A multi-user heat pump measurement and control network structure and data-driven control system

Abstract

The purpose of the present invention is to provide a multi-user heat pump measurement and control network structure and a data-driven control system, including a multi-user heat pump measurement and control network, a network dynamic behavior correlation analysis module, a system failure mode evaluation and response module, a network data-driven control module and a human-machine interface. It can obtain the multi-layer network connection distribution of sensors, actuators, control units, etc. in the heat pump system and the interactive relationship between each component, and identify the response changes and potential vulnerabilities of system components, obtain available data from complex network systems and realize data-driven control of complex networks, which will provide enterprises with parameter optimization and performance control strategies, improve system economy, and save enterprise operating costs.

Description

Technical Field

[0001] The invention relates to a heat pump measurement and control network structure, and in particular to a multi-user heat pump measurement and control network structure and a data driven control system. Background Art

[0002] As the scale of users increases, the multi-layer network connection distribution of sensors, actuators, control units, etc. in heat pump systems and the interactive relationship of components become more and more complex. It has become increasingly difficult to control, predict and evaluate the operation process and equipment. In addition, enterprises generate and store a large amount of production, equipment and process data every day, but they cannot effectively use these massive data, resulting in a large waste of resources. How to effectively integrate the existing automation system and make full use of the relevant data in the enterprise to improve the scheduling level, and realize the optimization control, prediction and evaluation of the production process and equipment under the condition that it is difficult to establish a more accurate mechanism model of the controlled system, has become an urgent problem to be solved in system control optimization. However, in the Internet of Things data platform and the temperature, pressure and other test sensor networks, due to the large scale of the data network and the uncertainty of the transmission state between sensors, the topological structure of the network cannot be mastered in real time. Although it can identify unknown structures, the calculation amount is large, requiring large memory and computing time, which affects the functional behavior of the network.

[0003] Different network structures have different anti-attack and self-recovery capabilities. Therefore, a test network structure suitable for multi-user heat pump systems needs to be proposed. The dynamic behavior of the heat pump system network is the key to exploring how the entire system responds to changes and potential failures. Most existing systems are single-layer network structures. In actual operation, the heat pump network system based on the Internet of Things cloud platform exists in the form of a multi-layer network. The dynamic behavior of the multi-layer network structure is far more complex than that of the single-layer network. How to achieve control of the multi-layer network is more challenging. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a multi-user heat pump measurement and control network structure and a data-driven control system, which can obtain the multi-layer network connection distribution of sensors, actuators, control units, etc. in the heat pump system and the interactive relationship between each component, and identify the response changes and potential vulnerabilities of system components, obtain available data from complex network systems and realize data-driven control of complex networks, which will provide enterprises with parameter optimization and performance regulation strategies, improve system economy, and save enterprise operating costs.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] A multi-user heat pump measurement and control network structure and data-driven control system, including a multi-user heat pump measurement and control network, a network dynamic behavior correlation analysis module, a system failure mode evaluation and response module, a network data-driven control module and a human-machine interface;

[0007] The multi-user heat pump measurement and control network includes a physical network structure and an information network structure. The physical network structure includes nodes such as compressors, condensers, evaporators, compressors and valves and physical connections between nodes. The information network structure includes nodes such as temperature sensors, pressure sensors, flow meters and actuators and logical connections between nodes.

[0008] The network dynamic behavior correlation analysis module analyzes the overall dynamic behavior of the multi-user heat pump measurement and control network, uses the persistence graph to encode information features, mines the topological information contained in the data sequence cloud graph, and evaluates the correlation between the network topology structure and the system dynamic behavior;

[0009] The system failure mode evaluation and response module is based on the response characteristics of the system under different operating conditions (conversion of the heat pump system from a stopped state to an operating state, conversion of the heat pump system from an operating state to a stopped state, and changes in the operating state of the heat pump system under different load conditions), selects characteristic parameters of system failure modes (temperature, pressure and other sensors cannot normally collect data, pipes in the heat pump system are blocked from fluid flow due to foreign matter or scaling, and actuators cannot normally execute control instructions), quantifies the connectivity loss of the evaluation system when nodes and connections fail, and transmits signals to the control system;

