Digital Twin-Based State Estimation, Deduction, and Evaluation Method for Industrial Equipment Health

Abstract

The present invention discloses a method for estimating, deducing, and evaluating the sound state of industrial equipment based on a digital twin. In the case of a data center infrastructure scenario (for example, an industrial Internet application scenario), although it is considered that the probabilities of failures of equipment in different environments are related to various factors, conventional solutions can only issue an alarm after a failure occurs in the equipment, and it is impossible to estimate, deduce, and evaluate the sound state of equipment that may have a failure. Therefore, they are hardly useful for prevention before a failure and cause analysis after a failure of the equipment. Accordingly, the present invention uses digital twin technology to model the equipment in the scenario, communicate the physical equipment with a digital model, and perform estimation, deduction, and evaluation based on the operating state data of the digital equipment model, thereby achieving the purpose of early warning for equipment in an abnormal operating state, improving the reliability and stability of the equipment, and ensuring the operation and safety of the data center. A solution for estimating, deducing, and evaluating the sound state of industrial equipment is proposed.

Description

[Technical Field]

[0001] The present invention relates to the field of digital twins, and in particular to a method for state estimation, deduction and evaluation in digital twins. [Background technology]

[0002] Digital twin refers to the use of digital technology to build a "digital twin" model that corresponds to an actual physical system in a virtual environment, and complete real-world simulation, analysis, and prediction through data interconnection with the actual physical system, thereby improving operational efficiency and accuracy in the real world.

[0003] In data center infrastructure scenarios (such as industrial Internet application scenarios), equipment failures have a significant impact on the normal operation and business security of the data center. Therefore, the development of digital twin estimation, deduction, and evaluation technologies is crucial to ensure the reliability and stability of data center equipment and improve the operational efficiency and business security of data centers.

[0004] Traditional monitoring methods often only issue an alarm after an equipment failure occurs, preventing early warning and early detection of abnormal changes in the equipment's health status. Digital twin inference, deduction, and evaluation technology connects physical equipment to a digital model, constructs the digital model as a virtual image of the physical equipment, and performs inference, deduction, and evaluation using the digital equipment model's operating status data to achieve early warning and fault analysis for equipment in abnormal operating conditions. At the same time, digital twin technology can also help optimize equipment design and operating parameters, improving equipment performance and efficiency, thereby reducing data center energy consumption and maintenance costs.

[0005] In short, digital twin estimation, deduction and evaluation technology has wide application potential in the early fault warning and management optimization of data center equipment, which can improve the reliability, stability and operating efficiency of data centers and provide an important guarantee for the safe and stable operation of data centers. Summary of the Invention

[0006] To address the above issues, the present invention provides a method for industrial equipment health state estimation, deduction, and evaluation based on digital twins, which applies digital twins to data center scenarios (e.g., industrial Internet application scenarios) to achieve the goal of improving the reliability, stability, and operating efficiency of data centers, thereby providing a foundation for the safe and stable operation of data centers.

[0007] The specific operation steps of the digital twin-based method for estimating, deducing, and evaluating the health state of industrial equipment are as follows:

[0008] Step 1: Combine with digital twin technology to establish a digital virtual model for the physical equipment in the data center scenario, communicate and interact with the physical equipment, and ensure that the real-time operating status information and real-time environmental information of the physical equipment in the data center scenario are synchronized with the digital virtual model.

[0009] Step 2: The past operational state information and environmental information of the physical equipment in the data center scenario are taken as empirical sample data, and a state sequence (S1, S2, ..., S) is generated to describe the different states of the equipment during operation using the statistical histogram method. k ) and the useful life of the equipment corresponding to each state is (d1,d2,...,d k ) and the state transition probability matrix P k×k and decision matrix X 6×4 Build.

[0010] Step 3: Using the equipment operating time length L, the equipment current strength I, and the equipment voltage level U as parameters, and considering environmental factors such as temperature T, humidity H, and the quality A of dust in the air, a condition evaluation formula is established that results in the health status of the equipment.

