CN117091273A - Control method and device for self-adaptive learning data model of central air conditioning system - Google Patents

Control method and device for self-adaptive learning data model of central air conditioning system Download PDF

Info

Publication number
CN117091273A
CN117091273A CN202311079805.5A CN202311079805A CN117091273A CN 117091273 A CN117091273 A CN 117091273A CN 202311079805 A CN202311079805 A CN 202311079805A CN 117091273 A CN117091273 A CN 117091273A
Authority
CN
China
Prior art keywords
data model
neural network
network data
air conditioning
central air
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202311079805.5A
Other languages
Chinese (zh)
Inventor
许泽锋
李维娜
凌靖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Mushroom Iot Technology Xiamen Co ltd
Original Assignee
Mushroom Iot Technology Xiamen Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Mushroom Iot Technology Xiamen Co ltd filed Critical Mushroom Iot Technology Xiamen Co ltd
Priority to CN202311079805.5A priority Critical patent/CN117091273A/en
Publication of CN117091273A publication Critical patent/CN117091273A/en
Pending legal-status Critical Current

Links

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F11/00—Control or safety arrangements
    • F24F11/89—Arrangement or mounting of control or safety devices
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F11/00—Control or safety arrangements
    • F24F11/62—Control or safety arrangements characterised by the type of control or by internal processing, e.g. using fuzzy logic, adaptive control or estimation of values
    • F24F11/63—Electronic processing
    • F24F11/64—Electronic processing using pre-stored data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Abstract

The invention provides a control method and a device for a self-adaptive learning data model of a central air-conditioning system. A control method of a self-adaptive learning data model of a central air conditioning system comprises the following steps: s1, monitoring and collecting data of each key point in a central air conditioning system; s2, preprocessing the acquired data to obtain a data set; s3, selecting and determining the structure of a neural network data model according to the complexity of the task and the characteristics of the data; s4, training the neural network data model by using a data set to obtain a trained neural network data model; s5, evaluating the trained neural network data model by using a test set, calculating an evaluation index, obtaining an evaluation result, and judging whether the evaluation result meets the requirement; s6, if the evaluation result does not meet the requirements, optimizing the neural network data model according to the evaluation result, and repeating S5; and S7, if the evaluation result meets the requirement, updating the neural network data model and applying the neural network data model to the central air conditioning system.

