An Online Method for Calculating the Reliable Adjustment Potential of Flexible Loads

Abstract

This invention discloses an online calculation method for the reliable adjustment potential of flexible loads, belonging to the technical field of flexible load adjustment methods. The online calculation method for the reliable adjustment potential of flexible loads in this invention adopts a data-driven approach, characterizing the operation process of flexible loads as a hidden Markov process, and representing uncertain disturbance events through probabilistic representations of the load state transition process. It achieves a highly reliable adjustment potential calculation by incorporating uncertain disturbance factors faced during the adjustment process, based on the existing maximum adjustment potential. This invention solves the problem of insufficient generalization in model-driven methods and the fusion of disturbance probabilities by establishing a data-driven reliable adjustment potential evaluation model, achieving a unified probabilistic representation of the flexible load operation process. This invention proposes an online evaluation method for the reliable adjustment potential of variable frequency air conditioners. Unlike evaluation methods based on macroscopic influencing factors, this invention combines microscopic probabilistic factors such as room temperature deviation and work-rest patterns to accurately evaluate the reliable adjustment potential of air conditioners under different operating conditions.

Description

Technical Field

[0001] This invention belongs to the technical field of flexible load regulation methods, and is particularly related to an online calculation method for the reliable regulation potential of flexible loads. Background Technology

[0002] Existing reliability studies on flexible load regulation processes primarily focus on the reliable capacity perspective. This involves introducing the concept of reliable capacity from renewable energy units to assess the equivalent capacity of load control or demand response, calculating the potential increase / decrease based on historical load data, and then considering the reliability of flexible load capacity through historical average availability or the average availability of generators across the entire system. However, the reliability of the regulation process is rarely considered during the operation of renewable energy units, making it difficult to directly apply to the reliability analysis of inherently unreliable control objects. In addition, some studies model the characteristics and capabilities of load demand response by introducing uncertainty factors, analyzing the uncertainty of price-based and incentive-based demand responses from different perspectives. However, these uncertainty factors are naturally coupled with load state changes, making it difficult to accurately assess the reliability of their regulation potential when independently characterized. Summary of the Invention

[0003] To address the problems existing in the aforementioned background technology, this invention proposes an online method for calculating the reliable adjustment potential of flexible loads.

[0004] Therefore, the present invention adopts the following technical solution: an online calculation method for the reliable adjustment potential of flexible loads, comprising the following steps:

[0005] Step 1: Establish a data-driven, reliable adjustment potential assessment model

[0006] Using a data-driven approach, the operation process of flexible loads is characterized as a hidden Markov model, and a Gaussian Mixture Model (GMM) is used to model the disturbance probability. The Gaussian Mixture Model is obtained by weighted summation of multiple independent Gaussian models.

[0007] Design a reliable adjustment potential assessment model, such as Figure 1 As shown, the steps include:

[0008] The first step: assessing the flexibility of load adjustment potential requires determining the initial state of the load, referring to the baseline and the time interval to be assessed.

[0009]

[0010] In the formula: S init The initial probability distribution π ins An example of this is the initial state of the time interval T to be evaluated; y refer As a potential benchmark, PL (t) is the baseline load curve; t s and t e These are the start and end times of the assessment, respectively.

[0011] Step 2: Based on the initial state, extrapolate the observation time series and the hidden danger time series, given the initial state S. init The implicit state time series representation of the expected load is as follows:

[0012] Q(S init )={S init (t0),S init (t1),L,S init (t E (2)

[0013] In the formula: Q(S) init ) represents the initial state S init The implicit state sequence within time T, S init (t1) represents the state-time sequence Q(S) init The implicit state at time t1;

[0014] Step 3: Calculate the maximum adjustment potential, assuming the initial state and given the adjustment behavior F. adj This causes the load state to change from S init Transformation S chg Then the adjusted load state timing Q(S) chg ) is represented as:

[0015]

[0016] In the formula: S chg (t1) represents the state-time sequence Q(S) chg The hidden state of N at time t1; F The total number of load adjustment commands; for each adjustment command, there is a corresponding state.

[0017] By iterating through all the regulatory behaviors, the maximum regulatory potential of period T is obtained as follows:

[0018]

[0019] In the formula: F max The adjustment instruction for maximum adjustment potential; S max To adjust the initial state corresponding to the instruction, H(Q(S) max The initial state is S. max State-series load curve; Γ max This represents the maximum adjustment potential value.

