A distributed event-driven elastic control method for microgrids
Through consistency cooperation indicators and trust propagation mechanisms, update the communication edge weights, and design an event-driven distributed control protocol, solve the problem of bad data injection attack detection and isolation in the microgrid, and realize the recovery and security of microgrid frequency and voltage Sex improvement.
Patent Information
- Application Number
- CN202210890508.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-07-27
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art is difficult to effectively detect and isolate bad data injection attacks in the microgrid while reducing the communication volume, and it is impossible to ensure that the recovery of nodes after the attack joins the microgrid, affecting the stability and security of the power system.
Through consistency cooperation indicators, detect whether the distributed power generation unit is subject to a network attack, update the communication edge weights based on local and public trust values, and design an event-driven distributed elastic control protocol to achieve microgrid frequency and voltage recovery under network attacks.
Effectively detect and isolate bad data injection attacks, ensuring that nodes can rejoin the microgrid after the attack, improving the operational elasticity and security of the microgrid, and having active defense and scalability.
Smart Images

Figure CN115241914B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of microgrid security, and in particular to a distributed event-driven elastic control method for a microgrid. Background Art
[0002] With the gradual depletion of conventional energy sources and the increasing environmental pollution problems, countries around the world have begun to pay attention to environmentally friendly and flexible distributed power generation methods represented by wind power and photovoltaic power generation. Connecting distributed power generation units to the large power grid in the form of a microgrid for grid-connected operation and supporting each other with the large power grid is the best way to exert the efficiency of distributed power generation. Microgrid technology represents the development trend of future distributed power energy supply systems and is an important component and support part of future intelligent power distribution and utilization systems.
[0003] There are two typical operating modes for a microgrid: under normal circumstances, it is grid-connected with the conventional distribution network and is called the grid-connected mode; when a grid fault or power quality does not meet the requirements is detected, the microgrid will be disconnected from the grid in time and operate independently, which is called the island mode. Due to the loss of the support of the main power grid, the dynamic characteristics of the island microgrid usually adopt a hierarchical control structure to control at different time scales. Among them, in the distributed secondary control framework proposed and developed in recent years, the distributed power generation units only interact with neighboring units. Compared with the centralized and decentralized control frameworks, it has the advantages of low communication cost, strong reliability, and good scalability.
[0004] With the deep integration of the information side and the physical side of the microgrid, on the one hand, it has brought about the improvement of the informatization and intelligence of the microgrid. However, at the same time, the large-scale application of power electronic devices has also increased the possibility of the microgrid being attacked by the network. Due to the particularity of the microgrid cyber-physical system, in addition to stealing information, the attacker may cause large-scale accidents such as power supply interruption and serious economic losses. Among them, the bad data injection attack tamps the transmitted data according to the attacker's purpose, destroys the integrity and reliability of the information in the system, and seriously threatens the stability and security of the microgrid system and the entire power system.
[0005] There has been some work on distributed secondary security control. Among them, reference [1] (see S. Abhinav, H. Modares, F. L. Lewis, F. Ferrese, and A. Davoudi, "Synchrony in networked microgrids under attacks," IEEE Transactions on Smart Grid, vol. 9, no. 6, pp. 6731-6741, Nov. 2018) proposed a distributed resilient synchronization control for inverter-based microgrids under attacks. However, it is assumed that the information between distributed generation units is continuously transmitted, which does not conform to the actual network transmission situation. Moreover, with the continuous expansion of the network scale, the problem of limited network bandwidth resources continues to emerge. Reference [2] (see M. Zhai, Q. Sun, B. Wang, R. Wang, Z. Liu, and H. Zhang, "Distributed secondary voltage control of microgrids with actuators bias faults and directed communication topologies: Event-triggered approaches," International Journal of Robust and Nonlinear Control, vol. 32, no. 7, pp. 4422-4437, Feb. 2022) considered the distributed event-driven voltage control problem, where the actuators of distributed generation units are affected by faults with known upper bounds. The faults assumed in this strategy need to know their upper bounds in advance and cannot be applied to the scenario of malicious bad data attacks with unknown upper bounds. Reference [3] (see X. Meng, N. Zhou, and Q. Wang, "Improved distributed event-triggered control for inverter-based AC microgrids under deceptive cyberattacks," International Journal of Electrical Power & Energy Systems, vol. 120, Art. no. 106000, Sept. 2020) studied the distributed event-triggered control of islanded AC microgrids under deceptive attacks, where the interaction information in the distributed control protocol is affected by bounded attack signals with Bernoulli probability distribution, and the boundedness of the error in the mean square sense of the distributed event-triggered protocol is given. However, the attack signals in practice are carefully designed by attackers and often do not satisfy a specific probability distribution.In view of the above problems, it is of great practical significance to design a more efficient distributed resilient control protocol for microgrids while reducing communication volume. Summary of the Invention
[0006] The technical problem to be solved by the present invention is to provide a distributed event-driven resilient control method for a microgrid, which can detect and isolate bad data injection attacks in time while reducing communication burden, and ensure that nodes recovered from the attacks can rejoin the microgrid, improving the operational resilience performance of the microgrid and thus enhancing the security of the power system.