Description

Control method and device for self-adaptive learning data model of central air conditioning system
Technical Field
The invention relates to the field of central air conditioning systems, in particular to a control method and a device for a self-adaptive learning data model of a central air conditioning system.
Background
The design of a central air conditioning system usually calculates its maximum cooling load in terms of the extreme climatic conditions of the building site and determines therefrom the installed capacity of the air conditioning main unit and its corresponding chilled water pump and cooling water pump capacities. In fact, the heating, ventilation and air conditioning refrigeration system is operated under partial load (much smaller than its rated capacity) for most of the time, which undoubtedly results in a great amount of wasted energy. On the other hand, the air conditioning load has variability. The load of the heating, ventilation and air conditioning refrigerating system has the characteristics of fluctuation and inconstancy due to the influence of various factors such as season alternation, climate change, round-the-clock rotation, use change, production load and the like, so the method can provide the central air conditioning system to effectively adjust along with the change of the end load and has advanced predictability.
The control mode of the current central air conditioning system in the aspect of self control is mainly PID control, and the control mode has the following disadvantages facing the multi-coupling system:
1. parameter adjustment is difficult: the performance of the PID control algorithm is highly dependent on the adjustment of the parameters. Determining the appropriate PID parameters requires actual system testing and debugging, which can take a significant amount of time and effort.
2. The adaptability is poor: PID control algorithms are typically based on fixed parameters and for conditions of large system dynamics, PID control may not be able to accommodate these changes, resulting in reduced control performance.
3. The response speed is limited: the response speed of the PID control algorithm to system disturbances is relatively slow, and more complex control algorithms may be required to increase the response speed, especially when there is a large nonlinearity and time-variability in the system.
4. The robustness is poor: the PID control algorithm is sensitive to the change and disturbance of the system parameters, and once the system parameters change, the performance of PID control can be influenced, and parameter adjustment needs to be performed in time.
5. It is difficult to cope with complex systems: for complex central air conditioning systems, PID control algorithms may not provide adequate control performance. In the face of nonlinear, multivariable, time-lapse, and other complex systems, more advanced control algorithms need to be employed.
Disclosure of Invention
The invention aims to provide a control method and a device for a self-adaptive learning data model of a central air conditioning system.
The invention aims to solve the problems in the prior art.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
the invention discloses a first aspect, which provides a control method of a self-adaptive learning data model of a central air-conditioning system, comprising the following steps: s1, monitoring and collecting data of each key point in a central air conditioning system; s2, preprocessing the acquired data to obtain a data set; s3, selecting and determining the structure of a neural network data model according to the complexity of the task and the characteristics of the data; s4, training the neural network data model by using the data set, and adjusting the weight and bias of the neural network data model by using an optimization algorithm until the error between the predicted result and the actual result is minimized, so as to obtain a trained neural network data model; s5, evaluating the trained neural network data model by using a test set, calculating an evaluation index, obtaining an evaluation result, and judging whether the evaluation result meets the requirement; s6, if the evaluation result does not meet the requirements, optimizing the neural network data model according to the evaluation result, and repeating S5; and S7, if the evaluation result meets the requirement, updating the neural network data model and applying the neural network data model to a central air conditioning system.
In a second aspect of the present disclosure, a control device for a central air conditioning system adaptive learning data model is provided, including: the acquisition module is configured to monitor and acquire data of each key point in the central air conditioning system; the preprocessing module is configured to preprocess the acquired data to obtain a data set; the training module is configured to select and determine the structure of the neural network data model according to the complexity of the task and the characteristics of the data; training the neural network data model by using the data set, and adjusting the weight and bias of the neural network data model by using an optimization algorithm until the error between the predicted result and the actual result is minimized, so as to obtain a trained neural network data model; the evaluation module is configured to evaluate the trained neural network data model by using the test set, calculate an evaluation index and obtain an evaluation result, and judge whether the evaluation result meets the requirement; if the evaluation result does not meet the requirements, optimizing the neural network data model according to the evaluation result; and the updating module is configured to update and apply the neural network data model to a central air conditioning system if the evaluation result meets the requirement.
The beneficial effects of the invention are as follows:
the invention realizes the self-adaptive learning of the neural network data model, optimizes the energy efficiency of the central air conditioning system, and realizes more accurate control and adjustment by learning the historical data and the real-time environment information so as to minimize the energy consumption and improve the energy efficiency; through self-adaptive learning neural network data model, can predict and satisfy user's comfortable demand more accurately, ensure parameters such as indoor temperature, humidity in ideal scope, provide better travelling comfort experience.
Drawings
Fig. 1 is a schematic diagram of a control method of a self-adaptive learning data model of a central air conditioning system according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of a control device of a central air conditioning system adaptive learning data model according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, based on the embodiments of the invention, which are apparent to those of ordinary skill in the art without inventive faculty, are intended to be within the scope of the invention.
In the description of the present invention, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
The invention takes energy consumption data as a control model core index preferentially, and records working condition environment simultaneously: temperature, pressure and outdoor temperature and humidity, and adjusting the running frequency of the circulating water pump, the running frequency of the tower fan and the outlet water temperature setting parameters of the refrigeration host.
The system self-learns the energy consumption modes and the efficiency under different environmental conditions, and predicts the optimal energy consumption adjustment strategy according to the real-time data and the environmental parameters, so that the reduction of the energy consumption is realized.
The real-time data and environmental parameters include: chilled water outlet temperature, chilled water return temperature, cooling water outlet temperature, cooling water return temperature, chilled water outlet pressure, chilled return pressure, outdoor temperature value and outdoor humidity value;
the energy consumption data includes: energy consumption of the circulating water pump, energy consumption of a cooling tower fan and energy consumption of a refrigerating machine;
the adjustment data includes: the operation frequency of the freezing/cooling water pump, the operation frequency of the fan of the cooling tower and the outlet water temperature set value of the refrigeration host;
the establishment of the neural network data model of the central air conditioning system is realized through the following three steps.
And (3) data acquisition:
1. temperature data: temperature data of each key point in the central air conditioning system is monitored, including condenser outlet temperature, evaporator outlet temperature, chilled water supply return water temperature and the like.
2. Pressure data: pressure data in the central air conditioning system, such as suction pressure, discharge pressure, cooling water supply and return water pressure and the like of the refrigerant compressor are recorded.
3. Flow data: flow data in the central air conditioning system, such as cooling water flow, chilled water flow, etc., are tracked.
4. Energy consumption data: the energy consumption of the central air conditioning system, including power consumption, cooling water consumption, etc., is monitored in order to evaluate the energy utilization efficiency of the system.
5. Operation mode data: and recording the running modes of the central air conditioning system, such as the running modes of the big and small machines, the number of the machines and the like, so as to adjust the running modes according to the needs.
6. Fault alarm data: and monitoring fault alarm information of the central air conditioning system, and timely finding and solving possible problems to ensure normal operation of the system.
Modeling data:
1. data cleaning and pretreatment: the data is preprocessed, including data cleaning, feature selection, feature scaling, data balancing, etc., to ensure the quality and applicability of the data.
2. And (3) network architecture design: a suitable neural network architecture is selected, including the number of layers of the network, the number of neurons per layer, the activation function, etc. The structure of the network is determined according to the complexity of the task and the characteristics of the data.
3. Model training: the neural network is trained using the prepared data set. In the training process, the weight and bias of the network are adjusted through an optimization algorithm to minimize the error between the predicted result and the actual result.
4. Model evaluation: and evaluating the trained model by using the test set, and calculating an evaluation index to evaluate the performance and generalization capability of the model.
5. Model optimization: and optimizing the model according to the evaluation result. Attempts may be made to improve model performance by adjusting network architecture, adjusting training parameters, adding regularization, etc.
6. Model application: the new unknown data is predicted or classified using the trained model. And deploying the model into practical application to perform real-time prediction or decision.
And (3) data packaging:
1. defining input and output interfaces: and determining the input and output interfaces of the model according to the data processing result.
2. Packaging data processing code: the data processing code and associated functional code are packaged into a single data processing module or function.
3. Adding pretreatment and post-treatment steps: and adding preprocessing and post-processing steps into the model package according to the requirement, so as to ensure that the input data meets the requirement of the model, and reasonably post-processing the output result.
4. Testing and verifying: and testing and verifying the data packaging module to ensure that the input data can be correctly processed and generate a reasonable output result.
Referring to fig. 1, a control method for a central air conditioning system adaptive learning data model includes:
s1, monitoring and collecting data of each key point in a central air conditioning system;
s2, preprocessing the acquired data to obtain a data set;
s3, selecting and determining the structure of a neural network data model according to the complexity of the task and the characteristics of the data;
s4, training the neural network data model by using the data set, and adjusting the weight and bias of the neural network data model by using an optimization algorithm until the error between the predicted result and the actual result is minimized, so as to obtain a trained neural network data model;
s5, evaluating the trained neural network data model by using a test set, calculating an evaluation index, obtaining an evaluation result, and judging whether the evaluation result meets the requirement;
s6, if the evaluation result does not meet the requirements, optimizing the neural network data model according to the evaluation result, and repeating S5;
and S7, if the evaluation result meets the requirement, updating the neural network data model and applying the neural network data model to a central air conditioning system.
In the step S1, data preprocessing is performed, and relevant operation parameters of a system are acquired through data acquisition; assume that temperature data 27℃and pressure data 2.1bar and flow data 1100m are collected 3 And/h, extracting features to obtain an input vector x= [ 27.2.1 1100]]。
In step S2, it includes:
the forward propagation mode is used for calculating the output of the model, and the final prediction result is obtained through the processing of the weighting and activation functions of the input features through each layer;
the weight matrix W of the neural network data model is:
the bias vector b is:
the weighted input z is:
the output a of the metapoint is:
where z is the weighted input, W is the weight matrix, x is the input feature vector, b is the bias vector, g is the sigmoid activation function, and a is the output of the metapoint.
In step S4, it includes:
the back propagation is used for updating the weight and bias of the neural network data model, so that the adaptive learning of the neural network data model is realized;
defining a loss function: assuming that the predicted value of the neural network data model is y_pred, the true value is y_true, and the Mean Square Error (MSE) loss function is:
wherein y_pred1, y_pred2, y_pred3 are the prediction results of the neural network data model, and correspond to three elements of a, namely g (z 1), g (z 2) and g (z 3), respectively.
In step S4, further includes:
calculating an error term delta: the error term delta, namely the gradient of the loss function MSE to the weighted input z, can be obtained through calculation by a chain rule;
where g' (z) is the derivative of the sigmoid function.
In step S4, further includes:
calculating the gradient of the parameters:
wherein x is an input feature vector;
for bias vector b:
wherein δ is the error term;
step S4, further comprising:
the final output model is the calculated activation output a in the forward propagation process, i.e
Outputting a predicted value:
wherein g (z 1), g (z 2) and g (z 3) are respectively output after the sigmoid activation function is processed, and represent the prediction result of the neural network data model on the input vector x= [ 27.2.1 ] respectively. The prediction results can be used for predicting the energy consumption of the central air conditioner, so that the optimal control of the central air conditioner system is realized.
Referring to fig. 2, a control device for a central air conditioning system adaptive learning data model is characterized by comprising:
the acquisition module is configured to monitor and acquire data of each key point in the central air conditioning system;
the preprocessing module is configured to preprocess the acquired data to obtain a data set;
the training module is configured to select and determine the structure of the neural network data model according to the complexity of the task and the characteristics of the data; training the neural network data model by using the data set, and adjusting the weight and bias of the neural network data model by using an optimization algorithm until the error between the predicted result and the actual result is minimized, so as to obtain a trained neural network data model;
the evaluation module is configured to evaluate the trained neural network data model by using the test set, calculate an evaluation index and obtain an evaluation result, and judge whether the evaluation result meets the requirement; if the evaluation result does not meet the requirements, optimizing the neural network data model according to the evaluation result;
and the updating module is configured to update and apply the neural network data model to a central air conditioning system if the evaluation result meets the requirement.
The above examples are only for illustrating the technical scheme of the present invention and are not limiting. It will be understood by those skilled in the art that any modifications and equivalents that do not depart from the spirit and scope of the invention are intended to be within the scope of the appended claims.