[0020] Step 4: Introduce perturbation probabilities and calculate credible adjustment potential; after introducing perturbation probabilities, each adjustment behavior F... adj The corresponding Q(S) chg There is a possibility of deviation from expectations, and different probability distribution characteristics will produce different deviation trajectories Q(S). unt At this point, the credible adjustment potential of the flexible load is defined as a two-dimensional space corresponding to the reference benchmark, time interval, and deviation trajectory. The maximum credible adjustment potential evaluation index Γ for the flexible load is defined. acfl for:

[0021]

[0022] In the formula: S uct E represents the initial state corresponding to the credible adjustment potential. acfl For the time interval, reference base, and deviation trajectory Q(S) uct The area integral of ) represents the reliably adjustable charge; |T| is the size of the time interval.

[0023] Step 2: Online Assessment Method for Reliable Regulation Potential of Variable Frequency Air Conditioners

[0024] Using the Viterbi algorithm to deduce reliable hidden state paths The solution process is as follows: Figure 4 As shown, the main steps include:

[0025] Step 1: Propose a modeling method for the observation set and state set of a variable frequency air conditioner. The main factors affecting the operating state of a variable frequency air conditioner include outdoor temperature, indoor temperature, and set temperature. Among these, the temperature difference between the outdoor and indoor temperatures affects the cooling power of the air conditioner, and the difference between the indoor temperature and the set temperature is the target of the air conditioner's cooling. Define the observation state of the variable frequency air conditioner as a binary tuple:

[0026]

[0027] In the formula: T gap Outdoor temperature T outd With indoor temperature T ind The difference, T goal Indoor temperature and set temperature T set The difference;

[0028] The implicit state represents the actual operating state of the air conditioner, which can be represented by a power segment. The operating state includes steady state and transient state, transitioning between each other with a certain probability. A small atomic time interval is defined, within which the state is unaffected by external interference. The direction of state transition is determined after the atomic interval ends. Simultaneously, a state can transition to itself, meaning a steady state can transition to another steady state without transient switching. Therefore, the implicit state of a variable frequency air conditioner is represented using a triplet.

[0029] S i =(t atom ,R fit ,ε fit (7)

[0030] In the formula: t atom R represents the atomic time interval size. fit The fitting parameters for the state power characteristics and ε fit Noise parameters. Through feature extraction from the sample data, the implicit state set of the inverter air conditioner is obtained, including 6 steady states and 6 transient states ("high-frequency cooling", "medium-frequency cooling", "low-frequency cooling", "inverter ramp-up cooling", "inverter ramp-down cooling", "standby fan") and 6 transient states ("on", "off", "cooling up", "cooling down", "cooling standby", "cooling restart"). Figure 5 As shown.

[0031] Step 2: Probabilistically characterize the uncertainties and disturbances in both room temperature deviation and work-rest patterns;

[0032] Assume T set,s and T set,e The minimum and maximum temperatures that can be set for air conditioning cooling, T set,b If the optimal temperature for human comfort is defined as the temperature range within which the human body feels comfortable, then the comfortable temperature range can be approximately characterized by a temperature range of approximately 100°C (T). set,b The distribution follows a Gaussian distribution with a mean of 1: In this case, the probability of disturbance is low in the comfort zone, and high in the deviation from the comfort zone; the density function can be expressed as:

[0033]

[0034] In the formula: E(x) tpt Let E(y) be the probability density function of human comfort temperature preference. Subtracting 1 from this density function and then renormalizing it yields the probability density function E(y) of the disturbance. tpt The mean and variance obtained at this time are μ, respectively. tpt 'and σ' tpt Probability distribution characteristics such as Figure 2 .

[0035] For the probability distribution of uncertain work-rest patterns, a Geometric Model (GMM) can be used for modeling when the sample set is small. With sufficient sample data, by traversing the dataset and statistically analyzing the on / off times, a Bernoulli distribution can be used for modeling, with parameters θ... rt The maximum likelihood method can be used for estimation:

[0036]

[0037] Where: N rtn is the number of samples; rt Time interval The frequency of human intervention. The probability distribution of human intervention in an office of a certain teaching building is as follows: Figure 3 As shown;

[0038] Step 3: Establish an online calculation method for the reliable regulation potential of variable frequency air conditioners.