[0007] To solve the above technical problems, the present invention provides a distributed event-driven resilient control method for a microgrid, including the following steps:
[0008] Step 1, primary droop control of the microgrid;
[0009] Step 2, determine whether a bad data injection attack occurs during the process of determining the interaction voltage and frequency status of distributed generation units;
[0010] Step 3, update the local trust value based on the consistency cooperation index;
[0011] Step 4, obtain the comprehensive trust value based on the trust propagation mechanism, and based on this, update the communication link weight between distributed generation units;
[0012] Step 5, establish an event-driven distributed resilient control protocol based on the dynamic communication link weights to achieve the frequency and voltage recovery of the microgrid under network attack conditions.
[0013] Preferably, in Step 1, the primary droop control of the microgrid is carried out according to the following formula:
[0014]
[0015]
[0016]
[0017] In the formula, V i , ω i are respectively the AC voltage amplitude and angular frequency output by distributed generation unit i, and are respectively the droop coefficients of voltage and angular frequency, P i and Q i are respectively the active power and reactive power of distributed generation unit i, ω ni , V ni are respectively the rated values of the primary frequency regulation angular frequency and amplitude.
[0018] Preferably, in step 2, when determining whether a bad data injection attack occurs during the process of the interactive voltage and frequency status of the distributed generation unit, it is carried out according to the following formula:
[0019]
[0020]
[0021]
[0022]
[0023] wherein, is the consistency cooperation index held by generation unit i for whether a voltage bad data attack occurs on generation unit j, is the consistency cooperation index held by generation unit i for whether a frequency bad data attack occurs on generation unit j, is the event-driven moment of generation unit i, is generation unit j at the event-driven moment The voltage status information transmitted to i, is generation unit j at the event-driven moment The frequency status information transmitted to i, if or It indicates that the transmission process from generation unit j to i is under a bad data injection attack.
[0024] Preferably, in step 3, updating the local trust value based on the consistency cooperation index is carried out according to the following formula:
[0025]
[0026]
[0027]
[0028]
[0029] wherein, T ij (t) is the local trust value held by generation unit i for generation unit j, are respectively the counting variables of generation unit j without network attack and suffering from network attack, is the index variable held by generation unit i for whether generation unit j is attacked at the trigger moment If generation unit j is not attacked at the trigger moment then otherwise φ∈[0,1],ψ∈[0,1] are forgetting factors, γ is the attenuation rate in the attack index.
[0030] Preferably, in step 4, the comprehensive trust value is obtained based on the trust propagation mechanism, and then the communication link weight between distributed generation units is updated as follows:
[0031]
[0032]
[0033] where A(t) = a ij (t) is the adjacency matrix of the communication network of each distributed generation unit in the microgrid, R ij (t) is the comprehensive trust value of distributed generation unit i to j determined by combining the local trust value and the trust value of common neighbor m, R min is the trust threshold, χ i > 0, i = 1, 2 are the weights of the local trust value and the neighbor trust value to the comprehensive trust value respectively. The larger the value of the weight χ 1 , the greater the influence of the local trust value on the comprehensive trust value. Conversely, the greater the influence of the common neighbor on the comprehensive trust value. Set the initial value R ij (0) = 1.
[0034] Preferably, in step 5, the distributed elastic control protocol based on event-driven and dynamic communication links is as follows:
[0035]
[0036]
[0037]
[0038]
[0039] or
[0040]
[0041]
[0042]
[0043] In the formula, is the voltage control input of generation unit i, is the frequency control input of generation unit i, i = 1, 2 are the corresponding control gains, represents the neighbor set of distributed generation unit i at time t, p i (t) is the pinning gain of generation unit i, pi (t)≠0 means that the power generation unit i can obtain the reference voltage V ref and the reference frequency ω ref , otherwise p i (t) = 0. is the time interval at the event-driven moment of the power generation unit i, υ i >0 is a small positive constant to avoid the Zeno phenomenon in event-driven control, σ i >0, i = 1, 2 are event-driven parameters, and inf represents taking the lower bound of the elements in a set.
