CN111855219A - A method for predicting security parameters of diesel engine lubricating oil entering the engine based on grey theory - Google Patents

Abstract

The invention is a method for predicting the security parameters of diesel engine lubricating oil entering the engine based on grey theory. The invention belongs to the technical field of diesel engine safety prediction. First, the operation data of the lubricating oil inlet pressure and temperature of the diesel engine are collected, and Kalman filter processing is performed on the operation data; and then the prediction of the diesel engine operation data is realized according to the modeling steps of the gray prediction model. function, and obtain the prediction results; use the relative error test method to test the gray prediction model. If the four-level test index is not met, use the residual error method to optimize the gray model until it meets the four-level and above-level indicators, and output accurate prediction data. ; On the basis of the relative error verification of the predicted data, the predicted result and the diesel engine alarm threshold are logically discriminated, and the predicted state result is output to the upper computer. The invention can effectively predict the changing trend of the diesel engine lubricating oil inlet parameters and reasonably predict the future working state of the diesel engine lubricating oil system.

Description

A method for predicting security parameters of diesel engine lubricating oil entering the engine based on grey theory

technical field

The invention relates to the technical field of diesel engine safety prediction, and relates to a method for predicting safety parameters of diesel engine lubricating oil entering the engine based on grey theory.

Background technique

With the rapid development of the level of industrialization in today's society, diesel engines, as the most commonly used power machinery equipment, are widely used in oil mines, stationary power generation, railway traction, construction machinery and special ships, etc., and diesel engines are increasingly large and high-speed. , The direction of precision development, the work performance is continuously improved, and the degree of automation is getting higher and higher. On the one hand, the use of advanced diesel engines will greatly improve labor productivity, improve product quality, and reduce production costs and energy consumption; but on the other hand, the problem of safe operation of diesel engines is becoming more and more serious. With the development of high load and high speed, the danger of diesel engine safety accidents is Further improvement, once a part or a link of the continuous overrun operation and no treatment is taken, it will often cause equipment damage, cause major safety accidents, and even endanger personal safety. Carrying out research on diesel engine safety protection technology can not only help us prevent safety accidents and ensure the personal safety of staff, but also avoid potential huge economic and social losses.

The task of the traditional diesel engine safety protection system is to monitor the important state parameters of the diesel engine, such as the diesel engine lubricating oil inlet pressure and temperature, to determine whether it exceeds the set threshold, and then to determine whether the diesel engine is safe to operate, and to take emergency shutdown measures before danger occurs. Ensure the safety of personnel and equipment. Although this method can avoid accidents to a certain extent, its drawbacks are also very obvious. When a diesel engine running normally under rated load suddenly stops operation (especially sudden unloading from full load), it will cause stress to moving parts In the process of the lubricating oil pressure gradually deviating from the normal range, the lubricating effect of the moving parts such as the body, piston and crankshaft is greatly reduced, causing huge irreversible damage to the equipment and seriously affecting the working life of the equipment.

Through the literature search of the prior art, it is found that the published document "Speed Measurement and Security System for Marine Diesel Engines" proposes a security system based on the rotational speed of the diesel engine. System. The user sets the alarm threshold according to the actual usage through the industrial touch screen. The alarm threshold is preset for each detection signal, and the alarm threshold of each detection signal can be set as analog quantity or switch value. The user sets the alarm threshold of each detection signal through the industrial touch screen as a switch value or an analog value. The main control module judges the detection signal through the alarm threshold and the detection signal exceeds the alarm threshold, then The main control module sends a stop signal to the diesel engine, and alarms through the remote control box that the detection signal does not exceed the alarm threshold, then no action is taken. The disadvantage is that the strategy adopted by the security system is: Setting a single threshold for real-time data can only be monitored and shut down after the equipment runs out of limit, which will still cause huge damage to the diesel engine.

SUMMARY OF THE INVENTION

In order to realize the method for predicting the lubricating oil inlet parameters of the diesel engine safety protection, the invention realizes the monitoring of the diesel engine running state by collecting the running data of the lubricating oil inlet parameters of the diesel engine in real time, and then processes the data and predicts the change trend. And before the dangerous situation occurs, an alarm signal is sent to the upper computer, and feedback measures can be taken in advance to avoid the situation of equipment overrunning and improve the reliability and safety of the equipment. The invention provides a diesel engine lubricating oil intake based on gray theory. A method for predicting aircraft security parameters, the present invention provides the following technical solutions:

A method for predicting security parameters of diesel engine lubricating oil entering the engine based on grey theory, comprising the following steps:

Step 1: Collect the pressure data of the diesel engine through the pressure sensor, the pressure data is the running data of the lubricating oil inlet pressure of the diesel engine, obtained from the pressure sensor of the diesel engine lubricating oil inlet pipeline through the data acquisition card, and store the data in Excel middle;

Step 2: Data preprocessing is performed on the collected operating data of the lubricating oil inlet pressure of the diesel engine to filter out the influence of environmental noise;

Step 3: According to the data after noise filtering, the gray prediction data of diesel engine lubricating oil inlet pressure is processed to obtain the prediction result;

Step 4: Compare the predicted result with the alarm threshold of the diesel engine safety protection system, judge the running state of the diesel engine, and issue an alarm according to the running state of the diesel engine.

