Energy-saving control method and system for refrigeration system based on pipe network simulation

Abstract

The application discloses a kind of based on pipe network simulation's refrigeration system energy-saving control method and system, comprising: target refrigeration system is divided into working area, based on pipe characteristic and pipe resistance characteristic Construction pipe network resistance simulation formula, and combined with random forest algorithm Construction pipe network resistance prediction model;Refrigeration system operation control information is obtained, input into the pipe network resistance prediction model in pipe network resistance prediction is carried out, and according to the prediction result Energy-saving control scheme is formulated;The operation monitoring information of target refrigeration system is obtained, and system operation anomaly detection is carried out according to the operation monitoring information, and operation anomaly detection information is obtained;According to the operation anomaly detection information, abnormal cause analysis and abnormal tracing are carried out, and abnormal early warning is carried out.The energy utilization efficiency of refrigeration system is improved, the operation energy consumption thereof is reduced, and operation anomaly monitoring is carried out, the refrigeration system operation stability and energy utilization efficiency are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy-saving control of refrigeration systems, and in particular to a refrigeration system energy-saving control method and system based on pipe network simulation. BACKGROUND

[0002] The pipe characteristic curve refers to the curve of the change relationship between water pressure drop and flow rate in the pipe system. For a fixed pipe, the water pressure drop (pipe resistance) increases with the increase of the flow rate, and the relationship between the pipe resistance and the flow rate is , wherein is the pipe network resistance, is the impedance coefficient, is the water flow rate. In the refrigeration system, the pipe network mainly includes the resistance of the host side, the pipe resistance, and the resistance of the air conditioner terminal side. The host side has multiple hosts in parallel, and the air conditioner terminal side also has multiple terminals in parallel. When the total resistance of the host side resistance, the pipe resistance, and the air conditioner terminal side resistance is added, the pipe characteristic curve cannot meet the entire pipe network system and can only meet the characteristics of the pipe resistance. In the actual application of the refrigeration system, only the total resistance of the host side resistance, the pipe resistance, and the air conditioner terminal side resistance under the full load (i.e., the maximum water flow rate) is used to select the flow rate and head of the water pump, which cannot predict or judge the system resistance change under variable flow rate. Without knowing the system resistance change under variable flow rate, the water pump head change of the entire refrigeration system cannot be simulated, and the energy-saving control of the water pump cannot be accurately controlled.

[0003] At the same time, based on the simulation and prediction of the pipe network resistance of the refrigeration system, the running change of the refrigeration system can be better judged to realize the monitoring of the system running. Therefore, how to better simulate and predict the pipe network resistance of the refrigeration system to better accurately control the refrigeration system and realize energy-saving control is an important problem. SUMMARY

[0004] The present application overcomes the defects of the prior art and provides a refrigeration system energy-saving control method and system based on pipe network simulation, which aims to improve the energy utilization efficiency of the refrigeration system and realize energy-saving control.

[0005] To achieve the above purpose, the first aspect of the present application provides a refrigeration system energy-saving control method based on pipe network simulation, comprising:

[0006] dividing the target refrigeration system into working areas, constructing a pipe network resistance simulation formula based on the pipe characteristics and the pipe resistance characteristics, and constructing a pipe network resistance prediction model in combination with a random forest algorithm;

[0007] Obtaining the operation control information of the refrigeration system, inputting into the pipe network resistance prediction model to predict the pipe network resistance, and formulating the energy-saving regulation scheme according to the prediction result;

[0008] Obtaining the operation monitoring information of the target refrigeration system, detecting the system operation abnormity according to the operation monitoring information to obtain the operation abnormity detection information;

[0009] Analyzing the abnormity reason and tracing the abnormity source according to the operation abnormity detection information, and performing the abnormity early warning.

[0010] In the scheme, the target refrigeration system is divided into working areas, a pipe network resistance simulation formula is constructed based on pipe characteristics and pipe resistance characteristics, and a pipe network resistance prediction model is constructed in combination with a random forest algorithm, and specifically includes:

[0011] Obtaining the structure information of the refrigeration system, dividing the target refrigeration system into working structures, dividing the target refrigeration system into three structures of refrigeration host, cooling water system and chilled water system to obtain the working structure division information;

[0012] Obtaining the working coverage area of the target refrigeration system through the structure information of the refrigeration system, and dividing the target refrigeration system into working areas according to the working structure division information;

[0013] Dividing the working area of the target refrigeration system into multiple sub-areas according to the coverage range of the refrigeration host to obtain the working area division information;

[0014] Constructing a pipe network resistance simulation formula of the refrigeration system variable flow based on pipe characteristics and pipe resistance characteristics, and constructing a pipe network resistance prediction model according to the pipe network resistance simulation formula in combination with a random forest.

[0015] In the scheme, the pipe network resistance simulation formula of the refrigeration system variable flow based on pipe characteristics and pipe resistance characteristics specifically includes:

[0016] Pipe network resistance under refrigeration system variable flow The simulation formula is as follows: Wherein, is the maximum water pressure drop of each refrigeration unit under full flow in operation, is the water pressure drop of the pipe under full flow, is the rated water pressure drop of the air conditioner terminal or cooling tower, is the rated flow rate of the refrigeration unit under the maximum water pressure drop under full flow, is the real-time flow rate of the refrigeration unit under variable flow under the maximum water pressure drop under full flow, is the real-time flow rate of the pipe under variable flow, is the rated flow rate of the pipe under full flow.