[0044] The beneficial effects of the present invention are as follows: The present invention detects whether the neighbor distributed power generation units are suffering from network attacks through the consistency cooperation index, combines the trust values of the common neighbors and the local trust values, gives the comprehensive trust value of the neighbor power generation units, reflects the reliability of the neighbor power generation units, and based on this, designs a dynamic update strategy for the edge weights of the neighbor communication connections; according to the dynamic communication link update strategy, designs an event-driven distributed control protocol without the Zeno phenomenon, which can effectively reduce the communication burden while timely detecting and isolating the bad data injection attack, and can ensure that the nodes recovered from the attack can rejoin the microgrid, with certain active defense and scalability, and has certain theoretical and practical application values. Description of the Drawings
[0045] Figure 1 is the schematic diagram of the method flow of the present invention.
[0046] Figure 2 is the initial communication topology diagram of the distributed power generation units in the microgrid adopted by the present invention.
[0047] Figure 3 is the local trust value of power generation units 1, 2, 4 to power generation unit 3 after the microgrid of the present invention suffers from a bad injection attack.
[0048] Figure 4 is the comprehensive trust value of power generation units 1, 2, 4 to power generation unit 3 after the microgrid of the present invention suffers from a bad injection attack.
[0049] Figure 5 is the voltage change curve diagram of each distributed power generation unit of the present invention.
[0050] Figure 6 is the frequency change curve diagram of each distributed power generation unit of the present invention. Detailed Embodiments
[0051] As Figure 1 shown, a microgrid distributed event-driven elastic control method includes the following steps:
[0052] Step 1: Primary droop control of the microgrid;
[0053] Step 2: Local detection of bad data injection attacks in the interaction process of distributed generation units;
[0054] Step 3: Update the local trust value based on the consistency cooperation index;
[0055] Step 4: Obtain the comprehensive trust value based on the trust propagation mechanism, and based on this, update the communication link weight between distributed generation units;
[0056] Step 5: Based on the dynamic communication connection edge weight, establish an event-driven distributed resilient control protocol to achieve the restoration of the microgrid frequency and voltage under network attack scenarios.
[0057] Specifically, for the primary droop control of the microgrid in Step 1, it is carried out according to the following formula:
[0058]
[0059]
[0060]
[0061] where V i , ω i are respectively the AC voltage amplitude and angular frequency output by the distributed generation unit i, and are respectively the voltage and angular frequency droop coefficients, P i and Q i are respectively the active power and reactive power of the distributed generation unit i, ω ni , V ni are respectively the rated values of the primary frequency regulation angular frequency and amplitude.
[0062] Specifically, in Step 2, to determine whether a bad data injection attack occurs during the process of determining the interaction voltage and frequency status of the distributed generation unit, it is carried out according to the following formula:
[0063]
[0064]
[0065] where is the consistency cooperation index held by the generating unit i regarding whether a voltage bad data attack occurs on the generating unit j, is the consistency cooperation index held by the generating unit i regarding whether a frequency bad data attack occurs on the generating unit j, is the event-driven moment of the generating unit i, is the voltage status information transmitted by the generating unit j to i at the event-driven moment ; The frequency status information transmitted to i at the event-driven moment of power generation unit j. If or or it indicates that the transmission process from power generation unit j to i is under a bad data injection attack.
[0066] Specifically, in step 3, the local trust value is updated based on the consistency cooperation index, as follows:
[0067]
[0068]
[0069]
[0070]
[0071] where T ij (t) is the local trust value of power generation unit i for power generation unit j, are the counting variables of power generation unit j without network attack and under network attack respectively, is the index variable of whether power generation unit j is under attack at the trigger moment If power generation unit j is not under attack at the trigger moment then otherwise φ∈[0,1], ψ∈[0,1] are forgetting factors, γ is the attenuation rate in the attack index.