[0010] The network data driven control module includes a data time series causal decomposition module, a network driving factor analysis module, a fault diagnosis control module, a user interaction management module and an execution device; the data time series causal decomposition module uses the empirical mode decomposition method and the Hilbert-Huang transform to solve the instantaneous phase coherence and analyze the causal relationship and strength between time series; the network driving factor analysis module combines the results of causal decomposition and analyzes the main driving factors of the network system based on the data mining method of principal component analysis and correlation analysis, quantifies the impact of environmental parameters and operating parameters on the system and identifies the main driving factors, extracts the main features in the data through the principal component analysis, reduces the amount of data testing, and improves the data transmission efficiency, and uses the correlation analysis to identify the correlation between parameters and screen out relevant parameters, and the correlation degree between the relevant parameters and the heat exchange amount, the energy efficiency ratio (COP) of the heat pump system and the input power is not less than 0.5; the fault diagnosis control module combines the information transmitted by the system fault mode evaluation and response module to identify the fault location and transmit it to the control system; the user interaction management module judges the result again, issues a control instruction to the control system, and controls through the execution device;

[0011] The human-machine interface provides a visual tool for people to check the system status, control the operation of the heat pump system, and check the analysis results.

[0012] Furthermore, the data mining method based on principal component analysis and correlation analysis includes the following steps:

[0013] (I) Select the following actual data input variable values ​​of the heat pump system in normal operation: inlet air dry bulb temperature, inlet air wet bulb temperature, inlet air relative humidity, user side inlet temperature, user side outlet temperature, user side inlet and outlet temperature difference, user side three-way valve back temperature, user side flow, user side inlet and outlet pressure difference, cooling tower return water temperature, test machine A phase voltage, test machine B phase voltage, test machine C phase voltage, test machine average voltage, test machine A phase current, test machine B phase current, test machine C phase current, test machine average current, input power, power factor, power supply frequency, atmospheric pressure, heat pump system high pressure side pressure, heat pump system low pressure side pressure, plate heat exchanger outlet temperature, plate heat exchanger inlet temperature, distributor inlet flow, suction, exhaust, oil pressure, upper air inlet, lower air inlet, upper air outlet, lower air outlet data, water side heat exchange, heat efficiency ratio (COP) of the heat pump system and water resistance data;

[0014] (ii) Conduct a preliminary screening of the variables, remove variables that have not changed, remove derivative variables of voltage, current, user-side temperature, and environmental variables that have a minor impact on the operating status of the heat pump (the impact is less than 5%);

[0015] (III) Standardize the original input data, use the standardized variable data to perform principal component analysis, and process the input variable values:

[0016] The original data matrix is:

[0017]

[0018] The raw data were standardized as follows:

[0019]

[0020] in,

[0021] (IV) Calculate the correlation coefficient matrix of sample data. Assuming that the original data is still represented by X after standardization, the correlation coefficient of the data after standardization is:

[0022]

[0023] in,

[0024] (V) Calculate the eigenvalues ​​(λ1, λ2, ..., λp ), the corresponding eigenvectors are:

[0025] a i =(a i1 ,a i2 ,…,a ip ),i=1,2,…p

[0026] (VI) Select important principal components and write the principal component expression. The contribution rate refers to the proportion of the variance of a principal component to the total variance, that is, the proportion of a certain eigenvalue to the total of all eigenvalues:

[0027]

[0028] The larger the contribution rate, the stronger the information of the original variable contained in the principal component. The number of principal components k is determined by the cumulative contribution rate of the principal components, which is required to be no less than 85%.

[0029] (VII) Analyze the correlation between input variable data and use the Pearson correlation coefficient γ to express the degree of relationship between the two. Its definition is as follows:

[0030]

[0031] Where: f i is the i-th sample f; is the mean of sample f, g i is the i-th sample, is the mean of sample g;

[0032] The correlation analysis adopts the following correlation degree judgment criteria:

[0033] When 0.7≤|γ|<1, variables f and g are highly correlated;

[0034] When 0.4≤|γ|<0.7, variables f and g are moderately correlated;

[0035] When 0.2≤|γ|<0.4, variables f and g are poorly correlated;

[0036] When |γ|<0.2, variables f and g are extremely poorly correlated or close to zero correlation.