[0039] Statistical analysis of the sample data yields the implicit state transition probability matrix A and the observed state transition probability matrix B. The initial probability π has only one unique value, representing the transient state start. Given a confidence level of 1-α for room temperature deviation and sleep-wake cycle patterns... tpt and 1-α rt The Viterbi algorithm is used to solve for reliable hidden state deduction paths. Because both room temperature fluctuations and daily routines can simultaneously interfere with the air conditioning's temperature adjustment process, the former will cause users to reset the cooling temperature, while the latter may turn the air conditioning on or off. Clearly, the latter has higher priority than the former. The process is as follows: Figure 4

[0040] As a supplement and improvement to the above technical solution, the present invention also includes the following technical features.

[0041] In step 2, the first step involves blurring the two differences, T. gap The values ​​"small," "medium," and "large" indicate the difference between outdoor and indoor temperatures, respectively. goal The terms "less than the set temperature," "slightly greater than the set temperature," and "greater than the set temperature" indicate that the indoor temperature has reached the set temperature, is close to the set temperature, and is far from the set temperature, respectively. The probability of unpredictable disturbances related to room temperature deviation and daily routines is represented.

[0042] The six steady states specifically include "high-frequency cooling", "medium-frequency cooling", "low-frequency cooling", "inverter ramp cooling", "inverter ramp cooling", and "standby fan", while the six transient states specifically include "on", "off", "cooling up", "cooling down", "cooling standby", and "cooling restart".

[0043] This invention achieves the following beneficial effects: The online calculation method for the reliable adjustment potential of flexible loads adopts a data-driven approach, characterizing the flexible load operation process as a Hidden Markov Process (HMM), and representing uncertain disturbance events through probabilistic representations of the load state transition process. This achieves a highly reliable adjustment potential calculation by incorporating uncertain disturbance factors encountered during the adjustment process, building upon the existing maximum adjustment potential. This invention solves the problem of insufficient generalization in model-driven methods and the fusion of disturbance probabilities by establishing a data-driven reliable adjustment potential evaluation model, achieving a unified probabilistic representation of the flexible load operation process. This invention proposes an online evaluation method for the reliable adjustment potential of variable frequency air conditioners. Unlike evaluation methods based on macroscopic influencing factors, this invention combines microscopic probabilistic factors such as room temperature deviation and work-rest patterns to accurately evaluate the reliable adjustment potential of air conditioners under different operating conditions. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the flexible load reliable adjustment potential assessment process of the present invention.

[0045] Figure 2 This is a schematic diagram of the room temperature offset perturbation probability distribution of the present invention.

[0046] Figure 3 This is a schematic diagram of the probability distribution of the disturbance to the work-rest pattern in this invention.

[0047] Figure 4 This is a schematic diagram of the online calculation process for the reliable regulation potential of variable frequency air conditioners according to the present invention.

[0048] Figure 5 This is a schematic diagram illustrating the basic features of the variable frequency air conditioner of the present invention. Detailed Implementation

[0049] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. The described embodiments are only for illustration and explanation of the present invention and do not constitute the only limitation of the present invention.

[0050] like Figures 1-5 As shown, the online calculation method for the reliable adjustment potential of flexible loads of the present invention includes the following steps:

[0051] Step 1: Establish a data-driven, reliable adjustment potential assessment model

[0052] Using a data-driven approach, the operation process of flexible loads is characterized as a hidden Markov model, and a Gaussian Mixture Model (GMM) is used to model the disturbance probability. The Gaussian Mixture Model is obtained by weighted summation of multiple independent Gaussian models.

[0053] Design a reliable adjustment potential assessment model, such as Figure 1 As shown, the steps include:

[0054] The first step: assessing the flexibility of load adjustment potential requires determining the initial state of the load, referring to the baseline and the time interval to be assessed.

[0055]

[0056] In the formula: S init The initial probability distribution π ins An example of this is the initial state of the time interval T to be evaluated; y refer As a potential benchmark, P L (t) is the baseline load curve; t s and t e These are the start and end times of the assessment, respectively.

