CN111030299A - Side channel-based power grid embedded terminal safety monitoring method and system - Google Patents
Side channel-based power grid embedded terminal safety monitoring method and system Download PDFInfo
- Publication number
- CN111030299A CN111030299A CN201911296619.0A CN201911296619A CN111030299A CN 111030299 A CN111030299 A CN 111030299A CN 201911296619 A CN201911296619 A CN 201911296619A CN 111030299 A CN111030299 A CN 111030299A
- Authority
- CN
- China
- Prior art keywords
- sample
- power consumption
- equipment
- power grid
- samples
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
Abstract
The invention discloses a side channel-based power grid embedded terminal safety monitoring method and system, and belongs to the field of intelligent power grid safety. The method comprises the steps of equipment running state information acquisition, data preprocessing, feature extraction, equipment normal running state model establishment, equipment next running state prediction and equipment running state anomaly detection; the operation state information of the power grid embedded terminal equipment is classified, power consumption is selected as the state information of the equipment during operation, and the operation state of the equipment is monitored in real time and compared with the normal operation state model of the equipment for analysis, so that whether the equipment is abnormal in operation is judged. The invention can realize the monitoring of the running state of the embedded terminal equipment of the power grid and improve the safety performance of the terminal equipment.
Description
Technical Field
The invention belongs to the field of intelligent power grid safety, and relates to a side channel-based power grid embedded terminal safety monitoring method and system.
Background
The power system is closely related to our lives, and as a platform for electric energy production and transmission, the power system needs to meet the requirements of reliability, flexibility and economy. With the rapid improvement of the informatization degree of the power system, the power grid is continuously developed towards a more intelligent direction. The embedded technology has become an indispensable part for controlling and monitoring a power system as one of the technologies with the widest application range at present, is widely applied to various links of an intelligent power grid, such as a PLC, an RTU, an HMI and the like in a power engineering control system, and plays a crucial role in the development of the intelligent power system. The power grid embedded terminal equipment brings more security risks while the power grid is more networked, intelligent and multifunctional. Because embedded equipment is widely deployed in the privacy sensitive and safety field, once damaged, the safety of the electric power engineering control system is greatly influenced, the electric power equipment is in failure, the normal operation of the smart grid is threatened, and the national and social safety are seriously threatened. The safety monitoring of the embedded terminal equipment of the intelligent power grid is beneficial to timely discovering the abnormal condition of the equipment, so that the system can predict and intercept the attack before being subjected to illegal attack, and the safety and stability operation of the power system can be guaranteed.
At present, security research aiming at the embedded terminal equipment of the smart power grid mainly focuses on the aspects of access control and security evaluation models, and few researches are carried out on the security monitoring problem of the embedded terminal equipment. The method based on the side channel provided by the invention is used for monitoring the running state of the equipment by utilizing side channel information leakage such as time consumption, power consumption and the like in the running process of the equipment. At present, more side channel information is used, such as power consumption, electromagnetic signals, sound, time and the like, and the information can often reflect some internal information of equipment. Aiming at the characteristics of the embedded terminal of the intelligent power grid, the invention aims to adopt the power consumption information of a CPU in the equipment to carry out safety monitoring on the embedded equipment of the intelligent power grid.
Disclosure of Invention
The invention discloses a side channel-based power grid embedded terminal equipment safety monitoring method, which realizes monitoring of the running state of embedded terminal equipment by analyzing the state information of the terminal during running. And establishing a normal operation state model of the equipment through operation state information acquisition, data preprocessing and feature extraction, thereby realizing the safety monitoring of the power grid terminal equipment. The invention can realize the monitoring of the running state of the embedded terminal equipment of the power grid and improve the safety performance of the terminal equipment.
The method adopts a recurrent neural network based on a long and short memory unit to model the running state of the embedded equipment of the power grid. Long Short-Term Memory (LSTM) is a special case of Recurrent Neural Network (RNN). The method overcomes the problem of gradient disappearance of the recurrent neural network, can learn long-term dependence information, processes long-sequence data, and has wide application in the industry. The LSTM can well capture the characteristics of the time sequence and predict the value of the next time point, so that the prediction fitting degree is high, and the prediction problem required to be solved by the invention is met.