Preferably, the step 2 is specifically:

Step 2.1: Initialize the diesel engine, determine the input signal w _k , the observation noise v _k , Q _k of the output signal, and the product specified error R _k , and use the Kalman filter method to perform data prediction on the collected operating data of the lubricating oil inlet pressure of the diesel engine. Process, filter out the influence of environmental noise, establish the state equation and output equation, and express the state equation x _k+1 and the output equation y _k by the following formula:

x _k+1 =A _k x _k +w _k

y _k =C _k x _k +v _k

where k is the time, w _k is the input signal, v _k is the observation noise of the output signal, A is the gain matrix between the state variables, x _k is the state variable at time k, and C is the difference between the state variable and the output signal gain matrix;

Determine the input signal w _k , the observed noise v _k , Q _k of the output signal and the product specification error R _k ;

Step 2.2: Set the initial true temperature as x ₍₀₎ =T ₀ and the covariance P ₍₀₎ at the initial moment;

Step 2.3: Read the optimal estimated value T _k-1 at time k-1, the measured value t _{k at time k} , calculate the gain factor, and express the gain factor by the following formula:

H _k =P _(k-1) /(P _(k-1) +R _k )

T _k =T _k-1 +H _k (t _k -T _k-1 )

The optimal-to-noise covariance at time k is expressed as:

P _(k) = (1-H _k )P _(k-1)

Step 2.4: Read the optimal estimated value T _k at the kth time, the measured value t _k+ 1 at the k+1 time, calculate the gain factors H _k+1 and T _k+1 , and the optimal noise at the k+1 time Covariance P _(k+1) .

Step 2.5: Repeat steps 2.3 to 2.4 to estimate the optimal value to obtain valid data after filtering out environmental noise.

Preferably, when the actual temperature of the diesel engine is constant, Q _k =0; when the temperature of the diesel engine varies with the operating state, Q _k =0.01.

Preferably, the step 3 is specifically:

Step 3.1: Establish a gray prediction model GM(2,1), and express the differential equation of the gray prediction model GM(2,1) by the following formula:

其中，p⁽¹⁾为经过一次累加运算生成的累加数列；t为时间；a₁，a₂，u为待估参数，分别为发展灰系数和内生控制灰系数；

Among them, p ⁽¹⁾ is the accumulated sequence generated by one accumulation operation; t is the time; a ₁ , a ₂ , u are the parameters to be estimated, which are the development gray coefficient and the endogenous control gray coefficient respectively;

Step 3.2: According to the data after filtering out the noise, establish the original pressure data series, and express the series P ⁽⁰⁾ by the following formula:

P ⁽⁰⁾ = (p ⁽⁰⁾ (1), p ⁽⁰⁾ (2), p ⁽⁰⁾ (3)…p ⁽⁰⁾ (n))

Among them, p ⁽⁰⁾ (i) is the time-series data of lubricating oil inlet pressure parameters, i=1, 2, ... n;

Step 3.3: According to P ⁽⁰⁾ , determine the size of the grade ratio σ ⁽⁰⁾ (i) of P ⁽⁰⁾ , carry out the grade ratio test, and express the grade ratio σ ⁽⁰⁾ (i) by the following formula:

Step 3.4: When the level ratio σ ⁽⁰⁾ (i) meets the inspection criteria, perform 1-AGO accumulation and 1-IAGO accumulation on P ⁽⁰⁾ to obtain P ⁽¹⁾ and α ⁽¹⁾ P ⁽⁰⁾ The sequence of numbers, represented by the following equations P ⁽¹⁾ and α ⁽¹⁾ P ⁽⁰⁾ sequences:

P ⁽¹⁾ = (p ⁽¹⁾ (1), p ⁽¹⁾ (2), …p ⁽¹⁾ (n))

α ⁽¹⁾ P ⁽⁰⁾ =(α ⁽¹⁾ p ⁽⁰⁾ (2),α ⁽¹⁾ p ⁽⁰⁾ (3)…α ⁽¹⁾ p ⁽⁰⁾ (n))

α ⁽¹⁾ p ⁽⁰⁾ (i)=p ⁽⁰⁾ (i)-p ⁽⁰⁾ (i-1), i=2,3,...,n

Step 3.5: Check whether P ⁽¹⁾ has a quasi-exponential law, and calculate the quasi-exponent δ ⁽¹⁾ (i) by the following formula:

Steps 3, 6: When the quasi-exponent meets the test standard, perform an adjacent mean generation operation on P ⁽¹⁾ , and express the adjacent mean by the following formula:

z ⁽¹⁾ (i)=0.5p ⁽¹⁾ (i)+0.5p ⁽¹⁾ (i-1)(i=2, 3...n)

Z ⁽¹⁾ = (z ⁽¹⁾ (2), z ⁽¹⁾ (3), z ⁽¹⁾ (4), z ⁽¹⁾ (5))

