a data type of the operating data, wherein the data type includes conditional variable data and state variable data;

[0038] The data processing unit is used to eliminate abnormal data from the operating data according to the data type of the operating data, and obtain steady-state data from the operating data, wherein the steady-state data includes steady-state data of conditional variables and steady-state data of state variables.

[0039] In one implementation, the data processing unit includes:

[0040] The data filtering subunit is configured to calculate a geometric weighted mean and variance of the operating data, and perform steady-state filtering on the operating data to obtain steady-state data from the operating data.

[0041] In one implementation, the first feature extraction module includes:

[0042] a first data extraction unit, configured to extract the variable operating condition features of the steady-state data based on a deep variational information bottleneck method to obtain variable operating condition feature data of the steady-state data;

[0043] The data elimination unit is used to eliminate the variable working condition characteristic data of the steady-state data to obtain working condition irrelevant data in the steady-state data.

[0044] In one implementation, the first data extraction unit includes:

[0045] The first data extraction subunit is used to input the conditional variable steady-state data in the steady-state data into the deep variational information bottleneck model, extract the variable operating condition features in the conditional variable steady-state data, and obtain the variable operating condition feature data of the steady-state data.

[0046] In one implementation, the data elimination unit includes:

[0047] The data elimination subunit is used to make a residual of the variable working condition characteristic data of the steady-state data and the steady-state data of the state variables in the steady-state data, and obtain the residual value as the working condition-independent data in the steady-state data.

[0048] In one implementation, the second feature extraction module includes:

[0049] The second feature extraction unit is used to extract the slowly degrading performance characteristic data from the working condition-independent data based on the deep variational information bottleneck method to obtain the slowly degrading performance characteristic data.

[0050] In one implementation, the second feature extraction unit includes:

[0051] The second data extraction subunit is used to input the working condition-independent data into the encoder model of the deep variational information bottleneck to extract the slowly degrading performance characteristic data from the working condition-independent data to obtain the slowly degrading performance characteristic data.

[0052] In one implementation, the evaluation result acquisition module includes:

[0053] a standard determination unit, configured to set a monitoring time window and determine a reference benchmark value based on the slow degradation performance characteristic data within the monitoring time window;

[0054] The evaluation result acquisition unit is used to calculate the distance between the mean value of the slowly degrading performance characteristic data in the current time window and the reference benchmark value, and obtain the health status evaluation result of the chiller according to the distance, wherein the distance is a health status indicator reflecting the degradation performance of the chiller.

[0055] In one implementation, the system further includes:

[0056] The alarm prompt module is used to issue an alarm maintenance message if the health status indicator exceeds the set threshold, notifying the operation and maintenance personnel to clean the equipment in time.

[0057] In the third aspect, an embodiment of the present invention also provides a terminal device, wherein the terminal device is a commercial display terminal, and the terminal device includes a memory, a processor, and a program of a variable operating condition chilled water host health status assessment method stored in the memory and runnable on the processor. When the processor executes the program of the variable operating condition chilled water host health status assessment method, the steps of the variable operating condition chilled water host health status assessment method of any one of the above-mentioned schemes are implemented.

[0058] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein a program of a variable operating condition chilled water host health status assessment method is stored on the computer-readable storage medium. When the program of the variable operating condition chilled water host health status assessment method is executed by a processor, the steps of the variable operating condition chilled water host health status assessment method described in any one of the above-mentioned schemes are implemented.

[0059] The beneficial effects of the present invention are as follows: Compared with the prior art, the present invention provides a method and system for assessing the health status of a chiller under variable operating conditions. The method first obtains the chiller's operating data and preprocesses the operating data to obtain steady-state data from the operating data. Then, the variable operating condition features of the steady-state data are extracted and eliminated to obtain operating condition-independent data from the operating data. Slowly degrading performance feature data is then extracted from the operating condition-independent data to reflect the performance degradation of the chiller. Finally, a chiller health status assessment result is obtained based on the slowly degrading performance feature data. By extracting and eliminating characteristic information related to the chiller's operating condition changes, the present invention effectively avoids the influence of changes in external environmental factors and operating conditions on the extraction of performance degradation-related features. Features related to equipment degradation performance are further extracted to reduce the presence of information unrelated to degraded equipment performance. Based on these features, a performance health status indicator is constructed to enable online monitoring of equipment performance. By eliminating operating condition processing and extracting features related to equipment degradation performance, it is possible to identify changing trends in the health status of the chiller under long-term variable operating conditions and assess its health status. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0061] Figure 1 It is a flow chart of a method for evaluating the health status of a chiller under variable operating conditions provided by an embodiment of the present invention.