[0017] In the scheme, the refrigeration system operation control information is obtained, input into the pipe network resistance prediction model for pipe network resistance prediction, and an energy-saving regulation scheme is formulated according to the prediction result, specifically including:

[0018] The historical operation control information of the target refrigeration system is obtained, the corresponding historical pipe network resistance is calculated according to the pipe network resistance simulation formula, the working area division information is obtained, the calculated historical pipe network resistance is associated with the corresponding working area, and an instance data set is formed;

[0019] The Markov chain is introduced for model optimization, and a training data set is constructed according to the instance data set for model training;

[0020] The state space is constructed according to the input training data, the state area of each working area is divided through the working area corresponding to each training data, and the state transition probability is calculated to generate a state transition matrix;

[0021] The state prediction result is obtained based on the state transition matrix, the state prediction result is taken as the input of the random forest algorithm, and a plurality of decision trees are generated, and the pruning operation is performed through the constraint condition formed by the state transition matrix;

[0022] The prediction results of each decision tree are obtained according to the pruning result, the pipe network resistance prediction result is obtained by weighted average, the pipe network resistance prediction result is evaluated, the model parameter optimization and adjustment are performed according to the evaluation result, and the pipe network resistance prediction model meeting the expectation is obtained;

[0023] The refrigeration system operation control information is obtained, input into the pipe network resistance prediction model for pipe network resistance prediction, and the pipe network resistance prediction information is obtained, the pipe network resistance prediction information including the pipe network resistance prediction result of each working area;

[0024] The pipe network resistance distribution diagram of each working area is generated according to the pipe network resistance prediction information, the required water pump head in each working area is calculated, and an energy-saving regulation scheme is generated, and the target refrigeration system is dynamically regulated.

[0025] In the scheme, the operation monitoring information of the target refrigeration system is obtained, and system operation anomaly detection is performed according to the operation monitoring information to obtain operation anomaly detection information, specifically including:

[0026] An operation anomaly detection model is constructed based on the isolated forest algorithm, the operation monitoring information of the target refrigeration system is obtained, and data anomaly detection is performed in the operation anomaly detection model;

[0027] The RANSAC algorithm is introduced for model optimization, a plurality of sample data are randomly selected according to the input operation monitoring information, linear fitting is performed on the plurality of sample data, the error between the predicted value and the actual value is calculated, and the model with the minimum error is selected as the final RANSAC regression model;

[0028] The fitting predicted value is obtained according to the final RANSAC regression model, the deviation between the fitting predicted value and the actual observation value is calculated, and the residual feature is generated;

[0029] The input operation monitoring information is subjected to feature extraction, the residual feature is combined to construct a feature matrix, which is input into the operation anomaly detection model for anomaly detection, and a feature space is constructed according to the input feature matrix;

[0030] The data points with low distribution density or isolation in the feature space are detected by constructing a random binary search tree, the path length of each data point in the isolation forest is calculated and the average value is obtained, whether the corresponding data point is an abnormal point is judged according to the obtained average value, and operation anomaly detection information is obtained.

[0031] In the scheme, the operation anomaly detection information is used for anomaly cause analysis and anomaly tracing, and anomaly early warning is performed, specifically including:

[0032] Various refrigeration system operation anomaly instances are obtained based on big data retrieval, feature extraction is performed on each operation anomaly instance to obtain anomaly instance features, the Pearson correlation coefficients between each anomaly instance feature and the corresponding operation anomaly instance are calculated to form a Pearson correlation coefficient matrix;

[0033] The Pearson correlation coefficient matrix is judged with a preset threshold, the correlation degree between each operation anomaly instance and each anomaly feature is analyzed according to the judgment result, and anomaly feature screening is performed to obtain anomaly feature screening information;

[0034] The Pearson correlation coefficients of each screened anomaly feature are obtained through the Pearson correlation coefficient matrix and are normalized, each screened anomaly feature is valued according to the normalization result, and an operation anomaly knowledge graph is constructed according to the anomaly instance, the screened anomaly feature and the valuation of the screened anomaly feature;

[0035] Obtain operation anomaly detection information, extract anomaly feature information from the operation anomaly detection information, and calculate the similarity between the anomaly feature information and the operation anomaly knowledge graph;

[0036] According to the similarity calculation result, the anomaly instance corresponding to the similar anomaly feature and the valuation of the similar anomaly feature are obtained, the valuations of the similar anomaly features corresponding to each anomaly instance are summed, the summation result is taken as the support degree for sorting, and the operation anomaly cause is analyzed according to the sorting result to obtain operation anomaly cause analysis information.

[0037] Abnormal tracing is performed in combination with the operation abnormality detection information and the operation abnormality cause analysis information, regional positioning is performed according to the source characteristics of the operation abnormality data, and the abnormal tracing information is obtained by combining the abnormal equipment in the corresponding region according to the operation abnormality cause analysis.

[0038] An operation abnormality report is generated according to the operation abnormality cause analysis information and the abnormal tracing information to perform operation abnormality early warning.

[0039] The second aspect of the application provides a refrigeration system energy-saving control system based on pipe network simulation, which comprises a memory and a processor, the memory contains a refrigeration system energy-saving control method based on pipe network simulation, and the refrigeration system energy-saving control method based on pipe network simulation is executed by the processor to realize the following steps: dividing the working area of a target refrigeration system, constructing a pipe network resistance simulation formula based on pipe characteristics and pipe resistance characteristics, and constructing a pipe network resistance prediction model in combination with a random forest algorithm.

[0040] Obtaining refrigeration system operation control information, inputting the refrigeration system operation control information into the pipe network resistance prediction model to perform pipe network resistance prediction, and formulating an energy-saving control scheme according to the prediction result.