进行最小二乘估计，得到a₁，a₂，u：Step 3.7: Parameter column

Perform least squares estimation to get a ₁ , a ₂ , u:

Step 3.8: Solve the whitening equation for GM(2,1) to obtain the time response formula for GM(2,1), which is expressed by the following formula:

Solving for the simulated value of P ⁽¹⁾ , the simulated value of P ⁽¹⁾ is expressed by:

The analog value of P ⁽⁰⁾ is determined according to the restoration of the analog value, and the analog value of P ⁽⁰⁾ is expressed by the following formula:

Use the simulated value of P ⁽⁰⁾ as the prediction result

Step 3.9: Perform error test on the model, and express the error detection value ε by the following formula:

The inspection indicators are: Level 1: the critical value of the indicator is 0.01; Level 2: the critical value of the indicator is 0.05; Level 3: the critical value of the indicator is 0.10; Level 4: the critical value of the indicator is 0.20; value, the residual optimization grey prediction model GM(2,1) is used.

Preferably, the judgment level ratio σ ⁽⁰⁾ (i) satisfies the inspection standard specifically:

When the grade ratio is within the range of the above-mentioned test criteria, it is satisfied.

Preferably, checking whether the quasi-exponential has a quasi-exponential law is specifically:

δ ⁽¹⁾ (i)∈(1, 1.5)

When the quasi-index is within the range of the above-mentioned test criteria, it is satisfied.

Preferably, after the prediction result is compared with the alarm threshold of the diesel engine safety protection system, according to the predicted lubricating oil inlet pressure will be lower than the pressure threshold of the security system after 80 minutes, then it will be determined that the future operating state of the diesel engine is the lubricating oil inlet pressure If it is too low, at this time, the security system will send an alarm signal to the upper computer through the program, and send the predicted data to the upper computer in the form of graphs for further analysis.

The present invention has the following beneficial effects:

The method first collects the pressure and temperature data of the lubricating oil entering the diesel engine; the collected data is processed by the Kalman filter technology to form the original parameter sequence of the lubricating oil entering pressure and temperature; according to the modeling steps of the gray theory prediction model, the The original pressure sequence is accumulated to generate the accumulated sequence, and then the data quasi-smoothness test is carried out for the accumulated sequence. Under the condition of satisfying the quasi-smoothness, a gray prediction GM(2,1) model is established for prediction calculation, and the relative error test method is used. Carry out model inspection, and it is considered qualified if it reaches the fourth-level standard and above. If it is qualified, the prediction result will be output. If it is not qualified, the residual sequence can be used to build a model, and the original model can be revised until it meets the inspection standard. On the basis of the model meeting the inspection standard, output The prediction result can realize the trend prediction of the pressure and temperature of the diesel engine lubricating oil; and then compare the prediction result with the safety alarm threshold of the diesel engine to realize the status judgment of whether the lubricating oil inlet pressure and temperature are normal; when the prediction result is abnormal, that is If the predicted result of the lubricating oil inlet pressure is lower than the specified threshold or the lubricating oil inlet temperature is higher than the specified threshold, an alarm signal will be output to the upper computer, so that feedback measures can be taken in advance.

The method of the invention directly obtains the operation data of the lubricating oil inlet parameters from the diesel engine, and simultaneously uses the grey theory prediction method to predict the trend of the lubricating oil inlet parameters, and judges the future operation state of the diesel engine through the prediction results. If it is judged that the diesel engine is about to be in an abnormal operation state, it can remind the upper computer and take feedback measures in advance to avoid overrun and sudden shutdown, so as to achieve the purpose of diesel engine safety protection.

According to the prediction method of the present invention, the change of the lubricating oil inlet pressure in the future can be effectively predicted, safety protection measures can be taken for the diesel engine as soon as possible, the reliability of the equipment can be improved, and the sudden shutdown of the diesel engine can be minimized under the condition of ensuring the safe operation of the diesel engine. the potential for harm.

Description of drawings

Fig. 1 is a flowchart of a method for predicting the security parameters of diesel engine lubricating oil entering the engine based on grey theory;

Fig. 2 is a graph showing the prediction result of the lubricating oil inlet pressure during the operation of the diesel engine.

Detailed ways

The present invention is described in detail below with reference to specific embodiments.

Specific embodiment one:

As shown in FIG. 1, the present invention provides a method for predicting the security parameters of diesel engine lubricating oil entering the engine based on the grey theory, specifically:

A method for predicting security parameters of diesel engine lubricating oil entering the engine based on grey theory, comprising the following steps:

Step 1: Collect the pressure data of the diesel engine through the pressure sensor, the pressure data is the running data of the lubricating oil inlet pressure of the diesel engine, obtained from the pressure sensor of the diesel engine lubricating oil inlet pipeline through the data acquisition card, and store the data in Excel middle;

Step 2: Data preprocessing is performed on the collected operating data of the lubricating oil inlet pressure of the diesel engine to filter out the influence of environmental noise;

The step 2 is specifically:

Step 2.1: Initialize the diesel engine, determine the input signal w _k , the observation noise v _k , Q _k of the output signal, and the product specified error R _k , and use the Kalman filter method to perform data prediction on the collected operating data of the lubricating oil inlet pressure of the diesel engine. Process, filter out the influence of environmental noise, establish the state equation and output equation, and express the state equation x _k+1 and the output equation y _k by the following formula:

x _k+1 =A _k x _k +w _k

y _k =C _k x _k +v _k

where k is the time, w _k is the input signal, v _k is the observation noise of the output signal, A is the gain matrix between the state variables, x _k is the state variable at time k, and C is the difference between the state variable and the output signal gain matrix;

Determine the input signal w _k , the observed noise v _k , Q _k of the output signal and the product specification error R _k ;

Step 2.2: Set the initial true temperature as x ₍₀₎ =T ₀ and the covariance P ₍₀₎ at the initial moment;

Step 2.3: Read the optimal estimated value T _k-1 at time k-1, the measured value t _{k at time k} , calculate the gain factor, and express the gain factor by the following formula:

H _k =P _(k-1) /(P _(k-1) +R _k )

T _k =T _k-1 +H _k (t _k -T _k-1 )

The optimal-to-noise covariance at time k is expressed as:

P _(k) = (1-H _k )P _(k-1)

Step 2.4: Read the optimal estimated value T _k at the kth time, the measured value t _k+ 1 at the k+1 time, calculate the gain factors H _k+1 and T _k+1 , and the optimal noise at the k+1 time Covariance P _(k+1) .

Step 2.5: Repeat steps 2.3 to 2.4 to estimate the optimal value to obtain valid data after filtering out environmental noise.

Preferably, when the actual temperature of the diesel engine is constant, Q _k =0; when the temperature of the diesel engine varies with the operating state, Q _k =0.01.

Step 3: According to the data after noise filtering, the gray prediction data of diesel engine lubricating oil inlet pressure is processed to obtain the prediction result;

The step 3 is specifically:

Step 3.1: Establish a gray prediction model GM(2,1), and express the differential equation of the gray prediction model GM(2,1) by the following formula:

其中，p⁽¹⁾为经过一次累加运算生成的累加数列；t为时间；a₁，a₂，u为待估参数，分别为发展灰系数和内生控制灰系数；

Among them, p ⁽¹⁾ is the accumulated sequence generated by one accumulation operation; t is the time; a ₁ , a ₂ , u are the parameters to be estimated, which are the development gray coefficient and the endogenous control gray coefficient respectively;

Step 3.2: According to the data after filtering out the noise, establish the original pressure data series, and express the series P ⁽⁰⁾ by the following formula:

P ⁽⁰⁾ = (p ⁽⁰⁾ (1), p ⁽⁰⁾ (2), p ⁽⁰⁾ (3)…p ⁽⁰⁾ (n))

Among them, p ⁽⁰⁾ (i) is the time-series data of lubricating oil inlet pressure parameters, i=1, 2, ... n;

Step 3.3: According to P ⁽⁰⁾ , determine the size of the grade ratio σ ⁽⁰⁾ (i) of P ⁽⁰⁾ , carry out the grade ratio test, and express the grade ratio σ ⁽⁰⁾ (i) by the following formula:

Step 3.4: When the level ratio σ ⁽⁰⁾ (i) meets the inspection criteria, perform 1-AGO accumulation and 1-IAGO accumulation on P ⁽⁰⁾ to obtain P ⁽¹⁾ and α ⁽¹⁾ P ⁽⁰⁾ The sequence of numbers, represented by the following equations P ⁽¹⁾ and α ⁽¹⁾ P ⁽⁰⁾ sequences:

P ⁽¹⁾ = (p ⁽¹⁾ (1), p ⁽¹⁾ (2), …p ⁽¹⁾ (n))

α ⁽¹⁾ P ⁽⁰⁾ =(α ⁽¹⁾ p ⁽⁰⁾ (2),α ⁽¹⁾ p ⁽⁰⁾ (3)…α ⁽¹⁾ p ⁽⁰⁾ (n))

α ⁽¹⁾ p ⁽⁰⁾ (i)=p ⁽⁰⁾ (i)-p ⁽⁰⁾ (i-1), i=2,3,...,n

Judgment grade ratio σ ⁽⁰⁾ (i) meets the inspection standard. Specifically:

When the grade ratio is within the range of the above-mentioned test criteria, it is satisfied.

Step 3.5: Check whether P ⁽¹⁾ has a quasi-exponential law, and calculate the quasi-exponent δ ⁽¹⁾ (i) by the following formula:

To test whether the quasi-exponential has a quasi-exponential law is as follows:

δ ⁽¹⁾ (i)∈(1, 1.5)

When the quasi-index is within the range of the above-mentioned test criteria, it is satisfied.