[0011] Step 4: According to the state evaluation formula, calculate the state evaluation value Q0 of the current equipment, and then find the closest state in the state sequence to find the corresponding equipment's

number

[0012] Step 5: Based on Q0, obtain the transition state sequence according to the state transition probability matrix, and perform a first estimation and deduction to obtain the first estimation and deduction state evaluation matrix Q1, and then calculate the corresponding state probability matrix

number

number

[0013] Step 6: If the expected service life of the equipment is within the safety limit d s Repeat step 5 to repeat the estimation and deduction until the numth estimated and deduced state evaluation matrix Q num , and the corresponding state probability matrix P rnum and Q num For each state rating in the

number

[0014] Step 7: Statistic the expected useful life of the equipment throughout the estimation and deduction process until the estimation and deduction is completed.

number

[0015] Step 8: Actual equipment

number

[0016] Furthermore, the state transition probability matrix P k×k and the decision matrix X 6×4 Specifically, the expression is as follows:

number

[0017] Furthermore, the expression for the state evaluation formula in step 3 is as follows:

number

[0018] Furthermore, the weight of the mth evaluation factor, w m The expression is as follows:

number

[0019] Furthermore, step 4 is specifically as follows:

number

[0020] Furthermore, the transition state sequence for step 5 is

number

[0021] Furthermore, the first estimated / deduced state evaluation matrix Q1 and the corresponding state probability matrix Q2 in step 5 are

number

[0022] For each condition evaluation value in Q1, the closest state in the condition sequence is matched to obtain the expected service life of the corresponding equipment, specifically, as follows:

number

[0023] Furthermore, the estimated and deduced state evaluation matrix Q num and the corresponding state probability matrix P rnum , equipment

number

number

[0024] Furthermore, the inferences and deductions of step 6

number

[0025] Furthermore, in step 7, the actual

number

number

[0026] Beneficial effects Conventional monitoring methods often only issue an alarm after a fault occurs in the equipment, making it difficult to provide early warnings or detect abnormal changes in the health status of the equipment.The digital twin estimation, deduction, and evaluation method proposed in this patent communicates with the physical equipment and builds the digital model as a virtual image of the physical equipment, and performs estimation, deduction, and evaluation using the operating status data of the digital equipment model, thereby realizing early warnings and fault analysis for equipment in abnormal operating states. [Brief explanation of the drawings]

[0027] [Figure 1] 1 is a flowchart of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0028] The following further illustrates the technical solution of the present invention in combination with drawings and specific examples.

[0029] The method for estimating, deducing, and evaluating the health state of industrial equipment based on digital twins is shown in Figure 1, and the specific operation steps are as follows:

[0030] Step S101: Combine with digital twin technology to establish a digital virtual model for physical equipment in a data center scenario (e.g., an industrial Internet application scenario), communicate and interact with the physical equipment, and ensure that the real-time operating status information and real-time environmental information of the physical equipment in the data center scenario are synchronized with the digital virtual model.

[0031] First, the equipment in the data center scenario is modeled, and the physical characteristics and operating conditions of the model are matched with the physical equipment. Next, real-time data of the environment and equipment is collected by sensors, the data is converted into a digital format, and data interaction is performed with the digital virtual model. Next, the mathematical model is fused with the real-time data converted into a digital format to establish a digital twin model.

[0032] Step S102: Using a statistical histogram method on the empirical sample data (past operating state information of the equipment and environmental data), a state sequence (S1, S2, ..., S3) is generated to describe different states of the equipment during operation. k ) and S krefers to the k-th state of the equipment during operation. A piece of equipment undergoes a gradual evolution process from a healthy state to an abnormal state prone to failure. This process is divided into k states, and each of these k states corresponds to a useful life of the equipment. The useful life sequence of the equipment (d1, d2, ..., d k ) and a state transition probability matrix P k×k and decision matrix X 6×4 The decision matrix takes into account three internal influence factors of the equipment (the equipment's operating time length L, the equipment's current strength I, the equipment's voltage level U) and three external environmental influence factors (temperature T, humidity H, the quality of dust in the air A), and classifies each influence factor into four levels: none, low, medium, and high, depending on the degree of impact.