[0062] Figure 2 This is a framework diagram of a variable-operating-condition chiller health status assessment method provided by an embodiment of the present invention.

[0063] Figure 3 It is a directed graph structure of a Deep VIB model of a variable-operating-condition chiller health status assessment method provided by an embodiment of the present invention.

[0064] Figure 4 This is a flowchart of eliminating working condition related information of the variable working condition chiller health status assessment method provided by an embodiment of the present invention.

[0065] Figure 5 This is a diagram of the process of extracting degradation performance-related features based on Deep VIB in the variable-operating-condition chiller health status assessment method provided by an embodiment of the present invention.

[0066] Figure 6 This is a graph showing the changing trend of health status indicators of a chiller according to a method for evaluating the health status of a chiller under variable operating conditions provided by an embodiment of the present invention.

[0067] Figure 7 This is a principle block diagram of a variable operating condition chiller health status assessment system provided by an embodiment of the present invention.

[0068] Figure 8 This is a block diagram of the internal structure of a variable-condition chiller health status assessment device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical solutions and advantages of the present invention more clear and distinct, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0070] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indications will also change accordingly.

[0071] This embodiment provides a method for evaluating the health status of a chilled water main unit under variable operating conditions, through which online monitoring of the health status of the chilled water main unit can be achieved. In specific implementation, first, the operating data of the chilled water main unit is obtained and the operating data is preprocessed, and steady-state data in the operating data is obtained; then, the variable operating condition characteristics of the steady-state data are extracted and eliminated to obtain operating condition-independent data in the operating data; further, the slowly degraded performance characteristic data in the operating condition-independent data is extracted, and the slowly degraded performance characteristic data is used to reflect the performance degradation of the chilled water main unit; finally, the health status evaluation result of the chilled water main unit is obtained based on the slowly degraded performance characteristic data. The present invention extracts and eliminates characteristic information related to the operating condition changes of the chilled water main unit, and further extracts characteristics related to the degradation performance of the chilled water main unit to construct a performance health status index, thereby realizing monitoring of the changing trend of the health status of the chilled water main unit under variable operating conditions and health status evaluation.

[0072] For example, to monitor and evaluate the health status of a chiller, the chiller's operating data is obtained and preprocessed to obtain steady-state data. The variable operating condition features in the obtained steady-state data are then extracted and removed to obtain operating condition-independent data. The slowly degrading performance feature data in the obtained operating condition-independent data is then extracted. Finally, the chiller's health status assessment result is obtained based on the extracted slowly degrading performance feature data. The present invention removes operating condition-related information to extract features related to the chiller's degraded performance from the operating data, thereby identifying and evaluating the health change trend of the chiller's health status under variable operating conditions.

[0073] Exemplary Methods

[0074] This embodiment provides a method for evaluating the health status of a chiller under variable working conditions. The method can be applied to a terminal device for monitoring the health of a chiller. Figure 1 As shown, the method includes:

[0075] Step S100: Acquire operating data of a chiller, pre-process the operating data, and acquire steady-state data from the operating data.

[0076] In this embodiment, the operating data of the chiller is obtained and preprocessed to obtain steady-state data that meets the requirements. In a specific implementation, first, the operating data of the chiller is obtained and the data type of the operating data is determined, which includes conditional variable data and state variable data. Then, based on the data type of the operating data, abnormal data is removed from the operating data to obtain steady-state data from the operating data, which includes conditional variable steady-state data and state variable steady-state data. By classifying the operating data of the chiller and performing data processing, interference from abnormal data is avoided, and the steady-state data can more accurately reflect the operating performance of the chiller.

[0077] In one implementation, the chiller operating data is divided into conditional variable data and state variable data according to data type. The conditional variable data is generally an input variable or a set variable. The input variable is a variable that acts on the chiller from the outside and is not directly affected by the chiller's operating state, such as the chilled water inlet temperature, chilled water flow rate, etc. The set variable is a set value issued to the chiller by the user or the intelligent control system, such as the chilled water outlet temperature. The state variable is some measured variable during the operation of the chiller. The state variable is affected by the conditional variable on the one hand, and on the other hand, it is also affected by changes in the performance of the device itself, such as the evaporator approach temperature, evaporator pressure, condenser approach temperature, condenser pressure, chiller power, cooling water outlet temperature, etc. The classification of the chiller measurement variables is shown in Table 1.