[0072] Specifically, in step 4, the comprehensive trust value is obtained based on the trust propagation mechanism, and then the communication link weight between distributed power generation units is updated, as follows:
[0073]
[0074]
[0075] where A(t) = a ij (t) is the adjacency matrix of the communication network of each distributed power generation unit in the microgrid, R ij (t) is the comprehensive trust value of distributed power generation unit i for j determined by combining the local trust value and the trust value of common neighbor m, R min is the reputation threshold, χ i >0, i = 1, 2 are the weights of the local trust value and the neighbor trust value for the comprehensive trust value respectively, and the weight χ 1The larger the value, the greater the influence of the local trust value on the comprehensive trust value. Conversely, the greater the influence of the common neighbors on the comprehensive trust value. Set the initial value R ij (0) = 1.
[0076] Specifically, in step 5, the distributed resilient control protocol based on event-driven and dynamic communication links is carried out through the following formula:
[0077]
[0078]
[0079]
[0080]
[0081] or
[0082]
[0083]
[0084]
[0085] In the formula, is the voltage control input of the power generation unit i, is the frequency control input of the power generation unit i, i = 1, 2 are the corresponding control gains, represents the neighbor set of the distributed power generation unit i at time t, p i (t) is the pinning gain of the power generation unit i, p i (t) ≠ 0 indicates that the power generation unit i can obtain the reference voltage V ref and the reference frequency ω ref of the information, otherwise is the time interval of the event-driven moment of the power generation unit i, υ i > 0 is a small positive constant to avoid the Zeno phenomenon in the event-driven control, σ i > 0, i = 1, 2 are the event-driven parameters, and inf represents taking the lower bound of the elements in a set.
[0086] Applying the above-designed technical solution to the actual situation, considering 4 distributed power generation units, DG1, DG2, DG3, DG4, the initial communication topology between the distributed power generation units is as Figure 2 shown, where the numerical values on the links represent the initial connection weights of the communication links between the DGs. Set that DG1 and DG2 can receive the reference signals V ref and ωref , p 1 (t) = p 2 (t) = 1. Based on MATLAB, a simulation model is built to simulate the control and isolation effects of the microgrid under bad data injection attacks, and the effectiveness of the method of the present invention is verified. The initial values of the voltage and frequency of the 4 DGs are set as: V 1 (0) = 377V, V 2 (0) = 371V, V 3 (0) = 383V, V 4 (0) = 388V, ω 1 (0) = 49.4Hz, ω 2 (0) = 49Hz, ω 3 (0) = 50.2Hz, ω 4 (0) 49.8Hz. The control parameters are set as follows: The event-driven parameters are set as: σ 1 = σ 2 = 0.05, υ i = 0.0012s, and the trust value update parameters are set as: φ = 0.5, ψ = 0.5, γ = 1, χ 1 = 20, χ 2 = 18.
[0087] In the simulation example, the attacker launches an attack on DG3 at 1s and ends at 3s. In t ∈ [1, 3]s, the frequency status information sent by DG3 to DG1 is tampered by the attacker to 60 + 5t Hz, and the voltage status information is tampered to 390 + 5t V. The frequency status information sent by DG3 to DG4 is tampered to 40 - 5t Hz, and the voltage status information is tampered to 370 - 5t V. What DG3 sends to DG2 is the real frequency and voltage information.
[0088] Figure 3 In, the attack occurs at 1s. Based on the consensus cooperation index, the local trust values T 13 (t), T 43 (t) of DG1 and 4 for DG3 drop rapidly. Figure 4 In, based on the trust propagation mechanism, the comprehensive trust values R 13 (t), R 43 (t) of DG1 and 4 for DG3 drop rapidly and are lower than the set threshold R min = 0.93. Thus, DG1 and 4 set the communication link weight between them and DG3 to 0 to isolate the impact of the attack. At the same time, since DG3 has been hijacked by the attacker, DG2 obtains the comprehensive trust value R 23(t), the communication link with the attacked DG3 can be effectively disconnected in a timely manner to prevent potential attack impacts. When the attack ends at t = 3 s, the comprehensive trust values of DG 1 and 4 for DG 3 will return to 1, and DG 3 can rejoin the microgrid.
[0089] Figure 5 and Figure 6 In it, under event-driven elastic control, it is respectively shown that the voltage and frequency of the distributed generation unit in the attack scenario can be restored to the reference values, reducing the impact of bad data attacks on the entire microgrid and improving the resilience and security of the microgrid operation. The method proposed by the present invention has good control effects.
[0090] After adopting the control method of the present invention, the voltage and frequency restoration of the power generation unit in the microgrid can be realized. While reducing the communication burden, network attacks can be effectively detected and isolated, and it is ensured that the nodes recovered from the attacks can rejoin the microgrid, taking into account both normal communication and network attack scenarios, and further improving the security of the microgrid operation.