Step 3.6: When the quasi-index satisfies the test standard, perform an adjacent mean generation operation on P ⁽¹⁾ , and express the adjacent mean by the following formula:

z ⁽¹⁾ (i)=0.5p ⁽¹⁾ (i)+0.5p ⁽¹⁾ (i-1)(i=2, 3...n)

Z ⁽¹⁾ = (z ⁽¹⁾ (2), z ⁽¹⁾ (3), z ⁽¹⁾ (4), z ⁽¹⁾ (5))

进行最小二乘估计，得到a₁，a₂，u：Step 3.7: Parameter column

Perform least squares estimation to get a ₁ , a ₂ , u:

Step 3.8: Solve the whitening equation for GM(2,1) to obtain the time response formula for GM(2,1), which is expressed by the following formula:

Solving for the simulated value of P ⁽¹⁾ , the simulated value of P ⁽¹⁾ is expressed by:

The analog value of P ⁽⁰⁾ is determined according to the restoration of the analog value, and the analog value of P ⁽⁰⁾ is expressed by the following formula:

Use the simulated value of P ⁽⁰⁾ as the prediction result

Step 3.9: Perform error test on the model, and express the error detection value ε by the following formula:

The inspection indicators are: Level 1: the critical value of the indicator is 0.01; Level 2: the critical value of the indicator is 0.05; Level 3: the critical value of the indicator is 0.10; Level 4: the critical value of the indicator is 0.20; value, the residual optimization grey prediction model GM(2,1) is used.

Step 4: Compare the predicted result with the alarm threshold of the diesel engine safety protection system, judge the running state of the diesel engine, and issue an alarm according to the running state of the diesel engine.

After comparing the prediction result with the alarm threshold of the diesel engine safety protection system, according to the predicted lubricating oil inlet pressure will be lower than the pressure threshold of the security system after 80 minutes, then it will be determined that the diesel engine's future operating state is that the lubricating oil inlet pressure is too low. At this time, the security system will send an alarm signal to the upper computer through the program, and send the predicted data to the upper computer in the form of graphs for further analysis.

The method firstly collects the operation data of diesel engine lubricating oil inlet pressure and temperature, and performs Kalman filter processing on the operation data; then, according to the modeling steps of the gray prediction model, the prediction function of the diesel engine operation data is realized, and the prediction result is obtained; Use the relative error test method to test the gray prediction model. If it does not meet the four-level test indicators, use the residual method to optimize the gray model until it meets the four-level and above indicators, and output accurate prediction data; On the basis of being qualified, the prediction result and the diesel engine alarm threshold are logically discriminated, and then the prediction state result is output to the upper computer. Through the above steps, the changing trend of the diesel engine lubricating oil inlet parameters can be effectively predicted, and the future working state of the diesel engine lubricating oil system can be reasonably predicted. Based on the grey theory, the invention predicts the change of the security parameters of the diesel engine lubricating oil entering the engine, and can assist the control system to take effective feedback measures before the equipment runs out of limit, thereby improving the safety and reliability of the equipment as a whole.

Specific embodiment two:

The invention includes the following steps: data collection of diesel engine lubricating oil inlet pressure, data preprocessing of diesel engine lubricating oil inlet pressure, calculation of gray prediction data of diesel engine lubricating oil inlet pressure, and judgment of diesel engine future working state. details as follows:

Step 1: Data collection of diesel engine lubricating oil inlet pressure: The operating data of the diesel engine lubricating oil inlet pressure is obtained by the data acquisition card from the pressure sensor of the diesel engine lubricating oil inlet pipeline, and the operating data is stored in Excel , so that it can be read and used by subsequent data preprocessing steps and gray prediction model building steps.

Step 2: Data preprocessing of diesel engine lubricating oil inlet pressure: Preprocessing the above collected diesel engine operation data can effectively filter out the noise caused by the sensor itself or the surrounding environment, thereby generating effective data with low interference signals. The method selected here is the Kalman filter method, which can be designed by setting the initial state and variance, and the parameter estimation at the next moment; the estimation correction process calculates the gain according to the statistical law of the external disturbance of the diesel engine and the noise of the sensor; finally On the basis of the predicted state, the predicted parameter value is modified by the gain. Get the best parameter estimates. The specific implementation steps are as follows:

First, the relevant definition of Kalman filter is given:

Assuming that the state variable of the system at time k is x _k , the state equation and output equation are expressed as follows:

X _k+1 =A _k x _k +w _k #(1)

y _k =C _k x _k +v _k #(2)

Among them, k represents time, that is, the k-th iteration; the input signal w _k is white noise, and the observation noise v _k of the output signal is also white noise; A represents the gain matrix between state variables, which changes with time k; C represents the state variable The gain matrix between the output signal and the output signal varies with time k; assuming that both w _k and v _k are normal white noise with zero mean and variances Q and R, respectively.

Step 1: Determine w _k , v _k , Q _k and R _k from the initial conditions of the system. When the real temperature of the system under test is constant, Q _k =0. Here, since the temperature of the diesel engine will change with the operating state, take Q _k =0.01. v _k is the measurement noise of the temperature measuring equipment, the factory manual will mark the sensor error, and R _k can be the specified error of the product.

Step 2: Establish a Kalman filter model.

x _k+1 =A _k x _k +w _k

y _k =C _k x _k +v _k

In the formula, x _k is a one-dimensional temperature variable; A _k =1; C _k =1; the variances of w _k and v _k are Q _k and R _k .

