Intelligent regulation and control method and system for clean air conditioning system

Abstract

The invention relates to the technical field of clean air conditioning systems, in particular to an intelligent regulation and control method and system for a clean air conditioning system.The method comprises the steps that in response to the situation that real-time environment parameters are not located in a target parameter interval, the real-time environment parameters and an initial control sequence are input into a prediction model to obtain adjusted environment parameters after the initial control sequence is executed, the initial control sequence is a control parameter of each dimension in a preset time period; predicting the noise degree of the initial control sequence, and constructing a loss function; and iteratively updating the initial control sequence by using a gradient descent method, and stopping when the value of the loss function is smaller than a preset value to obtain a target control sequence so as to realize regulation and control of the clean air conditioning system. By means of the technical scheme, precise regulation and control of the clean air conditioning system can be achieved.

Description

Technical Field

[0001] The present application relates to the technical field of clean air conditioning systems, and in particular to an intelligent control method and system for clean air conditioning systems. Background Art

[0002] Clean air conditioning system is an air conditioning system specially used to maintain specific environmental conditions. Its design goal is to provide suitable temperature and humidity while ensuring high cleanliness of air. This system is often used in laboratories, production workshops or hospitals and other places with high requirements for air quality.

[0003] At present, a patent application document with publication number CN118960167A discloses an energy-saving control method and controller for a clean air-conditioning system based on deep learning, wherein the method includes: calibrating the target clean parameters of the current space, collecting the clean parameters of the current space, wherein the target clean parameters include: indoor and outdoor pressure difference, air particle size, temperature, humidity, air flow rate, and indoor pollution source data; calculating the target difference between the clean parameters of the current space and the target clean parameters, inputting the target difference into a speed prediction model, and obtaining the predicted speeds of the filter fan and the exhaust fan; calculating the current system energy consumption according to the predicted speeds of the filter fan and the exhaust fan of the clean air-conditioning system, evaluating the current system energy consumption and the real-time clean parameters, and obtaining a real-time evaluation value; judging whether the real-time evaluation value is the optimal evaluation value, if so, using the predicted speed for control, otherwise adjusting the speed of the filter fan and the exhaust fan of the clean air-conditioning system according to the difference between the real-time evaluation value and the optimal evaluation value.

[0004] The above method obtains the predicted speed of the filter fan and the exhaust fan through the speed prediction model, which can find the best balance between cleanliness and energy consumption, ensuring that energy consumption is minimized while meeting the cleanliness standard. However, the predicted speed output by the speed prediction model cannot guarantee that the cleanliness parameters of the current space can be adjusted to the target cleanliness parameters within a reasonable time, and ignores the noise pollution generated during the operation of the clean air-conditioning system, resulting in inaccurate control results of the clean air-conditioning system. Summary of the invention

[0005] In order to solve the technical problem of inaccurate control results of a clean air-conditioning system, the present application provides an intelligent control method and system for a clean air-conditioning system, which can achieve precise control of the clean air-conditioning system.

[0006] In a first aspect, the present application provides an intelligent control method for a clean air conditioning system, the control method comprising: in response to a real-time environmental parameter not being within a target parameter interval, inputting the real-time environmental parameter and an initial control sequence into a prediction model to obtain an adjusted environmental parameter after executing the initial control sequence, the initial control sequence being the control parameters of each dimension within a preset time period; predicting the noise level of the initial control sequence, and constructing a loss function : ; To adjust the environmental parameters, is the minimum value of the target parameter interval, is the maximum value of the target parameter interval, is the noise level, is the noise level threshold; the initial control sequence is iteratively updated using the gradient descent method, and the process stops when the value of the loss function is less than a preset value, so as to obtain a target control sequence to achieve regulation of the clean air-conditioning system.