[0057] Step 2: Based on the initial state, extrapolate the observation time series and the hidden danger time series, given the initial state S. init The implicit state time series representation of the expected load is as follows:

[0058] Q(S init )={S init (t0),S init (t1),L,S init (t E (2)

[0059] In the formula: Q(S) init ) represents the initial state S init The implicit state sequence within time T, S init (t1) represents the state-time sequence Q(S) init The implicit state at time t1;

[0060] Step 3: Calculate the maximum adjustment potential, assuming the initial state and given the adjustment behavior F. adj This causes the load state to change from S init Transformation S chg Then the adjusted load state timing Q(S) chg ) is represented as:

[0061]

[0062] In the formula: S chg (t1) represents the state-time sequence Q(S) chg The hidden state of N at time t1; F The total number of load adjustment commands; for each adjustment command, there is a corresponding state.

[0063] By iterating through all the regulatory behaviors, the maximum regulatory potential of period T is obtained as follows:

[0064]

[0065] In the formula: F max The adjustment instruction for maximum adjustment potential; S max To adjust the initial state corresponding to the instruction, H(Q(S) max The initial state is S. max State-series load curve; Γ max This represents the maximum adjustment potential value.

[0066] Step 4: Introduce perturbation probabilities and calculate credible adjustment potential; after introducing perturbation probabilities, each adjustment behavior F... adj The corresponding Q(S) chg There is a possibility of deviation from expectations, and different probability distribution characteristics will produce different deviation trajectories Q(S). unt At this point, the credible adjustment potential of the flexible load is defined as a two-dimensional space corresponding to the reference benchmark, time interval, and deviation trajectory. The maximum credible adjustment potential evaluation index Γ for the flexible load is defined. acfl for:

[0067]

[0068] In the formula: S uct E represents the initial state corresponding to the credible adjustment potential. acfl For the time interval, reference base, and deviation trajectory Q(S) uct The area integral of ) represents the reliably adjustable charge; |T| is the size of the time interval.

[0069] Step 2: Online Assessment Method for Reliable Regulation Potential of Variable Frequency Air Conditioners

[0070] The Viterbi algorithm is used to deduce the reliable hidden state path Q*={S q,1 ,L,S q,tE The solution is performed as follows: Figure 4 As shown, the main steps include:

[0071] Step 1: Propose a modeling method for the observation set and state set of a variable frequency air conditioner. The main factors affecting the operating state of a variable frequency air conditioner include outdoor temperature, indoor temperature, and set temperature. Among these, the temperature difference between the outdoor and indoor temperatures affects the cooling power of the air conditioner, and the difference between the indoor temperature and the set temperature is the target of the air conditioner's cooling. Define the observation state of the variable frequency air conditioner as a binary tuple:

[0072]

[0073] In the formula: T gap Outdoor temperature T outd With indoor temperature T ind The difference, Tgoal Indoor temperature and set temperature T set The difference between the two differences is then blurred, T. gap The values ​​"small," "medium," and "large" indicate the difference between outdoor and indoor temperatures, respectively. goal The terms "less than the set temperature," "slightly greater than the set temperature," and "greater than the set temperature" indicate that the indoor temperature has reached the set temperature, is close to the set temperature, and is far from the set temperature, respectively. The probability of unpredictable disturbances related to room temperature deviation and daily routines is represented.

[0074] The implicit state represents the actual operating state of the air conditioner, which can be represented by a power segment. The operating state includes steady state and transient state, transitioning between each other with a certain probability. A small atomic time interval is defined, within which the state is unaffected by external interference. The direction of state transition is determined after the atomic interval ends. Simultaneously, a state can transition to itself, meaning a steady state can transition to another steady state without transient switching. Therefore, the implicit state of a variable frequency air conditioner is represented using a triplet.

[0075] S i =(t atom ,R fit ,ε fit (7)

[0076] In the formula: t atom R represents the atomic time interval size. fit The fitting parameters for the state power characteristics and ε fit Noise parameters. Through feature extraction from the sample data, the implicit state set of the inverter air conditioner is obtained, including 6 steady states ("high-frequency cooling", "medium-frequency cooling", "low-frequency cooling", "inverter ramp-up cooling", "inverter ramp-down cooling", "standby fan") and 6 transient states ("on", "off", "cooling up", "cooling down", "cooling standby", "cooling restart"). Figure 5 As shown.