[0078] Table 1. Classification of chiller measurement variables

[0079]

[0080] In one implementation, steady-state data from the operating data is obtained. The geometric weighted mean and variance of the operating data are calculated, and the operating data is subjected to steady-state filtering to obtain the steady-state data from the operating data. Due to the influence of external factors such as communication equipment and sensors, the collected operating data of the chiller may have problems such as missing data and outliers. First, it is necessary to perform abnormal data preprocessing on the operating data to eliminate abnormal data that does not meet the requirements to ensure the accuracy and effectiveness of the data so that accurate results can be obtained later. The operating state of the chiller can be divided into transient and steady-state. The transient state is mainly a transition stage in which the chiller switches from one stable stage to another during the startup, shutdown, or host load rate adjustment phase of the chiller. The steady state is a balance relationship maintained between the operating states of the chiller, and the change amplitude of each variable is small at this time. The use of steady-state data can more accurately reflect the operating performance of the chiller, so it is necessary to perform steady-state judgment and screening on the operating data.

[0081] In one implementation, the steady-state determination method is based on the fluctuation of key parameters that characterize the operating state of the unit. The steady-state filtering uses the calculation of geometric weighted mean and variance. For example, for a given set of steady-state data {x1, x2, ..., x n} Perform steady-state filtering and calculate the geometric weighted average of the steady-state data. The geometric weighted average calculation formula is as shown in formula (1):

[0082]

[0083] Calculate the variance of the steady-state data. The calculation formula of the variance is as follows:

[0084]

[0085]

[0086] Where, is the geometric weight value, n is the number of steady-state data, β is the geometric weight factor with a value between [0,1], x k is the kth steady-state data, τ ss is the finite time window length, Δt is the sampling interval, S n () is the mean square error of the steady-state data.

[0087] In one implementation, when the geometrically weighted mean square error S of the running data n () exceeds the set threshold, it means that the current operation is in the non-steady state stage.

[0088] Step S200 : extracting and eliminating the variable operating condition features of the steady-state data to obtain operating condition-independent data in the operating data.

[0089] In this embodiment, it is necessary to obtain operating-condition-independent data from the chiller's steady-state data. In specific implementation, the variable operating-condition characteristics of the steady-state data are first extracted using the deep variational information bottleneck method to obtain variable operating-condition characteristic data. Then, the variable operating-condition characteristic data from the steady-state data are removed to obtain operating-condition-independent data from the steady-state data. By removing the variable operating-condition characteristic data, the influence of variable operating conditions on the chiller's performance assessment can be eliminated.

[0090] In one implementation, the conditional variable steady-state data in the steady-state data is input into a deep variational information bottleneck model, and the variable operating condition features in the conditional variable steady-state data are extracted to obtain the variable operating condition feature data of the steady-state data. The Deep Variational Information Bottleneck method (Deep VIB) is developed on the theoretical basis of the Information Bottleneck method (IB), which can extract features related to the target output and remove irrelevant information from the extracted features. Figure 3 As shown, V represents the features extracted from X that are related to Y. The IB criterion is to maximize the mutual information I(V,Y) under the constraint of minimizing the mutual information I(X,V) to obtain the simplest statistics. The learning goal of the IB criterion is formula (4):

[0091] maxI(V,Y;θ),stI(V,X;θ)≤I c , (4)

[0092] Where θ is the parameter to be learned, I c is a constraint condition. In a continuous random process, the mutual information is expressed as formula (5),

[0093]

[0094] Based on the learning objective, the Information Bottleneck Lagrangian (IBL) optimization function is expressed as Equation (6):

[0095] J IB (θ) = I(V, Y; θ) - βI(X, V; θ), (6) where β is a coefficient used to compromise between the adequacy of V in representing X for Y and the minimization of I(X, V).