In order to achieve the purpose, the invention adopts the following technical scheme:
a side channel-based power grid embedded terminal safety monitoring method comprises the following steps:
1) acquiring a CPU power consumption signal of the embedded equipment by adopting a mutual inductance mode;
2) filtering direct current signals in the collected CPU power consumption signals, and slicing the filtered time domain power consumption information, wherein each section is used as a sample to obtain a power consumption sample set;
3) extracting initial characteristics of the samples from the power consumption sample set, and designing a characteristic screening algorithm to screen the initial characteristics to obtain an optimal characteristic set; the feature screening algorithm comprises the following steps:
s1: setting a linear classifier: f ═ WTx + b, where f is the estimated classmark number for a sample, x is an m-dimensional feature vector, W is the coefficient corresponding to each dimension of the feature in the classifier, and b is the bias constant;
s2: selecting a loss function:
where n denotes the number of input samples, yiRepresenting the output of the model;
s3: in the feature screening process, the redundant feature coefficients which are not selected are zero, and the optimization target formula is expressed as follows:
wherein | | W | | ceiling0Denotes xiThe weight of (c);
s4: solving an optimized target formula by using a least square method, wherein the concrete solving process is as follows: selecting MpA positive sample and MnEach sample is an m-dimensional vector and a real label of each sample is marked, wherein the label of the positive sample is +1, and the label of the negative sample is-1;
4) constructing a long and short memory unit neural network, which comprises an input layer with n multiplied by m units, an output layer with m units and two hidden layers, wherein m is the number of the characteristics in the optimal characteristic set, and n represents the number of samples; the two hidden layers are all connected, namely each hidden unit in the first hidden layer is connected with each hidden unit in the second hidden layer through feedforward connection; training the long and short memory unit neural networks by adopting the optimal feature set obtained in the step 3) to obtain trained long and short memory unit neural network models;
5) predicting an input signal to be detected by using a trained long and short memory unit neural network model, comparing and judging the input signal with an actually acquired signal at the next moment, and outputting a result that whether the equipment normally operates or not, wherein the method specifically comprises the following steps:
5.1) feature value prediction at the next moment: setting a sample window, and predicting by using a long and short memory unit neural network model according to the collected power consumption samples at the previous n-100 moments to obtain m-dimensional predicted power consumption characteristics of the power consumption samples at the next moment
5.2) collecting the actual value at the next moment: setting a sample window, acquiring and obtaining an actual power consumption sample at the next moment and obtaining actual power consumption characteristics
Calculating to obtain the actual power consumption characteristic x(t+1)And predicting power consumption characteristic X′(t+1)Error existing between
Obtaining an error vector
5.3) error comparison: when the error vector satisfies
If so, regarding the power consumption sample at the (t +1) moment as an abnormal sample, otherwise, regarding the power consumption sample as a normal sample, and the tau is a preset threshold value; updating the sample window, if the current sample obtained by detection is a normal sample, taking the current sample as the sample at the 100 th moment, and combining the samples before the current sample is detected as a normal sampleThe next moment is continuously predicted by 99 samples; if the current sample is an abnormal sample, the predicted power consumption characteristic X is obtained′(t+1)As the 100 th sample, prediction is continued for the next time.
And 5.4) when a plurality of abnormal samples are continuously detected, judging that the current running state is abnormal, and giving an alarm.
In order to realize the method, the invention also discloses a side channel-based power grid embedded terminal safety monitoring system, which comprises the following steps:
the side channel signal acquisition module is used for acquiring a CPU power consumption signal of the embedded equipment;
the preprocessing module is used for filtering direct current signals in the collected CPU power consumption signals and slicing the filtered time domain power consumption information;
the characteristic extraction module is used for extracting characteristics from the sliced power consumption data and screening the characteristics;
the equipment real-time monitoring module comprises a model establishing unit and an abnormality detecting unit; the model establishing unit is used for establishing a long and short memory unit neural network and training the long and short memory unit neural network, the abnormity detection unit is used for loading the trained long and short memory unit neural network model, predicting an input signal to be detected, comparing and judging the input signal with an actually acquired signal at the next moment, and outputting a result that whether the equipment normally operates or not.