[0041] Obtaining operation monitoring information of the target refrigeration system, performing system operation abnormality detection according to the operation monitoring information, and obtaining operation abnormality detection information.

[0042] Performing abnormality cause analysis and abnormal tracing according to the operation abnormality detection information, and performing abnormality early warning.

[0043] The application discloses a refrigeration system energy-saving control method and system based on pipe network simulation, which comprises the following steps: dividing the working area of a target refrigeration system, constructing a pipe network resistance simulation formula based on pipe characteristics and pipe resistance characteristics, and constructing a pipe network resistance prediction model in combination with a random forest algorithm; obtaining refrigeration system operation control information, inputting the refrigeration system operation control information into the pipe network resistance prediction model to perform pipe network resistance prediction, and formulating an energy-saving control scheme according to the prediction result; obtaining operation monitoring information of the target refrigeration system, performing system operation abnormality detection according to the operation monitoring information, and obtaining operation abnormality detection information; and performing abnormality cause analysis and abnormal tracing according to the operation abnormality detection information, and performing abnormality early warning. The energy utilization efficiency of the refrigeration system is improved, the operation energy consumption is reduced, operation abnormality monitoring is performed, and the operation stability and energy utilization efficiency of the refrigeration system are improved. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments or examples of the present application, the drawings needed to be used in the embodiments or examples will be briefly introduced. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0045] Figure 1 A pipe network simulation-based refrigeration system energy-saving control method flow chart provided by an embodiment of the present application;

[0046] Figure 2 A refrigeration system energy-saving control flow chart provided by an embodiment of the present application;

[0047] Figure 3 A pipe network simulation-based refrigeration system energy-saving control system block diagram provided by an embodiment of the present application;

[0048] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0049] In order to more clearly illustrate the above-mentioned purposes, features and advantages of the present application, the present application will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0050] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, however, the present application can also be implemented in other ways different from those described herein, therefore, the protection scope of the present application is not limited by the specific embodiments disclosed below.

[0051] Figure 1 A pipe network simulation-based refrigeration system energy-saving control method flow chart provided by an embodiment of the present application;

[0052] As shown in Figure 1 , the present application provides a pipe network simulation-based refrigeration system energy-saving control method flow chart, which comprises:

[0053] S102, the target refrigeration system is divided into working areas, a pipe network resistance simulation formula is constructed based on pipe characteristics and pipe resistance characteristics, and a pipe network resistance prediction model is constructed in combination with a random forest algorithm;

[0054] S104, the refrigeration system operation control information is acquired and input into the pipe network resistance prediction model for pipe network resistance prediction, and an energy-saving regulation and control scheme is formulated according to the prediction result;

[0055] S106, obtain operation monitoring information of the target refrigeration system, perform system operation anomaly detection according to the operation monitoring information, and obtain operation anomaly detection information;

[0056] S08, perform anomaly cause analysis and anomaly tracing according to the operation anomaly detection information, and perform anomaly early warning.

[0057] It should be noted that the present application provides a refrigeration system energy-saving control method and system based on pipe network simulation. The target refrigeration system is divided into different working areas. A variable flow pipe network resistance simulation formula of the refrigeration system is constructed based on pipe characteristics and pipe resistance characteristics, and a pipe network resistance prediction model is constructed combined with a random forest algorithm. The refrigeration system operation control information is obtained and input into the pipe network resistance prediction model for pipe network resistance prediction. The Markov chain is introduced for model optimization, the time sequence property of the pipe network resistance is considered, so that the prediction result is more accurate and close to the actual situation. Then, an energy-saving control scheme is developed according to the prediction result, the corresponding pump head is calculated, accurate control is realized, energy waste is avoided, and energy-saving control is realized. Then, the operation monitoring information of the target refrigeration system is obtained, an operation anomaly detection model is constructed according to the isolation forest algorithm for anomaly detection, residual features are introduced for detection optimization, the interference of normal operation fluctuations on anomaly detection is reduced, and the accuracy of anomaly detection is improved. Finally, anomaly cause analysis and anomaly tracing are performed and anomaly early warning is performed to ensure the stability of the refrigeration system operation, so as to avoid the increase of energy consumption caused by operation anomaly and realize energy-saving control of the refrigeration system.

[0058] Further, in a preferred embodiment of the present application, the working area of the target refrigeration system is divided, a pipe network resistance simulation formula is constructed based on pipe characteristics and pipe resistance characteristics, and a pipe network resistance prediction model is constructed combined with a random forest algorithm, which specifically includes:

[0059] Obtain refrigeration system structure information, divide the working structure of the target refrigeration system, divide the target refrigeration system into three structures of refrigeration host, cooling water system and chilled water system, and obtain working structure division information;

[0060] Obtain the working coverage area of the target refrigeration system through the refrigeration system structure information, and divide the working area of the target refrigeration system according to the working structure division information;

[0061] Divide the working area of the target refrigeration system into a plurality of sub-areas according to the coverage range of the refrigeration host, and obtain working area division information;

[0062] Construct a variable flow pipe network resistance simulation formula of the refrigeration system based on pipe characteristics and pipe resistance characteristics, and construct a pipe network resistance prediction model combined with a random forest according to the pipe network resistance simulation formula.