[0007] When the real-time environmental parameters are not in the target parameter interval, it means that the working state of the clean air-conditioning system can no longer ensure that the environmental parameters of the indoor area are in a comfortable state. The real-time environmental parameters and the randomly generated initial control sequence are input into the prediction model to obtain the adjusted environmental parameters after executing the initial control sequence. By adjusting the environmental parameters, it can be determined whether the initial control sequence can adjust the environmental parameters to the target parameter interval within the preset time period; further, the noise level of the initial control sequence is predicted, and a loss function is constructed. The loss function evaluates the control effect of the initial control sequence from two aspects: whether the adjusted environmental parameters are in the target parameter interval and whether the noise level is less than the noise level threshold; the initial control sequence is iteratively updated using the gradient descent method, and the loss function stops when the value of the loss function is less than the preset value, to obtain the target control sequence. The target control sequence can adjust the environmental parameters to the target parameter interval within the preset time period, and ensure that the noise generated by the clean air-conditioning system is lower than the noise level threshold, thereby realizing precise regulation of the clean air-conditioning system.

[0008] Preferably, the prediction model includes a time series extraction sub-model and a regression sub-model. The time series extraction sub-model is used to extract time series features of the initial control sequence to obtain control features; the regression sub-model is used to regress the splicing results of the control features and real-time environmental parameters to the adjustment environmental parameters.

[0009] An end-to-end prediction model is proposed, which can accurately output the environmental parameters of the indoor area after executing the initial control sequence under real-time environmental parameters.

[0010] Preferably, the training method of the prediction model includes: collecting real environmental parameters at historical moments, and taking the historical moments as the starting point, collecting real control sequences within a preset time period to obtain a set of training samples, and using the real environmental parameters after the preset time period as label values; inputting the training samples into the prediction model to obtain prediction results; and training the prediction model using the absolute value of the difference between the prediction result and the label value as the loss function.

[0011] Preferably, predicting the noise level of the initial control sequence includes: inputting the control parameters at any time in the initial control sequence into a noise assessment model to obtain the noise decibel at each time; collecting real-time environmental sounds, decomposing the real-time environmental sounds into multiple sound components using an independent component analysis algorithm, dividing each sound component into air-conditioning noise and non-air-conditioning noise, and taking the sum of the decibel values ​​of all non-air-conditioning noises as a decibel threshold; and calculating the proportion of moments when the noise decibel is greater than the decibel threshold within a preset time period to obtain the noise level.

[0012] Since the clean air-conditioning system is in operation, the real-time ambient sound is a mixture of air-conditioning noise and various non-air-conditioning noises. The sum of the decibel values ​​of non-air-conditioning noises is the noise level in the indoor area without considering the clean air-conditioning system. Therefore, the decibel threshold is adaptively determined based on the decibel values ​​of all non-air-conditioning noises to ensure that the noise decibels generated by the clean air-conditioning system will not have a significant impact on people in the indoor area, thereby achieving accurate quantification of the noise level.

[0013] Preferably, the noise assessment model is: , is the number of control parameters, For the moment Control parameters The value of For control parameters The undetermined coefficients of For the moment The noise decibel is determined by the least square method.

[0014] Preferably, dividing each sound component into air-conditioning noise and non-air-conditioning noise includes: calculating the similarity between the spectral characteristics of each sound component and the standard spectral characteristics of air-conditioning noise, taking the sound component corresponding to the maximum similarity as air-conditioning noise, and taking the remaining sound components as non-air-conditioning noise.

[0015] Preferably, iteratively updating the initial control sequence using the gradient descent method includes: back-propagating the prediction model according to the loss function to obtain the gradient value of each control parameter in the initial control sequence; the updating process of the initial control sequence is: , is any control parameter in the initial control sequence, is the update coefficient, For control parameters The gradient value of is the updated control parameter.

[0016] The gradient descent method can update the initial control sequence in the direction of reducing the loss function and update the initial control sequence multiple times. When the value of the loss function is less than the preset value, it means that the initial control sequence at this time can achieve better control effect. Taking the initial control sequence at this time as the target control sequence can achieve precise regulation of the clean air-conditioning system.

[0017] Preferably, the control method further comprises: in response to the number of iterations being greater than a preset number, the value of the loss function is not less than a preset value, increasing the noise level threshold, and continuing to iteratively update the initial control sequence.

[0018] When the number of iterations is greater than the preset number, the value of the loss function is still not less than the preset value, indicating that the initial control sequence cannot achieve a good control effect within the preset time period. In order to ensure that the clean air-conditioning system can adjust the environmental parameters of the indoor area to the target parameter range within the preset time period, the noise level threshold in the loss function can be increased, that is, a certain amount of noise is allowed in the clean air-conditioning system, thereby giving priority to ensuring the control effect of the environmental parameters.

