Energy efficiency optimization system based on adaptive control algorithm

Abstract

The invention discloses an energy efficiency optimization system based on a self-adaptive control algorithm, and the system comprises a consumption prediction module which carries out the training of collected actual operation data, i.e., a sample set, through a polynomial regression model, analyzes the abnormality of an energy consumption mode through the model, and predicts a potential equipment fault. The temperature control module is used for searching the optimal temperature setting in a given temperature range, so that the energy consumption is minimized while the requirement is met; an adaptive control module; a user side; and cloud. According to the energy efficiency optimization system based on the self-adaptive control algorithm, the self-adaptive control module takes predicted energy consumption data as a basis through the PID control algorithm and the temperature control module in cooperation with the temperature control module, and dynamically adjusts the heating power in combination with the target temperature and the current temperature, so that the energy consumption is minimized while the cooking requirement is met; and the energy consumption caused by the device is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of energy efficiency optimization, and in particular to an energy efficiency optimization system based on an adaptive control algorithm. Background Art

[0002] The existing energy efficiency optimization is to reasonably allocate energy, which is mainly aimed at central air conditioning and computer rooms, such as the publication number CN116384700A. The present invention relates to the field of energy efficiency optimization of air conditioning systems and the field of data mining, and specifically discloses a method for optimizing the energy efficiency of central air conditioning systems based on data mining; first, the parameter characteristics are analyzed and processed by preprocessing and feature engineering the data stored in the database. Then, the XGBoost algorithm is used to sort the feature importance, and the loss function gain and the decision tree branch attribute frequency are used as evaluation functions. After that, the frequent pattern growth algorithm is used to analyze the operating data to obtain the potential correlation between each parameter, and the niche method is used to improve the particle swarm optimization algorithm to improve its convergence. At the same time, the deep residual network model is used as the objective function of the optimization model, and the normal operating range of the parameters and the mined association rules are used as optimization constraints. Finally, the energy efficiency ratio of the central air conditioning system is optimized using the improved particle swarm optimization algorithm to obtain the relatively optimal working point under different load rates of the chiller. Its main feature is that it can reduce air-conditioning energy consumption on the basis of meeting actual working conditions, thereby achieving energy conservation and emission reduction of building energy consumption.

[0003] To sum up, when the central air conditioner and the computer room are working, the peak load is kept within a certain range, and the range and amplitude of change are small. When cooking in the kitchen, multiple devices are mostly superimposed or operated alternately, which can easily cause energy waste due to frequent and drastic changes in peak loads, and it is also easy to fail. That is, the existing energy efficiency optimization system is not suitable for kitchen usage scenarios. In view of the above problems, the existing system needs to be improved. Summary of the invention

[0004] The purpose of the present invention is to provide an energy efficiency optimization system based on an adaptive control algorithm to solve the problem raised in the above background technology that the peak load of the central air conditioner and the machine room is kept within a certain range when working, and the range and amplitude of change are small, while when cooking in the kitchen, most of the time, multiple devices work in superposition or alternately, which is easy to cause energy waste due to frequent and drastic changes in peak load, and is prone to failure, that is, the existing energy efficiency optimization system is not suitable for kitchen usage scenarios.

[0005] To achieve the above object, the present invention provides the following technical solution: an energy efficiency optimization system based on an adaptive control algorithm, comprising:

[0006] A consumption prediction module, which trains the collected actual operation data, i.e., the sample set, through a polynomial regression model, and analyzes the abnormality of the energy consumption pattern through the model to predict potential equipment failures;

[0007] A temperature control module, which searches for the best temperature setting within a given temperature range to meet demand while minimizing energy consumption;

[0008] An adaptive control module, wherein the adaptive control module dynamically adjusts the heating power according to the target temperature and the current temperature through a PID control algorithm;

[0009] A user terminal, which is used by staff to set and view parameters and recommend parameter setting suggestions based on the set parameters;

[0010] The cloud is used to store and deeply analyze the training data and results, that is, to push algorithm updates through the cloud to continuously improve system performance.

[0011] Preferably, the polynomial regression model adopts a BP neural network prediction model, the training of the BP neural network prediction model includes two aspects: forward transmission and backward transmission, and the construction steps of the BP neural network prediction model include data preprocessing, network structure setting, model training, model evaluation and parameter adjustment, and model application.

[0012] Preferably, the principle of the BP neural network prediction model is as follows: the learning process consists of two processes: forward propagation of the signal and reverse propagation of the error. During forward propagation, the pattern acts on the input layer, and after being processed by the hidden layer, the input error is transmitted in the reverse propagation stage, and the output error is returned to the input layer layer by layer through the hidden layer in a certain sub-form, and "shared" to all units of each layer, thereby obtaining the reference error or error signal of the units in each layer, which is used as the basis for modifying the weights of each unit in the process of continuous modification of the weights.