[0077] Step 2: Probabilistically characterize the uncertainties and disturbances in both room temperature deviation and work-rest patterns;

[0078] Assume T set,s and T set,e The minimum and maximum temperatures that can be set for air conditioning cooling, T set,b If the optimal temperature for human comfort is defined as the temperature range within which the human body feels comfortable, then the comfortable temperature range can be approximately characterized by a temperature range of approximately 100°C (T). set,b The distribution follows a Gaussian distribution with a mean of 1: In this case, the probability of disturbance is low in the comfort zone, and high in the deviation from the comfort zone; the density function can be expressed as:

[0079]

[0080] In the formula: E(x) tpt Let E(y) be the probability density function of human comfort temperature preference. Subtracting 1 from this density function and then renormalizing it yields the probability density function E(y) of the disturbance. tpt The mean and variance obtained at this time are μ, respectively. tpt 'and σ' tpt Probability distribution characteristics such as Figure 2 .

[0081] For the probability distribution of uncertain work-rest patterns, a Geometric Model (GMM) can be used for modeling when the sample set is small. With sufficient sample data, by traversing the dataset and statistically analyzing the on / off times, a Bernoulli distribution can be used for modeling, with parameters θ... rt The maximum likelihood method can be used for estimation:

[0082]

[0083] Where: N rt n is the number of samples; rt Time interval The frequency of human intervention. The probability distribution of human intervention in an office of a certain teaching building is as follows: Figure 3 As shown;

[0084] Step 3: Establish an online calculation method for the reliable regulation potential of variable frequency air conditioners.

[0085] Statistical analysis of the sample data yields the implicit state transition probability matrix A and the observed state transition probability matrix B. The initial probability π has only one unique value, representing the transient state start. Given a confidence level of 1-α for room temperature deviation and sleep-wake cycle patterns... tpt and 1-α rt The Viterbi algorithm is used to solve for reliable hidden state deduction paths. Because both room temperature fluctuations and daily routines can simultaneously interfere with the air conditioning's temperature adjustment process, the former will cause users to reset the cooling temperature, while the latter may turn the air conditioning on or off. Clearly, the latter has higher priority than the former. The process is as follows: Figure 4 .