[0096]

[0097] In the formula, H(Y) is determined only by the prior distribution p(y) and is a constant term. The variational upper bound of the mutual information I(X, V) in the IB objective function is expressed as formula (8):

[0098]

[0099]

[0100] In this embodiment, the distribution of collected data can be used to approximate the distribution p(x,v), and the historical operation data set is composed of input data and output data The optimization objective function of Deep VIB is approximately as follows:

[0101]

[0102] In one implementation, residuals are made for the variable operating condition characteristic data of the steady-state data and the state variable steady-state data in the steady-state data, and the residual values ​​are obtained as the operating condition-independent data in the steady-state data. Figure 4 As shown, the Deep VIB model is trained using the steady-state data of the conditional variables as input and the predicted data of the state variables as output. This Deep VIB model enables state variable prediction based on the conditional variables. The predicted values ​​essentially encompass all relevant information about the operating conditions, resulting in the residual between the actual measured and predicted values ​​of the state variables being largely independent of the conditional variables. This residual is used to represent information unrelated to the chiller's operating conditions, eliminating information related to the operating conditions.

[0103] Step S300: extracting slowly degraded performance characteristic data from the working condition-independent data, wherein the slowly degraded performance characteristic data is used to reflect the performance degradation of the chiller.

[0104] In this embodiment, the slowly degrading performance characteristic data is extracted from the operating condition-independent data using a deep variational information bottleneck method to obtain the slowly degrading performance characteristic data. The heat exchange tube structure of the chiller condenser undergoes a slowly degrading process, and this degradation trend is strongly correlated with time. By extracting time-dependent features from the chiller operating data, the performance change trend of the equipment can be better reflected.

[0105] In one implementation, the working condition-independent data is input into a deep variational information bottleneck encoder model to extract the slowly degrading performance feature data from the working condition-independent data to obtain the slowly degrading performance feature data. The Deep VIB method is used to extract features related to degradation performance, such as Figure 5As shown, the data of multiple operating cycles are input into the encoder model of Deep VIB according to a sliding window for a certain duration, and the encoder model outputs latent variable features. LSTM is used to obtain slowly changing information about the operation process of the chiller and reduce the interference of system noise at a single moment on the extraction of features related to degradation performance. The LSTM (Long Short-Term Memory) is a long short-term memory network, a time recursive neural network, suitable for processing and predicting events with relatively long intervals and delays in time series. Multiple operating cycles share an encoder model, and a separate decoder model is set for each operating cycle to predict the running time within a cycle based on the latent variable features. The model is trained using random batch sampling. Since the feature layer retains information related to the output and removes information irrelevant to the output, the extracted features can well reflect the performance degradation of the chiller.

[0106] In one implementation, the chiller's operating cycle is considered to be the period from the last descaling to the next descaling. The first power-on after descaling is completed is considered to be 0, and the timer begins counting. The chiller's operating time is represented as the total operating duration from the first power-on after descaling to the current moment, excluding power-off time. The time of a single operating cycle is normalized to a range of (0-1). Historical chiller operating data can be divided into multiple independent operating cycles.

[0107] Step S400: Obtain a chiller health status assessment result based on the slowly degrading performance characteristic data.

[0108] In this embodiment, to obtain the health status result of the chiller, first, set a monitoring time window, and determine a reference benchmark value based on the slowly degrading performance characteristic data within the monitoring time window; then, calculate the distance from the mean of the slowly degrading performance characteristic data within the current time window to the reference benchmark value, and obtain the health status assessment result of the chiller based on the distance, which is a health status indicator reflecting the degradation performance of the chiller.

[0109] In one implementation, the extracted performance degradation feature data is multidimensional and cannot intuitively reflect device performance degradation. Therefore, a single health status indicator is constructed using this data. First, a fixed-size monitoring time window is set. The feature mean within the time window corresponding to the initial time of a cycle is used as a reference baseline value. The distance between the feature mean within the time window at the current moment and the reference baseline value is calculated, and this distance value is used as the health status indicator reflecting the performance degradation. Figure 6The figure shows the changes in health indicators within a specific operating cycle. Based on the historical operating data, it is assumed that the equipment's health was at its worst before cleaning during at least one operating cycle, indicating the need for timely maintenance. Based on this setting, the health indicator is calculated from the historical data. The 1% of data with the largest index values ​​are selected as abnormal equipment health data, and the health indicator control limits are determined.

