A method for evaluating load adjustable potential

Abstract

The present invention discloses a method for evaluating load adjustable potential. First, historical load data of electricity of power users in various time periods are collected, and the adjustment factors of each time period are calculated using the historical load data. The cloud model vector of the adjustment factor and the adjustable potential is calculated based on the reverse cloud. Secondly, based on the forward cloud, the generated adjustable potential values ​​are screened by considering the operating conditions of the system, the adjustment factor and the Markov state transition matrix obtained from the historical adjustable potential to obtain multiple groups of numerical sequences that meet the requirements. Finally, a distance matrix is ​​calculated based on the dynamic time planning adjustment algorithm DTW, so that the sequence obtained by the cloud model has the greatest similarity with the historical data. Based on this, the adjustable potential of power users is evaluated, which helps power users to better participate in demand response and improve the stability of power system operation.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power dispatching, and in particular to a method for evaluating load adjustable potential. Background Art

[0002] Current demand response potential assessments can be roughly divided into the following three categories: (1) "bottom-up" statistical assessment methods, which evaluate the potential of a single typical user and then aggregate it to obtain the potential of the entire statistical area; (2) statistical analysis methods based on big data, which evaluate the demand response potential of load users through historical data or statistical data; and (3) potential prediction methods that consider user electricity consumption behavior, which consider user indicator characteristics, etc. Due to the large uncertainty of demand response and the ambiguity of the amount of demand response, traditional demand response potential assessment methods cannot effectively evaluate the demand response of power users, and therefore cannot ensure the stable and efficient operation of the power system. Summary of the Invention

[0003] In order to solve the deficiencies mentioned in the above background technology, the purpose of the present invention is to provide a method for evaluating load adjustable potential. Based on the historical load data and historical adjustable potential data of power users collected, the adjustable potential of power users is evaluated based on a two-dimensional chain cloud model and a dynamic time planning adjustment algorithm, thereby achieving the purpose of the invention of quantitatively evaluating the adjustable potential of users under the two characteristics of uncertainty and fuzziness of demand response.

[0004] The object of the present invention can be achieved by the following technical solution: a method for evaluating load adjustable potential, the method comprising the following steps:

[0005] Obtain historical load data of power users in each period and calculate the historical average load in each period;

[0006] The baseline load calculation method considering the adjustment factor uses the historical average load and the historical load data of each period to obtain the baseline load of each period, and calculates the historical adjustable potential data of each period through the baseline load and historical load data of each period;

[0007] The Gaussian inverse cloud algorithm is used to calculate the historical load data and the historical adjustable potential data to obtain the cloud model vector of the power user load adjustment factor and adjustable potential;

[0008] According to the power user load adjustment factor and the cloud model vector of the adjustable potential, the expectation, entropy and super entropy are calculated respectively, and based on the calculated expectation, entropy and super entropy, a normal distribution model is used to generate a random number of the adjustable potential for each time period that meets the system operation conditions and the Markov historical state transition matrix screening conditions;

[0009] The average value of the historical data of the adjustable potential of power users in each time period after optimization of the adjustment factors is calculated. The average value of the historical data of the adjustable potential is used as the benchmark sequence. The distance matrix is ​​calculated based on the dynamic time planning adjustment algorithm, so that the obtained random number sequence of the adjustable potential has the highest similarity with the average value sequence of the historical data of the adjustable potential, which is the adjustable potential evaluation value.

[0010] Preferably, the calculation process of the power user load adjustment factor includes the following steps:

[0011] First, calculate the historical average load for each time period: P a (h) represents the average load at time h, L i (h) represents the actual load value at time h on day i;

[0012] The calculation formula for the adjustment factor is: σ(i,h) is the adjustment factor for correcting the average load at hour h on day i; P r (i,h-1) and P r (i,h-2) are the actual load values ​​at h-1 and h-2 hours on day i; P a (i,h-1) and P a (i,h-2) are the average load values ​​corresponding to h-1 and h-2 hours on the i-th day respectively; if it is the first hour, the actual load values ​​of the 23rd and 24th hours on the i-1th day are used for calculation; if it is the second hour, the actual load values ​​of the 24th hour on the i-1th day and the first hour on the i-th day are used for calculation. The adjustment factor ranges from 0.8 to 1.2.