The invention has the following beneficial effects: the invention realizes the monitoring of the state of the side channel information of the embedded terminal equipment of the power grid during operation by an LSTM model training and predicting method based on the side channel information of the equipment, thereby judging whether the equipment is in a normal operation state. The power consumption analysis has the advantages that the information quantity is rich, the environmental interference is relatively small, and the analysis effect is better compared with other side channel analysis technologies. A feature screening algorithm is designed to screen the initial features, redundant features are eliminated, and the model calculation amount is greatly reduced. By adopting the intrusion detection method based on the power consumption side channel, the problem that the firmware of the power distribution terminal unit of the intelligent power grid is formulated by a manufacturer when leaving a factory and intrusion detection software cannot be installed is solved.
Drawings
FIG. 1 is a schematic diagram of the general structure of the present invention;
fig. 2 is a schematic structural diagram of a real-time monitoring module of the device.
Detailed Description
In order to make the contents and effects of the present invention more apparent, preferred embodiments of the present invention will be described in detail below.
The invention provides a side-channel-based power grid embedded terminal equipment safety monitoring method which comprises the steps of equipment running state information acquisition, data preprocessing, feature extraction, equipment normal running state model establishment, equipment next running state prediction and equipment running state abnormity detection. The operation state information of the power grid embedded terminal equipment is classified, power consumption is selected as the state information of the equipment during operation, and the operation state of the equipment is monitored in real time and compared with the normal operation state model of the equipment for analysis, so that whether the equipment is abnormal in operation is judged. The device running state model adopts a recurrent neural network based on a long and short memory unit (LSTM), can overcome the defect that the number of attack samples of the embedded terminal device of the smart grid is small, can better predict the device running state under the condition of only positive samples, and then realizes the real-time monitoring of the device running state by comparing whether the state information of the terminal device in running is consistent with the model prediction state. The invention provides a side channel-based power grid embedded terminal equipment safety monitoring method, which is a real-time monitoring method, and the general structural schematic diagram of the invention is shown in figure 1, and the method comprises intelligent power grid embedded terminal equipment, a side channel signal acquisition module, a preprocessing module, a feature extraction module and an equipment real-time monitoring module. Wherein:
the side channel signal acquisition module is mainly used for acquiring signals for monitoring the state of the embedded equipment of the smart grid. The side channel information of the embedded terminal equipment of the smart grid collected in the invention is power consumption information, in particular to the power consumption information of a CPU (central processing unit) of the embedded equipment, because the power consumption information of the CPU reflects the change of an internal operation instruction. In a specific implementation of the invention, a mutual inductance mode is adopted to acquire CPU power consumption signals of embedded equipment, and the acquisition method is a non-embedded method, namely, under the condition of not damaging or accessing original equipment, a mutual inductance coil or a mutual inductance sensor with high precision and high sensitivity is used for acquiring power consumption information.
The preprocessing module is mainly used for filtering direct current signals in the acquired power consumption information and slicing the filtered time domain power consumption information. In one implementation of the present invention, the extracted time-domain power consumption information is sliced in units of 5 seconds.
And the feature extraction module is used for extracting features of the preprocessed information. In one specific implementation of the present invention, the extracted initial features include 133 time-domain features, such as a maximum value, a minimum value, an average value, etc., basic frequency-domain features, such as a spectrum average value, etc., and probability density distribution features capable of better expressing signal characteristics. In order to reduce the computational complexity, the extracted initial features need to be screened. The invention adopts an algorithm for screening the optimal characteristics, and the specific implementation steps are as follows:
step 1: setting a linear classifier: f ═ WTx + b, where f is the estimated classmark number for a sample, x is an m-dimensional feature vector, W is the coefficient corresponding to each dimension of the feature in the classifier, and b is the bias constant;
step 2: selecting a loss function:
where n denotes the number of input samples, yiRepresenting the output of the model;
and step 3: in the feature screening process, the redundant feature coefficients which are not selected are zero, and the optimization target formula is expressed as follows:
wherein | | W | | ceiling0Denotes xiThe weight of (c);
and 4, step 4: solving an optimized target formula by using a least square method, wherein the concrete solving process is as follows: selecting MpA positive sample and MnNegative samples, each sample is an m-dimensional vector, and the real label of each sample is marked as +1(positive samples), -1 (negative samples). Finally, 12 optimal features are selected by solving the above equation.