[0033]

number

[0034] Step S103: By monitoring the internal influence factors of the digital virtual model equipment in real time, the equipment operating time length L, the equipment current intensity I, the equipment voltage level U are used as parameters, and the environmental influence factors of temperature T, humidity H, and dust quality A in the air are considered, and a state evaluation formula is established to obtain the current equipment state evaluation value Q0 according to this formula, and the closest state in the state sequence is matched to the corresponding equipment state.

number

[0035]

number

[0036] Weight of the mth evaluation factor w m is the decision matrix X 6×4 is obtained by the entropy weight method.

number

[0037] Based on the information entropy of the internal state influence factors of each device, the corresponding weight w m Calculate.

number

[0038] Based on the obtained Q0, find the closest state in the complete state sequence and calculate the device's

number

number

[0039] Step S104: Based on Q0, the transition state sequence of the device is determined according to the state transition probability matrix.

number

[0040]

number

[0041]

number

[0042]

number

[0043] Step S105, if the expected service life of the equipment is within the safety value d s Step S104 is repeated to perform estimation and deduction until the number of devices estimated and deduced for the numth time is smaller than 1×k according to the following tree branch structure estimation and deduction formula: num The state evaluation matrix Q num and the corresponding 1×k num The state probability matrix P rnum and Q num For each state estimate in the

number

[0044]

number

[0045] Each time estimation and deduction is performed, the expected useful life of a piece of equipment is obtained, and when probability factors are taken into account, the expected useful life of this equipment becomes smaller the more times estimation and deduction are performed, so it will always be smaller than the safety value.At this point, the final number of estimations and deductions is obtained, where num is any number of estimations and deductions in the process until the estimation and deduction is completed, and Num is the number of estimations and deductions when it is completed.

[0046] Step 106: Calculate the total expected useful life of the equipment in the estimation and deduction process until the estimation and deduction is completed, and calculate the estimated and deduced

number

number

[0047] Step S107, the actual

number

number

[0048] The above description of the embodiments is intended to facilitate understanding of the method and core idea of ​​the present application. It should be noted that those skilled in the art may make some improvements and modifications to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the scope of protection of the claims of the present application.

Claims

1. A method for estimating, deducing, and evaluating the health state of industrial equipment based on a digital twin, comprising: The specific steps of this method are as follows: Step 1: Combine with digital twin technology to establish a digital virtual model for the physical equipment in the data center scenario, and communicate with the physical equipment to ensure that the real-time operating status information and real-time environment information of the physical equipment in the data center scenario are synchronized with the digital virtual model; Step 2: The past operating state information and environment information of the physical equipment in the data center scenario are taken as empirical sample data, and the state sequence (S １ , S ２ , ..., S ｋ ) and S ｋ denotes the k-th state of the equipment in operation, and the corresponding equipment life sequence is (d １ , d ２ , ..., d ｋ ) and the state transition probability matrix P ｋ×ｋ and decision matrix X ６×４ Build Step 3: Establish a state evaluation formula that uses the operating time length L of the equipment, the current intensity I of the equipment, and the voltage level U of the equipment as parameters, and considers environmental factors such as temperature T, humidity H, and the quality A of dust in the air, and determines the health state of the equipment as a result; Step 4: According to the state evaluation formula, the current state evaluation value Q of the equipment ０ and match the closest state in the state sequence to the corresponding device's [Equation 36] and Step 5, Q ０ Based on this, the transition state sequence is obtained according to the state transition probability matrix, and is estimated and deduced once, and the first estimated and deduced state evaluation matrix Q １ and obtain the corresponding first state probability matrix [Equation 37] and Q １ For each state rating in the [Equation 38] and Step 6: The expected service life of the equipment is the safety value d ｓ Repeat step 5 to repeat the estimation and deduction until the numth estimated and deduced state evaluation matrix Q ｎｕｍ , and the corresponding state probability matrix P ｒｎｕｍ and Q ｎｕｍ For each state rating in the [0.39] and Step 7: Statistic the expected useful life of the equipment throughout the entire estimation / deduction process until the estimation / deduction is completed, and calculate the estimated / deduction [Equation 40] and Step 8: Actual equipment [Equation 41] Based on this, the condition assessment matrix is ​​fed back and revised. The state transition probability matrix P k×k and the decision matrix X 6×4 in step 2 are expressed as follows: [0.001] Xmn indicates that the mth evaluation factor is at the nth level, and the evaluation factors are L, I, U, T, H, and A, respectively, and the evaluation levels include four levels: [none, low, medium, high], where 1≦m≦6 and 1≦n≦4. The weight w m of the mth evaluation factor is expressed as follows: [Equation 43] Here, m = 1, 2, ..., 6, n = 1, 2, ..., 4, X mn indicates that the m-th evaluation factor is at the n-th level, the evaluation factors are L, I, U, T, H, and A, respectively, and the evaluation levels include four levels: [none, low, medium, high]. This is a method for estimating, deducing, and evaluating the health state of industrial equipment based on a digital twin.