[0063] It should be noted that the structure information of the refrigeration system is acquired first, and the overall working structure of the target refrigeration system is divided. The target refrigeration system is divided into three structures of refrigeration host, cooling water system and chilled water system, so as to obtain the division information of the working structure. The functions of each component of the system are clarified, which lays a foundation for subsequent optimization and control. After acquiring the structure information of the refrigeration system, the working coverage area of the target refrigeration system is further acquired. According to the working structure division information, the working area of the target refrigeration system is divided, and the working area of the entire refrigeration system is divided into multiple sub-areas, so as to understand the different areas of the refrigeration system more carefully. Next, based on the pipe characteristics and pipe resistance characteristics, a variable flow pipe network resistance simulation formula of the refrigeration system is constructed, which is used to describe the resistance change in the pipe network of the refrigeration system under different flow conditions. In order to improve the accuracy and practicability of the prediction, a random forest algorithm is used to construct a pipe network resistance prediction model.

[0064] Further, in a preferred embodiment of the present application, the variable flow pipe network resistance simulation formula of the refrigeration system based on the pipe characteristics and the pipe resistance characteristics specifically comprises:

[0065] Pipe network resistance under variable flow of refrigeration system The simulation formula is as follows: Wherein, is the maximum water pressure drop of each refrigeration unit under full flow during operation, is the water pressure drop of the pipe under full flow, is the rated water pressure drop of the air conditioning terminal or cooling tower, is the rated flow rate of the refrigeration unit with the maximum water pressure drop under full flow, is the real-time flow rate of the refrigeration unit with the maximum water pressure drop under variable flow under full flow, is the real-time flow rate of the pipe under variable flow, is the rated flow rate of the pipe under full flow.

[0066] It should be noted that the derivation process of the variable flow pipe network resistance simulation formula of the refrigeration system is as follows:

[0067] In the refrigeration system, the pipe network is mainly divided into two parts, one is the chilled water system, mainly composed of three parts of host side resistance, pipe resistance and air conditioning terminal side resistance; the other is the cooling water system, mainly composed of three parts of host side resistance, pipe resistance and cooling tower side resistance (mainly tower height); the resistance composition of the two pipe networks is similar, and in the whole system application, the air conditioning terminal resistance in the chilled water system and the resistance of the cooling tower do not change with the change of the system flow, which can be regarded as a constant value. The resistance of the pipe conforms to the change characteristics of , and the resistance of the host side is related to the number of hosts, is the pipe network resistance, is the impedance coefficient, For water flow.

[0068] In the entire pipeline, the pipeline resistance is divided into two parts. One is the resistance along the pipeline, also known as friction resistance, and its calculation formula is: in, is the resistance along the way, the unit is Pa; is the friction resistance coefficient, in m; is the length of the straight pipe section, in m; is the pipe diameter, in m; is the density of water, in kg / m3; is the water flow rate in m / s.

[0069] In the system, since the material, diameter and fluid medium of the pipeline remain unchanged, the formula for the resistance along the pipeline can be simplified as follows: in is a fixed coefficient.

[0070] The other part of pipeline resistance is local resistance. When water flows through a pipeline and encounters various pipe fittings such as elbows, tees, valves, etc., the energy loss caused by friction and eddy current is called local resistance. The calculation formula is: in, is the local resistance coefficient of the pipe fittings;

[0071] Because the local resistance coefficient is also a fixed coefficient for the corresponding components in the system, the local resistance formula can be simplified to in, is a fixed coefficient.

[0072] In the refrigeration system, the resistance of the chilled water network or cooling water network on the host side is a local resistance, so the resistance of the host is , For the host, it is a fixed coefficient; the pipeline resistance is , For the pipeline, it is a fixed coefficient; since the resistance of the air conditioning terminal or cooling tower is , is a fixed value, and express.

[0073] Therefore, the total resistance of the entire chilled water network or cooling water network is .

[0074] Based on the above content, the final formula is derived:

[0075] There are n parallel refrigeration units, and their pipe diameters are , full load flow is , the flow rate under full load flow is , the resistance of each host under full flow is (i.e. ); the pipe diameter of the pipeline is , the full load flow is , the flow rate under full load flow is , the resistance of the pipeline under full flow is (i.e. ); at this time, for each parallel branch of the host, the corresponding pipe network resistance is: wherein, The maximum value in is the most unfavorable resistance of the pipe network, and this value is the basis for selecting the pump head.

[0076] Further, when the pipe network of the refrigeration system is in variable flow, the flow rate of each host is , the resistance of each host is , according to , , ;

[0077] When the pipe network of the refrigeration system is in variable flow, the flow rate of the pipe network is , the resistance of the pipe network is , according to , , .

[0078] Therefore, for each parallel branch of the host, the corresponding pipe network resistance is , i.e. ; wherein, The maximum value in is the most unfavorable resistance of the pipe network, and this value is the basis for selecting the pump head.

[0079] As described above, the pipe network resistance simulation formula of the variable flow of the refrigeration system is obtained as wherein, is the maximum water pressure drop of each refrigeration unit under full flow during operation, is the water pressure drop of the pipeline under full flow, is the rated water pressure drop of the air conditioning terminal or cooling tower, is the rated flow rate of the refrigeration unit under the maximum water pressure drop under full flow, is the real-time flow rate of the refrigeration unit under variable flow under the maximum water pressure drop under full flow, is the real-time flow rate of the pipeline under variable flow, is the rated flow rate of the pipeline under full flow.

