JP3392010B2 - Filter method for periodic rpm fluctuation of rotating body and signal processing device therefor - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a filter method adaptable to periodic fluctuations of rpm every time a rotating body rotates. In this filter method, during each rotation, a large number (N) of substantially evenly distributed through the rotation path of the rotating body.
Measuring the actual rpm at each measurement position (i), or a value equivalent thereto, generating an electronic signal value of rpm for each of the rpm measurements, and each of the rpm signal values To store continuously, and rpm for each measurement position during each rotation
Processing the rpm signal value to generate a signal value.

The present invention is not limited solely to the above, but primarily improves the measurement of rpm, whereby
It enables cruise control of multi-cylinder diesel engines for ships and land.

[0003]

BACKGROUND OF THE INVENTION As a result of such engine operating conditions, including the effects of combustion pressure on pistons and the like, and the external effects of load on the engine, the crankshaft and output shaft of the engine must operate in a uniform operating condition. Instead, one or more vibration motions with different orders are added.

For regulation purposes, in modern computer controlled engines, the control of numerous engine component actuations, such as exhaust valve opening and closing, fuel pump control, lubrication, and speed changes, is Because it must be done based on uniform engine operation,
It is desirable to isolate uniform engine operation, including aperiodic variations due to load variations and the like.

In order to adjust the idling rpm of the internal combustion engine, the control parameters acting on the adjusting element are converted into electrically dependent variables, so that an adaptive adjusting characteristic for the continuously acting adjusting element is obtained. It is known from EP 0,136,449. This electrically dependent variable influences the offset or trend of its characteristic by means of multiplication or summation by at least one stored value. The stored value represents the output signal from the control loop that is excited during a particular operating condition. By comparing its control parameter with the actual measured value,
An output signal is obtained. This output signal stores one of the stored values by changing one of the stored values in the direction in which the deviation of the system is relatively small and ending the operating condition.

This storage is done in an integrator with a large time constant so that the actual idling speed can adapt to the offset or trend of the characteristics that are not affected.

From EP 0,429,819 a device for detecting cyclically changing entities of an internal combustion engine is known. This detector, periodically changing the output signal from the sensor, the center frequency is introduced into the variable band suppression filter. The output signal of the filter is representative of the dynamic reproduction of the mean value of the periodically varying entities.

[0008]

SUMMARY OF THE INVENTION With the above as a background, one object of the present invention is to provide a method for reducing the periodical rpm fluctuation of a rotating body to the first order, and a signal processing apparatus therefor. To eliminate aperiodic fluctuations, and thus without disturbing the "true" rpm signal that needs to be adjusted, for example, the torsional vibration of an engine crankshaft, or a screw shaft. It is possible to filter.

[0009]

In order to achieve the above object, the method according to the invention is based on the electronic rpm signal value.
The measured rpm represented Ri, estimated rpm deviation ([delta]
As a difference from (i)), the rpm signal value (ω _out ) obtained at each measurement position (i) during each rotation is obtained, and the estimated rpm is calculated at each measurement position () for a predetermined rotation (k + 1). i) updated as a function of its estimated rpm deviation value and the actual rpm deviation value for the same measurement position (iN) at the immediately preceding rotation (k), the actual r
The pm deviation value is the actual measurement rpm (ω _in , _k ) at the measurement position (1-N) and the calculated average rpm.
It is characterized in that it is obtained as a difference from (ω _av , _k ).

Periodic rp for a crankshaft rotating at a relatively low speed in a marine diesel engine
By using this method to make a filter adaptable to the m deviation value, the number of measurement positions is equal to, for example, twice the number of cylinders, ie 12 positions for a 6 cylinder engine. Becomes a number. This method uses a relatively small number of simple mathematical operations for each measurement point,
An rpm signal is generated that represents a dynamic value that is very closely approximated to rpm, which represents a uniform velocity with the effects of first order periodic variations removed. In order to calculate rpm for each measurement position during each rotation, one subtraction may be performed in real time, and the estimated rpm deviation value is calculated based on the previous preceding revolution.

In the preferred embodiment, the estimated rpm deviation value at a given measurement position during a given rotation is estimated based on the following algorithm.

[0012]

Δ (i) _{k + 1} = e (i) _k -e (i) _k exp (-1 / c) + δ (i) _k exp (-1 / c) where c is a constant, e (i) _k is the actual rpm, and δ
(I) _k is the estimated rpm deviation value, which is the value at the same measurement position (i) during rotation (k), respectively.