[0013] Preferably, the BP neural network prediction model is a multi-layer network composed of an input layer, an output layer and one or more hidden layer nodes interconnected, so that the multi-layer feedforward network establishes a suitable linear or nonlinear relationship between input and output without limiting the network output to between -1 and 1.

[0014] Preferably, the prediction principle process of the BP neural network prediction model is:

[0015] Step a: Initialize the population, where each individual represents the weight and bias of a BP neural network;

[0016] Step b: For each individual, use the BP neural network to train and calculate its fitness. The fitness function can be the sum of squares of the prediction error.

[0017] Step c: Use the selection operation to select the parent individual according to the fitness function;

[0018] Step d: Use the crossover operation to cross the parent individuals to generate new individuals;

[0019] Step e: Use mutation operation to mutate the new individual and introduce new genetic information;

[0020] Step f: Add new individuals to the population and delete individuals with lower fitness;

[0021] Step g: Repeat steps b to f until the stop condition is reached;

[0022] Step h: Select the individual with the highest fitness as the final solution, that is, the BP neural network with optimal weights and biases.

[0023] Preferably, the PID control algorithm relies on hardware module control, and the hardware module includes a PID controller, a temperature acquisition module, a button module, an LCD display module and a buzzer warning module.

[0024] Preferably, the PID controller is composed of a proportional unit, an integral unit and a differential unit, and the analytical expression of the relationship between the input e(t) and the output u(t) is:

[0025] u(t)=kp[e(t)dt+1 / Tl∫e(t)dt+TD*de(t) / dt]

[0026] Among them, u(t) is the given temperature value, y(t) is the temperature value output by the system, e is the error value, KP, TD, Tl are PID parameters, and t is the current time.

[0027] Preferably, the upper and lower limits of the integral in the analytical expression of the relationship are 0 and t, respectively, thereby obtaining the transfer function:

[0028] G(s)=U(s) / E(s)=kp[1+1 / (Tl*s)+TD*s]

[0029] Where kp is the proportional gain adjustment coefficient; TI is the integral gain adjustment constant; TD is the differential gain adjustment constant.

[0030] Preferably, the error value e=set value SP-feedback value PV, the kp proportional control takes the current error into consideration, and the TI integral control takes the past error into consideration.

[0031] Preferably, the TD differential control takes into account future errors, and TD adjusts the control output according to the temperature change rate.

[0032] Compared with the prior art, the beneficial effects of the present invention are: the energy efficiency optimization system based on the adaptive control algorithm,

[0033] The present invention can effectively solve the problem that when cooking in the kitchen, most of the time, multiple devices work in superposition or alternately, which is easy to cause energy waste due to frequent and drastic changes in peak load, and it is easy to fail, that is, the existing energy efficiency optimization system is not suitable for the use scenario of the kitchen. The BP neural network prediction model is used to predict the energy consumption of the kitchen in a unit time period and the existing hidden dangers and failures, so as to facilitate maintenance and repair of the failure. At the same time, targeted human intervention is carried out in advance on the load change to avoid failures due to large changes, thereby ensuring normal operation. Secondly, the adaptive control module uses the PID control algorithm to cooperate with the temperature control module to use the predicted energy consumption data as a basis, and dynamically adjusts the heating power in combination with the target temperature and the current temperature, so that it can minimize energy consumption while meeting cooking needs, thereby reducing the energy consumption caused by the device. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is an overall schematic diagram of the system of the present invention;

[0035] Figure 2 This is a schematic diagram of the consumption prediction module of the present invention;

[0036] Figure 3 This is a principle block diagram of the simulated PID control system of the present invention;

[0037] Figure 4 This is the structural diagram of the PID control system of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] See also Figure 1-4 The present invention provides a technical solution: an energy efficiency optimization system based on an adaptive control algorithm, comprising:

[0040] The consumption prediction module trains the collected actual operation data, i.e., the sample set, through a polynomial regression model, and analyzes the abnormality of the energy consumption pattern through the model to predict potential equipment failures;

[0041] Temperature control module, the temperature control module searches for the best temperature setting within a given temperature range to meet demand while minimizing energy consumption;

[0042] Adaptive control module, which dynamically adjusts the heating power according to the target temperature and the current temperature through the PID control algorithm;

[0043] The user end is used by staff to set and view parameters, and recommend parameter setting suggestions based on the set parameters;

[0044] The cloud is used to store and deeply analyze training data and results, that is, to push algorithm updates through the cloud and continuously improve system performance.

[0045] The polynomial regression model adopts the BP neural network prediction model. The training of the BP neural network prediction model includes forward transmission and backward transmission. The construction steps of the BP neural network prediction model include data preprocessing, network structure setting, model training, model evaluation and parameter adjustment, and model application.

[0046] The principle of the BP neural network prediction model is: the learning process consists of two processes: forward propagation of the signal and reverse propagation of the error. During forward propagation, the pattern acts on the input layer. After being processed by the hidden layer, the input error is transmitted in the reverse propagation stage. The output error is returned to the input layer layer by layer through the hidden layer in a certain sub-form, and "shared" to all units in each layer, thereby obtaining the reference error or error signal of the units in each layer, which is used as the basis for modifying the weights of each unit.

[0047] The BP neural network prediction model is a multi-layer network composed of an input layer, an output layer and one or more hidden layer nodes interconnected, which enables the multi-layer feedforward network to establish a suitable linear or nonlinear relationship between input and output without limiting the network output to between -1 and 1.

[0048] The prediction principle process of the BP neural network prediction model is as follows:

[0049] Step a: Initialize the population, where each individual represents the weight and bias of a BP neural network;

[0050] Step b: For each individual, use the BP neural network to train and calculate its fitness. The fitness function can be the sum of squares of the prediction error.

[0051] Step c: Use the selection operation to select the parent individual according to the fitness function;

[0052] Step d: Use the crossover operation to cross the parent individuals to generate new individuals;

[0053] Step e: Use mutation operation to mutate the new individual and introduce new genetic information;

[0054] Step f: Add new individuals to the population and delete individuals with lower fitness;

[0055] Step g: Repeat steps b to f until the stop condition is reached;

[0056] Step h: Select the individual with the highest fitness as the final solution, that is, the BP neural network with optimal weights and biases.

[0057] The PID control algorithm relies on hardware module control, and the hardware module includes a PID controller, a temperature acquisition module, a button module, an LCD display module and a buzzer alarm module.

[0058] The PID controller consists of a proportional unit, an integral unit, and a differential unit, and the analytical expression of the relationship between the input e(t) and the output u(t) is:

[0059] u(t)=kp[e(t)dt+1 / Tl∫e(t)dt+TD*de(t) / dt]

[0060] Among them, u(t) is the given temperature value, y(t) is the temperature value output by the system, e is the error value, KP, TD, Tl are PID parameters, and t is the current time.

[0061] The upper and lower limits of the integral in the analytical expression are 0 and t, respectively, and the transfer function is obtained as follows:

[0062] G(s)=U(s) / E(s)=kp[1+1 / (Tl*s)+TD*s]

[0063] Where kp is the proportional gain adjustment coefficient; TI is the integral gain adjustment constant; TD is the differential gain adjustment constant.

[0064] Error value e = set value SP - feedback value PV, kp proportional control takes into account the current error, TI integral control takes into account the past error.

[0065] TD differential control takes future errors into account, and TD adjusts the control output according to the temperature change rate.

[0066] Specifically, in the actual working process, the staff can set the relevant parameters through the user end, and the system recommends the optimal solution close to the parameter to the staff based on the set value, that is, generates a detailed energy usage report to help the staff make decisions. The BP neural network algorithm predicts the kitchen's energy consumption today and the hidden dangers and faults of related equipment, so as to facilitate the pre-maintenance of related equipment and ensure the normal cooking work in the kitchen. In addition, the adaptive control module uses the PID algorithm to coordinate the heating power of various equipment through the predicted energy consumption, and simultaneously considers historical error data to improve the control accuracy, that is, considers the energy use of the entire kitchen and coordinates the operation of multiple devices. At the same time, the temperature control module searches for the optimal temperature setting within a given power or temperature range, thereby meeting cooking needs while minimizing energy consumption.

[0067] Embodiment 2

[0068] This embodiment is a further description of the above embodiment. It should be understood that this embodiment includes all the aforementioned technical features and is further described in detail.

[0069] The consumption prediction module can also be predicted by the random forest method, and the prediction process of the random forest method is:

[0070] Step 1: Parallel sampling of the original data set is performed on each node: the original training set D is sampled repeatedly using the Bootstrap sampling method to obtain k sub-training sets;

[0071] Step 2: Decision tree modeling at each node: Use the above k sub-training sets to build k decision models. It is worth noting that not all features are used in the process of tree building. Each tree randomly selects m features from all features for modeling;

[0072] Step 3: Obtain random forest model: collect k decision tree models to construct a random forest model;

[0073] Step 4: Vote for the predicted samples, and the one with the most votes is the final category.

