Harmonic distributed governance method based on self-learning algorithm

Abstract

The application discloses a harmonic distributed treatment method based on a self-learning algorithm, and belongs to the field of harmonic treatment of distribution networks. The method adopts a voltage detection active power filter (VDAPF) to treat harmonics, and the key to distributed treatment of a "double-high" distribution network is proposed. The application proposes a method for quantifying the global treatment effect and the degree of participation of VDAPF in collaborative treatment, respectively, to obtain a preliminary performance factor. The preliminary performance factor is converted into a final performance factor based on a self-learning algorithm, and the local control G-U lifting adjustment characteristics of the VDAPF are combined to realize self-organizing distributed optimization of treatment equipment in a distributed treatment system. The harmonic distributed treatment method based on the self-learning algorithm can solve the problem that scattered treatment is not applicable due to the intensification and decentralization of harmonic sources, and can change passive harmonic treatment into active harmonic treatment.

Description

Technical Field

[0001] This invention relates to harmonic mitigation in power distribution networks, and in particular to a distributed harmonic mitigation method based on a self-learning algorithm. Background Technology

[0002] To address the issues of large-scale grid connection of distributed generation and the massive penetration of modern power electronic loads in distribution networks, it is crucial to actively research remediation solutions and urgently utilize remediation equipment rationally and effectively to better address the power quality problems faced by distribution networks. Pollution sources in new energy distribution networks exhibit dynamic changes, multi-source pollution, and high-frequency harmonic pollution characteristics. Pollution emissions are greater than before, the uncertainty of pollution distribution is stronger, and the difficulty of remediation is correspondingly increased. For large-scale harmonic sources, one-to-one remediation of each harmonic source is no longer feasible. Improving the level and effectiveness of harmonic pollution remediation can be achieved by providing distributed remediation on the grid side, thus balancing economic development and environmental protection. Summary of the Invention

[0003] To address the above problems, this invention provides a distributed harmonic mitigation method based on a self-learning algorithm, which makes the entire distributed system more adaptable to situations where harmonic pollution worsens. It can not only ensure that the voltage distortion rate of nodes meets the standard on a short time scale, but also guarantee that the ideal mitigation effect can be achieved on a long time scale.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A harmonic distributed governance method based on a self-learning algorithm, applied to a new energy distribution network, is characterized by comprising the following steps:

[0006] S1. Quantify the two values ​​of VDAPF’s overall governance effect and VDAPF’s active participation in collaborative governance. Then, assign weights to the quantified values ​​and sum them to obtain the preliminary performance factors of each VDAPF.

[0007] S2. Input the preliminary performance factors obtained in S1 into the self-learning algorithm trained by harmonic current to obtain the final performance factors of each treatment device.

[0008] S3. Substitute the final performance factor obtained in S2 into the integral quantization process to obtain the conductivity regulation degree. Based on the local control GU of VDAPF, the spontaneous response of each VDAPF is realized.

[0009] A further improvement to the technical solution of the present invention is that S1 includes the following specific steps:

[0010] S11, Global governance effectiveness factor ε τ Calculation method:

[0011] The overall governance effectiveness factor ε of the τth assisted governance. τ The formula is as follows:

[0012]

[0013] in

[0014]

[0015]

[0016]

[0017] In the formula f THDv f is the sum of the weighted harmonic voltage distortion rates of all nodes in the system. THDvmax The upper limit of the sum of the weighted harmonic voltage distortion rates of all nodes in the system, i.e., the distortion rate value corresponding to the minimum standard for voltage compliance; N is the total number of nodes in the distributed governance system, μ j Weights set according to the different harmonic level requirements of different nodes; α j U is the sensitivity factor j of the controlled node; h,j U represents the effective value of the voltage at node j with respect to the h-th harmonic. 1,j This represents the effective value of the fundamental voltage at node j.

[0018] S12. The positive quantification of collaborative governance yields the individual performance factor ζ. di,τ Calculation method:

[0019] Individual performance factor ζ in the first instance of collaborative governance with di-level governance equipment di,τ The formula is as follows:

[0020]

[0021] k r,di =S′ VDAPF,di -S VDAPF,di

[0022] In the formula, n represents the number of VDAPFs participating in the collaborative governance process; k r,di To assist in compensating for capacity; S VDAPF,di S′ VDAPF,di , These represent the actual compensation capacity before, the actual compensation capacity after, and the rated capacity of each VDAPF unit, respectively.

[0023] The actual compensation capacity of VDAPF is as follows:

[0024]

[0025] In the formula G h,diU is the h-th harmonic conductance of VDAPF at control node di; h,di Let be the h-th harmonic voltage at node di.

[0026] The actual harmonic mitigation capacity of the VDAPF must meet the capacity limit requirements, and the constraints are as follows:

[0027]

[0028] S13, VDAPF performance factor v di,τ Calculation method:

[0029] v di,τ =ωε τ +(1-ω)ζ di,τ

[0030] In the formula, ω is the weighting coefficient.

[0031] A further improvement to the technical solution of the present invention is that S2 includes the following specific steps:

[0032] S21. Initialize the preliminary performance factors of each VDAPF, and set the actual assistance compensation capacity k′. r,di Assign a value of 0; initially express the factor v di,τ Similarly, assign the value 0; this will determine the optimal number of adjustments. Assign a value of 1; and set the step size for assisted governance (the amount of change in compensation capacity for each output);

[0033] S22. Set a reasonable number of training iterations and input the remaining governance capacity. q di The value of the treatment equipment's participation in assisting treatment is within the range of [0,1]. First, [the following is considered]... For k′ r,di The assignment is performed, and the objective function is to obtain the maximum initial performance factor value and find the corresponding auxiliary compensation capacity k′ of the treatment equipment at this time. r,di The process continues until the maximum remaining capacity is reached.

[0034]

[0035] S23. With the goal of obtaining a larger preliminary performance factor, and with the remaining governance capacity as a constraint, the constraint is expressed as follows:

[0036]

[0037] q di =1-b di ω

[0038] In the formula b diThe coefficient representing the positivity of a certain treatment device; where ω is the weighting coefficient; q di Representing positivity; ω and q di They are negatively correlated, and their relationship is linear.

[0039] S24. Each time the value of the harmonic compensation capacity is adjusted, the preliminary performance factor obtained from the latest training is calculated and compared with the optimal value of the preliminary performance factor recorded in the previous training. The size, if the current preliminary performance factor is greater than Then the current preliminary performance factor value v di,τ Give on the contrary Remain unchanged;

[0040] S25, Through the largest initial performance factor The number of steps τ for assisting in the effort can be determined, and finally k can be obtained. r,di This is also determined. The initial performance factor of each VDAPF can be used to obtain the final performance factor value through the assistance compensation capacity value determined by the self-learning process.

[0041] A further improvement to the technical solution of the present invention is that S3 includes the following specific steps:

[0042] S31. Substitute the final performance factor into the individual performance factor ζ. di,τ The calculation process yields the h-th harmonic command conductance G′. h,di .

[0043] S32. The VDAPF (Variable Energy Discharge Pulse Control) device spontaneously responds based on its GU (Guaranteed Induction Voltage) rise regulation characteristic. At the h-th harmonic frequency, the expression for the GU rise regulation characteristic of the VDAPF is:

[0044] G′ h,di =G h,di -b h,di (U h,di -U′ h,di )

[0045] In the formula, G′ h,di and G h,di These are the h-th harmonic command conductance required for successful participation in assisted governance by the VDAPF connected to node di, and the initial value of the h-th harmonic conductance before participation in assisted governance; b h,di U is the h-th harmonic conductance regulation of the VDAPF at this node; h,di and U′ h,di These represent the initial h-th harmonic voltage value before adjustment and the actual h-th harmonic voltage value after adjustment at node di where VDAPF is located.

[0046] The technological advancements achieved by this invention due to the adoption of the above technical solutions are as follows:

[0047] This invention proposes a distributed harmonic governance method based on a self-learning algorithm. It quantifies the global governance effect of the VDAPF (Vibration Distributed Power Generation Platform) and the VDAPF's active participation in collaborative governance. The introduction of the concept of governance device activity enables each governance device to actively contribute to obtain integrals, achieving self-organized optimization. Simultaneously, the self-learning algorithm achieves fully distributed harmonic governance. This invention differs from centralized optimization methods for harmonic governance, alleviating the computational pressure on the central control center, achieving self-organized optimization, and proactively carrying out collaborative governance work, thus providing a new approach to achieving fully distributed harmonic governance. Attached Figure Description

[0048] Figure 1 This is a flowchart illustrating an embodiment of the present invention;

[0049] Figure 2 This is a flowchart illustrating the preliminary performance factor acquisition process of an embodiment of the present invention;

[0050] Figure 3 This is a flowchart illustrating the final performance factor acquisition process of an embodiment of the present invention.

[0051] Figure 4 This is a flowchart of the governance system response according to an embodiment of the present invention. Detailed Implementation

[0052] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings:

[0053] like Figure 1 As shown, a distributed harmonic governance method based on a self-learning algorithm includes the following steps:

[0054] Step S1: As Figure 2 As shown, the two values ​​of the VDAPF's overall governance effect and the VDAPF's active participation in collaborative governance are quantified separately. The quantified values ​​are weighted and summed to obtain the preliminary performance factor of each VDAPF: Step S11: Confirm the initial operating status of the VDAPF and the value of the initial harmonic compensation capacity.

[0055] Step S12: Quantify the global governance effect factor, and calculate it using the formula shown in (1):

[0056]

[0057] In the formula f THDv f is the sum of the weighted harmonic voltage distortion rates of all nodes in the system. THDvmax f is the upper limit of the sum of the weighted harmonic voltage distortion rates of all nodes in the system, which corresponds to the minimum standard for voltage compliance. To ensure the power quality of the distributed governance system, f... THDvThe value is less than or equal to f THDvmax The value of ε always holds true, therefore we know that ε τ Within the range [0,1], f THDv As shown in equation (2).

[0058]

[0059]

[0060] In the formula, N is the total number of nodes in the distributed governance system, and μ j The weights set according to the differentiated requirements of different nodes for harmonic levels, as shown in formula (3), can be obtained through the sensitivity factor α. j The degree of voltage distortion requirement for node j is characterized by the stringency of the load connected to the node. j The larger the value, the greater α becomes. j denoted as j, the sensitivity factor of the controlled node, whose value is taken as the value corresponding to the highest harmonic tolerance level of the load equipment connected to that node. THDv,j The total voltage distortion rate of the controlled node j is shown in formula (4).

[0061]

[0062] In the formula U h,j U represents the effective value of the voltage at node j relative to the h-th harmonic, which can be obtained through power flow calculations for the h-th harmonic; 1,j This represents the effective value of the fundamental voltage at node j.

[0063] Step S13: Quantify the individual performance factors of collaborative governance, as shown in formula (5):

[0064]

[0065] k r,di =S′ VDAPF,di -S VDAPF,di (6)

[0066] In the formula, n represents the number of VDAPFs participating in the collaborative governance process; k r,di To assist in compensating for capacity; S VDAPF,di S′ VDAPF,di , These represent the actual compensation capacity before, the actual compensation capacity after, and the rated capacity of each VDAPF unit, respectively.

[0067] The actual compensation capacity of VDAPF can be obtained from equation (7):

[0068]

[0069] Where: Gh,di U is the h-th harmonic conductance of VDAPF at control node di; h,di Let be the h-th harmonic voltage at node di.

[0070] The actual harmonic mitigation capacity of VDAPF must meet the capacity limit requirements, and the constraint conditions are as shown in equation (8):

[0071]

[0072] Step S14: Taking into account both the overall governance effect and the individual collaborative governance performance of each governance device, the preliminary performance factor v of the di VDAPF under the τth assisted governance is given. di,τ The calculation method is used as the basis for judging the performance of each harmonic control device in each harmonic control operation, as shown in formula (9):

[0073] v di,τ =ωε τ +(1-ω)ζ di,τ (9)

[0074] In the formula, ω is the weighting coefficient, which takes the value of [0,1]. It can be seen that the larger ω is, the more the governance score focuses on the governance effect; otherwise, it focuses more on the individual governance performance of the governance equipment.

[0075] Step S2: As Figure 3 As shown, the preliminary performance factors obtained in step 1 are input into a self-learning algorithm trained by harmonic current to obtain the final performance factors of each treatment device. The specific steps are as follows:

[0076] Step S21: Initialize the preliminary performance factor of each VDAPF, and set the actual assistance compensation capacity k′. r,di Assign a value of 0; initially express the factor v di,τ Similarly, assign the value 0; this will determine the optimal number of adjustments. Assign a value of 1; and set the step size for assisted governance (the amount of change in compensation capacity for each output);

[0077] Step S22: Set a reasonable number of training iterations and input the remaining governance capacity. q di The value of the treatment equipment's participation in assisting treatment is within the range of [0,1]. First, [the following is considered]... For k′ r,di The assignment is shown in formula (10). The objective function is to obtain the maximum initial performance factor value and find the actual assistance compensation capacity k′ corresponding to the treatment equipment at this time. r,di The process continues until the maximum remaining capacity is reached.

[0078]

[0079] Step S23: With the goal of obtaining a larger preliminary performance factor, and with the remaining governance capacity as a constraint, the constraint is as shown in equation (11):

[0080]

[0081] q di =1-b di ω (12)

[0082] In the formula b di The coefficient representing the positivity of a certain treatment device; where ω is the weighting coefficient; q di Representing positivity; ω and q di The negative correlation is shown in equation (12), and the two are linearly related.

[0083] Step S24: Each time the value of the harmonic compensation capacity is adjusted, the preliminary performance factor obtained from the latest training is calculated and compared with the optimal value of the preliminary performance factor recorded in the previous training. The size, if the current preliminary performance factor is greater than Then the current preliminary performance factor value v di,τ Give on the contrary Remain unchanged;

[0084] Step S25: Through the largest initial performance factor The number of steps τ for assisting in the effort can be determined, and finally k can be obtained. r,di This is also determined. The initial performance factor, through the assistance compensation capacity value determined by the self-learning process, yields the final performance factor value.

[0085] Step S3: As Figure 4 As shown, the final performance factor obtained in step 1 is substituted into the integral quantization process to obtain the conductance regulation. Based on the local control GU of the VDAPF, the spontaneous response of each VDAPF is realized. The specific steps are as follows:

[0086] Step S31: Each VDAPF receives a signal requesting collaborative governance.

[0087] Step S32: Substitute the final performance factor into the individual performance factor ζ di,τ The calculation process yields the h-th harmonic command conductance G′. h,di .

[0088] Step S33: The VDAPF (Variable Energy Discharge Pulse Control) device spontaneously responds based on its GU (Guaranteed Induction Voltage) rise regulation characteristic. At the h-th harmonic frequency, the expression for the GU rise regulation characteristic of the VDAPF is:

[0089] G′ h,di=G h,di -b h,di (U h,di -U′ h,di )

[0090] In the formula, G′ h,di and G h,di These are the h-th harmonic command conductance required for successful participation in assisted governance by the VDAPF connected to node di, and the initial value of the h-th harmonic conductance before participation in assisted governance; b h,di U is the h-th harmonic conductance regulation of the VDAPF at this node; h,di and U′ h,di These represent the initial h-th harmonic voltage value before adjustment and the actual h-th harmonic voltage value after adjustment at node di where VDAPF is located.

[0091] Step S34: After each VDAPF performs collaborative governance, it broadcasts and shares its collaborative output status to all nodes in the distribution network.

[0092] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A distributed harmonic mitigation method based on a self-learning algorithm, applied to the harmonic mitigation scenario of power distribution networks, characterized in that, The method includes the following steps: S1. Quantify the overall governance effect of VDAPF and the enthusiasm of VDAPF in participating in collaborative governance respectively. The quantified values ​​are weighted and summed to obtain the preliminary performance factor of each VDAPF. S2. Input the preliminary performance factor obtained in S1 into the self-learning algorithm trained by harmonic current to obtain the final performance factor of each treatment device. The final performance factor is the final auxiliary compensation capacity. S3. Substitute the final performance factor obtained in S2 into the integral quantization process to obtain the conductivity regulation degree. Based on the local control unit (GU) adjustment characteristics of the VDAPF, realize the spontaneous response of each VDAPF. S1 includes the following specific steps: S11, Global governance effectiveness factor ε τ Calculation method: The factor ε of the overall governance effect of the τth assistance τ The formula is as follows: in: In the formula f THDv f is the sum of the weighted harmonic voltage distortion rates of all nodes in the system. THDvmax The upper limit of the sum of the weighted harmonic voltage distortion rates of all nodes in the system, i.e., the distortion rate value corresponding to the minimum standard for voltage compliance; N is the total number of nodes in the distributed governance system, μ j Weights set according to the different harmonic level requirements of different nodes; α j Let j be the sensitivity factor of the controlled node; W THDv,j U is the total voltage distortion rate of the controlled node j. h,j U represents the effective value of the voltage at node j with respect to the h-th harmonic. 1,j This represents the effective value of the fundamental voltage at node j; S12. The positive quantification of collaborative governance yields individual performance factors. Calculation method: Individual performance factors in the τth instance of collaborative governance with di-level governance equipment The formula is as follows: In the formula, n represents the number of VDAPFs participating in the collaborative governance process; k r,di To assist in compensating for capacity; , , These are the actual compensation capacity before assisted treatment, the actual compensation capacity after assisted treatment, and the rated capacity of each VDAPF unit. The actual compensation capacity of VDAPF is as follows: In the formula G h,di U is the h-th harmonic conductance of VDAPF at control node di; h,di Let h be the harmonic voltage at node di; The actual harmonic mitigation capacity of the VDAPF must meet the capacity limit requirements, and the constraints are as follows: S13, VDAPF's preliminary performance factor v di,τ Calculation method: In the formula, ω is the weighting coefficient.

2. The harmonic distributed governance method based on a self-learning algorithm according to claim 1, characterized in that: Step S2 includes the following specific steps: S21 initializes the initial performance factor of each VDAPF, and sets the actual assistance compensation capacity. Assign a value of 0; initially express the factor v di,τ Similarly, assign the value 0; this will determine the optimal number of adjustments. Assign a value of 1; and set the step size for assisted governance, which is the change in compensation capacity for each output. S22. Set a reasonable number of training iterations and input the remaining governance capacity. q di The value represents the level of enthusiasm of the treatment equipment to participate in the treatment, and is within the range of [0,1]. First, the min( , )right The assignment is performed, and the objective function is to obtain the maximum initial performance factor value and find the corresponding actual auxiliary compensation capacity of the treatment equipment at this time. The process continues until the maximum remaining capacity is reached. : S23. With the goal of obtaining a larger preliminary performance factor, and with the remaining governance capacity as a constraint, the constraint is expressed as follows: In the formula b di q represents the positivity coefficient of a certain treatment device; where ω is the weighting coefficient; di Representing positivity; ω and q di Negative correlation; the two have a linear relationship. S24. Each time the value of the harmonic compensation capacity is adjusted, the preliminary performance factor obtained from the latest training is calculated and compared with the optimal value of the preliminary performance factor recorded in the previous training. The size, if the current preliminary performance factor is greater than Then the current preliminary performance factor value Give ,on the contrary Remain unchanged; S25, Through the largest initial performance factor The number of steps τ for assisting in the output can be determined, and the final k r,di This also determines the initial performance factor of each VDAPF, which is the value of the assistance compensation capacity determined through the self-learning process, and thus the final performance factor value.

3. The harmonic distributed governance method based on a self-learning algorithm as described in claim 1, characterized in that, Step S3 includes: S31. The final performance factor is determined by obtaining the h-th harmonic command conductance. ; S32. The VDAPF (Variable Energy Discharge Pulse Control) device spontaneously responds based on its GU (Guaranteed Induction Voltage) rise regulation characteristic. At the h-th harmonic frequency, the expression for the GU rise regulation characteristic of the VDAPF is: In the formula, and G h,di These are the h-th harmonic command conductance required for successful participation in assisted governance by the VDAPF connected to node di, and the initial value of the h-th harmonic conductance before participation in assisted governance; b h,di U is the h-th harmonic conductance regulation of the VDAPF at this node; h,di and These represent the initial h-th harmonic voltage value before adjustment and the actual h-th harmonic voltage value after adjustment at node di where VDAPF is located.

Images