[0086] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. An online method for calculating the reliable adjustment potential of flexible loads, characterized in that... The online calculation method for reliable adjustment potential includes the following steps: Step 1: Establish a data-driven, reliable adjustment potential assessment model Using a data-driven approach, the operation process of flexible loads is characterized as a hidden Markov model, and a Gaussian mixture model is used to model the disturbance probability. The Gaussian mixture model is obtained by weighted summation of multiple independent Gaussian models. Designing a reliable adjustment potential assessment model includes the following steps: The first step: assessing the flexibility of load adjustment potential requires determining the initial state of the load, referring to the baseline and the time interval to be assessed. In the formula: S init The initial probability distribution π ins An example of this is the initial state of the time interval T to be evaluated; y refer As a potential benchmark, P L (t) is the baseline load curve; t s and t e These are the start and end times of the assessment, respectively. Step 2: Based on the initial state, extrapolate the observation time series and the hidden danger time series, given the initial state S. init The implicit state time series representation of the expected load is as follows: Q(S init )={S init (t0),S init (t1),L,S init (t E )} (2) In the formula: Q(S) init ) represents the initial state S init The implicit state sequence within time T, S init (t1) represents the state-time sequence Q(S) init The implicit state at time t1; Step 3: Calculate the maximum adjustment potential, assuming the initial state and given the adjustment behavior F. adj This causes the load state to change from S init Transformation S chg Then the adjusted load state timing Q(S) chg ) is represented as: In the formula: S chg (t1) represents the state-time sequence Q(S) chg The hidden state of N at time t1; F The total number of load adjustment commands; for each adjustment command, there is a corresponding state. By iterating through all the regulatory behaviors, the maximum regulatory potential of period T is obtained as follows: In the formula: F max The adjustment instruction for maximum adjustment potential; S max To adjust the initial state corresponding to the instruction, H(Q(S) max The initial state is S. max State-series load curve; Γ max The maximum adjustment potential value; Step 4: Introduce perturbation probabilities and calculate credible adjustment potential; after introducing perturbation probabilities, each adjustment behavior F... adj The corresponding Q(S) chg There is a possibility of deviation from expectations, and different probability distribution characteristics will produce different deviation trajectories Q(S). unt At this point, the credible adjustment potential of the flexible load is defined as a two-dimensional space corresponding to the reference benchmark, time interval, and deviation trajectory. The maximum credible adjustment potential evaluation index Γ of the flexible load is defined. acfl for: In the formula: S uct E represents the initial state corresponding to the credible adjustment potential. acfl For the time interval, reference base, and deviation trajectory Q(S) uct The area integral of ) represents the reliably adjustable charge; |T| is the size of the time interval. Step 2: Online Assessment Method for the Reliable Regulation Potential of Variable Frequency Air Conditioners The Viterbi algorithm is used to deduce the reliable hidden state path Q*={S q,1 ,L,S q,tE The solution process includes the following steps: Step 1: Propose a modeling method for the observation set and state set of a variable frequency air conditioner. The main factors affecting the operating state of a variable frequency air conditioner include outdoor temperature, indoor temperature, and set temperature. Among these, the temperature difference between the outdoor and indoor temperatures affects the cooling power of the air conditioner, and the difference between the indoor temperature and the set temperature is the target of the air conditioner's cooling. Define the observation state of the variable frequency air conditioner as a binary tuple: In the formula: T gap Outdoor temperature T outd With indoor temperature T ind The difference, T goal Indoor temperature and set temperature T set The difference; The implicit state represents the actual operating state of the air conditioner, which can be represented by a power segment. The operating state includes steady state and transient state, transitioning between each other with a certain probability. An atomic time interval is defined for each state; within this interval, the state is unaffected by external interference. The direction of state transition is determined after the atomic interval ends. Furthermore, a state can transition to itself, meaning that a steady state can transition to another steady state without needing a transient switch. Therefore, the implicit state of a variable frequency air conditioner is represented using a triplet. S i =(t atom ,R fit ,ε fit ) (7) In the formula: t atom R represents the atomic time interval size. fit The fitting parameters for the state power characteristics and ε fit Noise parameters; through feature extraction from sample data, the implicit state set of the variable frequency air conditioner is obtained, which includes 6 steady states and 6 transient states; Step 2: Probabilistically characterize the uncertainties and disturbances in both room temperature deviation and work-rest patterns; Assume T set,s and T set,e The minimum and maximum temperatures that can be set for air conditioning cooling, T set,b If the optimal temperature for human comfort is defined as the temperature range within which the human body feels comfortable, then the comfortable temperature range can be approximately characterized by a temperature range of approximately 100°C (T). set,b The distribution follows a Gaussian distribution with a mean. In this case, the probability of disturbance is low in the comfort zone, and high in the deviation from the comfort zone. The density function can be expressed as: In the formula: E(x) tpt Let y be the probability density function of human comfort temperature preference; by subtracting the density function from 1 and then renormalizing, we can obtain the density function of the disturbance probability E(y). tpt The mean and variance obtained at this time are μ, respectively. tpt 'and σ' tpt ; For the probability distribution of uncertain work-rest patterns, a Geometric Model (GMM) can be used for modeling when the sample set is small. With sufficient sample data, by traversing the dataset and statistically analyzing the on / off times, a Bernoulli distribution can be used for modeling, with parameters θ... rt The maximum likelihood method can be used for estimation: Where: N rt n is the number of samples; rt Time interval The frequency of human intervention in the process; Step 3: Establish an online calculation method for the reliable regulation potential of variable frequency air conditioners; The implicit state transition probability matrix A and the observed state transition probability matrix B can be obtained from the sample data. The initial probability π has only one unique value, i.e., the transient state start; given the confidence level 1-α for room temperature deviation and work-rest pattern. tpt and 1-α rt The Viterbi algorithm is used to solve for reliable hidden state deduction paths.

2. The online calculation method for the reliable adjustment potential of a flexible load according to claim 1, characterized in that: In step 2, the first step involves blurring the two differences, T. gap "Small," "Medium," and "Large" indicate small, medium, and large differences between outdoor and indoor temperatures, respectively; T goal The values ​​"less than the set temperature", "slightly greater than the set temperature", and "greater than the set temperature" indicate that the indoor temperature has reached the set temperature, is close to the set temperature, and is far from the set temperature, respectively. The probability of uncertain disturbance factors in both room temperature deviation and daily routine is represented.

3. The online calculation method for the reliable adjustment potential of a flexible load according to claim 2, characterized in that: The six steady states specifically include "high-frequency cooling", "medium-frequency cooling", "low-frequency cooling", "inverter ramp cooling", "inverter ramp cooling", and "standby fan", while the six transient states specifically include "on", "off", "cooling up", "cooling down", "cooling standby", and "cooling restart".

Images