[0080] Further, in a preferred embodiment of the present application, the refrigeration system operation control information is obtained, input into the pipe network resistance prediction model for pipe network resistance prediction, and an energy-saving regulation scheme is formulated according to the prediction result, specifically comprising:

[0081] The historical operation control information of the target refrigeration system is obtained, the corresponding historical pipe network resistance is calculated according to the pipe network resistance simulation formula, the working area division information is obtained, the calculated historical pipe network resistance is associated with the corresponding working area, and an example data set is constructed;

[0082] A Markov chain is introduced for model optimization, a training data set is constructed according to the example data set for model training;

[0083] The state space is constructed according to the input training data, the state area of each working area is divided through the working area corresponding to each training data, the state transition probability is calculated, and a state transition matrix is generated;

[0084] The state prediction result is obtained based on the state transition matrix, the state prediction result is taken as the input of the random forest algorithm and a plurality of decision trees are generated, and the pruning operation is performed through the constraint condition constituted by the state transition matrix;

[0085] The prediction results of each decision tree are obtained according to the pruning result, the pipe network resistance prediction result is obtained by weighted average, the pipe network resistance prediction result is evaluated, the model parameter optimization and adjustment are performed according to the evaluation result, and a pipe network resistance prediction model meeting the expectation is obtained;

[0086] The refrigeration system operation control information is obtained, input into the pipe network resistance prediction model for pipe network resistance prediction, and a pipe network resistance prediction information is obtained, the pipe network resistance prediction information comprising the pipe network resistance prediction result of each working area;

[0087] The pipe network resistance distribution map of each working area is generated according to the pipe network resistance prediction information, the required water pump head in each working area is calculated, an energy-saving regulation scheme is formulated, and the target refrigeration system is dynamically regulated.

[0088] It should be noted that, first, the historical operation control information of the target refrigeration system is acquired, and the corresponding historical pipe network resistance is calculated according to the pipe network resistance simulation formula. Then, combined with the division information of the working area, the calculated historical pipe network resistance is associated with the corresponding working area to form an instance data set containing the corresponding relationship between the pipe network resistance in the historical operation and each working area. A Markov chain is introduced for model optimization. The training data set is constructed by using the instance data set, and the model is trained. By inputting the training data, a state space is constructed, and each working area is divided into different state areas according to the working area corresponding to each training data. At the same time, the state transition probability of each state area is calculated, and a state transition matrix is generated. Based on the state transition matrix, the state prediction result is obtained as the input of the random forest algorithm, and a plurality of decision trees are generated. By constructing the constraint condition through the state transition matrix, the prediction result that does not conform to the state transition matrix is punished, the invalid result is removed, and the decision tree is pruned to simplify the model and improve the prediction accuracy and efficiency. Finally, the operation control information of the refrigeration system is acquired, which is input into the optimized pipe network resistance prediction model to predict the pipe network resistance, and the pipe network resistance prediction result of each working area is obtained. Based on the pipe network resistance prediction information, the pipe network resistance distribution diagram of each working area is generated, and the required water pump head in each working area is calculated to generate an energy-saving control scheme to dynamically control the target refrigeration system to achieve the energy-saving goal of the system, thereby realizing the energy-saving control of the refrigeration system.

[0089] Further, in a preferred embodiment of the present application, the operation monitoring information of the target refrigeration system is acquired, and system operation anomaly detection is performed according to the operation monitoring information to obtain operation anomaly detection information, specifically comprising:

[0090] An operation anomaly detection model is constructed based on the isolated forest algorithm, the operation monitoring information of the target refrigeration system is input into the operation anomaly detection model for data anomaly detection;

[0091] A RANSAC algorithm is introduced for model optimization, a plurality of sample data are randomly selected according to the input operation monitoring information, linear fitting is performed on the selected sample data, the error between the predicted value and the actual value is calculated, and the model with the smallest error is selected as the final RANSAC regression model;

[0092] The fitting predicted value is obtained according to the final RANSAC regression model, the deviation between the fitting predicted value and the actual observation value is calculated, and a residual feature is generated;

[0093] The input operation monitoring information is feature extracted, a feature matrix is constructed in combination with the residual feature, and is input into the operation anomaly detection model for anomaly detection, and a feature space is constructed according to the input feature matrix;

[0094] The running abnormality detection information is obtained by constructing a random binary search tree to detect data points with low distribution density or isolation in a feature space, calculating path lengths of each data point in the isolation forest, and obtaining an average value.

[0095] It should be noted that, first, in the construction of the running abnormality detection model based on the isolation forest algorithm, the isolation forest algorithm can effectively identify data points with low distribution density or isolation by constructing multiple random binary search trees, thereby detecting potential abnormal points. The RANSAC algorithm is introduced, and the random sample consensus method is used to randomly select a plurality of sample data from the input running monitoring information for linear fitting. By calculating the error between the predicted value of the fitting model and the actual value, the model with the smallest error is selected as the final RANSAC regression model. The robustness and accuracy of the model for detecting abnormal data are improved. The final RANSAC regression model is used to fit the input monitoring information to obtain a fitting predicted value. Then, the deviation between the fitting predicted value and the actual observed value is calculated to generate residual features. The residual features reflect the deviation degree of the model prediction from the actual situation and are an important indicator for identifying abnormalities. Next, feature extraction is performed on the input running monitoring information, and a feature matrix is constructed in combination with the generated residual features. The feature matrix is input into the running abnormality detection model for further abnormality detection. In the feature space, a random binary search tree is constructed to detect data points with low distribution density or isolation. The isolation forest algorithm calculates the path length of each data point in the forest and obtains its average value. Data points with short path lengths are usually located in high-density areas, while data points with long path lengths may be abnormal points. According to the average value of these path lengths, it can be determined whether the corresponding data point is an abnormal point, thereby obtaining the running abnormality detection information of the refrigeration system.