[0110] In one implementation, if a health indicator exceeds a set threshold, an alarm is issued to notify maintenance personnel to clean the equipment promptly. This prompts maintenance personnel to clean the equipment promptly to avoid equipment failure caused by long-term cleaning or a significant reduction in cooling efficiency due to scaling, which wastes resources.

[0111] Exemplary Systems

[0112] Based on the above embodiments, the present invention also provides a health status assessment system for a variable operating condition chiller. Figure 7 As shown, the system in this embodiment includes a data acquisition module 10, a first feature extraction module 20, a second feature extraction module 30, and an evaluation result acquisition module 40. Specifically, the data acquisition module 10 is used to obtain the operating data of the chiller and preprocess the operating data, and obtain steady-state data in the operating data; the first feature extraction module 20 is used to extract and eliminate the variable operating condition features of the steady-state data to obtain operating condition-independent data in the operating data; the second feature extraction module 30 is used to extract slowly degraded performance characteristic data from the operating condition-independent data, and the slowly degraded performance characteristic data is used to reflect the performance degradation of the chiller; the evaluation result acquisition module 40 is used to obtain the health status evaluation result of the chiller based on the slowly degraded performance characteristic data.

[0113] In one implementation, the data acquisition module includes:

[0114] a type determination unit, configured to obtain operating data of the chiller and determine a data type of the operating data, wherein the data type includes conditional variable data and state variable data;

[0115] The data processing unit is used to eliminate abnormal data from the operating data according to the data type of the operating data, and obtain steady-state data from the operating data, wherein the steady-state data includes steady-state data of conditional variables and steady-state data of state variables.

[0116] In one implementation, the data processing unit includes:

[0117] The data filtering subunit is configured to calculate a geometric weighted mean and variance of the operating data, and perform steady-state filtering on the operating data to obtain steady-state data from the operating data.

[0118] In one implementation, the first feature extraction module includes:

[0119] a first data extraction unit, configured to extract the variable operating condition features of the steady-state data based on a deep variational information bottleneck method to obtain variable operating condition feature data of the steady-state data;

[0120] The data elimination unit is used to eliminate the variable working condition characteristic data of the steady-state data to obtain working condition irrelevant data in the steady-state data.

[0121] In one implementation, the first data extraction unit includes:

[0122] The first data extraction subunit is used to input the conditional variable steady-state data in the steady-state data into the deep variational information bottleneck model, extract the variable operating condition features in the conditional variable steady-state data, and obtain the variable operating condition feature data of the steady-state data.

[0123] In one implementation, the data elimination unit includes:

[0124] The data elimination subunit is used to make a residual of the variable working condition characteristic data of the steady-state data and the steady-state data of the state variables in the steady-state data, and obtain the residual value as the working condition-independent data in the steady-state data.

[0125] In one implementation, the second feature extraction module includes:

[0126] The second feature extraction unit is used to extract the slowly degrading performance characteristic data from the working condition-independent data based on the deep variational information bottleneck method to obtain the slowly degrading performance characteristic data.

[0127] In one implementation, the second feature extraction unit includes:

[0128] The second data extraction subunit is used to input the working condition-independent data into the encoder model of the deep variational information bottleneck to extract the slowly degrading performance characteristic data from the working condition-independent data to obtain the slowly degrading performance characteristic data.

[0129] In one implementation, the evaluation result acquisition module includes:

[0130] a standard determination unit, configured to set a monitoring time window and determine a reference benchmark value based on the slow degradation performance characteristic data within the monitoring time window;

[0131] The evaluation result acquisition unit is used to calculate the distance between the mean value of the slowly degrading performance characteristic data in the current time window and the reference benchmark value, and obtain the health status evaluation result of the chiller according to the distance, wherein the distance is a health status indicator reflecting the degradation performance of the chiller.

[0132] In one implementation, the system further includes:

[0133] The alarm prompt module is used to issue an alarm maintenance message if the health status indicator exceeds the set threshold, notifying the operation and maintenance personnel to clean the equipment in time.

[0134] The working principles of each module in the variable operating condition chiller health status assessment system of this embodiment are the same as the principles of each step in the above method embodiment, and will not be repeated here.