Step 3: Set the initial real temperature value x ₍₀₎ = T ₀ and the covariance P ₍₀₎ at the initial moment

Step 4: Read the optimal estimated value T _k-1 at time k-1, the measured value t _{k at time k} , and calculate the gain factor

H _k =P _(k-1) /(P _(k-1) +R _k )#(3)

but

T _k =T _k-1 +H _k (t _k -T _k-1 )#(4)

Optimal value noise covariance at time k

P _(k) = (1-H _k )P _(k-1) #(5)

Step 5: Read the optimal estimated value T _k at the kth time, the measured value t _k+ 1 at the k+1 time, calculate the gain factors H _k+1 and T _k+1 , and the optimal value at the k+1 time Noise covariance P _(k+1) .

Step 6: Repeat steps 4 and 5 to estimate the optimal temperature value.

Step 3: Calculation of gray prediction data of diesel engine lubricating oil inlet pressure: The gray modeling trend prediction step of the diesel engine is the core of the present invention, and the effective data obtained after the preprocessing step is predicted and calculated according to the construction step of the gray prediction model and get the prediction result. The development environment in this step uses Matlab software.

First, the relevant definitions of the grey prediction model GM(2,1) are given:

The gray prediction model GM(2,1) reflects the second-order differential function of a variable against time, and its corresponding differential equation can be expressed as

Among them, p ⁽¹⁾ is the cumulative sequence generated by one cumulative operation; t is the time; a ₁ , a ₂ , and u are the parameters to be estimated, which are the development gray coefficient and the endogenous control gray coefficient, respectively.

The establishment and calculation steps of grey GM(2,1) model

Step 1: According to the valid data obtained after preprocessing, establish the original pressure data series.

Let P ⁽⁰⁾ = (p ⁽⁰⁾ (1), p ⁽⁰⁾ (2), p ⁽⁰⁾ (3)…p ⁽⁰⁾ (n))

where p ⁽⁰⁾ (i) (i=1, 2, ... n) is the time series data of lubricating oil inlet pressure parameters

Step 2: For a given sequence P ⁽⁰⁾ , calculate the magnitude of the rank ratio σ ⁽⁰⁾ (i) of P ⁽⁰⁾ , and perform a rank ratio test. The formula for calculating the grade ratio is

The inspection standard is

If the inspection criteria are met, the subsequent modeling steps of GM(2,1) can be performed.

The third step: perform 1-AGO accumulation operation and 1-IAGO accumulation and subtraction on P ⁽⁰⁾ to obtain P ⁽¹⁾ and α ⁽¹⁾ P ⁽⁰⁾ sequence

P ⁽¹⁾ = (p ⁽¹⁾ (1), p ⁽¹⁾ (2), …p ⁽¹⁾ (n))

α ⁽¹⁾ P ⁽⁰⁾ =(α ⁽¹⁾ p ⁽⁰⁾ (2),α ⁽¹⁾ p ⁽⁰⁾ (3)…α ⁽¹⁾ p ⁽⁰⁾ (n))

in

α ⁽¹⁾ p ⁽⁰⁾ (i)=p ⁽⁰⁾ (i)-p ⁽⁰⁾ (i-1), i=2,3,...,n#(10)

Step 4: Check whether P ⁽¹⁾ has a quasi-exponential law. The formula for calculating the quasi-index is

The inspection standard is

δ ⁽¹⁾ (i)∈(1, 1.5)

If the inspection criteria are met, the subsequent modeling steps of GM(2,1) can be performed.

Step 5: Perform an adjacent mean generation operation on P ⁽¹⁾ . make

z ⁽¹⁾ (i)=0.5p ⁽¹⁾ (i)+0.5p ⁽¹⁾ (i-1)(i=2,3...n)#(12)

get Z ⁽¹⁾ = (z ⁽¹⁾ (2), z ⁽¹⁾ (3), z ⁽¹⁾ (4), z ⁽¹⁾ (5))

进行最小二乘估计。得a₁，a₂，u值。Step 6: Parameter column

Perform a least squares estimation. Get a ₁ , a ₂ , u values.

in

Step 7: Determine the solution to the model and equations

Regarding the solution of the GM(2,1) whitening equation, if P ^(1)* is

是对应齐次方程special solution,

is the corresponding homogeneous equation

为GM(2,1)白化方程的通解。general solution, then

is the general solution of the GM(2,1) whitening equation.

The general solution of the whitening equation: when the characteristic equation r ² +a ₁ r+a ₂ =0 has unequal real roots r ₁ , r ₂ ,

When the characteristic equation r ² +a ₁ r+a ₂ =0 has the same real root r,

When the characteristic equation r ² +a ₁ r+a ₂ =0 has complex conjugate roots r ₁ =α+iβ, r ₂ =α-iβ,

The special solution of the whitening equation: when zero is not the root of the characteristic equation, P ^(1)* =C;

When the single root of the characteristic equation at zero time, P ^(1)* =Cp;

When the multiple roots of the characteristic equation at zero time, P ^(1)* =Cp ² ;

又因为零不是特征方程的跟，则P^(1)*＝C。For example: when a ₁ =-3, a ₂ =2, u = 10, r ² -3r+2=0 has two different real roots, then the general solution of the corresponding homogeneous equation is

And since zero is not the heel of the characteristic equation, then P ^(1)* =C.