[0013] Preferably, the cloud model vector of adjustable potential includes three digital characteristic values ​​of expectation, entropy and super entropy of the load adjustment factor value of each power user in each time period and three digital characteristic values ​​of expectation, entropy and super entropy of the adjustable potential value in each time period.

[0014] Preferably, the calculation process of the value of the adjustable potential is as follows:

[0015]

[0016]

[0017] Where, It represents the random value of the adjustment factor of the cloud model at time t on the i-th day, generated by the normal distribution model. NORE() represents the normal distribution model. To adjust the first-order absolute center distance in the factor cloud model vector, that is, the expectation, are random numbers of entropy and superentropy generated by normal distribution in the cloud model vector, and To adjust the entropy in the factor cloud model vector, To adjust the super entropy in the factor cloud model vector; represents the random value of the adjustable potential of the cloud model at time t on day i, generated by the normal distribution model, is the first-order absolute center distance in the adjustable potential cloud model vector, i.e., the expectation, is the random number with entropy and super entropy in the adjustable potential vector, is the entropy in the adjustable potential cloud model vector, is the super entropy in the adjustable potential cloud model vector.

[0018] Preferably, the constraint adjustment of the value of the adjustable potential is:

[0019] and The upper and lower limits of the load adjustable potential are specified; P min and P max The upper and lower limits of the specified load operation process; the constraint on the adjustment factor is: 0.8≤σ pre ≤1.2.

[0020] Preferably, the method for constructing the Markov historical state transition matrix is:

[0021] Setting the historical data of adjustable potential P pot The number of transitions between two adjacent time periods is n T -1, the statistical transition probability of various states is expressed as: In the formula, M and N represent the interval P in the historical data of adjustable potential pot The probability / number of times that the value of is converted from M to M, from M to N, from N to N, and from N to M in two adjacent time periods, and Therefore, the Markov historical state transition matrix is

[0022] Preferably, the screening conditions based on the Markov historical state transition matrix are:

[0023] Let M and N represent the interval Take M and N as two state spaces, and take the first period of each day, that is, 0 o'clock, as the base period. The load at 0 o'clock is generally in interval N, so its state component in N space is 1, and its state component in M ​​space is 0. Therefore, the initial distribution is P(0) = (0, 1), so the absolute distribution of the first point is expressed as: P(1) = P(0)·P = (0, 1)·P = (p1(M), p1(N)), and the predicted state satisfies Therefore, the absolute distribution of the kth point is expressed as P(k)=P(0)·P k =(p k (M),p k (N)), the predicted state satisfies According to the interval corresponding to the state at each time point, the data generated at each time point that does not conform to the interval corresponding to the predicted state is eliminated, and the filtered data is arranged and combined to obtain several groups of chain time series.

[0024] Preferably, the process of calculating the distance matrix based on the dynamic time planning adjustment algorithm is:

[0025]

[0026] Where, Indicates the cumulative value of the Euclidean distance at each time point in a day, n T =24, P t pot represents the historical average value of adjustable potential at time point t, P t pot,pre represents the random value generated by the adjustable potential at time point t.

[0027] Preferably, a device comprises:

[0028] one or more processors;

[0029] a memory for storing one or more programs;

[0030] When one or more of the programs are executed by one or more of the processors, the one or more processors implement the load adjustment potential evaluation method as described above.

[0031] Preferably, a storage medium contains computer executable instructions, which, when executed by a computer processor, are used to perform the load adjustment potential assessment method as described above.

[0032] Beneficial effects of the present invention:

[0033] The present invention calculates the load adjustment factor of each time period based on the collected historical load data and historical adjustable potential of users, uses the inverse Gaussian cloud model to generate cloud models representing the uncertain characteristics of the adjustment factor and the adjustable potential, and then uses the forward two-dimensional chain cloud algorithm to divide the output data into random values ​​as the adjustment factor and the adjustable potential.

[0034] (2) The present invention adopts a two-dimensional chain cloud model to construct a random value of the adjustable potential in each time period, adopts a Markov state transition matrix for screening, takes the average value of the adjustable potential historical data of each time period corrected by the adjustment factor as the benchmark sequence, adopts the dynamic time planning adjustment algorithm (DTW) to calculate the similarity matrix between the random value sequence and the benchmark sequence, takes the time sequence with the maximum similarity as the evaluation value of the adjustable potential of the power user, and can realize the qualitative and quantitative evaluation of the adjustable potential of the power user based on the known historical data.