[0135] Based on the above embodiment, the present invention further provides a terminal device, the principle block diagram of the terminal device can be as follows: Figure 8 The terminal device may include one or more processors 100 ( Figure 8 Only one is shown), memory 101, and a computer program 102 stored in memory 101 and executable on one or more processors 100, for example, a program for a method for assessing the health status of a variable-condition chiller. When one or more processors 100 execute computer program 102, each step of the method embodiment of the method for assessing the health status of a variable-condition chiller can be implemented. Alternatively, when one or more processors 100 execute computer program 102, the functions of each module / unit in the apparatus embodiment of the method for assessing the health status of a variable-condition chiller can be implemented, without limitation herein.

[0136] In one embodiment, the processor 100 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0137] In one embodiment, the memory 101 may be an internal storage unit of an electronic device, such as a hard disk or memory of the electronic device. The memory 101 may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device. Furthermore, the memory 101 may also include both an internal storage unit of the electronic device and an external storage device. The memory 101 is used to store computer programs and other programs and data required by the terminal device. The memory 101 may also be used to temporarily store data that has been output or is about to be output.

[0138] Those skilled in the art will understand that Figure 8 The principle block diagram shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the terminal device to which the solution of the present invention is applied. The specific terminal device includes Figure 8 More or fewer components may be shown, or some components may be combined, or the components may be arranged differently.

[0139] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, operating database or other media used in the embodiments provided by the present invention may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0140] In summary, the present invention discloses a method and system for evaluating the health status of a variable-operating-condition chiller, the method comprising: obtaining the operating data of the chiller and preprocessing the operating data, and obtaining steady-state data in the operating data; extracting and eliminating the variable-operating-condition features of the variable-operating-condition data of the steady-state data to obtain operating-condition-independent data in the operating data; extracting slow-return performance characteristic data from the operating-condition-independent data, the slow-degradation performance characteristic data being used to reflect the performance degradation of the chiller; and obtaining a chiller health status evaluation result based on the slow-degradation performance characteristic data. The present invention extracts and eliminates characteristic information related to operating-condition changes of the chiller by adopting a deep variational information bottleneck method, effectively avoiding the influence of changes in external environmental factors and operating conditions on the extraction of performance-degradation-related features, further adopting a deep variational information bottleneck method to extract features related to equipment degradation performance, reducing the existence of information irrelevant to the performance of the degraded equipment, and constructing a performance health status indicator based on this to achieve online monitoring of equipment performance. By removing the operating condition processing and extracting the relevant features of equipment degradation performance, the changing trend identification and health status assessment of the chiller under long-term variable operating conditions can be achieved.

[0141] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for evaluating the health status of a variable-operating-condition chiller, characterized in that: The method comprises: Acquiring operating data of the chiller, preprocessing the operating data, and acquiring steady-state data from the operating data; Extracting and eliminating the variable operating condition features of the steady-state data to obtain operating condition-independent data in the operating data; Extracting slowly degraded performance characteristic data from the working condition-independent data, wherein the slowly degraded performance characteristic data is used to reflect the performance degradation of the chiller; Obtaining a chiller health status assessment result based on the slowly degrading performance characteristic data; The step of obtaining the operating data of the chiller, preprocessing the operating data, and obtaining steady-state data from the operating data includes: Acquire operating data of the chiller and determine a data type of the operating data, wherein the data type includes condition variable data and state variable data; Eliminate abnormal data from the operating data according to the data type of the operating data, and obtain steady-state data from the operating data, wherein the steady-state data includes steady-state data of conditional variables and steady-state data of state variables; The step of extracting and eliminating the variable operating condition features of the steady-state data to obtain operating condition-independent data in the operating data includes: Extracting the variable operating condition features of the steady-state data based on a deep variational information bottleneck method to obtain variable operating condition feature data of the steady-state data; Eliminating the variable operating condition characteristic data of the steady-state data to obtain operating condition-irrelevant data in the steady-state data; The step of extracting the variable operating condition features of the steady-state data based on the deep variational information bottleneck method to obtain variable operating condition feature data of the steady-state data includes: Inputting the conditional variable steady-state data in the steady-state data into a deep variational information bottleneck model, extracting the variable operating condition features in the conditional variable steady-state data, and obtaining variable operating condition feature data of the steady-state data; Eliminating the variable operating condition characteristic data of the steady-state data and obtaining operating condition-irrelevant data in the steady-state data includes: Making a residual of the variable working condition characteristic data of the steady-state data and the steady-state data of the state variables in the steady-state data, and obtaining the residual value as the working condition-independent data in the steady-state data; The extracting the slowly degrading performance characteristic data from the operating condition-independent data includes: The slowly degrading performance characteristic data in the working condition-independent data is extracted based on a deep variational information bottleneck method to obtain the slowly degrading performance characteristic data.