[0019] In a second aspect of the present application, an intelligent control system for a clean air-conditioning system is also provided, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent control method for a clean air-conditioning system according to the first aspect of the present application is implemented.

[0020] The technical solution of this application has the following beneficial technical effects: First, when the real-time environmental parameters are not in the target parameter interval, it means that the working state of the clean air-conditioning system can no longer ensure that the environmental parameters of the indoor area are in a comfortable state. The real-time environmental parameters and the randomly generated initial control sequence are input into the prediction model to obtain the adjusted environmental parameters after executing the initial control sequence. By adjusting the environmental parameters, it can be determined whether the initial control sequence can adjust the environmental parameters to the target parameter interval within the preset time period; further, the noise level of the initial control sequence is predicted, and a loss function is constructed. The loss function evaluates the control effect of the initial control sequence from two aspects: whether the adjusted environmental parameters are in the target parameter interval and whether the noise level is less than the noise level threshold; the initial control sequence is iteratively updated using the gradient descent method, and the loss function stops when the value of the loss function is less than the preset value, and the target control sequence is obtained. The target control sequence can adjust the environmental parameters to the target parameter interval within the preset time period, and ensure that the noise generated by the clean air-conditioning system is lower than the noise level threshold, thereby realizing precise regulation of the clean air-conditioning system. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present application will become easy to understand. In the accompanying drawings, several embodiments of the present application are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 is a flow chart of an intelligent control method for a clean air conditioning system according to an embodiment of the present application; Figure 2 It is a structural block diagram of an intelligent control system for a clean air conditioning system according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0023] It should be understood that when the terms "first", "second", etc. are used in the claims, specification and drawings of the present application, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the specification and claims of the present application indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0024] According to the first aspect of the present application, the present application provides an intelligent control method for a clean air-conditioning system. Figure 1 1 is a flow chart of an intelligent control method for a clean air conditioning system according to an embodiment of the present application. Figure 1 As shown, the intelligent control method of the clean air conditioning system includes steps S101 to S103, which are described in detail below.

[0025] S101, in response to the real-time environmental parameter not being within the target parameter interval, inputting the real-time environmental parameter and an initial control sequence into a prediction model to obtain an adjusted environmental parameter after executing the initial control sequence, wherein the initial control sequence is the control parameter of each dimension within a preset time period.

[0026] In one embodiment, the real-time environmental parameter may be the temperature, humidity or PM value of an indoor area, and the indoor area may be a production workshop, an office area, a laboratory, etc., which is not limited in this application. The target parameter interval is preset. For example, the comfortable temperature range of the office area is 22 degrees Celsius to 26 degrees Celsius. Therefore, when the real-time environmental parameter is the temperature of the office area, the target parameter interval is .

[0027] In response to the real-time environmental parameters not being within the target parameter interval, it indicates that the current working state of the clean air-conditioning system can no longer guarantee that the environmental parameters of the indoor area are in a comfortable state, and the control parameters of the clean air-conditioning system need to be adjusted so that the environmental parameters of the indoor area can be adjusted to the target parameter interval within a preset time period. Among them, the shorter the preset time period, the shorter the time it takes to adjust the environmental parameters of the indoor area to the target parameter interval, and the better the effect of the clean air-conditioning system on improving the air quality; on the contrary, the longer the preset time period, the longer the time it takes to adjust the environmental parameters of the indoor area to the target parameter interval, and the worse the effect of the clean air-conditioning system on improving the air quality. In the embodiment of the present application, the preset time period is 10 seconds, which means that the clean air-conditioning system is required to adjust the environmental parameters of the indoor area to the target parameter interval within 10 seconds.

[0028] The input of the prediction model is the real-time environmental parameters and the initial control sequence. The initial control sequence is the control parameters of multiple dimensions within a preset time period. The control parameters of multiple dimensions include temperature, speed of the filter fan and the exhaust fan, and other control parameters. The prediction model includes a time series extraction sub-model and a regression sub-model. The time series extraction sub-model is used to extract the time series features of the initial control sequence to obtain the control features; the regression sub-model is used to regress the splicing results of the control features and the real-time environmental parameters to the adjustment environmental parameters.