In order to carry out the present invention, the present invention comprises a sensor device (4) for measuring each actual rpm or a value equivalent to each actual rpm at a large number of measuring positions for each rotation of the rotating body. A signal processing device is provided. The measurement positions are substantially evenly distributed over the entire rotation path of the rotating body, the processing device generates an electronic rpm signal (ω _in ) for each of the measurement positions, and Is obtained at each measurement position based on the measurement result.
It comprises a signal processor (5) for calculating a pm value ( ω _out ).

According to the present invention, such a signal processing device is
First storage device for storing the above rpm signal (ω _in )
(7), and the actual rpm value (e (i)) at each measurement position (i ) is the same during each rotation.
Measured rpm signal value (ω _in ) and average rpm _in (i)
First subtraction step, calculated as the difference from (ω _av , _k ).
A second storage device (12) including (11) and storing the actual rpm deviation thus obtained for each measurement position, and an estimated rpm deviation value for each measurement position.
A calculation unit under program control for calculating (δ (i)) is provided, and the estimated rpm deviation value is calculated as the immediately preceding rotation (k).
Rotation (k + ) as a function of the estimated rpm and the actual rpm deviation value (δ (i)) at the same measurement position (i) in
1) A third storage device (1 ) for updating the estimated rpm deviation value (δ) at all the measurement positions.
7) and a second calculation of rmp (ω _out ) obtained as a difference between the actual rpm (ω _in ) and the estimated rpm deviation value (δ (i)) at a predetermined measurement position during a predetermined rotation. And a subtraction step (16).

[0015]

DETAILED DESCRIPTION OF THE INVENTION The present invention will be described in more detail below with reference to the schematic drawings.

FIG. 1 shows, for example, a screw shaft 2
And a 6-cylinder marine diesel engine 1 connected with a screw 3. A sensor device 4 is attached to the screw shaft 2. The sensor device 4 measures the shaft rpm at each of a number of measurement positions N and calculates an electronic rpm signal value (ω _in ) for each of these measurements.

The sensor device 4 is, for example, a prior art of a tachometer type, and is designed to be capable of measuring a value equivalent to rpm as a moving time of a measuring position passing through the sensor element, for example. You can

The number N of measurement positions is typically equal to twice the number of cylinders, for example N = 12 for a 6-cylinder engine.

The signal from the sensor device 4 is introduced into a signal processor 5, which for each measuring position eliminates the periodic fluctuations, a correction or adjustment rpm.
Calculate (ω _out ) . This signal is used as information to a computer-based engine controller 6 which is responsible for controlling different engine components.

As is apparent from the flow diagram of FIG.
From the adaptive filter realized by the method according to the invention for periodic rpm fluctuations, a given rotation k + 1
For each measurement position i (where i = 1 ... N), the actual rpm, ie the signal value ω _in (i), which is the signal provided by the sensor device 4, is measured and Subtract the estimated rpm deviation value δ (i) from the measured rpm ω _in ,
The rpm value (ω _out ) is calculated by a simple subtraction method.

[0021]

Ω _out (i) = ω _in (i) − δ (i) (1) This subtraction is the only operation that causes a delay in measurement for each measurement position during each rotation, and Up to
The estimated deviation value δ (i) is calculated based on the measured value and the calculated value during the rotation immediately after the new measurement .

Also, for each measurement position during rotation k,
The rpm ω _in (i) is measured. Based on these measurements, the average rpm value ω _av (i) for each of the positions i is calculated by summing the rpm ω _in for a number of measurement positions covering at least part of the total rotation. The calculation of this average value is typically done by measuring the actual measured position i
Is used as an intermediate value, and the rpms at all the measurement positions are summed up, and the number of the measurement positions is adjusted to be uniform, and the sum is expressed by the following equation.

[0023]

[Equation 5] For actual measurement position i, actual rpm deviation value e
(I) is the measured value rpmω _in(I) and average valueω
_avCalculated from (i).

[0024]

E (i) = ω _in (i) − ω _AV (i) (3) with the estimated rpm deviation value δ (i) _k calculated for the same measurement position during the previous rotation k Using the actual rpm deviation value e (i) calculated during the previous revolution k,
Update the estimated rpm deviation value to the value δ (i) _{k + 1} used during the next rotation k + 1. This calculation is performed according to the following equation.