2. The health status assessment method for a variable operating condition chiller according to claim 1, characterized in that: Eliminating abnormal data from the operating data according to the data type of the operating data to obtain steady-state data from the operating data includes: A geometric weighted mean and variance of the operating data are calculated, and steady-state filtering is performed on the operating data to obtain steady-state data in the operating data.

3. The health status assessment method for a variable operating condition chiller according to claim 1, characterized in that: The method of extracting the slowly degrading performance characteristic data from the working condition-independent data based on the deep variational information bottleneck method to obtain the slowly degrading performance characteristic data includes: The working condition-independent data is input into a deep variational information bottleneck encoder model to extract slowly degrading performance characteristic data from the working condition-independent data to obtain slowly degrading performance characteristic data.

4. The health status assessment method for a variable operating condition chiller according to claim 1, characterized in that: The obtaining of a chiller health status assessment result based on the slowly degrading performance characteristic data includes: Setting a monitoring time window, and determining a reference baseline value based on the slow degradation performance characteristic data within the monitoring time window; The distance between the mean value of the slowly degrading performance characteristic data in the current time window and the reference benchmark value is calculated, and the health status assessment result of the chiller is obtained according to the distance, where the distance is a health status indicator reflecting the degradation performance of the chiller.

5. The health status assessment method for a variable operating condition chiller according to claim 1, characterized in that: The method further comprises: If the health status indicator exceeds the set threshold, an alarm maintenance message will be issued to notify the operation and maintenance personnel to clean the equipment in time.

6. A variable operating condition chiller health status assessment system, characterized in that: The system comprises: A data acquisition module, configured to acquire operating data of the chiller, pre-process the operating data, and acquire steady-state data from the operating data; A first feature extraction module is used to extract and eliminate the variable operating condition features of the steady-state data to obtain operating condition-irrelevant data in the operating data; a second feature extraction module, configured to extract slowly degrading performance feature data from the operating condition-independent data, wherein the slowly degrading performance feature data is used to reflect the performance degradation of the chiller; An evaluation result acquisition module, configured to acquire a health status evaluation result of the chiller based on the slowly degrading performance characteristic data; The data acquisition module includes: a type determination unit, configured to obtain operating data of the chiller and determine a data type of the operating data, wherein the data type includes conditional variable data and state variable data; a data processing unit, configured to remove abnormal data from the operating data according to a data type of the operating data, and obtain steady-state data from the operating data, wherein the steady-state data includes steady-state data of conditional variables and steady-state data of state variables; The first feature extraction module includes: a first data extraction unit, configured to extract the variable operating condition features of the steady-state data based on a deep variational information bottleneck method to obtain variable operating condition feature data of the steady-state data; a data elimination unit, configured to eliminate characteristic data of variable working conditions from the steady-state data, and obtain data irrelevant to working conditions from the steady-state data; The first data extraction unit includes: a first data extraction subunit, configured to input the conditional variable steady-state data in the steady-state data into a deep variational information bottleneck model, extract variable operating condition features from the conditional variable steady-state data, and obtain variable operating condition feature data of the steady-state data; The data elimination unit includes: A data elimination subunit is used to make a residual of the variable working condition characteristic data of the steady-state data and the steady-state data of the state variables in the steady-state data, and obtain the residual value as the working condition irrelevant data in the steady-state data; The second feature extraction module includes: The second feature extraction unit is used to extract the slowly degrading performance characteristic data from the working condition-independent data based on the deep variational information bottleneck method to obtain the slowly degrading performance characteristic data.

7. A terminal device, characterized in that: The terminal device includes a memory, a processor, and a variable operating condition chilled water host health status assessment method program stored in the memory and runnable on the processor. When the processor executes the variable operating condition chilled water host health status assessment method program, it implements the steps of the variable operating condition chilled water host health status assessment method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program for the health status assessment method of a variable operating condition chiller. When the program is executed by the processor, the steps of the health status assessment method of a variable operating condition chiller are implemented as described in any one of claims 1 to 5.

Images