[0096] Further, in a preferred embodiment of the present application, the abnormality cause analysis and abnormality tracing based on the running abnormality detection information, and the abnormality early warning specifically include:

[0097] Based on big data retrieval, various refrigeration system running abnormality instances are obtained, feature extraction is performed on each running abnormality instance to obtain abnormality instance features, the Pearson correlation coefficients between each abnormality instance feature and the corresponding running abnormality instance are calculated to form a Pearson correlation coefficient matrix;

[0098] The Pearson correlation coefficient matrix is judged with a preset threshold, the correlation degree between each running abnormality instance and each abnormality feature is analyzed according to the judgment result, and abnormality feature screening is performed to obtain abnormality feature screening information;

[0099] The Pearson correlation coefficients of the screened abnormal features are obtained through a Pearson correlation coefficient matrix and normalized, the screened abnormal features are valued according to the normalized results, and an operation abnormal knowledge graph is constructed according to the abnormal instances, the screened abnormal features and the values of the screened abnormal features;

[0100] Obtaining operation abnormality detection information, performing feature extraction on the operation abnormality detection information to obtain abnormal feature information, and performing similarity calculation on the abnormal feature information and the operation abnormality knowledge graph;

[0101] According to the similarity calculation result, the abnormal instances corresponding to the similar abnormal features and the values of the similar abnormal features are obtained, the values of the similar abnormal features corresponding to each abnormal instance are summed, the sum result is taken as the support degree for sorting, the operation abnormality reason is analyzed according to the sorting result, and operation abnormality reason analysis information is obtained;

[0102] Abnormal tracing is performed in combination with the operation abnormality detection information and the operation abnormality reason analysis information, regional positioning is performed according to the source characteristics of the operation abnormality data, and abnormal tracing information is obtained in combination with the abnormal equipment in the region corresponding to the operation abnormality reason analysis.

[0103] An operation abnormality report is generated according to the operation abnormality reason analysis information and the abnormal tracing information for operation abnormality early warning.

[0104] It should be noted that after the abnormality detection, the running abnormality detection information is obtained, the abnormality reason analysis is performed according to the detected abnormal data, various refrigeration system running abnormality instances are obtained through big data retrieval, and the characteristics of each abnormality instance are obtained. Next, the Pearson correlation coefficient between each abnormality instance characteristic and the corresponding running abnormality instance is calculated to form a Pearson correlation coefficient matrix. Then, the Pearson correlation coefficient matrix is compared with a preset threshold value, the correlation degree between each running abnormality instance and each abnormality characteristic is analyzed according to the judgment result, and the abnormality characteristic is screened to obtain abnormality characteristic screening information. The screened abnormality characteristic retains the characteristics highly related to the abnormality instance, which helps to improve the accuracy of subsequent analysis. The Pearson correlation coefficient of each screened abnormality characteristic is obtained through the Pearson correlation coefficient matrix and is normalized. According to the normalization result, each screened abnormality characteristic is valued, and an operation abnormality knowledge graph is constructed according to the abnormality instance, the screened abnormality characteristic and the value thereof. Next, the corresponding abnormality characteristic is obtained from the operation abnormality detection information, the similarity calculation is performed with the operation abnormality knowledge graph, the abnormality instance corresponding to the similar abnormality characteristic and the value of the similar abnormality characteristic are obtained, the values of the similar abnormality characteristics corresponding to each abnormality instance are summed, and each abnormality instance is sorted according to the sorting result. According to the sorting result, the corresponding operation abnormality reason is analyzed. The abnormality is traced according to the operation abnormality detection information and the operation abnormality reason analysis information, the source characteristics of the operation abnormality data are analyzed for regional positioning, and the abnormal equipment in the corresponding region is combined with the operation abnormality reason analysis, and finally the abnormality tracing information is obtained. Finally, the operation abnormality report is generated according to the operation abnormality reason analysis information and the abnormality tracing information, the operation abnormality warning is performed, the energy consumption caused by abnormal operation is avoided, the economic loss is caused, and thus the energy saving control of the target refrigeration system is realized.

[0105] Figure 2 A refrigeration system energy saving control flowchart is provided for an embodiment of the present application;

[0106] As Figure 2 shown, the present application provides a refrigeration system energy saving control flowchart, which comprises:

[0107] S202, obtain refrigeration system running control information, input into a pipe network resistance prediction model for pipe network resistance prediction, and obtain pipe network resistance prediction information;

[0108] S204, generate a pipe network resistance distribution map of each working area according to the pipe network resistance prediction information, and calculate the required water pump head in each working area;

[0109] S206, generate a control scheme according to the calculated water pump head, and adjust the scheme according to the real-time change of pipe network resistance, and perform refrigeration system energy saving control;

[0110] S208, obtain the operation monitoring information of the target refrigeration system, input into the operation anomaly detection model for data anomaly detection, and obtain operation anomaly detection information;

[0111] S210, perform anomaly cause analysis and anomaly tracing according to the operation anomaly detection information, and generate an operation anomaly report for early warning.

[0112] In addition, in the refrigeration system energy-saving control method based on pipe network simulation provided by the application, the following steps are further included:

[0113] The historical energy consumption curve of the target refrigeration system is analyzed to obtain real-time energy consumption monitoring information, and energy transaction analysis and energy leakage are performed.