[0029] The time series extraction sub-model can adopt a recurrent neural network structure such as LSTM, GRU or Transformer; the regression sub-model adopts a fully connected neural network. The adjusted environmental parameters are the environmental parameters of the indoor area after the clean air conditioning system is controlled according to the initial control sequence under the real-time environmental parameters.

[0030] Wherein, the initial control sequence is randomly generated.

[0031] In order to enable the prediction model to output accurate adjustment environment parameters, the prediction model needs to be trained. Specifically, the prediction model training method includes: collecting real environment parameters at historical moments, and taking the historical moments as the starting point, collecting real control sequences within a preset time period to obtain a set of training samples, and using the real environment parameters after the preset time period as label values; inputting the training samples into the prediction model to obtain prediction results; and training the prediction model using the absolute value of the difference between the prediction result and the label value as the loss function.

[0032] In this way, the real-time environmental parameters and the initial control sequence are input into the trained prediction model, which can output the adjusted environmental parameters after executing the initial control sequence; the control effect of the initial control sequence can be judged by adjusting the environmental parameters. If the adjusted environmental parameters are within the target parameter range, it means that the environmental parameters of the indoor area can be adjusted to the target parameter range within the preset time period through the initial control sequence, and the control effect of the initial control sequence is better.

[0033] S102, predicting the noise level of the initial control sequence and constructing a loss function.

[0034] In one embodiment, noise is generated during the operation of the clean air conditioning system, and the noise level is related to the initial control sequence. Therefore, the noise level generated during the operation of the clean air conditioning system can be predicted according to the initial control sequence.

[0035] Specifically, predicting the noise level of the initial control sequence includes: inputting the control parameters at any time in the initial control sequence into a noise evaluation model to obtain the noise decibel at each time; collecting real-time environmental sounds, using an independent component analysis algorithm to decompose the real-time environmental sounds into multiple sound components, dividing each sound component into air-conditioning noise and non-air-conditioning noise, and taking the sum of the decibel values ​​of all non-air-conditioning noises as a decibel threshold; and counting the proportion of moments when the noise decibel is greater than the decibel threshold within a preset time period to obtain the noise level.

[0036] Wherein, the noise assessment model may adopt a polynomial model. The noise assessment model is: , is the number of control parameters, For the moment Control parameters The value of For control parameters The undetermined coefficients of For the moment After determining the unknown coefficients in the noise assessment model using the least squares method, the noise decibels at each moment can be accurately obtained based on the noise assessment model.

[0037] Among them, the independent component analysis algorithm (ICA algorithm) can extract multiple independent source signals (i.e. sound components) from the observation signal (i.e. real-time environmental sound). The independent component analysis algorithm is a well-known technology for those skilled in the art and will not be described in detail here.

[0038] After obtaining multiple independent sound components, it is necessary to find the sound component corresponding to the air-conditioning noise from the multiple sound components. Specifically, the sound components are divided into air-conditioning noise and non-air-conditioning noise, including: calculating the similarity between the spectral characteristics of each sound component and the standard spectral characteristics of air-conditioning noise, taking the sound component corresponding to the maximum similarity as air-conditioning noise, and the remaining sound components as non-air-conditioning noise. The standard spectral characteristics of air-conditioning noise are obtained in advance; the similarity is negatively correlated with the Euclidean distance.

[0039] It can be understood that since the clean air-conditioning system is in operation, the real-time ambient sound is a mixture of air-conditioning noise and various non-air-conditioning noises. The sum of the decibel values ​​of non-air-conditioning noises is the noise level in the indoor area without considering the clean air-conditioning system. Therefore, the sum of the decibel values ​​of all non-air-conditioning noises is used as the decibel threshold. It is only necessary to ensure that the decibel of the noise generated by the clean air-conditioning system is not greater than the decibel threshold, so that the decibel of the noise generated by the clean air-conditioning system will not have a significant impact on people in the indoor area.