Claims (8)

1. The control method of the self-adaptive learning data model of the central air conditioning system is characterized by comprising the following steps of:
s1, monitoring and collecting data of each key point in a central air conditioning system;
s2, preprocessing the acquired data to obtain a data set;
s3, selecting and determining the structure of a neural network data model according to the complexity of the task and the characteristics of the data;
s4, training the neural network data model by using the data set, and adjusting the weight and bias of the neural network data model by using an optimization algorithm until the error between the predicted result and the actual result is minimized, so as to obtain a trained neural network data model;
s5, evaluating the trained neural network data model by using a test set, calculating an evaluation index, obtaining an evaluation result, and judging whether the evaluation result meets the requirement;
s6, if the evaluation result does not meet the requirements, optimizing the neural network data model according to the evaluation result, and repeating S5;
and S7, if the evaluation result meets the requirement, updating the neural network data model and applying the neural network data model to a central air conditioning system.
2. The control method of a central air conditioning system adaptive learning data model according to claim 1, wherein in step S1, the number of collected temperatures is assumedAccording to 27 ℃, the pressure data is 2.1bar, and the flow data is 1100m 3 And/h to obtain an input vector x= [ 27.2.1 1100]]。
3. The method for controlling an adaptive learning data model of a central air conditioning system according to claim 2, wherein step S2 includes:
the weight matrix W of the neural network data model is:
the bias vector b is:
the weighted input z is:
the output a of the metapoint is:
where z is the weighted input, W is the weight matrix, x is the input feature vector, b is the bias vector, g is the sigmoid activation function, and a is the output of the metapoint.
4. A control method for a central air conditioning system adaptive learning data model according to claim 3, wherein in step S4, the method comprises:
defining a loss function: assuming that the predicted value of the neural network data model is y_pred, the true value is y_true, and the Mean Square Error (MSE) loss function is:
wherein y_pred1, y_pred2, y_pred3 are the prediction results of the neural network data model, and correspond to three elements of a, namely g (z 1), g (z 2) and g (z 3), respectively.
5. The method for controlling an adaptive learning data model of a central air conditioning system according to claim 4, wherein in step S4, further comprising:
calculating an error term delta:
the error term delta, namely the gradient of the loss function MSE to the weighted input z, can be obtained through calculation by a chain rule;
where g' (z) is the derivative of the sigmoid function.
6. The method for controlling an adaptive learning data model of a central air conditioning system according to claim 5, wherein in step S4, further comprising:
calculating the gradient of the parameters:
wherein x is an input feature vector;
for bias vector b:
wherein δ is the error term;
7. the method for controlling an adaptive learning data model of a central air conditioning system according to claim 6, wherein step S4 further comprises:
outputting a predicted value:
wherein g (z 1), g (z 2) and g (z 3) are respectively output after the sigmoid activation function is processed, and represent the prediction result of the neural network data model on the input vector x= [ 27.2.1 ] respectively.
8. A control device for a central air conditioning system adaptive learning data model, comprising:
the acquisition module is configured to monitor and acquire data of each key point in the central air conditioning system;
the preprocessing module is configured to preprocess the acquired data to obtain a data set;
the training module is configured to select and determine the structure of the neural network data model according to the complexity of the task and the characteristics of the data; training the neural network data model by using the data set, and adjusting the weight and bias of the neural network data model by using an optimization algorithm until the error between the predicted result and the actual result is minimized, so as to obtain a trained neural network data model;
the evaluation module is configured to evaluate the trained neural network data model by using the test set, calculate an evaluation index and obtain an evaluation result, and judge whether the evaluation result meets the requirement; if the evaluation result does not meet the requirements, optimizing the neural network data model according to the evaluation result;
and the updating module is configured to update and apply the neural network data model to a central air conditioning system if the evaluation result meets the requirement.
CN202311079805.5A 2023-08-25 2023-08-25 Control method and device for self-adaptive learning data model of central air conditioning system Pending CN117091273A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311079805.5A CN117091273A (en) 2023-08-25 2023-08-25 Control method and device for self-adaptive learning data model of central air conditioning system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311079805.5A CN117091273A (en) 2023-08-25 2023-08-25 Control method and device for self-adaptive learning data model of central air conditioning system