So the whitening model

的解为p⁽¹⁾(t)＝c₁e^t+c₂e^2t+C.

The solution is p ⁽¹⁾ (t)=c ₁ e ^t +c ₂ e ^2t +C.

Then use the boundary conditions P ⁽¹⁾ = (p ⁽¹⁾ (1), p ⁽¹⁾ (2), ... p ⁽¹⁾ (n)) to solve c ₁ , c ₂ , C. The time response formula for GM(2,1) can be obtained

Step 8: Find the analog value of P ⁽¹⁾

The ninth step: restore the analog value of P ⁽⁰⁾ . Depend on

have to

Step 10: Check for errors.

There are three methods for model testing: relative error, correlation, and mean square error. In general, the most common is the relative error test:

The specific test indicators are: Level 1: the critical value of the indicator is 0.01; Level 2: the critical value of the indicator is 0.05; Level 3: the critical value of the indicator is 0.10; Level 4: the critical value of the indicator is 0.20. If the output result does not meet the critical value of the four-level index, the residual optimization gray prediction model is used.

Step 4: Judgment of the future working state of the diesel engine: after comparing the prediction result with the alarm threshold of the diesel engine safety protection system, make a logical judgment on the running state of the diesel engine, and feed the judgment result back to the upper computer. Fig. 2 is the prediction result of the lubricating oil inlet pressure during the operation of a certain type of diesel engine according to the present invention. As shown in Figure 2, the diesel engine lubricating oil inlet pressure has a downward trend. The predicted lubricating oil inlet pressure based on these operating data will be lower than the pressure threshold of the security system after 80 minutes, then the diesel engine will be judged to be lubricated in the future. Oil inlet pressure is too low. At this time, the security system will send an alarm signal to the upper computer through the program, and send the predicted data to the upper computer in the form of graphs for further analysis.

The above is only a preferred embodiment of a method for predicting the security parameters of diesel engine lubricating oil based on grey theory. , all the technical solutions under this idea belong to the protection scope of the present invention. It should be pointed out that for those skilled in the art, some improvements and changes without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims (7)

1. A method for predicting the security parameters of diesel engine lubricating oil entering a machine based on a grey theory is characterized by comprising the following steps: the method comprises the following steps:

step 1: acquiring pressure data of the diesel engine through a pressure sensor, wherein the pressure data is lubricating oil feeding pressure operation data of the diesel engine, acquiring the pressure data from the pressure sensor of a lubricating oil feeding inlet pipeline of the diesel engine through a data acquisition card, and storing the data into Excel;

step 2: carrying out data preprocessing on collected lubricating oil feeding pressure operation data of the diesel engine, and filtering out environmental noise influence;

and step 3: according to the data after noise filtering, the diesel engine lubricating oil enters the engine pressure gray prediction data to be processed, and a prediction result is obtained;

And 4, step 4: and comparing the prediction result with an alarm threshold value of the diesel engine safety protection system, judging the running state of the diesel engine, and giving an alarm according to the running state of the diesel engine.

2. The diesel engine lubricating oil inlet security parameter prediction method based on the gray theory as claimed in claim 1, characterized in that: the step 2 specifically comprises the following steps:

step 2.1: initializing filter parameters and determining input signal w_kObservation noise v of output signal_k、Q_kAnd product specification error R_kPreprocessing the collected pressure operation data of the lubricating oil of the diesel engine by adopting a Kalman filtering method, filtering the environmental noise influence, establishing a state equation and an output equation, and expressing a state equation x by the following formula_k+1And output equation y_k：

x_k+1＝A_kx_k+w_k

y_k＝C_kx_k+v_k

Wherein k is time, w_kFor input signal, v_kFor the observed noise of the output signal, A is the gain matrix between the state variables, x_kC is a gain matrix between the state variable and the output signal;

determining an input signal w_kObservation noise v of output signal_k、Q_kAnd product specification error R_k；

Step 2.2: setting the initial true temperature to x₍₀₎＝T₀And covariance P of initial time₍₀₎；

Step 2.3: reading the optimal estimated value T at the k-1 time _k-1Measured value t at time k_kCalculating a gain factor, the gain factor being represented by:

H_k＝P_(k-1)/(P_(k-1)+R_k)

T_k＝T_k-1+H_k(t_k-T_k-1)

the optimal to noise covariance at time k is represented by:

P_(k)＝(1-H_k)P_(k-1)

step 2.4: reading the optimal estimated value T at the k-th moment_kAt time k +1Measured value t_k+1Calculating a gain factor H_k+1And T_k+1The optimum noise covariance P at time k +1_(k+1)；

Step 2.5: and (5) repeating the steps 2.3 to 2.4, estimating an optimal value, and obtaining effective data after the environmental noise is filtered.