[0035] (3) The present invention takes into account the uncertainty of electricity users' participation in demand response and the constraints of the safe and stable operation of the system. It uses historical data as a benchmark to evaluate the adjustable potential of electricity users. On the one hand, it reduces the pressure on grid regulation and is conducive to safer and more stable operation of the grid; on the other hand, it helps electricity users better participate in demand response and improve utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0037] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0039] like Figure 1 As shown, a method for evaluating load adjustable potential includes the following steps:

[0040] Step 1: Calculate the adjustment factor

[0041] Collect historical load data of power users, calculate the baseline load of each time period using the baseline load method that takes into account the adjustment factor, and thereby obtain the historical adjustable potential data of each time period.

[0042] Before calculating the adjustment factor, first calculate the average load for each time period: P a (h) represents the average load at time h, L i (h) represents the actual load value at time h on day i; on this basis, the adjustment factor is calculated as follows: σ(i,h) is the adjustment factor for correcting the average load at hour h on day i; P r (i,h-1) and P r (i,h-2) are the actual load values ​​at h-1 and h-2 hours on day i; P a (i,h-1) and P a (i,h-2) are the average load values ​​corresponding to hours h-1 and h-2 on day i, respectively. If the value is hour 1, the actual load values ​​at hours 23 and 24 on day i-1 are used for calculation. If the value is hour 2, the actual load values ​​at hours 24 on day i-1 and hour 1 on day i are used for calculation. The adjustment factor generally ranges from 0.8 to 1.2.

[0043] Adjustable potential data can be expressed as: P pot (h) = P r (h)-P p (h), P r (h) represents the actual load value at the time h on the calculation day, P p (h) represents the baseline load value at time h, P pot (h) represents the calculated load reduction, i.e. the load adjustment potential.

[0044] Step 2: Build cloud model vector

[0045] Collect the load adjustment factor σ of the power load i in each period of the calculation day his and load adjustable potential P pot Historical values, generate n i Two-dimensional historical data pairs

[0046] According to the reverse Gaussian cloud algorithm, the steps for calculating the expectation, entropy, and super entropy of the cloud model vector of the power user load adjustment factor and the historical values ​​of the adjustable potential are as follows:

[0047] (1) Calculate the average value of the historical values ​​of the load adjustment factor and adjustable potential of power users:

[0048] (2) Calculate the first-order absolute center distance between the load adjustment factor and the historical value of the adjustable potential of the power user, that is, the expectation:

[0049] (3) Calculate the second-order absolute center distance of the historical values ​​of the power user load adjustment factor and the adjustable potential, that is, the variance:

[0050] (4) Calculate the entropy of the historical values ​​of the power user load adjustment factor and adjustable potential:

[0051] (5) Calculate the super entropy of the historical values ​​of the load adjustment factor and adjustable potential of power users:

[0052] (6) From this, the cloud model vector representing the power user can be obtained as:

[0053] Step 3: Generate random values ​​for adjustment factors and adjustable potential

[0054] Based on the cloud model vector of the power user calculated in step 2, the forward Gaussian cloud algorithm is used to consider the screening conditions of system operation to generate a random value of the adjustable potential of each power user. Taking into account the system operation conditions and the Markov historical state transition matrix screening conditions, multiple groups of adjustable potential values ​​for each time period are output.

[0055] The screening condition means that the adjustment factor and the random value of the adjustable potential calculated by the cloud model algorithm should meet the requirements of system operation and be consistent with the historical load status of the power user. It is represented by the Markov historical state transition matrix and is divided into the following two parts:

[0056] (1) System operation requirements should meet:

[0057]

[0058] In the above formula, and The upper and lower limits of the load regulation potential specified for the system; P min and P max These are the upper and lower limits of the load specified for the system during operation; 0.8 and 1.2 are derived from historical experience.

[0059] (2) Markov historical state transition matrix:

[0060] In general, the load size and nature of the same period are basically the same, so the load interval state can be used to determine whether the demand is met. The present invention uses the adjustable potential value to determine; assuming that the historical data of the adjustable potential P pot The number of transitions between two adjacent time periods is n T -1, the statistical transition probability of various states is expressed as: M and N represent the interval P in historical data pot The probability / number of times that the value of is converted from M to M, from M to N, from N to N, and from N to M in two adjacent time periods, and So the state transition matrix is

[0061] After obtaining the state transition matrix, it can be screened. Taking the first period of each day, that is, 0 o'clock, as the base period, according to historical experience, the load at 0 o'clock is generally in the interval N, so its state component in the N space is 1, and its state component in the M space is 0, so the initial distribution is P(0) = (0,1), so the absolute distribution of the first point can be expressed as: P(1) = P(0)·P = (0,1)·P = (p1(M), p1(N)), and the predicted state satisfies Therefore, the absolute distribution of the kth point can be expressed as P(k) = P(0)·P k =(p k (M),p k (N)), the predicted state satisfies According to the interval corresponding to the state at each time point, the data generated at each time point that does not conform to the interval corresponding to the predicted state is eliminated, and the filtered data is arranged and combined to obtain several groups of chain time series.

[0062] The steps of generating random values ​​of the adjustable potential of power users by the forward Gaussian cloud algorithm considering the screening conditions are as follows:

[0063] (1) Input the cloud model vector of the power user According to entropy and super entropy Generate a random number that adjusts the entropy of the factor: According to entropy and super entropy Generate random numbers with adjustable potential entropy:

[0064] (2) According to the expectations in the cloud model vector and Generate random numbers for adjustment factors and adjustable potential respectively:

[0065] (3) Inspection and Whether the above screening conditions are met, if the conditions are met, the result is output, and the loop steps are repeated until the random number with adjustable potential is output.

[0066] Step 4: DTW algorithm corrects data

[0067] According to the random number sequence obtained in step 3, the distance matrix between the historical average data of the adjustable potential and the random number sequence generated by the cloud model is calculated using the dynamic time planning adjustment algorithm (DTW), which is expressed as: Select a set of sequences with the smallest distance from the obtained random number sequence. The smallest distance means the greatest similarity, so the obtained random number sequence is closer to the actual value.

[0068] Based on the same inventive concept, the present invention also provides a computer device, which includes: one or more processors, and a memory for storing one or more computer programs; the program includes program instructions, and the processor is used to execute the program instructions stored in the memory. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is used to implement one or more instructions, specifically for loading and executing one or more instructions in a computer storage medium to implement the above method.

[0069] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium having a computer program stored thereon, which executes the above method when executed by a processor. The storage medium can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electrical, magnetic, infrared, or semiconductor system, device or component, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or component.

[0070] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present disclosure. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0071] The above shows and describes the basic principles, main features and advantages of the present disclosure. Those skilled in the art should understand that the present disclosure is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present disclosure. Various changes and improvements may be made to the present disclosure without departing from the spirit and scope of the present disclosure, and such changes and improvements shall fall within the scope of the present disclosure.

Claims

1. A method for evaluating load adjustable potential, characterized in that: The method comprises the following steps: Obtain historical load data of power users in each period and calculate the historical average load in each period; The baseline load calculation method considering the adjustment factor uses the historical average load and the historical load data of each period to obtain the baseline load of each period, and calculates the historical adjustable potential data of each period through the baseline load and historical load data of each period; The Gaussian inverse cloud algorithm is used to calculate the historical load data and the historical adjustable potential data to obtain the cloud model vector of the power user load adjustment factor and adjustable potential; The calculation process of the power user load adjustment factor includes the following steps: First, calculate the historical average load for each time period: P a (h) represents the average load at time h, L i (h) represents the actual load value at time h on day i; The calculation formula for the adjustment factor is: σ(i,h) is the adjustment factor for correcting the average load at hour h on day i; P r (i,h-1) and P r (i,h-2) are the actual load values ​​at h-1 and h-2 hours on day i; P a (i,h-1) and P a (i,h-2) are the average load values ​​corresponding to h-1 and h-2 hours on the i-th day, respectively. If it is the 1st hour, the actual load values ​​of the 23rd and 24th hours on the i-1th day are used for calculation. If it is the 2nd hour, the actual load values ​​of the 24th hour on the i-1th day and the 1st hour on the i-th day are used for calculation. The adjustment factor ranges from 0.8 to 1.

2. According to the power user load adjustment factor and the cloud model vector of the adjustable potential, the expectation, entropy and super entropy are calculated respectively, and based on the calculated expectation, entropy and super entropy, a normal distribution model is used to generate a random number of the adjustable potential for each time period that meets the system operation conditions and the Markov historical state transition matrix screening conditions; The average value of the historical data of the adjustable potential of power users in each time period after optimization of the adjustment factors is calculated. The average value of the historical data of the adjustable potential is used as the benchmark sequence. The distance matrix is ​​calculated based on the dynamic time planning adjustment algorithm, so that the obtained random number sequence of the adjustable potential has the highest similarity with the average value sequence of the historical data of the adjustable potential, which is the adjustable potential evaluation value.

2. A method for evaluating load adjustable potential according to claim 1, characterized in that: The cloud model vector of the adjustable potential includes three digital characteristic values ​​of the expectation, entropy and super entropy of the load adjustment factor value of each power user in each time period and three digital characteristic values ​​of the expectation, entropy and super entropy of the adjustable potential value in each time period.

3. A method for evaluating load adjustable potential according to claim 2, characterized in that: The calculation process of the value of the adjustable potential is as follows: Where, It represents the random value of the adjustment factor of the cloud model at time t on the i-th day, generated by the normal distribution model. NORE() represents the normal distribution model. To adjust the first-order absolute center distance in the factor cloud model vector, that is, the expectation, are random numbers of entropy and superentropy generated by normal distribution in the cloud model vector, and To adjust the entropy in the factor cloud model vector, To adjust the super entropy in the factor cloud model vector; represents the random value of the adjustable potential of the cloud model at time t on day i, generated by the normal distribution model, is the first-order absolute center distance in the adjustable potential cloud model vector, i.e., the expectation, is the random number with entropy and super entropy in the adjustable potential vector, is the entropy in the adjustable potential cloud model vector, is the super entropy in the adjustable potential cloud model vector.

4. A method for evaluating load adjustable potential according to claim 3, characterized in that: The constraint adjustment of the value of the adjustable potential is: and The upper and lower limits of the load adjustable potential are specified; P min and P max The upper and lower limits of the specified load operation process; the constraint on the adjustment factor is: 0.8≤σ pre ≤1.

2.

5. The method for evaluating load adjustable potential according to claim 1, wherein: The method for constructing the Markov historical state transition matrix is: Setting the historical data of adjustable potential P pot The number of transitions between two adjacent time periods is n T -1, the statistical transition probability of various states is expressed as: In the formula, M and N represent the interval P in the historical data of adjustable potential pot The probability / number of times that the value of is converted from M to M, from M to N, from N to N, and from N to M in two adjacent time periods, and Therefore, the Markov historical state transition matrix is 6. A method for evaluating load adjustable potential according to claim 5, characterized in that: The screening conditions based on the Markov historical state transition matrix are: Let M and N represent the interval Take M and N as two state spaces, and take the first period of each day, that is, 0 o'clock, as the base period. The load at 0 o'clock is generally in interval N, so its state component in N space is 1, and its state component in M ​​space is 0. Therefore, the initial distribution is P(0) = (0, 1), so the absolute distribution of the first point is expressed as: P(1) = P(0)·P = (0, 1)·P = (p1(M), p1(N)), and the predicted state satisfies Therefore, the absolute distribution of the kth point is expressed as P(k)=P(0)·P k =(p k (M),p k (N)), the predicted state satisfies According to the interval corresponding to the state at each time point, the data generated at each time point that does not conform to the interval corresponding to the predicted state is eliminated, and the filtered data is arranged and combined to obtain several groups of chain time series.

7. A method for evaluating load adjustable potential according to claim 1, characterized in that: The process of calculating the distance matrix based on the dynamic time planning adjustment algorithm is as follows: Where, Indicates the cumulative value of the Euclidean distance at each time point in a day, n T =24, represents the historical average value of adjustable potential at time point t, represents the random value generated by the adjustable potential at time point t.

8. A device, characterized in that include: one or more processors; a memory for storing one or more programs; When one or more of the programs are executed by one or more of the processors, the one or more processors implement the load adjustment potential assessment method according to any one of claims 1 to 7.

9. A storage medium containing computer-executable instructions, characterized in that: When the computer executable instructions are executed by a computer processor, they are used to perform a load adjustment potential assessment method according to any one of claims 1 to 7.

Images