[0025]

(7) δ (i) _{k + 1} = e (i) _k -e (i) _k exp (-1 / c) + δ (i) _k exp (-1 / c) (4) where c Is a filter constant corresponding to a simple first rank filter.

In the signal processor 5, the signal value ω _in of the measured rpm is supplied to the storage device 7, which stores
In order to perform the above calculations, it is possible to include rpm for all measurement positions during about 1 1/2 revolution, for example 12 measurement positions 17 for storing the measured value of the measurement signal ω _in. It costs.

The sum of the rpm values at each measurement position is
As shown in the above equation (2), it is performed in the summing circuit 8,
After that, the actual average rpm is calculated by dividing by the total number N in the dividing circuit 9.

The average value ω _av thus obtained is stored in the storage device 10.

Then, for each measurement position i, the actual rpm deviation value e (i) described above is calculated according to the above equation (3). As a result of calculating the average value for a given measurement point i, after storing the measured rpms for a number of subsequent positions including position j = i + N / 2, the calculation of e (i) is performed in the subtraction circuit 11. Will be done.

Here, for a given measurement position, the estimated r
The pm deviation value δ (i) is updated so that the value δ (i) _k previously calculated for that position during the previous rotation and stored in the memory 12 is the same rotation according to equation (4) above. In the meantime, it is correlated with the actually measured deviation value e (i) and input to the portion of the measurement position in the storage device 12.

The update by the equation (4) is performed by the multiplication circuits 13 and 1
4 (the filter constant K = exp (−1 / c) is given to both of them) and the addition circuit 15.

During the next revolution k + 1, if the actual measured position is reached, the system obtains rpm ω obtained by a simple subtraction in subtraction circuit 16 according to equation (1) above.
_It is ready to calculate _out and store this signal in storage device 17.

For example, in order to record in the logbook of a ship, this system is preferably designed as follows.
That is, for the purpose of monitoring and testing the system,
Ω for each measuring point in the storage device 7, 12, 17
_The design is such that the stored value of _in , δ, or ω _out can be taken out and registered externally.

As can be seen from the examples shown in FIGS. 4 and 5 for the waveforms for unfiltered rpm (a) and filtered rpm (b) respectively, the adaptive filter according to the present invention allows simultaneous And for situations with aperiodic fluctuations or no such fluctuations,
For example, smoothing of the signal is achieved, in which periodic fluctuations resulting from torsional vibrations are eliminated or substantially reduced.

The method of updating the estimated rpm deviation value typically results in an inaccurate change being stored in storage device 12 by abruptly changing the waveform and / or abruptly starting the system. , With some kind of transition phenomenon.

By properly selecting the filter constant c, such a problem hardly occurs. However, the system could alternatively be modified, for example, by shutting down the engine and resetting the filters 1, 12, 12 or by disabling the signal processor until the engine 1 has been slightly rotated. It is possible.

Although the present invention has been described above as a control system for a marine diesel engine, the method and signal processing device according to the present invention mainly controls rpm of any type where it is desired to filter periodic signal variations. Also generally considered useful.

[Brief description of drawings]

FIG. 1 is a diagram of an application example of the present invention, which is a block diagram showing main parts of a computer-based control system for a marine diesel engine.

FIG. 2 is a flow diagram for explaining the method according to the present invention.

FIG. 3 is a diagram of one embodiment of a signal processor in a signal processing device according to the present invention.

FIG. 4 is a graphical representation of the actual measured rpm waveform superimposed with the twist signal and the rpm signal waveform filtered by the method of the present invention.

5 is a waveform diagram similar to FIG.

[Explanation of symbols]

1 Marine Diesel Engine 2 Screw Shaft 3 Screw 4 Sensor Device 5 Signal Processor 6 Engine Control Device 7 Memory Device 8 Summation Circuit 9 Dividing Circuit 10, 12, 17
Storage device 11, 16 Subtraction circuit 13, 14 Multiplication circuit 15 Addition circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-7-318574 (JP, A) JP-A-7-279738 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G05D 13/62 F02D 45/00 G01P 3/44

Claims (7)