Claims
1. A distributed event-driven elastic control method for a microgrid, characterized in that: The steps include: Step 1: primary droop control of microgrid; Step 2: determine whether a bad data injection attack occurs during the interactive voltage and frequency state of the distributed generation unit; Step 3: Update the local trust value based on the consistency cooperation indicator; follow the following formula: Among them, T ij (t) is the local trust value held by power generation unit i for power generation unit j, are the count variables of power generation unit j without cyber attack and under cyber attack, respectively. is the power generation unit i holds for power generation unit j at the triggering time The indicator variable of whether it is attacked. If the power generation unit j is at the triggering time If not attacked, otherwise φ∈[0,1],ψ∈[0,1] are forgetting factors, γ is the decay rate in the attack index; Step 4: Obtain a comprehensive trust value based on the trust propagation mechanism, and based on this, update the communication edge weights between distributed generation units; Step 5: Based on the dynamic communication connection edge weight, an event-driven distributed elastic control protocol is established to achieve microgrid frequency and voltage recovery under network attack conditions.
2. The microgrid distributed event-driven elastic control method according to claim 1, characterized in that: In step 1, the primary droop control of the microgrid is performed according to the following formula: In the formula, is the voltage control input of power generation unit i, is the frequency control input of power generation unit i, V i ,ω i are the AC voltage amplitude and angular frequency output by distributed generation unit i, and are the droop coefficients of voltage and angular frequency, P i and Q i are the active power and reactive power of distributed generation unit i, ω ni ,V ni They are the rated values of primary frequency modulation angular frequency and amplitude respectively.
3. The microgrid distributed event-driven elastic control method according to claim 1, characterized in that: In step 2, it is determined whether a bad data injection attack occurs during the interactive voltage and frequency state of the distributed generation unit, according to the following formula: in, is the consistency cooperation indicator held by power generation unit i on whether a voltage bad data attack occurs on power generation unit j, is the consistency cooperation indicator held by power generation unit i on whether a frequency bad data attack occurs on power generation unit j, is the event driving time of power generation unit i, is the power generation unit j at the event driving time The voltage status information transmitted to i, is the power generation unit j at the event driving time The frequency status information transmitted to i, if or This indicates that the transmission process from power generation unit j to i is attacked by bad data injection.
4. The microgrid distributed event-driven elastic control method according to claim 1, characterized in that: In step 4, the comprehensive trust value is obtained based on the trust propagation mechanism, and then the communication edge weights between distributed generation units are updated according to the following formula: Among them, A(t)=a ij (t) is the adjacency matrix of the communication network of each distributed generation unit in the microgrid, R ij (t) is the comprehensive trust value of distributed generation unit i to j determined by combining the local trust value and the trust value of the common neighbor m, R min is the trust threshold, χ i >0, i=1, 2 are the weights of local trust value and neighbor trust value on the comprehensive trust value. The larger the value of weight χ1, the greater the influence of local trust value on the comprehensive trust value. Conversely, the greater the influence of public neighbors on the comprehensive trust value. Set the initial value R ij (0)=1.
5. The microgrid distributed event-driven elastic control method according to claim 1, characterized in that: In step 5, the distributed elastic control protocol based on event-driven and dynamic communication edges is performed according to the following formula: or In the formula, is the voltage control input of power generation unit i, is the frequency control input of generating unit i, i=1,2 is the corresponding control gain, represents the neighbor set of distributed generation unit i at time t, p i (t) is the restraint gain of power generation unit i, p i (t)≠0 means that the power generation unit i can obtain the reference voltage V ref and reference frequency ω ref , otherwise p i (t) = 0, is the time interval between the event driving moments of power generation unit i, υ i >0 is a small positive constant to avoid Zeno phenomenon in event-driven control, σ i >0, i=1, 2 are event-driven parameters, and inf means taking the lower bound of the elements in a set.
Citation Information
Patent Citations
TREATED OFFSET PLATES FOR PREPARATION ON INK JET PLOTTER
AR106000A1
Method and apparatus for maintaining the concentration of a liquid developing solution, in particular for a developing process using the semi-wet method
EP0002020A1
Micro-grid distributed synchronous detection method for false injection attack
CN113285495A
Micro-grid distributed elastic control method based on confidence factors
CN113285496A
Distributed power supply elastic control method for network attack
CN114512983A