[0114] Obtain historical energy use information of the target refrigeration system, and construct a historical energy consumption curve of the target refrigeration system according to the historical energy use information;

[0115] Based on the constructed historical energy consumption curve, a total energy consumption portrait of the target refrigeration system is constructed, working area division information is obtained, the constructed energy consumption portrait is associated with each working area, and a regional energy consumption portrait is obtained;

[0116] According to the constructed total energy consumption portrait and regional energy consumption portrait, an energy consumption database is constructed, real-time energy consumption monitoring information is obtained, and similarity calculation is performed with the energy consumption portrait in the energy consumption database;

[0117] A preset judgment threshold is set, the calculated similarity value is compared with the judgment threshold, and whether the target refrigeration system energy consumption at the current time is abnormal is analyzed according to the judgment result, and energy consumption anomaly analysis information is obtained;

[0118] An energy consumption anomaly detection model is constructed, the real-time energy consumption monitoring information is input for anomaly detection, and energy consumption anomaly detection information is obtained;

[0119] According to the energy consumption anomaly detection information, abnormal data is extracted, the reference value corresponding to the abnormal data is obtained through the energy consumption database, the deviation between the abnormal data and the reference value is calculated, and deviation value information is obtained;

[0120] According to the energy consumption anomaly detection information, abnormal area tracing is performed, the deviation value information is combined to generate an energy leakage report, and the energy delivery of the abnormal area is controlled to perform energy leakage early warning.

[0121] It should be noted that based on the energy consumption habit of the target refrigeration system, whether the energy consumption is abnormal is judged from the periodicity, the increase of energy consumption caused by pipe leakage and other reasons is avoided, the energy utilization efficiency of the refrigeration system is reduced, and energy-saving control is realized.

[0122] Figure 3 A pipe network simulation-based refrigeration system energy-saving control system 3 is provided for an embodiment of the present application, and the system comprises a memory 31 and a processor 32, the memory 31 contains a pipe network simulation-based refrigeration system energy-saving control method program, and the pipe network simulation-based refrigeration system energy-saving control method program is executed by the processor 32 to implement the following steps:

[0123] The target refrigeration system is divided into working areas, a pipe network resistance simulation formula is constructed based on pipe characteristics and pipe resistance characteristics, and a pipe network resistance prediction model is constructed in combination with a random forest algorithm;

[0124] Obtain refrigeration system operation control information, input the information into the pipe network resistance prediction model to predict pipe network resistance, and develop an energy-saving control scheme according to the prediction result;

[0125] Obtain operation monitoring information of the target refrigeration system, perform system operation anomaly detection according to the operation monitoring information, and obtain operation anomaly detection information;

[0126] Perform anomaly cause analysis and anomaly tracing according to the operation anomaly detection information, and perform anomaly early warning.

[0127] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the various components shown or discussed can be through some interface, indirect coupling or communication connection of the devices or units, which can be electrical, mechanical or other forms.

[0128] The units described above as separate components can or can not be physically separated, and the components shown as units can or can not be physical units; they can be located in one place or distributed on multiple network units; some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0129] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be a unit alone, or two or more units can be integrated into one unit; the integrated unit can be realized in the form of hardware or in the form of hardware plus software functional unit.

[0130] Those skilled in the art can understand that all or part of the steps of the above-mentioned method embodiments can be completed by program instruction related hardware, the foregoing program can be stored in a computer readable storage medium, and the program executes the steps of the method embodiments when executed; and the foregoing storage medium includes a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disc or an optical disc, and various storage medium capable of storing program codes.

[0131] Alternatively, the integrated unit of the present application can be stored in a computer readable storage medium if it is realized in the form of a software function module and sold or used as an independent product. Based on such understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, and the computer software product is stored in a storage medium, includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the embodiments of the present application. The foregoing storage medium includes a mobile storage device, a ROM, a RAM, a magnetic disc or an optical disc, and various storage medium capable of storing program codes.

[0132] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A refrigeration system energy-saving control method based on pipe network simulation, characterized in that: include: The target refrigeration system is divided into working areas. A pipeline network resistance simulation formula is constructed based on pipeline characteristics and pipeline resistance characteristics. A pipeline network resistance prediction model is constructed in combination with the random forest algorithm. Obtaining refrigeration system operation control information, inputting it into the pipe network resistance prediction model to perform pipe network resistance prediction, and formulating an energy-saving control plan based on the prediction results; Acquiring operation monitoring information of a target refrigeration system, performing system operation abnormality detection based on the operation monitoring information, and obtaining operation abnormality detection information; Analyze the cause of the abnormality and trace the source of the abnormality based on the abnormal operation detection information, and issue an abnormality warning; The acquisition of refrigeration system operation control information, inputting it into the pipe network resistance prediction model to perform pipe network resistance prediction, and formulating an energy-saving control plan based on the prediction results, specifically includes: Obtain historical operation control information of the target refrigeration system, calculate the corresponding historical pipe network resistance according to the pipe network resistance simulation formula, obtain work area division information, and associate the calculated historical pipe network resistance with the corresponding work area to form an instance data set; Introducing a Markov chain to perform model optimization, and constructing a training data set based on the example data set to perform model training; Construct a state space based on the input training data, divide the state area of ​​each working area by the working area corresponding to each training data, calculate the state transition probability, and generate a state transition matrix; The state prediction is performed based on the state transition matrix to obtain the state prediction results, the state prediction results are used as the input of the random forest algorithm to generate several decision trees, and the pruning operation is performed by constraining the state transition matrix; The prediction results of each decision tree are obtained based on the pruning results, and the weighted average is performed to obtain the pipeline network resistance prediction result. The pipeline network resistance prediction result is evaluated, and the model parameters are optimized and adjusted based on the evaluation results to obtain a pipeline network resistance prediction model that meets the expectations; Acquire refrigeration system operation control information, input it into the pipe network resistance prediction model to perform pipe network resistance prediction, and obtain pipe network resistance prediction information, wherein the pipe network resistance prediction information includes pipe network resistance prediction results for each working area; Generate a pipe network resistance distribution map for each working area based on the pipe network resistance prediction information, calculate the required water pump head in each working area to generate an energy-saving control plan, and dynamically control the target refrigeration system; The target refrigeration system is divided into working areas, a pipeline network resistance simulation formula is constructed based on pipeline characteristics and pipeline resistance characteristics, and a pipeline network resistance prediction model is constructed in combination with a random forest algorithm, specifically including: Obtaining refrigeration system structure information, dividing the target refrigeration system into three types of structures: refrigeration host, cooling water system, and chilled water system, and obtaining working structure division information; Obtaining a working coverage area of ​​a target refrigeration system through the refrigeration system structure information, and dividing the working area of ​​the target refrigeration system according to the working structure division information; Divide the working area of ​​the target refrigeration system into multiple sub-areas according to the coverage of the refrigeration host, and obtain working area division information; Based on the pipeline characteristics and pipeline resistance characteristics, a pipeline network resistance simulation formula for a variable flow refrigeration system is constructed, and a pipeline network resistance prediction model is constructed based on the pipeline network resistance simulation formula and combined with random forest; The pipeline network resistance simulation formula for building a variable flow refrigeration system based on pipeline characteristics and pipeline resistance characteristics specifically includes: Pipeline network resistance under variable flow in refrigeration system The simulation formula is as follows: , in, is the maximum water pressure drop of each refrigeration unit at full flow during operation, is the water pressure drop in the pipeline at full flow, is the rated water pressure drop of the air conditioning terminal or cooling tower, is the maximum water pressure drop at full flow, the rated flow rate of the refrigeration unit, The maximum water pressure drop at full flow is the real-time flow rate of the refrigeration unit under variable flow, is the real-time flow velocity of the pipeline under variable flow, It is the rated flow rate of the pipeline at full flow.