Publications (1)

Publication Number Publication Date
CN117091273A true CN117091273A (en) 2023-11-21

Family

ID=88781063

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311079805.5A Pending CN117091273A (en) 2023-08-25 2023-08-25 Control method and device for self-adaptive learning data model of central air conditioning system

Country Status (1)

Country Link
CN (1) CN117091273A (en)

Similar Documents

Publication Publication Date Title
CN103912966B (en) A kind of earth source heat pump refrigeration system optimal control method
CN104698843B (en) A kind of data center's energy-saving control method based on Model Predictive Control
CN108168030B (en) Intelligent control method based on refrigeration performance curve
CN107940679B (en) Group control method based on performance curve of water chilling unit of data center
CN110458340B (en) Building air conditioner cold load autoregressive prediction method based on mode classification
CN113039506B (en) Causal learning-based data center foundation structure optimization method
CN107036238B (en) Intelligent energy-saving control method for dynamically predicting external air and load
WO2020199682A1 (en) Air conditioner control method, air conditioner control apparatus, storage medium, memory and air conditioner
CN111306706B (en) Air conditioner linkage control method and system
CN113405223A (en) Cold machine number control method based on GRNN and control system thereof
CN114279042B (en) Central air conditioner control method based on multi-agent deep reinforcement learning
CN112070353B (en) Method and system for accurately detecting energy efficiency of data center
CN111780384A (en) Central air-conditioning control system
KR20180138371A (en) Method for evaluating data based models and conducting predictive control of capsule type ice thermal storage system using the same
CN114923268A (en) Machine room air conditioner regulation and control method and device based on air conditioner load and regional temperature evaluation
CN114154677A (en) Air conditioner operation load model construction and prediction method, device, equipment and medium
CN113028610B (en) Method and device for global optimization and energy-saving control of dynamic load of central air conditioner
CN113268913B (en) Intelligent building air conditioner cooling machine system operation optimization method based on PSO-ELM algorithm
CN114046593A (en) Dynamic predictive machine learning type air conditioner energy-saving control method and system
CN114282729A (en) Load prediction-based ice storage air conditioner optimal scheduling method
CN117091273A (en) Control method and device for self-adaptive learning data model of central air conditioning system
CN110762739A (en) Data center air conditioner control method, device, equipment and storage medium
CN114811857A (en) Cold station system operation optimization method
CN111787764B (en) Energy consumption optimization method and device for multi-split refrigerating unit, electronic equipment and storage medium
WO2021234763A1 (en) Indoor temperature estimation device, program, and indoor temperature estimation method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination