A frequency conversion control system driven by a compressor

Abstract

The present application discloses a variable frequency control system driven by a compressor, which relates to the field of variable frequency technology for air conditioners, and includes: a data acquisition module for collecting environmental parameters; a storage module for storing historical operating data; an optimization module for using a greedy algorithm to calculate the balance point between user comfort and system energy consumption, the balance point indicating the compressor operating frequency when the system energy consumption is the lowest while the system reaches the required cooling capacity or heating capacity for user comfort, the compressor operating frequency indicating the start-stop cycle of the compressor, and outputting the calculated compressor operating frequency to the control module; the control module uses a PID algorithm to adjust the compressor frequency according to the collected environmental parameters and historical operating data, and uses the compressor operating frequency obtained by the optimization module to adjust the calculated operating frequency. In view of the problem that it is difficult to balance the energy efficiency of the air conditioning system and the user comfort in the prior art, the present application realizes the balanced optimization of user comfort and system energy consumption.

Description

Technical Field

[0001] The present application relates to the technical field of air-conditioning frequency conversion, and in particular to a frequency conversion control system driven by a compressor. Background Art

[0002] In today's society, as people's requirements for quality of life and environmental comfort increase, air conditioning systems have become an indispensable and important equipment in modern buildings. However, while pursuing user comfort, traditional air conditioning systems often ignore the effective use of energy, making it difficult to achieve an ideal balance between system energy efficiency and user comfort. Therefore, how to minimize system energy consumption while ensuring user comfort has become an important direction of current air conditioning system technology research.

[0003] Traditional air conditioning systems usually use compressors that run at fixed frequencies. Their working mode is simple and crude, and they cannot be flexibly adjusted according to actual environmental changes and user needs. This leads to low system efficiency and energy waste in some cases, while also failing to fully meet user comfort requirements. Therefore, there is an urgent need for a new control system that can intelligently adjust the compressor frequency to achieve a balanced optimization between user comfort and system energy consumption.

[0004] In the related art, for example, Chinese patent document CN116255725A provides a variable frequency control system driven by a compressor. A variable frequency control system driven by a compressor includes: a temperature monitoring module for real-time monitoring of indoor temperature and exhaust temperature of the compressor; a data processing module for processing the monitoring data obtained by the state monitoring module to obtain a processing result; and a frequency control module for regulating the frequency of the compressor according to the processing result. The present invention monitors the indoor temperature and regulates the operating frequency of the compressor through the frequency control module according to the difference between the target temperature selected by the user and the indoor temperature, so that the compressor can operate more stably in some overcooled or overheated environments and adapt to environmental changes. However, the scheme adopts a single control strategy based on the indoor temperature difference to adjust the operating frequency of the compressor. Although this strategy can make the system respond to temperature changes more stably, it may not be able to take into account the optimal balance between user comfort and energy efficiency. Because it only focuses on the deviation between the indoor temperature and the target temperature, and does not consider other important factors of user comfort, such as humidity, wind speed, etc. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In response to the problem of difficulty in balancing the energy efficiency and user comfort of air-conditioning systems in the prior art, the present application provides a compressor-driven variable frequency control system, which achieves balanced optimization of user comfort and system energy consumption by establishing a user comfort scoring function and a system energy consumption scoring function, and using a greedy algorithm to iteratively calculate the optimal operating frequency of the compressor.

[0007] 2. Technical solution

[0008] The purpose of this application is achieved through the following technical solutions.

[0009] The embodiment of the present specification provides a variable frequency control system driven by a compressor, including: a data acquisition module, used to collect environmental parameters, the environmental parameters include indoor temperature; a storage module, storing historical operation data, the historical operation data includes temperature and humidity and user comfort, the user comfort indicates the satisfaction degree of the air-conditioning operation effect fed back by the user; a control module, using a PID algorithm to adjust the compressor frequency according to the collected environmental parameters and historical operation data, the compressor frequency indicates the operating frequency of the compressor; an optimization module, using a greedy algorithm, calculating the balance point between user comfort and system energy consumption, the balance point indicates the compressor operating frequency when the system energy consumption is lowest while the system achieves the cooling capacity or heating capacity required for user comfort, the compressor operating frequency indicates the start and stop cycle of the compressor, and the calculated compressor operating frequency is output to the control module.

[0010] Specifically, the data acquisition module plays the role of obtaining environmental parameters in the system. Environmental parameters are important inputs for evaluating and controlling the state of the air conditioning system. One of them, the indoor temperature, is mentioned here because it is directly related to the user's thermal comfort and is one of the key parameters for air conditioning system control. The data acquisition module needs to be set in an appropriate position to detect the temperature signal representing the environmental state of the space. It is generally set at the return air outlet of the air conditioner. The data acquisition module uses a temperature sensor to detect the temperature. Commonly used sensors include thermocouples, thermistors, integrated circuit temperature sensors, etc. Signal amplification and filtering circuits are also required for signal processing. The collected temperature signal needs to be converted to a digital quantity for data transmission, or directly transmitted in a standardized digital signal format. The collected indoor temperature will be input into the control system of the air conditioner, and the operating parameters will be adjusted in real time according to the temperature feedback to meet the thermal comfort requirements.

[0011] Among them, user comfort is a subjective evaluation of the satisfaction of air-conditioning users on the operating effect of air-conditioning. It quantitatively reflects whether the operation of air-conditioning meets the user's expectations through the user's perception of environmental parameters such as temperature, humidity, and wind speed. In this application, the user's scoring feedback on the air-conditioning effect is collected through a mobile app or an Internet panel as user comfort. The user comfort data is associated with the measured values ​​of the environmental parameters for the corresponding time period and stored. The historical curve of user comfort scores under different control schemes is analyzed. User comfort is used as an optimization goal, and the control strategy is adjusted to improve user satisfaction. Compare the comfort evaluations of different users and provide personalized air-conditioning services. Continuously improve the air-conditioning control algorithm through user comfort feedback.

[0012] Specifically, the storage module plays the role of data storage in the system and saves historical operation data. Historical operation data is an important basis for evaluating the control effect of the system. Two types of data, temperature and humidity and user comfort, are specifically mentioned here. Temperature and humidity are important environmental parameters for air conditioning operation. By storing temperature and humidity curves at different time periods, environmental changes can be analyzed. User comfort is obtained through user feedback and is a quantitative representation of user subjective feelings. User ratings can be obtained in the form of questionnaires or buttons. User comfort is associated with temperature and humidity in the corresponding time period and stored. This can evaluate the impact of control strategies on user comfort. The storage module can use relational databases to implement data access management. And provide data interfaces for control modules to call. Analyzing historical operation data helps to obtain the direction of improvement of control strategies and improve user satisfaction.

[0013] Among them, the greedy algorithm is a heuristic algorithm for solving optimization problems. At each step, the best option for the current state is selected without considering whether it is the global optimal option. In this scheme, the compressor frequency that maximizes the local comfort is selected based on the user comfort scoring function. User comfort: It indicates the user's subjective satisfaction with the air-conditioning effect, which can be obtained through automatic questionnaires or user rating feedback. In this scheme, user comfort is used as the optimization goal and maximized through control. System energy consumption: It indicates the energy consumption of the air-conditioning system. It is related to the compressor frequency. In this scheme, the system energy consumption is used as a constraint, and the optimization control minimizes the energy consumption. Balance point: The compressor frequency operating state that minimizes the system energy consumption while meeting the user's comfort requirements. The optimal solution obtained by iterative solution of the greedy algorithm.

[0014] The start-stop cycle refers to the time interval between the start and stop of the compressor. The compressor adjusts the cooling / heating capacity by starting or stopping. In the present application, the shorter the start-stop cycle of the compressor, the more times it starts and stops per unit time, which is equivalent to the compressor running at a higher frequency. After the optimization module calculates the optimal operating frequency of the compressor, it needs to be converted into a control signal for the start-stop cycle in order to control the operation of the compressor. The calculated optimal compressor frequency is mapped to the corresponding start-stop cycle time, and a control signal is output to adjust the start and stop of the compressor. By adjusting the start-stop cycle, the operating frequency of the compressor can be controlled, thereby achieving the technical effect of optimizing user comfort and system energy consumption.

[0015] Specifically, the function of the optimization module is to calculate the optimal balance point between user comfort and system energy consumption. A greedy algorithm is used, with the user comfort score function as the optimization target and the system energy consumption score function as the constraint. Different combinations of compressor frequency parameters are iteratively calculated to select the local optimal solution. When the compressor frequency makes the user comfort meet the requirements and the system energy consumption is minimized, the global optimal solution balance point is obtained. The balance point represents the optimal operating frequency of the compressor at this time. It not only meets the user's cooling / heating needs, but also realizes efficient operation of the system. The optimal compressor frequency obtained by the solution is converted into a control signal of the corresponding start-stop cycle. The frequency control signal is output to the control module in the form of a start-stop cycle. The control module adjusts the operation of the compressor according to the signal to achieve energy-saving optimization.

[0016] Specifically, the control module collects environmental parameters (such as temperature) and historical operating data (such as temperature curve) in real time, and obtains the current operating frequency of the compressor. Compare the collected temperature data with the set temperature target value to find their deviation e. Input the deviation e into the PID algorithm to calculate the weight coefficients of the three control parameters of the PID. The proportional term in the control parameter is used to compensate for the current deviation; the integral term is used to compensate for the historical deviation; and the differential term is used to predict future deviations. The increment Δf of the compressor frequency is calculated by integrating the three PID control quantities. Add the increment Δf to the current compressor frequency to obtain a new frequency control quantity, and output a control signal to the compressor. The compressor receives the new frequency control signal, adjusts the output of the inverter, and realizes the adjustment of the compressor speed, thereby meeting the requirements of air conditioning temperature control. The control module continuously receives the feedback temperature, repeats the above calculation, and adjusts the compressor frequency in real time.

[0017] Furthermore, the storage module uses the CockroachDB distributed relational database to store historical operation data; wherein, the CockroachDB distributed relational database uses the Hint-based Raft consensus algorithm as the consensus protocol; and uses a token ring topology for load balancing. Among them, CockroachDB distributed relational database, CockroachDB is an open source distributed SQL database that stores data in tables by row, supports transaction ACID semantics, and provides a SQL interface similar to PostgreSQL. In this application, its distributed capabilities are used to achieve high availability of storage. Data model design is based on SQL to organize time series data such as temperature and humidity. Use its transaction support to maintain data consistency. Adjust the cluster size according to storage requirements. Read historical data to train the model.

[0018] Hint: relevant context information of the write request, such as the request key range, access mode, etc. In this solution, it is used as hint information to optimize Raft log replication. Raft: a distributed consistency consensus algorithm that achieves high availability through mechanisms such as supplier election and log replication. In this solution, Raft is used for distributed coordination between nodes. Consistency algorithm: an algorithm that keeps data copies of different nodes in a distributed system consistent to ensure strong data consistency. In this solution, Raft is a consistency algorithm. Consensus protocol: a protocol mechanism for nodes in a distributed system to reach a consensus on a certain state. Consensus algorithm is one of the methods to achieve consensus. This application uses Raft as the consensus protocol of CockroachDB. In this application, Hint information is used to optimize the log sorting, writing and other processes within Raft to improve its performance. The Raft algorithm keeps node data strongly consistent, and CockroachDB implements distributed coordination based on Raft. The optimized Raft is used as the consensus protocol to achieve high availability and data consistency of the database.

[0019] Among them, the token ring topology is a logical topology of a distributed database. The nodes form a logical ring, and request scheduling is achieved by passing a control token in sequence on the ring. In this application, the token is passed between nodes in sequence, and the node holding the token is responsible for processing the request. The request load is balanced by adjusting the token distribution.

[0020] Among them, load balancing is a technology that distributes the workload to multiple nodes so that the load is evenly distributed among the nodes. In this application, read and write requests are distributed on database nodes to prevent load overload. Token distribution is dynamically adjusted according to real-time load. The load is automatically redistributed when the node is expanded.

[0021] Furthermore, the control module includes: a measuring unit, used to measure the output frequency y of the compressor and the preset target frequency y'; a calculating unit, connected to the measuring unit, calculating the deviation e between the output frequency y and the target frequency y', and calculating the change ec of the deviation e; an increment calculating unit, connected to the calculating unit, calculating the positive increment value u+ and the negative increment value u-, respectively, wherein: when the change ec is greater than or equal to zero, the positive increment value u+ including the proportional term, the integral term and the differential term is calculated; when the change ec is less than zero, the negative increment value u- including the proportional term is calculated; fuzzy The unit is connected with the increment calculation unit, constructs the membership function of the deviation e and the change ec, calculates the membership degree corresponding to the deviation e and the change ec, and respectively uses the method of obtaining the maximum value and obtaining the minimum value to calculate the maximum fuzzy control quantity A1 and the minimum fuzzy control quantity B1 corresponding to the positive increment value and the negative increment value; the control unit is connected with the fuzzy unit, adopts the fold-integral method to convert the maximum fuzzy control quantity A into the control quantity A2, adopts the fold-integral method to convert the minimum fuzzy control quantity B1 into the control quantity B2, and adjusts the compressor frequency according to the control quantity A2 and the control quantity B2.

[0022] Among them, the output frequency y of the compressor: it represents the real-time operating speed of the compressor, in Hz. The compressor can continuously adjust the speed through frequency conversion technology, and y is its current output frequency value. In this application, the speed signal at the output end of the compressor is detected in real time by a sensor and converted into a frequency value y. y reflects the actual operating state of the compressor and is the process feedback value in the regulation loop. The detected y is compared with the target frequency y' to calculate the optimized control amount.

[0023] Among them, the preset target frequency y' is the expected output frequency setting value of the compressor, obtained according to the optimization calculation, and the unit is also Hz. In this application, the optimal frequency value calculated according to the load is set as the target frequency y' of the compressor. y' is input into the control system as the expected value of the compressor output. The control system adjusts the actual output frequency of the compressor according to the deviation between y and y'.

[0024] Specifically, an electronic speed sensor is installed at the output end of the compressor to detect the rotation speed of the compressor shaft. The measurement principle of the speed sensor is to use the Hall effect or electromagnetic induction to detect the rotation speed of the rotor. The speed signal detected by the sensor is converted into a corresponding frequency value as the current output frequency y of the compressor. The target frequency y' of the compressor is pre-configured into the measuring unit according to the optimization calculation results. The measuring unit has a frequency signal acquisition module, a signal conversion module and a data processing module. The acquisition module receives the analog signal of the speed sensor, and the conversion module converts it into a digital frequency value. The data processing module obtains the current frequency y and the target frequency y' and stores them in the memory. The measurement results are output to the control unit in real time for calculating the adjustment control amount of the compressor. Circuits including filtering, amplification, A / D conversion, etc. can also be set to improve the measurement accuracy. The entire measuring unit has a compact structure and can be directly integrated into the compressor system.

[0025] Specifically, the calculation unit is connected to the measuring unit, and can obtain the current frequency y and the target frequency y' of the compressor output by the measuring unit. The calculation unit includes a signal acquisition interface circuit for receiving the values ​​of y and y'. A microprocessor is also integrated in the calculation unit, which has a certain computing power. The microprocessor reads the y and y' values ​​obtained by the acquisition interface, performs a subtraction operation in the program, and obtains the difference e between the two. e is the deviation value between the current output frequency y and the target frequency y'. The microprocessor continues to calculate the difference between the value of e and the value of e at the previous moment, and obtains the change ec of the deviation e. ec reflects the dynamic error information in the process of adjusting the compressor frequency. The calculation unit outputs the calculated data such as e and ec to the control unit. The accuracy of the calculation unit is determined by the number of bits of the microprocessor and the algorithm design. The calculation unit has a simple and reliable structure, and the calculation is accurate and efficient.

[0026] Among them, the positive increment value u+: the amount by which the compressor frequency needs to be increased, calculated when the deviation ec is greater than or equal to zero. The negative increment value u-, the amount by which the compressor frequency needs to be reduced, is calculated when ec is less than zero. Proportional term: a linear control quantity based on the current deviation e, which responds faster to instantaneous errors. Integral term: a control quantity based on the accumulation of historical deviations, which can eliminate static errors. Differential term: a control quantity based on predicted future deviations to improve system stability. In this application, u+ and u- are calculated as control quantities for increasing and decreasing the compressor frequency, respectively. u+ contains all three terms, which optimizes the control effect; u- contains only the proportional term, which responds faster to negative feedback. The three terms compensate for the control errors from the present, past and future, respectively. The PID algorithm combines the advantages of the three terms to improve control accuracy and stability.

[0027] Among them, membership: The membership indicates the degree to which an element in fuzzy set theory belongs to a fuzzy set. It is calculated based on the membership function. In this application, according to the deviation e and the change ec, their corresponding membership functions are established. The membership function maps e and ec to the interval [0, 1], indicating the degree to which it belongs to a fuzzy set. Calculate the membership values ​​of e and ec under their respective membership functions. The membership of e indicates the degree to which it belongs to "positive deviation" or "negative deviation". The membership of ec indicates the degree to which it belongs to "positive change" or "negative change". According to the membership value, calculate the fuzzy control quantity, such as the maximum value A1 and the minimum value B1.

[0028] Among them, the calculation process of obtaining the maximum value is to select the maximum one in a set of values. The calculation process of obtaining the minimum value is to select the minimum one in a set of values. In the present application, the membership of the deviation e on the positive and negative membership functions is calculated to obtain its maximum value. The maximum membership indicates that e belongs most strongly to the corresponding set of the membership function. According to the maximum membership, the fuzzy control amount A1 is calculated to obtain the corresponding maximum positive increment. The membership of the change ec on the positive and negative membership functions is calculated to obtain its minimum value. The minimum membership indicates that ec belongs most weakly to the corresponding set of the membership function. According to the minimum membership, the fuzzy control amount B1 is calculated to obtain the corresponding minimum negative increment. By obtaining the maximum and minimum memberships, a fuzzy control strategy based on directionality is implemented.

[0029] Specifically, the fuzzy unit is connected behind the incremental calculation unit, and the deviation e and the change ec output by the incremental calculation unit can be obtained. The fuzzy unit includes a fuzzy control algorithm module and an interface circuit. The algorithm module stores a preset membership function for calculating the membership of e and ec. The interface circuit collects the e and ec signals and transmits them to the microprocessor to perform membership calculation. The microprocessor takes the horizontal coordinate on the corresponding membership function according to the e and ec values, and the vertical coordinate is the membership. The membership of e is calculated on the "positive deviation" and "negative deviation" membership functions. The membership of ec is calculated on the "positive change" and "negative change" membership functions. The membership values ​​of the deviation e on the "positive deviation" and "negative deviation" membership functions are calculated, which are μ1 and μ2 respectively. The membership values ​​of the change ec on the "positive change" and "negative change" are also calculated, which are μ3 and μ4. Match μ1 and μ3 to the rule for calculating positive increments, and take the maximum value Max(μ1,μ3) between the two, which is the membership degree of this rule. Similarly, use the minimum value M of μ2 and μ4 min(μ2,μ4) is used as the rule membership for calculating negative increments. Applying the maximum relationship to the membership of the positive increment rule can obtain the fuzzy control quantity A1 of the positive increment. Applying the minimum relationship to the membership of the negative increment rule can obtain the fuzzy control quantity B1 of the negative increment. Therefore, the maximum and minimum memberships correspond to different rules respectively, and are mapped to the corresponding control quantity output according to the shape of the membership function. Thus, the fuzzy strategy of calculating the increment control according to the deviation and the direction of change is realized.

[0030] Among them, the reduced integral method is a digital control method that combines integral control with broken line approximation so that the control quantity changes along the broken line planning target. In this application, the maximum control quantity A1 output by the fuzzy unit is converted to A2 through the reduced integral. The minimum control quantity B1 is converted to B2 through the reduced integral. In the reduced integral calculation, a target increment threshold is set, and the change in the control quantity cannot exceed the threshold. When the quantitative change exceeds the threshold, the control quantity remains unchanged and a step output is achieved. Through the broken line output characteristics, the control quantity is prevented from increasing or decreasing too quickly to ensure system stability. Finally, the compressor frequency is adjusted according to A2 and B2. The reduced integral method ensures both rapid response and control stability.

[0031] Specifically, the control unit is connected behind the fuzzy unit, and can obtain the maximum control quantity A1 and the minimum control quantity B1 output by the fuzzy unit. The control unit includes a microprocessor, an integral algorithm module and a drive module. The microprocessor receives A1 and B1, and calls the algorithm module to perform integral calculations. The algorithm module contains an integral function, sets the target increment threshold, and maintains the control quantity when the variable exceeds the threshold. The microprocessor outputs the integrated A2 and B2 to the drive module. The drive module converts them into corresponding analog or digital signals and connects them to the frequency converter. After receiving the control signal, the frequency converter adjusts the output frequency to the compressor. A2 and B2 determine the step and amplitude of the increase or decrease of the compressor frequency. The control unit updates the output in real time to achieve closed-loop control of the compressor frequency. The control accuracy of the control unit depends on the number of bits of the processor, the integral design, etc.

[0032] Furthermore, the membership function expression is as follows: Wherein, u(x) represents the membership function, which represents the membership corresponding to the input variable; x represents the input variable, which represents the deviation e or the change ec; η represents the peak value of the membership function, which ranges from 0 to 1; represents the center of the membership function, which is determined by the expected value of the input variable x; ρ represents the parameter of the membership function width, which is determined by the distribution range of the input variable x; ν represents the exponential parameter of the membership function shape, with a value range of 0 to 3.

[0033] Specifically, the membership function adopts a Gaussian function form with a peak value. u(x) represents the membership degree of x on the membership function, ranging from 0 to 1. η represents the peak height of the function curve, which can be set to 1 to obtain a standard Gaussian curve. Indicates the center position of the function, corresponding to the expected value of the input variable x. ρ determines the width of the function curve, which is related to the range of x. The larger ρ is, the wider the curve is. ν is called the exponential parameter, which determines the bell shape of the function curve. When ν is 0, it is a standard bell shape. The larger d is, the sharper the curve is. This function maps the input variable x to a membership μ of 0-1. By adjusting η, The parameters ρ and ν can be used to obtain membership functions of different centers, widths, and shapes. Substitute e and ec into the form to construct their positive and negative membership functions. Then calculate the membership degrees of e and ec on their respective membership functions.

[0034] Furthermore, the parameter ρ of the width of the membership function is expressed as follows: ρ = k × σ, where σ is the standard deviation of the input variable x and k is the proportionality coefficient. Specifically, in the membership function u(x), the parameter ρ determines the width of the function. ρ is related to the value range of the input variable x. The wider the range of x, the wider the function curve, so ρ needs to be associated with the statistical distribution of x. Here, the standard deviation σ of x is used to represent its value distribution range. The standard deviation σ reflects the scale of the degree of aggregation of x values. In order to establish the corresponding relationship between ρ and σ, a proportionality coefficient k is introduced. Let ρ be proportional to σ, that is, ρ = k × σ. Among them, k is usually taken as 1 to 3. In this way, the parameter ρ can be determined by the standard deviation σ of the input variable x and a suitable k value is given. When k takes a smaller value, a narrower membership function is obtained; when k takes a larger value, a wider membership function is obtained. By adjusting the value of k, membership functions of different widths can be adaptively obtained. Therefore, according to the statistical distribution of x, the width parameter ρ of the membership function can be reasonably set.

[0035] Furthermore, the expression of the fold integration method is as follows: y = w1 × A1 + w2 × B1 + λ × (A1-B1) 2 , where y represents the converted control quantity; w1 and w2 represent weight coefficients; λ represents the nonlinear correction coefficient; the expression of λ is as follows: Among them, A1 represents the maximum fuzzy control amount; B1 represents the minimum fuzzy control amount.

[0036] Specifically, the output control quantity y of the fold-integral method is composed of three parts: the linear term of the weight coefficient w1 and the maximum control quantity A1, the linear term of the weight coefficient w2 and the minimum control quantity B1, and λ×(A1-B1) 2Nonlinear correction term. The weight coefficients w1 and w2 balance the influence of A1 and B1 on y. λ is the nonlinear correction coefficient, which introduces nonlinear factors. In the λ expression, the denominator A1+B1 represents the total control amount, and the numerator A1-B1 represents the difference between the positive and negative control amounts. When A1>>B1, λ is close to 1, and the nonlinear term plays a major role. When A1≈B1, λ is close to 0, and the effect of the nonlinear term is very small. In this way, through the design of λ, nonlinear compensation is achieved when the positive and negative control amounts are unbalanced. The fold-integral method combines linear output and nonlinear correction to make the change of control amount more reasonable.

[0037] Furthermore, a greedy algorithm is used to calculate the balance point between user comfort and system energy consumption, including: establishing a user comfort scoring function between user comfort and environmental parameters based on the comfort data fed back by the user; establishing a system energy consumption scoring function between system energy consumption and environmental parameters and compressor frequency based on environmental parameters and compressor frequency; taking the user comfort scoring function as the optimization target and the system energy consumption scoring function as the constraint, iteratively calculating different compressor frequency parameter combinations through a simulated annealing algorithm, and selecting the compressor frequency that minimizes the system energy consumption under the premise of meeting the user comfort requirements; wherein, the compressor frequency parameter combination is used as the solution space, and the frequency parameters that optimize the objective function the most are selected one by one, and the local optimal solution is obtained iteratively; the iteration is repeated until the global optimal solution that satisfies the constraints is obtained; the local optimal solution represents the frequency parameter that maximizes the objective function under the current iteration step; the global optimal solution represents the compressor frequency that makes the objective function reach the optimal solution under the constraints in the entire solution space.

[0038] Specifically, collect users' subjective ratings of environmental comfort through questionnaires or feedback terminals. The ratings can be quantitatively expressed in a range of 1 to 5 points. Collect objective parameters of the user's environment, such as temperature, humidity, light, noise, etc. Create a data set with user rating data and corresponding parameters, and apply machine learning algorithms to find the relationship between the two. Linear regression, neural network and other methods can be used to build models and find a mathematical model between user ratings and environmental parameters. Obtain the user comfort rating function f(x1, x2, ..., xn), where xi represents environmental parameters. Based on this function, the user's comfort rating can be predicted by environmental parameters. The model is iteratively optimized through more user evaluation data to improve the accuracy of the rating function. The rating function combined with environmental sensors can evaluate user comfort in real time.

[0039] Specifically, the environmental parameter data during the operation of the air conditioner is collected through sensors, such as indoor and outdoor temperature and humidity, fresh air volume, etc. At the same time, the operating frequency data of the compressor is collected. The data on the power consumption of the system is collected, such as the power parameters recorded by the air conditioner power consumption meter. The data of environmental parameters, compressor frequency and power consumption are used to establish a data set, and each set of data corresponds to an operating state. Apply machine learning algorithms, take environmental parameters and frequency as input, and power consumption as output, to train the system energy consumption scoring model. Linear regression, neural network and other methods can be used for training to obtain the system energy consumption function f(x1,x2,......,xn,y). Among them, xi represents the environmental parameters, y represents the compressor frequency, and the function value represents the corresponding power consumption. The current system energy consumption can be estimated in real time through this scoring function. The function is iteratively optimized through more operating data to improve the scoring accuracy.

[0040] Among them, optimization target: the object function that needs to reach the optimal value, here is the user comfort scoring function f1. Constraints: the constraints that the optimization target needs to meet, here is the system energy consumption scoring function f2. Simulated annealing algorithm: an optimization algorithm based on random search, which gradually reaches the global optimum by controlling random parameters. Compressor frequency parameter combination: a combination of different frequency value options, in which the algorithm searches for the optimal solution. In this application, f1 is used as the optimization target to maximize user comfort. Use f2 as a constraint to limit system energy consumption. Use the simulated annealing algorithm to iteratively search for the optimal combination of frequencies. Select a combination from the frequency parameter options to achieve the best optimization target under the constraints.

[0041] Specifically, a user comfort scoring function f1 and a system energy consumption scoring function f2 have been established. Take f1 as the optimization target and f2 as the constraint. Set the value range space of the compressor frequency and randomly initialize the frequency parameter combination. Apply the simulated annealing algorithm to iteratively search for the optimal solution: Calculate the values ​​of f1 and f2 based on the current parameter combination. Determine whether f1 meets the comfort requirements and whether f2 meets the energy consumption constraints. Randomly adjust the parameters and repeat the calculation until a satisfactory solution is found. As the iterations gradually converge, a parameter combination of the compressor frequency is obtained. This combination minimizes the system energy consumption while satisfying the comfort requirements. A control strategy module can be established to control the compressor frequency in real time based on the optimization results. The optimization can be rerun periodically with new data to update the control strategy.

[0042] Among them, solution space: the set of solutions consisting of all possible compressor frequency combinations. The frequency value range and accuracy determine the size of the space. Local optimal solution: in the current iteration step, the frequency combination that makes the optimization objective function reach the optimal, but it may not be the global optimal. Global optimal solution: in the entire solution space, the frequency combination that makes the objective function reach the optimal under the constraints. In this application, the value range and accuracy of the compressor frequency are defined to construct the solution space. Initially, a frequency combination is randomly selected and the objective function value is calculated. Adjust the parameters of the combination, recalculate the objective function, and accept it if it is better. Repeat the operation until the local optimal solution is obtained. Starting from the local optimal solution, continue to iteratively search for the global optimal solution. When the constraints are met and the objective function cannot be further optimized, the final solution is obtained.

[0043] Specifically, define the range of compressor frequency as the solution space and set the precision. Randomly initialize the frequency parameter combination x and calculate the objective function f(x). Randomly perturb x in a small range to obtain x' and calculate f(x'). If f(x')>f(x), accept the new x', otherwise keep x. Repeat the perturbation and comparison until f(x) cannot be further improved and the local optimal solution x is obtained. * . From x * At the beginning, continue the perturbation search, but accept the new x with a lower score with a lower probability. Reduce the acceptance probability by "cooling down", avoid falling into the local optimum, and expand the search range. If f(x) cannot be improved after a certain number of perturbations, it is considered that the global optimal solution x_optimal is found. Check whether x_optimal meets the constraints. If not, continue searching. Until the global optimal solution that meets the constraints is obtained.

[0044] Furthermore, the user comfort score function expression is as follows: Among them, S represents the user comfort score; T represents the ambient temperature; H represents the ambient humidity; M represents the fresh air volume; a0, a1, a2, a3, b1, b2, b3, c1, c2, c3, d1, d2, d3, d4, d5, and d6 represent coefficients.

[0045] Specifically, the function adopts a polynomial regression model with ambient temperature T, humidity H and fresh air volume M as independent variables. S is the user's subjective rating of comfort as the dependent variable. It contains linear terms, quadratic terms and cubic terms of T, H and M, describing the individual effects of the three parameters on S. It contains cross terms of TH, TM and HM, describing the effect of the interaction of two parameters on S. Each coefficient a0, a1, a2, a3, b1, b2, b3, c1, c2, c3, d1, d2, d3, d4, d5 and d6 needs to be trained based on user rating data. Establish a data set to collect user ratings S under different environmental conditions. Substitute the data set into the function form and use the least squares method to estimate each coefficient. Obtain the polynomial correspondence between the user comfort score and the environmental parameters. Substitute the new environmental parameters T, H and M into the function to calculate the S value and evaluate the comfort.

[0046] Furthermore, the system energy consumption scoring function expression is as follows: Among them, P represents the system energy consumption score; f represents the compressor operating frequency; Q represents the cooling capacity or heating capacity; T represents the ambient temperature; g1, g2, g3, and G represent coefficients.

[0047] Specifically, the function adopts an exponential regression model with compressor frequency f, cooling / heating Q and ambient temperature T as independent variables. P is the system power consumption, which is used as the dependent variable. It contains linear terms of f, Q and T, describing the influence of the three parameters on P. The exponential form is introduced to reflect the nonlinear relationship between system energy consumption and parameter changes. The basic exponential part Ge represents the basic energy consumption of the system. The exponential part contains parameter linear terms, describing the influence of parameters on basic energy consumption. Each coefficient g1, g2, g3, and G needs to be trained based on operating data. Collect energy consumption data under different operating conditions to establish a data set. Substitute the data set into the function form and use the nonlinear regression algorithm to estimate each coefficient. Obtain the corresponding relationship between system energy consumption and influencing factors.

[0048] Furthermore, it also includes: a prediction module, which predicts the load of the compressor using a time series analysis method based on the collected environmental parameters and historical operating data, and outputs the prediction result to the control module.

[0049] Among them, the time series analysis method: a statistical analysis method for establishing a time series prediction model by analyzing the rules between time series data. The load of the compressor: refers to the amount of work that the compressor bears during operation, reflecting the operating status of the compressor. In this application, environmental parameters and load data are collected during the operation of the compressor. The data is organized into a time series form to observe the relationship between the data. Apply time series analysis methods, such as the ARIMA model, to establish a prediction model between environmental parameters and compressor load. When new environmental parameter data is collected, it is brought into the model to predict the compressor load in the future. The prediction results are output to the control module for adjusting the compressor operation strategy.

[0050] Specifically, the data acquisition module collects environmental parameters (temperature, humidity, etc.) and historical compressor load data in real time. Data preprocessing converts the data into a standard time series format, with regular sampling intervals to fill in missing values. Establish a time series prediction model: Find the data correlation between environmental parameters and load, and select input variables. Determine the time lag relationship and establish time series prediction models such as ARIMA. Use historical data to train the model and obtain model parameters. Model prediction: Obtain newly collected environmental parameter data. Bring in the trained time series model to predict the compressor load in the future. Output the prediction to the control module. The control module optimizes the compressor operation strategy based on the prediction results. Regularly retrain the model with new data for adaptive prediction.

[0051] 3. Beneficial effects

[0052] Compared with the prior art, the advantages of this application are:

[0053] Through real-time data collection and historical data analysis, the system can intelligently adjust the compressor frequency, thereby minimizing energy consumption while ensuring user comfort, and achieving a balance between energy efficiency and user comfort;

[0054] The CockroachDB distributed relational database is used, which has high availability and scalability, improves the stability and reliability of the system, and optimizes database performance by load balancing through the token ring topology;

[0055] The control module uses PID algorithm and fuzzy control method to accurately adjust the compressor frequency, so that the system can respond to environmental changes in a timely manner and improve the stability and performance of the air conditioning system;

[0056] The prediction module is based on time series analysis and can accurately predict the compressor load, providing more accurate parameters for the control module and further optimizing the system's operating efficiency;

[0057] By using greedy algorithms and simulated annealing algorithms, the optimal compressor frequency combination can be found in the complex space of multiple parameters, thereby maximizing user comfort while minimizing system energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0059] Figure 1 is an exemplary module diagram of a variable frequency control system driven by a compressor according to some embodiments of this specification;

[0060] Figure 2 is an exemplary flow chart for calculating the minimum compressor frequency according to some embodiments of the present specification. DETAILED DESCRIPTION

[0061] The method and system provided in the embodiments of this specification are described in detail below with reference to the accompanying drawings.

[0062] Figure 1 This is an exemplary module diagram of a compressor-driven variable frequency control system shown in some embodiments of this specification, and a compressor-driven variable frequency control system includes: a data acquisition module, which is used to collect environmental parameters, and the environmental parameters include indoor temperature; a storage module, which stores historical operation data, and the historical operation data includes temperature and humidity and user comfort, and the user comfort indicates the satisfaction of the air-conditioning operation effect fed back by the user; an optimization module, which uses a greedy algorithm to calculate the balance point between user comfort and system energy consumption, and the balance point indicates the compressor operating frequency when the system reaches the cooling capacity or heating capacity required for user comfort and the system energy consumption is the lowest, and the compressor operating frequency indicates the start-stop cycle of the compressor, and the calculated compressor operating frequency is output to the control module; the control module, according to the collected environmental parameters and historical operation data, uses the PID algorithm to adjust the compressor frequency, and the compressor frequency indicates the operating frequency of the compressor, and the compressor operating frequency obtained by the optimization module is used to adjust the calculated operating frequency. A prediction module, according to the collected environmental parameters and historical operation data, uses a time series analysis method to predict the load of the compressor, and outputs the prediction result to the control module.

[0063] Data acquisition module: collects environmental parameters, including indoor temperature. Install temperature sensors. You can choose temperature probes such as thermocouples, thermistors, and thermistors. Install the temperature probe at the indoor sampling point to avoid being affected by sunlight and equipment heat sources. Connect the probe to the data acquisition card and convert it into electrical signal acquisition. The data acquisition card has A / D conversion function and converts it into digital signals. Configure the sampling time interval of the acquisition card, such as collecting temperature values ​​once every 1min. Write a program to read the temperature digital value of the acquisition card sampling port at regular intervals. Transmit the temperature acquisition data to the storage module through the serial port, network, etc. The storage module can be stored in the temperature field of the relational database, or it can be stored as a csv file. Read the temperature data in the database or file in real time for subsequent analysis. The upper and lower alarm limits can be configured to detect temperature anomalies.

[0064] User feedback: Collect user satisfaction with the air conditioner operation effect and build a user comfort scoring function. Set up a user feedback terminal, such as a touch screen installed on the air conditioner control panel. Regularly prompt users to rate the air conditioner operation effect, such as 1-5 points to indicate satisfaction. Collect user scoring data and environmental parameter data at the corresponding time. Build a data set that includes user scores, temperature and humidity, wind speed and other data. Use machine learning algorithms to train the model with environmental parameters as input and user scores as output. You can use linear regression, neural networks and other methods to train a user comfort scoring function. Obtain real-time environmental parameters and bring them into the scoring function to calculate user satisfaction. Continuously iterate and optimize the scoring function to improve accuracy. The scoring function can be used to optimize air conditioning control strategies and improve user comfort.

[0065] Storage module, prepare 3-5 Linux servers as cluster nodes. Configure CPU, memory, network and other resources. Download and install CockroachDB software on each server. Use cockroachstart to initialize nodes. On one server, use cockroachinit to initialize the cluster. Generate a root certificate. Copy the root certificate to other servers, start the cluster nodes, and join the cluster. Use cockroachsql to connect to the cluster and create the database and table schema. Create temperature, humidity, and user rating tables and define field types. Use the ORM framework or CockroachDBdriver to develop applications. The application connects to the CockroachDB database and inserts data using SQL. Data is automatically replicated to multiple nodes to achieve storage redundancy. Monitor node status and add more nodes to expand the cluster. Each data range in CockroachDB corresponds to a Raft group. The group maintains data copy consistency. The Raft leader is responsible for processing write requests and replicating them to other followers. Followers persist write operations to local logs and send ACKs to the leader. After the leader receives most ACKs, it confirms that the log entry has been committed. Before the leader sends the log, a follower is assigned as a hint for the entry. HintFollower sends logs directly to Leader without broadcasting over the network. Leader also sends ACK directly to HintFollower without waiting for other followers. Improves high-throughput and low-latency write performance. Logs are broadcast to other followers only after writing is committed to achieve consistency. Each node in the cluster is evenly allocated a certain range of key value space, called the "token" of the key space. The nodes are logically arranged into a ring, and adjacent nodes manage adjacent key spaces. When creating data, it is mapped to the corresponding node according to the hash of the key value. When querying, it is also mapped to the corresponding node according to the key value to achieve automatic routing. When adding a new node, some tokens are obtained from other nodes and the key space is reallocated. As the number of nodes increases, requests can still be evenly distributed to avoid hot spots. When a node fails, its key space token will be reallocated to other nodes. The token ring will not have a single point, which enhances availability. Balance network traffic and prevent node overload. Create a table temperature_data with fields including timestamp timestamp and temperature temperature value. Create a table humidity_data with fields including timestamp and humidity humidity value. Insert temperature and humidity time series data at time intervals (such as 1 minute). Create table ac_feedback with fields including timestamp and comfort_score, which are user satisfaction ratings for air conditioners. After collecting user feedback, insert rating data.The application reads historical temperature, humidity and rating data from CockroachDB. It is used to train the temperature and humidity prediction model and can also be used to analyze the factors affecting comfort. CockroachDB provides high availability through data replication. The number of nodes can be expanded as needed to handle larger data volumes.

[0066] Prediction module: collect environmental parameters such as temperature, humidity, and inlet and outlet pressure through sensors. Install current and voltage sensors to calculate the real-time power of the compressor as the load parameter. The data acquisition module collects environmental parameters and load data at a fixed frequency. Clean the collected data and handle missing values ​​and outliers. Unify the time granularity and resample at intervals of 1 minute. Format the data so that each variable forms a standard time series structure, with timestamp as the time index. Draw the time series graph of each variable to analyze the correlation between the load and the environmental variables. Calculate the Pearson correlation coefficient between the two time series. Select the environmental variable with the strongest correlation with the load as the model input. Import the preprocessed environmental variables and load time series data into the time series analysis tool. Use methods such as ADF to test the stationarity of the time series. Perform differential processing on non-stationary series. Use ACF and PACF graphs to select the ARIMA model orders p and q. Establish an ARIMA (p, d, q) model, where d is the differential order. Divide the historical data into training data and test data. Use the training data to determine the ARIMA model parameters and obtain a trained model. Verify the model effect on the test data. If the effect is not good, adjust the p and q parameters and re-model. Obtain an ARIMA model with a small test error for prediction. The newly added data every week is used for incremental training to improve the model. Use 2 to 3 years of historical environmental data and load data to train the ARIMA model. Estimate model parameters, including constant terms, autoregressive coefficients, and moving average coefficients, through methods such as maximum likelihood estimation. Obtain an ARIMA model with optimal parameters for future predictions. Obtain new 15-minute environmental data every 15 minutes. Normalize and differentially process the new data to meet the model input requirements. Bring the processed new data into the trained ARIMA model. Based on this, the model calculates the predicted value of the compressor load within 1 hour from the current moment. Obtain the predicted load for each 15-minute period within 1 hour. Transmit the prediction results to the control module to optimize the control strategy. Transmit the compressor load at the 15-minute level in the next 1 hour predicted by the ARIMA model to the control module. The control module analyzes the predicted load curve and determines the peak and valley periods. In the period of low predicted load, appropriately reduce the compressor frequency to save energy. In the period of high predicted load, preheat in advance to ensure that the cooling / heating demand is met. Use newly added operating data to incrementally train the ARIMA model every week. Continuously optimize the model prediction accuracy through parameter adjustment and input variable adjustment. Collect feedback from the control module and evaluate the effect of model prediction on energy-saving control. Continue to iterate the model and prediction method to achieve automatic correction and continuous improvement.

[0067] Optimization Module: Collect the satisfaction scores of a large number of users under different temperature, humidity, and wind speed conditions. Use machine learning algorithms with temperature, humidity, and wind speed as features to train a user comfort score model. Linear regression, neural networks, etc. can be used to obtain a prediction function comfort_score(T, H, V). Collect the system energy consumption data under different frequencies and loads. Train a system energy consumption prediction model energy_score(f, L), where f is the frequency and L is the load. The higher the value of the comfort_score function, the more satisfied the user is. The lower the value of the energy_score function, the lower the system energy consumption. Integrate the two functions as the evaluation index for the greedy algorithm. Initialize the current frequency of the compressor f = f0. Gradually adjust the frequency in the allowed frequency range with a step size of delta_f. Each time the frequency is adjusted, calculate the user comfort comfort_score and the system energy consumption energy_score. If increasing the frequency results in a greater increase in comfort_score, adopt the new frequency. If decreasing the frequency results in a greater decrease in energy_score, also adopt the new frequency. Record the frequency f that minimizes the difference between the two functions. * After looping through all frequency combinations, obtain the optimal frequency f. * Take f * as the output of the optimal operating frequency of the compressor. Run the algorithm repeatedly to continuously optimize the frequency. Define an acceptable threshold for the difference between the comfort and energy consumption scores, for example, set it as delta_score. The current frequency is f, and increase it by one step to f_new. Calculate the difference c_diff between comfort_score(f) and comfort_score(f_new). Calculate the difference e_diff between energy_score(f) and energy_score(f_new). If c_diff > e_diff, adopt the new frequency f_new. If e_diff > c_diff, also adopt the new frequency f_new. Repeatedly compare different frequencies until Finding: c_diff - e_diff < delta_score. This frequency f * is the optimal balance frequency that minimizes the difference between the two scores. Return f * as the optimization result. The optimization module calculates the current optimal frequency f * . Through the API or database method, transmit the value f * to the control module. The control module calls the compressor driver program, sends a control instruction, and sets the frequency to f * . The compressor receives the instruction, adjusts the PWM signal, and changes the operating frequency of the scroll compressor. Monitor that the compressor reaches the expected f *After the frequency is reached, the feedback control cycle is completed. The optimization module will repeat every 10 minutes to calculate the new f * The control module receives the new f * Then adjust the compressor again. Circular feedback and control make the compressor frequency close to the optimal one.

[0068] Figure 2 This is an exemplary flow chart for calculating the minimum compressor frequency according to some embodiments of this specification. The control module: installs a hall speed sensor on the compressor shaft to detect the shaft speed n in real time. Convert the speed n to the corresponding frequency output y. Conversion formula: y = n / 60. The optimization module calculates the optimal frequency y' and sends it to the control module as the target frequency. Every sampling period Ts, the current actual frequency y of the compressor is obtained. Calculate the deviation e = y'-y. Calculate the deviation change: ec k =e k -e k-1. The current deviation e and the change ec are passed to the PID controller. The PID controller calculates the output control amount based on e and ec to adjust the compressor frequency y. In each sampling time Ts, the compressor frequency is detected as y, and the deviation e and the change ec are calculated. Determine the value of ec: If ec is greater than or equal to 0: calculate the proportional term P'=KP×e, KP is the proportional coefficient. Calculate the integral term I'=KI×∑e, KI is the integral coefficient, ∑e is the cumulative sum of e. Calculate the differential term D'=KD×ec, KD is the differential coefficient. Calculate the PID positive increment u+=P+I+D. u+ reflects the positive increment control amount of the frequency at the current moment. Pass u+ to the fuzzy control module to participate in the calculation of the final control amount. In the sampling period Ts, the compressor frequency y is detected, and the deviation e and the change ec are calculated. Determine the value of ec: If ec<0: only calculate the proportional term P'=KP×e, KP is the proportional coefficient. Do not calculate the integral term I and the differential term D. Calculate the PID negative increment u-=P. When ec<0, only proportional control is used to obtain a stable negative feedback. Integral control is not used to prevent integral saturation. u- is passed to the fuzzy control module to participate in the calculation of the final control quantity. Define the input variables e and ec within the scope of discussion. For example, e = [-10, 10], ec = [-5, 5]. According to empirical data, the type of membership function of e and ec is determined, and Gauss type function is often used. Set the relevant parameters of the membership function: center point b, peak a, width c. Construct three membership functions of e: μN(e), μZ(e), μP(e). Construct three membership functions of ec: μN(ec), μZ(ec), μP(ec). Substitute the specific e and ec values ​​obtained by sampling into the corresponding membership function to calculate the membership. For example, μN(e=1)=0.7, which means that when e=1, the membership of the membership function Z is 0.7. Finally, the membership values ​​of e and ec on each membership function are obtained. u+ and u- have been calculated, which represent the compressor frequency control quantity based on PID. In the fuzzy control space, define the range of discussion of the output variable, for example, the range is [-10, 10]. Construct three membership functions of the output variable: μN(u), μZ(u), μP(u). Map u+ and u- to the output space respectively: μ+→μP(μ+)=A1, μ-→μN(μ-)=B1, where A1 is the maximum membership of u+ mapped to the output P membership function. B1 is the maximum membership of u- mapped to the output N membership function. A1 and B1 are respectively used as the maximum control quantity and the minimum control quantity at this moment. A1 and B1 are passed into the fold-integral module to calculate the final control quantity. In the fuzzy control module, the maximum control quantity A1 and the minimum control quantity B1 have been obtained. Substitute A1 and B1 into the fold-integral formula to calculate the control quantity: A2=W1A1+W2B1+λ(A1-B1) 2 , B2=W1B1+W2A1+λ(B1-A1)2 , where w1 and w2 are weight coefficients, and λ is a nonlinear correction coefficient. Two sets of output quantities A2 and B2 are obtained for independent control. A threshold is set, and the final control quantity u is selected within the value range of A2 and B2. u is converted into a PWM control signal and sent to the compressor driver. The driver adjusts the operating frequency of the compressor motor according to the PWM signal. Real-time detection of frequency y, closed-loop control, and approaching the target frequency y'. Obtain the control quantities A2 and B2 calculated by the fold-integral method. Set the mapping relationship between the control quantity value range and the PWM signal, for example, A2 = [0, 10] corresponds to a PWM duty cycle of [20%, 80%]. According to the values ​​of A2 and B2, determine the duty cycle of the PWM duty cycle output value at the current moment. Send the duty cycle signal to the IGBT driver module. The driver module outputs a PWM waveform to control the motor operating frequency. The speed sensor detects the shaft speed n in real time and calculates the frequency y. Compare y with the target frequency y' and calculate the deviation e. If e exceeds the error range, enter the next round of PID+fuzzy control cycle. Accurately adjust the duty cycle value until y stabilizes near y'.

[0069] System energy consumption evaluation: Set up a user feedback module in the air conditioning / heat pump system, and provide a network or physical button so that users can score and evaluate comfort in real time. Set up temperature and humidity sensors in the air-conditioned room to collect temperature and humidity data regularly. Install a speed sensor on the compressor to monitor the compressor operating frequency. Set up a flow meter to count the amount of heat generated by cooling or heating. Collect a large amount of operating data, including user comfort scores, room temperature and humidity, compressor frequency, and cooling / heat. Clean and preprocess the data to remove missing values ​​and outliers. Divide the collected data into training data and test data. Use the training data to establish a user comfort scoring function and a system energy consumption scoring function. Use test data to evaluate the prediction effect of the function. Collect data samples containing temperature T, humidity H, and user score S. The form of the user comfort function is: Substitute sample data into the polynomial regression model and train to solve coefficients a0, a1, a2, a3, b1, b2, b3, c1, c2, c3, d1, d2, d3, d4, d5, d6. Ordinary least squares method can be used to establish the matrix format: Y = Xβ + ε, iterative calculation, so that the residual ε is minimized to obtain the regression coefficient β. β is the parameter a0 to d6 to be calculated. Get the polynomial regression function of user comfort and temperature and humidity. Use test data to verify the prediction effect of the function. Collect sample data including frequency f, cooling capacity Q, temperature T and energy consumption score P. The energy consumption function is in the form of: Substitute the sample data into the exponential function regression model to calculate the parameters g1, g2, g3, and G. Transform the problem into a linear regression form: lnP = lnG + g1f + g2Q + g3T. Use the ordinary least squares method to iteratively calculate to minimize the residual and obtain the regression coefficients. The regression coefficients are g1, g2, g3, and G. Obtain the exponential function relationship between the system energy consumption and frequency, cooling capacity, and temperature. Use the test data to verify the prediction effect of the function. According to the technical specifications of the compressor, determine the possible value range of the frequency f, for example, from 30 Hz to 60 Hz. Set the traversal step size of the frequency, for example, set it to 1 Hz. Initialize the frequency f = 30 Hz. Substitute the current f value into the comfort score function S and calculate S. Substitute the current f value into the energy consumption score function P and calculate P. Record the current f, S, and P values. f' = f + 1 Hz and enter the next iteration. Repeat until f traverses the entire value range. Finally, obtain all (f, S, P) combinations within the frequency value range. Initialize the frequency f and calculate the corresponding S and P. Set the initial temperature T and the cooling rate α. Generate a new frequency f' and calculate S' and P'. If S' > S, accept the new frequency f'. If S' < S, accept f' with a certain probability: Generate a random number rand. If rand < p, accept f' and reduce the temperature T according to the cooling rate α. Repeat until the temperature cools to the stop temperature. During the iteration, record the f = f that maximizes S and satisfies the P constraint * . f * is the optimal compressor frequency.

[0070] Set temperature and humidity sensors at the air outlet of the air conditioner to monitor the real-time ambient temperature and humidity. Install a user feedback button inside the air conditioner to collect the real-time comfort score of the user. Install a speed sensor on the compressor to monitor the operating frequency. Set a flow meter on the refrigeration pipeline to monitor the cooling capacity. Through the data acquisition card, collect the sensor signals, convert them into digital quantities, and transmit them to the control system. In the control system, input the temperature and humidity data into the user comfort score function to calculate the comfort score S. Input the frequency, cooling capacity, temperature, and humidity data into the system energy consumption score function to calculate the energy consumption score P. Determine the optimal operating frequency of the compressor according to S and P calculated by the controller. Define the value range [f min , f max and the traversal step size d f . Initialize the frequency f = f min . Input the current f value into the score function to calculate the comfort score S and the energy consumption score P. Set the initial temperature T0 and the termination temperature T f . Randomly generate a new frequency f' and calculate the corresponding S' and P'. If S' > S, accept the new frequency f'. If S' < S, accept f' with a probability . Reduce the temperature T *=a×T, α is the cooling rate. Repeat until T<T f . Record the maximum S encountered during the process and satisfy the P constraint f = f * .f * The optimal frequency that satisfies the constraints is calculated by the optimization module. * . * The value of f is sent to the compressor control module. The control module queries the corresponding table of frequency and PWM duty cycle. * Look up the table to get the corresponding PWM control signal duty cycle. Output the duty cycle signal to the compressor inverter. The inverter adjusts the motor operating frequency according to the duty cycle signal to control the compressor. Real-time detection of user comfort score S. If the change of S exceeds the set threshold, restart the optimization module to calculate the updated f * . Repeated real-time adjustment of compressor frequency to achieve intelligent energy-saving control.

[0071] The invention of the present application and its implementation methods are described schematically above. The description is not restrictive. Without departing from the spirit or basic features of the present application, the present application can be implemented in other specific forms. What is shown in the accompanying drawings is only one of the implementation methods of the invention of the present application. The actual structure is not limited to this, and any figure mark in the claims should not limit the claims involved. Therefore, if ordinary technicians in this field are inspired by it, without departing from the purpose of the present invention, they can creatively design structural methods and embodiments similar to the technical solution, which should all belong to the scope of protection of the present application. In addition, the word "including" does not exclude other elements or steps, and the word "one" before the element does not exclude the inclusion of "multiple" elements. The multiple elements stated in the product claim can also be implemented by one element through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any specific order.

Claims

1. A variable frequency control system driven by a compressor, comprising: A data acquisition module is used to collect environmental parameters, including indoor temperature; A storage module stores historical operation data, which includes temperature, humidity and user comfort. The user comfort indicates the satisfaction of the air conditioner operation effect as reported by the user. The optimization module uses a greedy algorithm to calculate the balance point between user comfort and system energy consumption. The balance point represents the compressor operating frequency when the system achieves the required cooling or heating capacity for user comfort and the system energy consumption is the lowest. The compressor operating frequency represents the start-stop cycle of the compressor, and the calculated compressor operating frequency is output to the control module; The control module adjusts the compressor frequency using the PID algorithm according to the collected environmental parameters and historical operation data, the compressor frequency represents the operating frequency of the compressor, and uses the compressor operating frequency obtained by the optimization module to adjust the calculated operating frequency; The storage module uses Cockroach DB distributed relational database to store historical operation data; Among them, Cockroach DB distributed relational database uses Hint-based Raft consensus algorithm as consensus protocol and uses token ring topology for load balancing; The control module includes: Measuring unit, used to measure the compressor output frequency y and the preset target frequency ; The calculation unit is connected to the measurement unit to calculate the output frequency y and the target frequency The deviation e of ; Incremental calculation unit, connected to the calculation unit, calculates the positive increment value and negative increment values ,in: When the change When greater than or equal to zero, calculate the positive incremental value including proportional, integral, and differential terms ; When the change When less than zero, calculates negative incremental values ​​including proportional terms ; The fuzzy unit is connected to the incremental calculation unit to construct the deviation e and the change Membership function, calculate the deviation e and change The corresponding membership degree is calculated by respectively using the method of obtaining the maximum value and obtaining the minimum value to calculate the maximum fuzzy control quantity A1 and the minimum fuzzy control quantity B1 corresponding to the positive increment value and the negative increment value; The control unit is connected to the fuzzy unit, and uses the folding integral method to convert the maximum fuzzy control quantity A into the control quantity A2, uses the folding integral method to convert the minimum fuzzy control quantity B1 into the control quantity B2, and adjusts the compressor frequency according to the control quantity A2 and the control quantity B2; The membership function expression is as follows: in, represents the membership function, which represents the membership corresponding to the input variable; x represents the input variable, which represents the deviation e or the change ; Indicates the peak value of the membership function, ranging from 0 to 1; represents the center of the membership function, which is determined by the expected value of the input variable x; The parameter representing the width of the membership function is determined by the distribution range of the input variable x; The exponential parameter representing the shape of the membership function ranges from 0 to 3; Parameter for the width of the membership function The expression is as follows: ,in is the standard deviation of the input variable x, k is the proportionality coefficient; The expression of the fold-integral method is as follows: Among them, y represents the control amount after conversion; w1 and w2 represent weight coefficients; represents the nonlinear correction coefficient; The expression is as follows: Among them, A1 represents the maximum fuzzy control amount; B1 represents the minimum fuzzy control amount.

2. The variable frequency control system for compressor driving according to claim 1, characterized in that: A greedy algorithm is used to calculate the balance point between user comfort and system energy consumption, including: According to the comfort data fed back by users, a user comfort scoring function between user comfort and environmental parameters is established; According to the environmental parameters and the compressor frequency, a system energy consumption scoring function between the system energy consumption and the environmental parameters and the compressor frequency is established; Taking the user comfort score function as the optimization target and the system energy consumption score function as the constraint condition, the simulated annealing algorithm is used to iteratively calculate different compressor frequency parameter combinations, and the compressor frequency that minimizes the system energy consumption while meeting the user comfort requirements is selected. Among them, the compressor frequency parameter combination is used as the solution space, and the frequency parameters that optimize the objective function the most are selected one by one, and the local optimal solution is obtained iteratively; the iteration is repeated until the global optimal solution that meets the constraints is obtained; the local optimal solution represents the frequency parameter that maximizes the objective function under the current iteration step; the global optimal solution represents the compressor frequency that makes the objective function reach the optimal solution in the entire solution space while meeting the constraints.

3. The variable frequency control system for compressor driving according to claim 2, characterized in that: The user comfort score function expression is as follows: Among them, S represents the user comfort score; T represents the ambient temperature; H represents the ambient humidity; M represents the fresh air volume; Represents the coefficient.

4. The variable frequency control system for compressor driving according to claim 3, characterized in that: The system energy consumption scoring function expression is as follows: Among them, P represents the system energy consumption score; f represents the compressor operating frequency; Q represents the cooling capacity or heating capacity; T represents the ambient temperature; Represents the coefficient.

5. The variable frequency control system for compressor driving according to claim 4, characterized in that: Also includes: The prediction module predicts the load of the compressor based on the collected environmental parameters and historical operation data using the time series analysis method, and outputs the prediction results to the control module.

Images