[0037] Compared with the prior art, the present invention has the following beneficial technical effects:

[0038] A multi-user heat pump measurement and control network structure and data-driven control system of the present invention can obtain the multi-layer network connection distribution of sensors, actuators, control units, etc. in the heat pump system and the interactive relationship between each component, and identify how system components respond to changes and potential vulnerabilities, obtain available data from complex network systems and realize data-driven control of complex networks, which will provide enterprises with parameter optimization and performance regulation strategies, improve system economy, and save enterprise operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Schematic diagram of the multi-user heat pump measurement and control network structure and data-driven control system

[0040] Figure 2 Schematic diagram of the principal component cumulative contribution rate results of the principal component analysis of the case of the present invention

[0041] Figure 3 Schematic diagram of the correlation coefficient results between variables in the case of the present invention

[0042] Figure 4 Schematic diagram of the correlation coefficient between the heat transfer and other medium-related variables in the case of the present invention DETAILED DESCRIPTION

[0043] Reference Figure 1 The present invention discloses a multi-user heat pump measurement and control network structure and a data-driven control system, comprising a multi-user heat pump measurement and control network (1), a network dynamic behavior correlation analysis module (2), a system failure mode evaluation and response module (3), a network data-driven control module (4) and a human-machine interface (5);

[0044] The multi-user heat pump measurement and control network (1) comprises a physical network structure and an information network structure, wherein the physical network structure comprises nodes such as compressors, condensers, evaporators, compressors and valves and physical connections between the nodes, and the information network structure comprises nodes such as temperature sensors, pressure sensors, flow meters and actuators and logical connections between the nodes;

[0045] The network dynamic behavior correlation analysis module (2) analyzes the overall dynamic behavior of the multi-user heat pump measurement and control network, uses the persistent graph encoding information features, mines the topological information contained in the data sequence cloud graph, and evaluates the correlation between the network topological structure and the system dynamic behavior;

[0046] The system failure mode evaluation and response module (3) selects characteristic parameters of system failure modes (temperature, pressure and other sensors cannot normally collect data, pipes in the heat pump system are blocked due to foreign matter or scaling, and actuators cannot normally execute control instructions) based on the response characteristics of the system under different operating conditions (conversion of the heat pump system from a stopped state to an operating state, conversion of the heat pump system from an operating state to a stopped state, and changes in the operating state of the heat pump system under different load conditions), quantitatively evaluates the connectivity loss of the system when nodes and connections fail, and transmits signals to the control system;

[0047] The network data driven control module (4) comprises a data time series causal decomposition module, a network driving factor analysis module, a fault diagnosis control module, a user interaction management module and an execution device; the data time series causal decomposition module adopts the empirical mode decomposition method and the Hilbert-Huang transform to solve the instantaneous phase coherence and analyze the causal relationship and strength between time series; the network driving factor analysis module combines the result of causal decomposition and analyzes the main driving factors of the network system based on the data mining method of principal component analysis and correlation analysis, quantifies the influence of environmental parameters and operating parameters on the system and identifies the main driving factors, extracts the main features in the data through the principal component analysis, reduces the amount of data testing, and improves the data transmission efficiency; uses the correlation analysis to identify the correlation between parameters and screen out relevant parameters, and the correlation degree between the relevant parameters and the heat exchange amount, the energy efficiency ratio (COP) of the heat pump system and the input power is not less than 0.5; the fault diagnosis control module combines the information transmitted by the system fault mode evaluation and response module (3) to identify the fault location and transmit it to the control system; the user interaction management module judges the result again, issues a control instruction to the control system, and performs control through the execution device;

[0048] The human-machine interface (5) provides a visual tool that allows people to conveniently view system status, control the operation of the heat pump system, and view analysis results.

[0049] The data mining method based on principal component analysis and correlation analysis in this embodiment includes the following steps:

[0050] (I) Select the following actual data input variable values ​​of the heat pump system in normal operation: inlet air dry bulb temperature, inlet air wet bulb temperature, inlet air relative humidity, user side inlet temperature, user side outlet temperature, user side inlet and outlet temperature difference, user side three-way valve back temperature, user side flow, user side inlet and outlet pressure difference, cooling tower return water temperature, test machine A phase voltage, test machine B phase voltage, test machine C phase voltage, test machine average voltage, test machine A phase current, test machine B phase current, test machine C phase current, test machine average current, input power, power factor, power supply frequency, atmospheric pressure, heat pump system high pressure side pressure, heat pump system low pressure side pressure, plate heat exchanger outlet temperature, plate heat exchanger inlet temperature, distributor inlet flow, suction, exhaust, oil pressure, upper air inlet, lower air inlet, upper air outlet, lower air outlet data, water side heat exchange, heat efficiency ratio (COP) of heat pump system, water resistance data;