(57) [Claims] 1. Whenever the rotating body rotates,rpm
A filter method adaptable to fluctuations, During each rotation, it is distributed almost evenly through the rotation path of the rotating body.
The actual number at each of the multiple (N) measurement positions (i)
measuring rpm or a value equivalent thereto, Rpm electronic measured for each of the rpm measurements
Signal (ω_in) Is generated, Continuously storing each of the rpm signals; Rpm signal value for each measurement position during each rotation
(Ω_out) Processing the rpm signal to generate
A filtering method, including steps and The rpm signal (ω_in) Actual measurement represented byrpm
And the estimated rpm deviation (δ (i)) as the difference between
Rpm signal obtained at each measurement position (i)
(Ω_out) Is calculated, and the estimated rpm value
For each measurement position (i) for (k + 1), its estimate
The constant rpm deviation value and the same value at the immediately preceding rotation (k)
Relationship with the actual rpm deviation value for the measurement position (i-N)
Updated as a number, The actualrpmDeviation value at the measurement position (i-N)
Actual measurement rpm (ω_in,_k) And the calculated average rp
m value (ω_av,_k) Is obtained as the difference between
How to filter. 2. The filter method according to claim 1,
As the actual measurement rpm for all measurement positions (N) during one rotation, where the predetermined measurement position is the intermediate position,
Said average rpm for a given measuring position (i)
A filter method characterized in that (ω _av , _k ) is obtained. 3. The estimated rpm deviation value (δ (i) used in the filter method according to claim 1 or 2 for a given measuring position (i) during a given rotation (k + 1).
_{k + 1} ) is calculated based on the following algorithm: ## EQU1 ## δ (i) _{k + 1} = e (i) _k -e (i) _k exp (-1 / c) + δ (i) _k exp (-1 / c) where c is a constant, e (i) _k is the actual rpm , and δ
(I) _k is the estimated rpm deviation value, which is the same measurement position (i) during the previous rotation (k).
A filtering method, characterized in that it is the value at. 4. The filter method according to any one of claims 1 to 3, wherein in addition to measuring the actual measurement rpm value, all measurement positions during rotation just completed are measured. The actual measured rpm value is summed for each of the measured positions (i) during each revolution (k), the summed value being the measured position (i-N / 2) located at the position of a half turn backward. , _Is used to determine the average rpm value (ω _av ), and is also used to determine the estimated rpm deviation value used for the measured position of the estimated rpm deviation value during the subsequent rotation (k + 1). Characterizing filter method. 5. A signal processing apparatus for carrying out the method according to claim 1, wherein a plurality of measurement positions (i) (N) are set for each rotation of the rotating body. ), A sensor device (4) for measuring each actual rpm, or a value equivalent thereto, wherein the measurement positions are substantially evenly distributed over the entire rotation path of the rotating body, and the processing device is , An electronic rpm signal value (ω _in ) is generated for each of the measurement positions, and the rpm obtained at each measurement position is based on the measurement result.
A signal processing device comprising a signal processor (5) for calculating a value (ω _out ), wherein the signal processing device (5) stores the rpm signal value (ω _in ). And the actual rpm deviation value (e (i)) at each measurement position (i) during each rotation is measured r at the same measurement position (i).
Includes a first subtraction step (11), which calculates as the difference between the pm measured value (ω _in ) and the average rpm (ω _av , _k ), and calculates the estimated rpm deviation value (δ (i)) for each measured position A program-controlled computing unit for calculating the estimated rpm deviation value as a function of the estimated rpm and the actual rpm deviation value (δ (i)) at the same measurement position (i) during the immediately preceding rotation (k). A second storage device (12) that is updated for each rotation (k + 1) and stores the estimated rpm deviation value (δ) at all the measurement positions, and an actual rpm (at a predetermined measurement position during a predetermined rotation). ω _in ) and estimated r
rpm obtained as the difference from the pm deviation value (δ (i))
A second subtraction step (16) for calculating (ω _out ). 6. The signal processing device according to claim 5,
The signal processor (5) calculates the sum of the actual rpm values for all measuring positions (N) during one revolution, the first summing device (8) and the divider device (9) for calculating the average value. ) And a signal processing device. 7. The signal processing device according to claim 5, wherein the calculation unit is configured to calculate the estimated deviation about a predetermined measurement position (i) during a predetermined rotation (k + 1) according to the following equation. It is a program control formula for calculating the value (δ (i) _{k + 1} ), where δ (i) _{k + 1} = e (i) _k -e (i) _k exp (-1 / c) + δ (i) _k exp (-1 / c) where c is a constant, N is the total number of measurement positions, and e (i) _k
Is the actual rpm, and δ (i) _k is the estimated rpm deviation value, and the rpm value is the value at the same measurement position (i) during the previous one revolution (k). And a signal processing device.