3. The diesel engine lubricating oil inlet security parameter prediction method based on the gray theory as claimed in claim 1, characterized in that: when the real temperature of the diesel engine is constant, Q_k0; when the temperature of the diesel engine changes along with the running state, Q_k＝0.01。

4. The diesel engine lubricating oil inlet security parameter prediction method based on the gray theory as claimed in claim 1, characterized in that: the step 3 specifically comprises the following steps:

step 3.1: establishing a gray prediction model GM (2,1), and expressing a differential equation of the gray prediction model GM (2,1) by the following formula:

wherein p is⁽¹⁾The accumulated number sequence is generated through one accumulation operation; t is time; a is₁，a₂U is a parameter to be estimated, namely a developing ash coefficient and an endogenous control ash coefficient;

step 3.2: establishing an original pressure data array according to the data after noise filtering, and expressing the array P by the following formula ⁽⁰⁾：

P⁽⁰⁾＝(p⁽⁰⁾(1)，p⁽⁰⁾(2)，p⁽⁰⁾(3)…p⁽⁰⁾(n))

Wherein p is⁽⁰⁾(i) Time series data of the pressure parameter of the lubricating oil inlet machine, i is 1, 2, … n;

step 3.3: according to P⁽⁰⁾Determining P⁽⁰⁾Step ratio of⁽⁰⁾(i) Is tested for the step ratio, and the step ratio σ is expressed by the following formula⁽⁰⁾(i)：

Step 3.4: current class ratio sigma⁽⁰⁾(i) When the inspection standard is satisfied, for P⁽⁰⁾Performing 1-AGO accumulation operation and 1-IAGO accumulation subtraction to obtain generated P⁽¹⁾And alpha⁽¹⁾P⁽⁰⁾Array, P is represented by the formula⁽¹⁾And alpha⁽¹⁾P⁽⁰⁾The sequence of the numbers:

P⁽¹⁾＝(p⁽¹⁾(1)，p⁽¹⁾(2)，…p⁽¹⁾(n))

α⁽¹⁾P⁽⁰⁾＝(α⁽¹⁾p⁽⁰⁾(2)，α⁽¹⁾p⁽⁰⁾(3)…α⁽¹⁾p⁽⁰⁾(n))

α⁽¹⁾p⁽⁰⁾(i)＝p⁽⁰⁾(i)-p⁽⁰⁾(i-1)，i＝2，3，…，n

step 3.5: test P⁽¹⁾If the standard index rule exists, the standard index is calculated by the following formula⁽¹⁾(i)：

Step 3.6: when the quasi-index meets the inspection standard, P is aligned⁽¹⁾Performing an immediate mean generation operation, wherein the immediate mean is represented by the following formula:

z⁽¹⁾(i)＝0.5p⁽¹⁾(i)+0.5p⁽¹⁾(i-1)(i＝2，3…n)

Z⁽¹⁾＝(z⁽¹⁾(2)，z⁽¹⁾(3)，z⁽¹⁾(4)，z⁽¹⁾(5))

step 3.7: for parameter column

Performing least square estimation to obtain a₁，a₂，u：

Step 3.8: solving a whitening equation for GM (2,1) to obtain a time response equation for GM (2,1), the time response equation being represented by:

solving for P⁽¹⁾Is represented by the following formula⁽¹⁾Analog value of (d):

determining P from analog value reduction⁽⁰⁾Is represented by the following formula⁽⁰⁾Analog value of (d):

will P⁽⁰⁾As a result of the prediction

Step 3.9: and carrying out error check on the model, and expressing the error detection value by the following formula:

the test indexes are as follows: first-stage: the index critical value is 0.01; and (2) second stage: the index critical value is 0.05; third-stage: the index critical value is 0.10; and (4) fourth stage: the index critical value is 0.20; and if the error detection value does not meet the critical value of the four-level index, adopting a residual error optimization gray prediction model GM (2, 1).

5. The method for predicting the security parameters of the diesel engine lubricating oil inlet based on the gray theory as claimed in claim 4, wherein the method comprises the following steps: determination of the step ratio sigma⁽⁰⁾(i) The method specifically meets the inspection standard:

when the grade ratio is within the above formula check criteria, then it is satisfied.

6. The method for predicting the security parameters of the diesel engine lubricating oil inlet based on the gray theory as claimed in claim 4, wherein the method comprises the following steps: the specific method for checking whether the quasi-index has the quasi-index rule is as follows:

⁽¹⁾(i)∈(1，1.5)

the quasi-index is satisfied when it is within the above-described criteria.

7. The diesel engine lubricating oil inlet security parameter prediction method based on the gray theory as claimed in claim 1, characterized in that: and after the prediction result is compared with the alarm threshold of the diesel engine safety protection system, the pressure threshold is lower than that of the security system after 80min according to the predicted oil feeding pressure, the future running state of the diesel engine is judged to be that the lubricating oil feeding pressure is too low, at the moment, the security system sends an alarm signal to the upper computer through a program, and the prediction data is simultaneously sent to the upper computer in a charted form for further analysis.

Images