[0051] (II) Preliminary screening of variables, removing the unchanged variable power frequency, removing the derived variables use side inlet and outlet temperature difference, the average voltage of the tested machine, the average current of the tested machine, and the environmental variables atmospheric pressure, high pressure side pressure of the heat pump system, and low pressure side pressure of the heat pump system that have a relatively small impact on the operation state of the heat pump (the impact amplitude is less than 5%), and 30 variables are obtained after preliminary screening. Then, the 30 parameter variables after preliminary screening are expressed, as shown in Table 1;

[0052] Table 1. Operation data set of heat pump system refrigeration process

[0053]

[0054]

[0055] (III) Standardize the original input data, use the standardized variable data to perform principal component analysis, and process the input variable values:

[0056] The original data matrix is:

[0057]

[0058] The raw data were standardized as follows:

[0059]

[0060] in,

[0061] (IV) Calculate the correlation coefficient matrix of sample data. Assuming that the original data is still represented by X after standardization, the correlation coefficient of the data after standardization is:

[0062]

[0063] in,

[0064] (V) Calculate the eigenvalues ​​(λ1, λ2, ..., λ p ), the corresponding eigenvectors are:

[0065] a i =(a i1 ,a i2 ,…,a ip ),i=1,2,…p

[0066] (VI) Select important principal components and write the principal component expression. The contribution rate refers to the proportion of the variance of a principal component to the total variance, that is, the proportion of a certain eigenvalue to the total of all eigenvalues:

[0067]

[0068] The larger the contribution rate, the stronger the information of the original variable contained in the principal component. The number of principal components k is determined by the cumulative contribution rate of the principal components, which is required to be no less than 85%.

[0069] (VII) Analyze the correlation between input variable data and use the Pearson correlation coefficient γ to express the degree of relationship between the two. Its definition is as follows:

[0070]

[0071] Where: f i is the i-th sample f; is the mean of sample f, g i is the i-th sample, is the mean of sample g;

[0072] like Figure 1 As shown, in the heat pump system performance test method and system based on principal components and correlation in this embodiment, the principal component analysis feature extraction method obtains new principal components sorted by variance contribution rate, and obtains several principal components containing most of the original data information. The principal component analysis transforms the original 30 parameter features, and the new principal component feature quantity is a first-order polynomial combination value of the 30 original parameter feature quantities, such as Figure 2 The first five principal component features reflect more than 85% of the information. Using these principal component feature groups to represent the operation greatly reduces the number of PCA principal component features compared to the original 30 features.

[0073] Table 2 reflects the parameters of the main components

[0074]

[0075] like Figure 3As shown in Table 3, correlation analysis was performed on all 30 variables. In order to more intuitively express the correlation between the heat exchange of the heat pump refrigeration process, the energy efficiency ratio (COP) of the heat pump system and the input power and other variables, the variables with an absolute value of the correlation coefficient greater than 0.5 were screened. The screening results are shown in Table 3.

[0076] Table 3 Correlation screening table

[0077]

[0078]

[0079] like Figure 3 , as shown in 4, in this embodiment, the correlation analysis is specifically performed according to the following strategy:

[0080] The correlation analysis adopts the following correlation degree judgment criteria:

[0081] When 0.7≤|γ|<1, variables f and g are highly correlated;

[0082] When 0.4≤|γ|<0.7, variables f and g are moderately correlated;

[0083] When 0.2≤|γ|<0.4, variables f and g are poorly correlated;

[0084] When |γ|<0.2, variables f and g are extremely low correlated or close to zero correlation;

[0085] 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 above embodiments, a person skilled in the art should understand that the specific implementation modes of the present invention can still be modified or replaced by equivalents, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be included in the scope of the claims.