The method comprises the steps of training a long-time normal operation sample of the equipment to obtain a normal operation state model of the equipment, predicting a characteristic vector of the power consumption information of the equipment at the next time point based on the model, and judging whether the newly acquired power consumption information of the sample is normal or abnormal by comparing the characteristic vector of the acquired sample with the predicted characteristic vector, so that real-time abnormal monitoring of the embedded equipment of the intelligent power grid is realized. Therefore, the real-time monitoring module of the device can be further divided into a model building unit and an abnormality detection unit, as shown in fig. 2.
The concrete implementation steps of the model building unit are as follows:
step 1: and (5) establishing a model. LSTM is a special neural network model, which includes three network structures, an input layer, a hidden layer and an output layer, wherein the hidden layer may have one or more layers. In one embodiment of the present invention, a structure of an n × m cell input layer, an m cell output layer, and two hidden layers is employed. The two hidden layers are all connected, namely each hidden unit in the first hidden layer is connected with each hidden unit in the second hidden layer through feedforward connection.
Step 2: and (5) training data acquisition. In one embodiment of the present invention, the CPU power consumption in the normal operation state is divided into one sample every 5 seconds, and the 12 features, i.e., m is 12, are extracted to obtain the time sequence X ═ { X ═ 12(1),x(2),…,x(n)In which x(t)Is a matrix of dimensions m in which,
a feature vector representing samples of power consumption at time t.
And step 3: and (5) training a model. And (4) taking part of the collected series of samples as training samples and taking part of the collected series of samples as correction samples, and training the LSTM model until the model training is finished.
An abnormality detection unit: after the model building unit trains the model of the equipment, the trained model is applied to the equipment monitoring module, the model predicts the input signal to be detected, compares the input signal with the actually acquired signal at the next moment, and outputs a result that whether the equipment normally operates or not. The specific implementation steps of the abnormality detection are as follows:
step 1: feature value prediction at the next time: setting a sample window, and predicting by using a long and short memory unit neural network model according to the collected power consumption samples at the previous n-100 moments to obtain m-dimensional predicted power consumption characteristics of the power consumption samples at the next moment
Step 2: and (3) acquiring an actual value at the next moment: setting a sample window, acquiring and obtaining an actual power consumption sample at the next moment and obtaining actual power consumption characteristics
Calculating to obtain the actual power consumption characteristic x(t+1)And predicting power consumption characteristic X′(t+1)Error existing between
Obtaining an error vector
And step 3: and (3) error comparison: when the error vector satisfies
If so, regarding the power consumption sample at the (t +1) moment as an abnormal sample, otherwise, regarding the power consumption sample as a normal sample, and the tau is a preset threshold value; updating a sample window, if the current sample is detected to be a normal sample, taking the current sample as a sample at the 100 th moment, and continuously predicting the next moment by combining the first 99 samples; if the current sample is an abnormal sample, the predicted power consumption characteristic X is obtained′(t+1)As the 100 th sample, prediction is continued for the next time.
And 4, step 4: and when a plurality of abnormal samples are continuously detected, judging that the current running state is abnormal, and giving an alarm.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (4)
1. A side channel-based power grid embedded terminal safety monitoring method is characterized by comprising the following steps:
1) acquiring a CPU power consumption signal of the embedded equipment by adopting a mutual inductance mode;
2) filtering direct current signals in the collected CPU power consumption signals, and slicing the filtered time domain power consumption information, wherein each section is used as a sample to obtain a power consumption sample set;
3) extracting initial characteristics of the samples from the power consumption sample set, and designing a characteristic screening algorithm to screen the initial characteristics to obtain an optimal characteristic set; the feature screening algorithm comprises the following steps:
s1: setting a linear classifier: f ═ WTx + b, where f is the estimated classmark number for a sample, x is an m-dimensional feature vector, W is the coefficient corresponding to each dimension of the feature in the classifier, and b is the bias constant;
s2: selecting a loss function:
where n denotes the number of input samples, yiRepresenting the output of the model;
s3: in the feature screening process, the redundant feature coefficients which are not selected are zero, and the optimization target formula is expressed as follows:
wherein | | W | | ceiling0Denotes xiThe weight of (c);
s4: solving an optimized target formula by using a least square method, wherein the concrete solving process is as follows: selecting MpA positive sample and MnEach sample is an m-dimensional vector and a real label of each sample is marked, wherein the label of the positive sample is +1, and the label of the negative sample is-1;
4) constructing a long and short memory unit neural network, which comprises an input layer with n multiplied by m units, an output layer with m units and two hidden layers, wherein m is the number of the characteristics in the optimal characteristic set, and n represents the number of samples; the two hidden layers are all connected, namely each hidden unit in the first hidden layer is connected with each hidden unit in the second hidden layer through feedforward connection; training the long and short memory unit neural networks by adopting the optimal feature set obtained in the step 3) to obtain trained long and short memory unit neural network models;
5) predicting an input signal to be detected by using a trained long and short memory unit neural network model, comparing and judging the input signal with an actually acquired signal at the next moment, and outputting a result that whether the equipment normally operates or not, wherein the method specifically comprises the following steps:
5.1) feature value prediction at the next moment: setting a sample window, and predicting by using a long and short memory unit neural network model according to the collected power consumption samples at the previous n-100 moments to obtain m-dimensional predicted power consumption characteristics of the power consumption samples at the next moment
5.2) collecting the actual value at the next moment: setting a sample window, acquiring and obtaining an actual power consumption sample at the next moment and obtaining actual power consumption characteristics
Calculating to obtain the actual power consumption characteristic x(t+1)And predicting power consumption characteristic X′(t+1)Error existing between
Obtaining an error vector
5.3) error comparison: when the error vector satisfies
If so, regarding the power consumption sample at the (t +1) moment as an abnormal sample, otherwise, regarding the power consumption sample as a normal sample, and the tau is a preset threshold value; updating a sample window, if the current sample is detected to be a normal sample, taking the current sample as a sample at the 100 th moment, and continuously predicting the next moment by combining the first 99 samples; if the current sample is an abnormal sample, the predicted power consumption characteristic X is obtained′(t+1)As the 100 th sample, continuing to predict the next moment;
and 5.4) when a plurality of abnormal samples are continuously detected, judging that the current running state is abnormal, and giving an alarm.
2. The side-channel-based power grid embedded terminal security monitoring method of claim 1, wherein in the signal acquisition process of step 1), a mutual inductance mode is adopted to obtain a CPU power consumption signal of the embedded device.
3. The side-channel-based power grid embedded terminal security monitoring method as claimed in claim 1, wherein the filtered time domain power consumption information is sliced in units of 5 seconds, that is, the filtered time domain power consumption information is taken as a sample in units of 5 seconds.
4. A side channel-based power grid embedded terminal security monitoring system, which is used for implementing the power grid embedded terminal security monitoring method of claim 1, and comprises:
the side channel signal acquisition module is used for acquiring a CPU power consumption signal of the embedded equipment;
the preprocessing module is used for filtering direct current signals in the collected CPU power consumption signals and slicing the filtered time domain power consumption information;
the characteristic extraction module is used for extracting characteristics from the sliced power consumption data and screening the characteristics;
the equipment real-time monitoring module comprises a model establishing unit and an abnormality detecting unit; the model establishing unit is used for establishing a long and short memory unit neural network and training the long and short memory unit neural network, the abnormity detection unit is used for loading the trained long and short memory unit neural network model, predicting an input signal to be detected, comparing and judging the input signal with an actually acquired signal at the next moment, and outputting a result that whether the equipment normally operates or not.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911296619.0A CN111030299A (en) | 2019-12-16 | 2019-12-16 | Side channel-based power grid embedded terminal safety monitoring method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911296619.0A CN111030299A (en) | 2019-12-16 | 2019-12-16 | Side channel-based power grid embedded terminal safety monitoring method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111030299A true CN111030299A (en) | 2020-04-17 |
Family
ID=70210953