2. The state evaluation formula for step 3 is as follows: [0.0000] ここで、 (A １ , A ２ , A ３ , A ４ , A ５ , A ６ )=(L,I,U,T,H,A) D is the distance between the equipment and the safety line, and D ｈ and D ｌ are the upper and lower safety limits, and μ ｍ is the expected safety value of the mth evaluation factor, and K １ is the amplification factor, and w ｍ is the weight of the m-th evaluation factor, and the evaluation factors are L, I, U, T, H, and A, respectively.

3. Step 4 is specifically as follows: [Equation 45] where r0 is the state sequence matched by Q ０ is the state number closest to S ｖ The method for health state estimation, deduction, and evaluation of industrial equipment based on a digital twin according to claim 1 , wherein represents a value in the constructed state sequence.

4. The 0th state probability matrix in step 5 is [Equation 46] State transition probability matrix P ｋ×ｋ is the r0th row of Q, where r0 is the matched state sequence. ０ The method for estimating, deducing, and evaluating a health state of industrial equipment based on a digital twin according to claim 1, wherein the state number is the closest to the state number.

5. Step 5: First estimated / deduced state evaluation matrix Q １ and the corresponding first state probability matrix [Equation 47] where τ(1) represents the first random value of the oscillatory function τ, and τ is a Gaussian distribution τ ∼ N(μ, σ ２ ) and μ, σ ２ are the average values ​​of this oscillatory function, respectively. The mean and expected variance are Q １ For each condition rating, the closest state in the state sequence is matched to obtain the corresponding expected service life of the equipment, specifically: [Number 48] The method for estimating, deducing, and evaluating the health state of industrial equipment based on a digital twin according to claim 4, characterized in that it is a transposition of

6. The estimated / deduced state evaluation matrix Q of the numth step in step 6 ｎｕｍ and the corresponding state probability matrix P ｒｎｕｍ , equipment [Number 49] Specifically, it is as follows: [Number 50] The method for estimating, deducing, and evaluating the health state of industrial equipment based on a digital twin as described in claim 1, characterized in that the threshold is a threshold value for the useful life of the equipment.

7. Step 7: Estimated / Deduced [0.51] The method for estimating, deducing, and evaluating the health state of industrial equipment based on a digital twin as described in claim 1, wherein NUM is the total number of estimations and deductions performed until the estimations and deductions are completed.

8. Step 8: Actual equipment [Number 52] Based on this, the equipment life threshold d is calculated according to the following formula: ｔｈ to provide feedback and correction to the condition assessment matrix: [Number 53] The method for estimating, deducing, and evaluating the health state of industrial equipment based on a digital twin as described in claim 7, characterized in that the threshold is a threshold for the useful life of the equipment after updating.

Images