Claims

1. A multi-user heat pump measurement and control network structure and data-driven control system, characterized in that: It includes a multi-user heat pump measurement and control network (1), a network dynamic behavior correlation analysis module (2), a system failure mode evaluation and response module (3), a network data driven control module (4) and a human-machine interface (5); The multi-user heat pump measurement and control network (1) comprises a physical network structure and an information network structure, wherein the physical network structure comprises compressor, condenser, evaporator, compressor and valve nodes and physical connections between nodes, and the information network structure comprises temperature sensor, pressure sensor, flow meter and actuator nodes and logical connections between nodes; The network dynamic behavior correlation analysis module (2) analyzes the overall dynamic behavior of the multi-user heat pump measurement and control network, uses the persistent graph encoding information features, mines the topological information contained in the data sequence cloud graph, and evaluates the correlation between the network topological structure and the system dynamic behavior; The system failure mode evaluation and response module (3) selects characteristic parameters of the system failure mode based on the response characteristics of the system under different operating conditions, quantitatively evaluates the connectivity loss of the system when the node and connection fail, and transmits a signal to the control system. The operating conditions include the conversion of the heat pump system from a stop state to a running state, the conversion of the heat pump system from a running state to a stop state, and the change of the running state of the heat pump system under different load conditions. The failure mode includes the temperature and pressure sensors failing to collect data normally, the fluid flow being blocked due to foreign matter or scaling in the pipeline of the heat pump system, and the actuator failing to execute the control command normally. The network data driven control module (4) comprises a data time series causal decomposition module, a network driven factor analysis module, a fault diagnosis control module, a user interaction management module and an execution device; the data time series causal decomposition module adopts an empirical mode decomposition method and a Hilbert-Huang transform to solve the instantaneous phase coherence and analyze the causal relationship and strength between time series; the network driven factor analysis module combines the result of causal decomposition and analyzes the main driving factors of the network system based on a data mining method of principal component analysis and correlation analysis, quantifies the influence of environmental parameters and operating parameters on the system and identifies the main driving factors, extracts the main features in the data through the principal component analysis, identifies the correlation between parameters by using the correlation analysis and selects the relevant parameters, and the correlation degree between the relevant parameters and the heat exchange amount, the energy efficiency ratio of the heat pump system and the input power is not less than 0.5; The fault diagnosis and control module combines the information transmitted by the system fault mode evaluation and response module (3) to identify the fault location and transmit it to the control system; The user interaction management module judges the result again, issues a control instruction to the control system, and performs control through the execution device; The human-machine interface (5) provides a visual tool that allows people to conveniently view system status, control the operation of the heat pump system, and view analysis results.

2. A multi-user heat pump measurement and control network structure and data driven control system according to claim 1, characterized in that: The data mining method based on principal component analysis and correlation analysis comprises the following steps: (I) Select the following actual data input variable values ​​of the heat pump system in normal operation: inlet air dry bulb temperature, inlet air wet bulb temperature, inlet air relative humidity, user side inlet temperature, user side outlet temperature, user side inlet and outlet temperature difference, user side three-way valve back temperature, user side flow, user side inlet and outlet pressure difference, cooling tower return water temperature, test machine A phase voltage, test machine B phase voltage, test machine C phase voltage, test machine average voltage, test machine A phase current, test machine B phase current, test machine C phase current, test machine average current, input power, power factor, power supply frequency, atmospheric pressure, heat pump system high pressure side pressure, heat pump system low pressure side pressure, plate heat exchanger outlet temperature, plate heat exchanger inlet temperature, distributor inlet flow, suction, exhaust, oil pressure, upper air inlet, lower air inlet, upper air outlet, lower air outlet data, water side heat exchange, heat efficiency ratio of heat pump system and water resistance data; (ii) Conduct a preliminary screening of the variables, remove variables that do not change, remove derivative variables of voltage, current, user-side temperature, and environmental variables that have an impact of less than 5% on the operating parameters of the heat pump; (III) Standardize the original input data, use the standardized variable data to perform principal component analysis, and process the input variable values: The original data matrix is: The raw data were standardized as follows: (IV) Calculate the correlation coefficient matrix of sample data. Assuming that the original data is still represented by X after standardization, the correlation coefficient of the data after standardization is: in, (V) Calculate the eigenvalues ​​(λ1, λ2, ..., λ p ), the corresponding eigenvectors are: a i =(a i1 ,a i2 ,…,a ip ),i=1,2,…p (VI) Select important principal components and write the principal component expression. The contribution rate refers to the proportion of the variance of a principal component to the total variance, that is, the proportion of a certain eigenvalue to the total of all eigenvalues: The larger the contribution rate, the stronger the information of the original variable contained in the principal component. The number of principal components k is determined by the cumulative contribution rate of the principal components, which is required to be no less than 85%. (VII) Analyze the correlation between input variable data and use the Pearson correlation coefficient γ to express the degree of relationship between the two. Its definition is as follows: Where: f i is the i-th sample f; f is the mean of sample f, g i is the i-th sample, is the mean of sample g; The correlation analysis adopts the following correlation degree judgment criteria: When 0.7≤|γ|<1, variables f and g are highly correlated; When 0.4≤|γ|<0.7, variables f and g are moderately correlated; When 0.2≤|γ|<0.4, variables f and g are poorly correlated; When |γ|<0.2, variables f and g are extremely poorly correlated or close to zero correlation.

Images