[0074] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An energy efficiency optimization system based on an adaptive control algorithm, characterized in that ,include: A consumption prediction module, which trains the collected actual operation data, i.e., the sample set, through a polynomial regression model, and analyzes the abnormality of the energy consumption pattern through the model to predict potential equipment failures; A temperature control module, which searches for the best temperature setting within a given temperature range to meet demand while minimizing energy consumption; An adaptive control module, wherein the adaptive control module dynamically adjusts the heating power according to the target temperature and the current temperature through a PID control algorithm; A user terminal, which is used by staff to set and view parameters and recommend parameter setting suggestions based on the set parameters; The cloud is used to store and deeply analyze the training data and results, that is, to push algorithm updates through the cloud to continuously improve system performance.

2. The energy efficiency optimization system based on the adaptive control algorithm according to claim 1, characterized in that: The polynomial regression model adopts a BP neural network prediction model. The training of the BP neural network prediction model includes two aspects: forward transmission and backward transmission. The construction steps of the BP neural network prediction model include data preprocessing, network structure setting, model training, model evaluation and parameter adjustment, and model application.

3. The energy efficiency optimization system based on the adaptive control algorithm according to claim 1, characterized in that: The principle of the BP neural network prediction model is as follows: the learning process consists of two processes: the forward propagation of the signal and the reverse propagation of the error. During the forward propagation, the pattern acts on the input layer. After being processed by the hidden layer, the input error is transmitted in the reverse propagation stage. The output error is returned to the input layer layer by layer through the hidden layer in a certain sub-form, and "shared" to all units in each layer, thereby obtaining the reference error or error signal of the units in each layer, which is used as the basis for modifying the weights of each unit in the process of continuous modification of the weights.

4. The energy efficiency optimization system based on the adaptive control algorithm according to claim 1, characterized in that: The BP neural network prediction model is a multi-layer network composed of an input layer, an output layer and one or more hidden layer nodes interconnected, so that the multi-layer feedforward network establishes a suitable linear or nonlinear relationship between input and output without limiting the network output to between -1 and 1.

5. The energy efficiency optimization system based on the adaptive control algorithm according to claim 1, characterized in that: The prediction principle process of the BP neural network prediction model is as follows: Step a: Initialize the population, where each individual represents the weight and bias of a BP neural network; Step b: For each individual, use the BP neural network to train and calculate its fitness. The fitness function can be the sum of squares of the prediction error. Step c: Use the selection operation to select the parent individual according to the fitness function; Step d: Use the crossover operation to cross the parent individuals to generate new individuals; Step e: Use mutation operation to mutate the new individual and introduce new genetic information; Step f: Add new individuals to the population and delete individuals with lower fitness; Step g: Repeat steps b to f until the stop condition is reached; Step h: Select the individual with the highest fitness as the final solution, that is, the BP neural network with optimal weights and biases.

6. The energy efficiency optimization system based on the adaptive control algorithm according to claim 1, characterized in that: The PID control algorithm relies on hardware module control, and the hardware module includes a PID controller, a temperature acquisition module, a button module, an LCD display module and a buzzer alarm module.

7. The energy efficiency optimization system based on the adaptive control algorithm according to claim 1, characterized in that: The PID controller is composed of a proportional unit, an integral unit and a differential unit, and the analytical expression of the relationship between the input e(t) and the output u(t) is: u(t)=kp[e(t)dt+1 / Tl∫e(t)dt+TD*de(t) / dt] Among them, u(t) is the given temperature value, y(t) is the temperature value output by the system, e is the error value, KP, TD, Tl are PID parameters, and t is the current time.

8. The energy efficiency optimization system based on the adaptive control algorithm according to claim 7, characterized in that: The upper and lower limits of the integral in the analytical expression of the relationship are 0 and t, respectively, and the transfer function is obtained as follows: G(s)=U(s) / E(s)=kp[1+1 / (Tl*s)+TD*s] Where kp is the proportional gain adjustment coefficient; TI is the integral gain adjustment constant; TD is the differential gain adjustment constant.

9. The energy efficiency optimization system based on adaptive control algorithm according to claim 7, characterized in that: The error value e=set value SP-feedback value PV, kp proportional control takes current error into consideration, and TI integral control takes past error into consideration.

10. The energy efficiency optimization system based on adaptive control algorithm according to claim 7, characterized in that: The TD derivative control takes into account future errors, and TD adjusts the control output according to the temperature change rate.

Images