2. The refrigeration system energy-saving control method based on pipe network simulation according to claim 1 is characterized in that: The acquiring of the operation monitoring information of the target refrigeration system, performing system operation abnormality detection based on the operation monitoring information, and obtaining operation abnormality detection information specifically includes: An operation anomaly detection model is constructed based on the isolation forest algorithm to obtain the operation monitoring information of the target refrigeration system and input it into the operation anomaly detection model to perform data anomaly detection; The RANSAC algorithm is introduced to optimize the model. Several sample data are randomly selected according to the input operation monitoring information. Linear fitting is performed on the selected sample data. The error between the predicted value and the actual value is calculated. The model with the smallest error is selected as the final RANSAC regression model. Obtaining a fitted prediction value based on the final RANSAC regression model, calculating the deviation between the fitted prediction value and the actual observed value, and generating a residual feature; Extracting features from the input operation monitoring information, constructing a feature matrix based on the residual features, inputting the feature matrix into the operation anomaly detection model for anomaly detection, and constructing a feature space based on the input feature matrix; By constructing a random binary search tree to detect data points with low distribution density or isolated in the feature space, the path length of each data point in the isolation forest is calculated and the average value is obtained. Based on the obtained average value, it is determined whether the corresponding data point is an outlier, and the operation anomaly detection information is obtained.

3. The energy-saving control method for a refrigeration system based on pipe network simulation according to claim 1, characterized in that: The abnormality cause analysis and abnormality tracing based on the operation abnormality detection information and abnormality warning are specifically performed as follows: Based on big data retrieval, various refrigeration system operation abnormality instances are obtained, feature extraction is performed on each operation abnormality instance to obtain abnormal instance features, and the Pearson correlation coefficient between each abnormal instance feature and the corresponding operation abnormality instance is calculated to form a Pearson correlation coefficient matrix; The Pearson correlation coefficient matrix is ​​judged against a preset threshold, and the correlation between each abnormal operation instance and each abnormal feature is analyzed according to the judgment result, and abnormal feature screening is performed to obtain abnormal feature screening information; Obtain the Pearson correlation coefficient of each screened abnormal feature through the Pearson correlation coefficient matrix and normalize it. Assign a value to each screened abnormal feature based on the normalization result. Construct an operation abnormality knowledge graph based on the abnormal instance, the screened abnormal feature, and the assigned value of the screened abnormal feature. Acquire operation anomaly detection information, perform feature extraction on the operation anomaly detection information to obtain anomaly feature information, and perform similarity calculation between the anomaly feature information and the operation anomaly knowledge graph; Obtain abnormal instances corresponding to similar abnormal features and values ​​assigned to similar abnormal features based on the similarity calculation results, sum the values ​​assigned to similar abnormal features corresponding to each abnormal instance, use the summed results as support for sorting, analyze the cause of the operation abnormality based on the sorting results, and obtain analysis information on the cause of the operation abnormality; Combine the abnormality detection information and the abnormality cause analysis information to trace the abnormality, locate the area according to the source characteristics of the abnormality data, analyze the abnormal equipment in the corresponding area in combination with the abnormality cause, and obtain abnormality tracing information; Generate an operation abnormality report based on the operation abnormality cause analysis information and abnormality tracing information to provide operation abnormality warning.

4. A refrigeration system energy-saving control system based on pipe network simulation, characterized in that: The system includes: a memory and a processor, wherein the memory contains a refrigeration system energy-saving control method program based on pipe network simulation, and when the refrigeration system energy-saving control method program based on pipe network simulation is executed by the processor, the steps of a refrigeration system energy-saving control method based on pipe network simulation as described in claims 1-3 are implemented.

Images