[0040] In one embodiment, after obtaining the noise level of the initial control sequence, a loss function is constructed, which evaluates the control effect of the initial control sequence from two aspects: whether the adjustment environment parameter is within the target parameter range and whether the noise level is less than the noise level threshold. for: ; To adjust the environmental parameters, is the minimum value of the target parameter interval, is the maximum value of the target parameter interval, is the noise level, is the noise level threshold.

[0041] Among them, the first part of the loss function Used to constrain the adjustment environment parameters to be within the target parameter range. When the value of the first part is equal to 0, it means that the adjustment environment parameters satisfy , that is, adjust the environmental parameters to the target parameter interval; the second part of the loss function Used to constrain the noise level to be less than the noise level threshold. When the value of the second part is equal to 0, it means the noise level , that is, the noise level is less than the noise level threshold, so when the loss function The closer the value of is to 0, the better the control effect of the initial control sequence is.

[0042] In this way, the decibel threshold of the air-conditioning noise is determined based on the real-time environmental sound, and then the noise level of the initial control sequence is calculated, which can accurately reflect the impact of the air-conditioning noise on people in the indoor area; and then a loss function is constructed, which evaluates the control effect of the initial control sequence from two aspects: whether the adjusted environmental parameters are within the target parameter range and whether the noise level is less than the noise level threshold.

[0043] S103, iteratively updating the initial control sequence using a gradient descent method, stopping in response to a loss function value being less than a preset value, and obtaining a target control sequence to achieve regulation of the clean air conditioning system.

[0044] In one embodiment, iteratively updating the initial control sequence using the gradient descent method includes: back-propagating the prediction model according to the loss function to obtain the gradient value of each control parameter in the initial control sequence; the updating process of the initial control sequence is: , is any control parameter in the initial control sequence, is the update coefficient, For control parameters The gradient value of is the updated control parameter, where the update coefficient is 0.5.

[0045] An update process will simultaneously update all control parameters in the initial control sequence, update the initial control sequence in the direction of reducing the loss function, and update the initial control sequence multiple times. When the value of the loss function is less than the preset value, it means that the initial control sequence at this time can achieve a good control effect. The initial control sequence at this time is used as the target control sequence to achieve the control of the clean air conditioning system. Among them, the preset value is 0.01.

[0046] In another embodiment, in response to the number of iterations being greater than a preset number, the value of the loss function is still not less than a preset value, indicating that the initial control sequence cannot achieve a good control effect within the preset time period. In order to ensure that the clean air-conditioning system can adjust the environmental parameters of the indoor area to the target parameter range within the preset time period, the noise level threshold in the loss function can be increased, that is, a certain amount of noise is allowed in the clean air-conditioning system, thereby giving priority to ensuring the control effect of the environmental parameters.

[0047] Specifically, the control method further includes: in response to the number of iterations being greater than a preset number, the value of the loss function is not less than a preset value, increasing the noise level threshold, and continuing to iteratively update the initial control sequence.

[0048] The preset number of times is 30, and when the noise level threshold is increased, the noise level threshold is increased in steps of 0.1.

[0049] The above introduces the technical principles and implementation details of a smart control method for a clean air conditioning system of the present application through specific embodiments. First, when the real-time environmental parameters are not in the target parameter interval, it means that the working state of the clean air conditioning system at this time can no longer ensure that the environmental parameters of the indoor area are in a comfortable state. The real-time environmental parameters and the randomly generated initial control sequence are input into the prediction model to obtain the adjusted environmental parameters after executing the initial control sequence. By adjusting the environmental parameters, it can be determined whether the initial control sequence can adjust the environmental parameters to the target parameter interval within a preset time period; further, the noise level of the initial control sequence is predicted, and a loss function is constructed. The loss function evaluates the control effect of the initial control sequence from two aspects: whether the adjusted environmental parameters are in the target parameter interval and whether the noise level is less than the noise level threshold; the initial control sequence is iteratively updated using the gradient descent method, and the loss function stops when the value of the loss function is less than the preset value, and the target control sequence is obtained. The target control sequence can adjust the environmental parameters to the target parameter interval within the preset time period, and ensure that the noise generated by the clean air conditioning system is lower than the noise level threshold, so as to achieve precise control of the clean air conditioning system.