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911296619.0A Withdrawn CN111030299A (en) | 2019-12-16 | 2019-12-16 | Side channel-based power grid embedded terminal safety monitoring method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111030299A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112306778A (en) * | 2020-11-20 | 2021-02-02 | 浙江大学 | Resource-limited terminal equipment safety monitoring method based on bypass |
| CN113657628A (en) * | 2021-08-20 | 2021-11-16 | 武汉霖汐科技有限公司 | Industrial equipment monitoring method and system, electronic equipment and storage medium |
| CN116342108A (en) * | 2023-04-06 | 2023-06-27 | 安徽配隆天环保科技有限公司 | Environment-friendly monitoring equipment operation state analysis system and analysis method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105652781A (en) * | 2016-03-12 | 2016-06-08 | 浙江大学 | PLC safety monitoring method based on bypass information |
| CN106888205A (en) * | 2017-01-04 | 2017-06-23 | 浙江大学 | A kind of non-intrusion type is based on the PLC method for detecting abnormality of power consumption analysis |
| CN108393892A (en) * | 2018-03-05 | 2018-08-14 | 厦门大学 | A kind of robot feedforward torque compensation method |
-
2019
- 2019-12-16 CN CN201911296619.0A patent/CN111030299A/en not_active Withdrawn
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105652781A (en) * | 2016-03-12 | 2016-06-08 | 浙江大学 | PLC safety monitoring method based on bypass information |
| CN106888205A (en) * | 2017-01-04 | 2017-06-23 | 浙江大学 | A kind of non-intrusion type is based on the PLC method for detecting abnormality of power consumption analysis |
| CN108393892A (en) * | 2018-03-05 | 2018-08-14 | 厦门大学 | A kind of robot feedforward torque compensation method |
Non-Patent Citations (1)
| Title |
|---|
| 肖玉珺: "非侵入式的基于功耗的PLC异常监測系统", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112306778A (en) * | 2020-11-20 | 2021-02-02 | 浙江大学 | Resource-limited terminal equipment safety monitoring method based on bypass |
| CN112306778B (en) * | 2020-11-20 | 2022-05-10 | 浙江大学 | Resource-limited terminal equipment safety monitoring method based on bypass |
| CN113657628A (en) * | 2021-08-20 | 2021-11-16 | 武汉霖汐科技有限公司 | Industrial equipment monitoring method and system, electronic equipment and storage medium |
| CN116342108A (en) * | 2023-04-06 | 2023-06-27 | 安徽配隆天环保科技有限公司 | Environment-friendly monitoring equipment operation state analysis system and analysis method |
| CN116342108B (en) * | 2023-04-06 | 2023-09-05 | 安徽配隆天环保科技有限公司 | Environment-friendly monitoring equipment operation state analysis system and analysis method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Feng et al. | Multi-level anomaly detection in industrial control systems via package signatures and LSTM networks | |
| CN106888205B (en) | Non-invasive PLC anomaly detection method based on power consumption analysis | |
| CN111030299A (en) | Side channel-based power grid embedded terminal safety monitoring method and system | |
| CN110933031A (en) | Intelligent power grid power distribution terminal unit intrusion detection method based on LSTM | |
| CN110619587B (en) | Method and system for foundation pit monitoring intelligent early warning and data evidence storage | |
| CN109361648B (en) | Method and device for detecting hidden attack of industrial control system | |
| CN115169479A (en) | Remote monitoring method, system and storage medium for sewage treatment process | |
| CN110865625A (en) | Process data anomaly detection method based on time series | |
| CN116823227A (en) | Intelligent equipment management system and method based on Internet of things | |
| CN114124420A (en) | Network flow abnormity detection method based on deep neural network | |
| CN113758652B (en) | Oil leakage detection method and device for converter transformer, computer equipment and storage medium | |
| CN116980958B (en) | Radio equipment electric fault monitoring method and system based on data identification | |
| CN116951328B (en) | Intelligent drainage pipeline operation monitoring system based on big data | |
| CN116614177B (en) | Optical fiber state multidimensional parameter monitoring system | |
| Liu et al. | Deep learning approach for sensor data prediction and sensor fault diagnosis in wind turbine blade | |
| Ribu Hassini et al. | A machine learning and deep neural network approach in industrial control systems | |
| Gu et al. | Research on prediction of remaining useful life of underwater turntable based on slow feature analysis | |
| CN115874993A (en) | Shale gas well production control method, equipment and system based on artificial intelligence | |
| CN115962428A (en) | Real-time online intelligent interpretability monitoring and tracing method for gas pipe network leakage | |
| CN110874506A (en) | Low-temperature equipment fault prediction method | |
| CN116798198A (en) | Sensor abnormality detection and early warning method based on multivariate time sequence prediction model | |
| CN112085043B (en) | Intelligent monitoring method and system for network security of transformer substation | |
| CN114006744A (en) | LSTM-based power monitoring system network security situation prediction method and system | |
| CN116700206B (en) | Industrial control system anomaly detection method and device based on multi-modal neural network | |
| Xu et al. | Design of Monitoring Algorithm for Abnormal Network Traffic Sequence Based on Wavelet Transform |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20200417 |
|
| WW01 | Invention patent application withdrawn after publication |