[0050] According to the second aspect of the present application, the present application also provides an intelligent control system for a clean air-conditioning system. Figure 2 is a structural block diagram of an intelligent control system for a clean air conditioning system according to an embodiment of the present application. Figure 2 As shown, the system 50 includes a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent control method for a clean air conditioning system according to the first aspect of the present application is implemented. The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, whose settings and functions are known in the art, so they are not described here.

[0051] In the present application, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in this application may be implemented using computer-readable / executable instructions that may be stored or otherwise maintained by such a computer-readable medium.

[0052] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0053] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. An intelligent control method for a clean air conditioning system, characterized in that: The control method comprises: In response to the real-time environmental parameter not being within the target parameter interval, inputting the real-time environmental parameter and the initial control sequence into the prediction model to obtain the adjusted environmental parameter after executing the initial control sequence, wherein the initial control sequence is the control parameter of each dimension within a preset time period; Predict the noise level of the initial control sequence and construct a loss function : ; To adjust the environmental parameters, is the minimum value of the target parameter interval, is the maximum value of the target parameter interval, is the noise level, is the noise level threshold; The initial control sequence is iteratively updated using the gradient descent method, and the process stops when the value of the loss function is less than a preset value, so as to obtain a target control sequence to achieve regulation of the clean air conditioning system.

2. The intelligent control method of a clean air conditioning system according to claim 1, characterized in that: The prediction model includes a time series extraction sub-model and a regression sub-model. The time series extraction sub-model is used to extract time series features from the initial control sequence to obtain control features; the regression sub-model is used to regress the splicing results of the control features and real-time environmental parameters to the adjustment environmental parameters.

3. The intelligent control method of a clean air conditioning system according to claim 2, characterized in that: The training method of the prediction model includes: Collect the real environment parameters at the historical moment, and take the historical moment as the starting point to collect the real control sequence within the preset time period to obtain a set of training samples, and use the real environment parameters after the preset time period as the label value; Input the training samples into the prediction model to obtain the prediction results; The prediction model is trained using the absolute value of the difference between the prediction result and the label value as the loss function.

4. The intelligent control method for a clean air conditioning system according to claim 1, characterized in that: Predicting the noise level of the initial control sequence includes: The control parameters at any time in the initial control sequence are input into the noise assessment model to obtain the noise decibel at each time; Collect real-time environmental sound, use independent component analysis algorithm to decompose the real-time environmental sound into multiple sound components, divide each sound component into air-conditioning noise and non-air-conditioning noise, and use the sum of the decibel values ​​of all non-air-conditioning noises as the decibel threshold; The proportion of moments when the noise decibel is greater than the decibel threshold within a preset time period is counted to obtain the noise level.

5. The intelligent control method of a clean air conditioning system according to claim 4, characterized in that: The noise assessment model is: , is the number of control parameters, For the moment Control parameters The value of For control parameters The undetermined coefficients of For the moment The noise decibel is determined by the least square method.

6. The intelligent control method of a clean air conditioning system according to claim 4, characterized in that: Dividing each sound component into air-conditioning noise and non-air-conditioning noise includes: calculating the similarity between the spectrum characteristics of each sound component and the standard spectrum characteristics of air-conditioning noise, taking the sound component corresponding to the maximum similarity as the air-conditioning noise, and taking the remaining sound components as the non-air-conditioning noise.

7. The intelligent control method for a clean air conditioning system according to claim 1, characterized in that: Iteratively updating the initial control sequence using the gradient descent method includes: Back-propagating the prediction model according to the loss function to obtain the gradient value of each control parameter in the initial control sequence; The update process of the initial control sequence is: , is any control parameter in the initial control sequence, is the update coefficient, For control parameters The gradient value of is the updated control parameter.

8. An intelligent control method for a clean air conditioning system according to any one of claims 1 to 7, characterized in that: The control method further comprises: In response to the fact that the number of iterations is greater than a preset number, the value of the loss function is not less than a preset value, the noise level threshold is increased, and the initial control sequence is continuously iterated and updated.

9. An intelligent control system for a clean air conditioning system, characterized in that: It comprises a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent control method for a clean air-conditioning system according to any one of claims 1 to 8 is implemented.

Images