EP2348307A1

EP2348307A1 - On-line diagnostics and control process of dielectric behavior of power transformers and device to implement the process

Info

Publication number: EP2348307A1
Application number: EP11000347A
Authority: EP
Inventors: Josef Altmann
Original assignee: Individual
Current assignee: Individual
Priority date: 2010-01-22
Filing date: 2011-01-18
Publication date: 2011-07-27
Also published as: CZ201051A3

Abstract

On-line diagnostic and control of dielectric behavior of power transformers is based on on-line reading of the relative humidity (RH) of the oil in oil filling (102) of main tank (10) of transformer (1) by humidity sensor (2). The on-line reading of temperature of this oil is performed by temperature sensor (3). Both sensors are immersed in oil leaving the upper part of the cellulose insulation system of transformer (1) where both measured variables are periodically sampled by proces control device (4). The value of the oil temperature is also transfered into temperature governor (6) of transformer (1). Proces control device (4), in the predetermined time-intervals, stores values of measured variables into its memory and calculates the Qw-value, which represents the water content in oil, and the Ud,t-value, which represents the theoretical dielectric strength of the oil. Proces control device (4) verifies the veracity and relevancy of calculated Ud,t-value by the comparison of this theoretical value with the Ud,lab-value acquired by the direct reading of the dielectric strength of oil in lab, at the moment of the sampling of oil from transformer (1).

The results of this verification process are futher utilized not only for the on-line diagnostics and control of the dielectric behavior of transformer (1), but also for the critical evaluation of mutual veracity of both Ud-reading methods: the theoretical one versus direct measurement Proces control device (4) compares the processed Ud,t-value with the Ud-value requested by the norm or by internal regulations of the provider, calculating the Tsoll-value, the value of the setting point of temperature governor (6). Temperature governor (6) then controls the temperature of transformer (1) in such a way, that the real dielectric strength of the oil never decreaeses under pre-determined limits.

A device to implement the on-line diagnostics and the control of the dielectric behavior of power transformers consists of transformer (1), to which is connected humidity sensor (2) of oil filling (102), temperature sensor (3), process control device (4) and temperature governor (6). The main tank (10) includes the active part of transformer (1), consisting of magnetic circuit (100) and winding (101) and both immersed in oil filling (102) of transformer(1).

The upper part of main tank (10) is connected by connecting tube (13) to conservator (12) and simultaneously right and left side of said main tank (10) is by upper sleeve (111) and bottom sleeve (112) hydraulically connected to oil coolers (11) where fan casting (113) and fan (114) is situated. The bottom part of said main tank (10) is also provided with sampling cock (14).

The right upper sleeve (111) is provided with well (31) of temperature sensor (3) and with sleeve (21) of humidity sensor (2) whereas process control device (4) is connected by first measuring line (30) to temperature sensor (3), by second measuring line (20) to humidity sensor (2), by first data line (41) to the control port (42), by second data line (43) to external PC, by third data line (46) to temperature governor (6) which is also connected by first measuring line (30) to temperature sensor (3). Said temperature governor (6) is connected by first control line (61) and by second control line (62) to fans (114) of oil coolers (11) of oil filling (102).

Description

Technical Field

The invention relates to the method and to the device to perform on-line diagnostics and the control of the dielectric behavior of power transformers.

State of the Art

The present, most frequent diagnostics of a dielectric behavior of power transformers is based on the single reading of the dielectric strength of transformer oil, the Ud-value (e.g. kV/2.5mm). If a Ud-value is higher then the norm-defined level can the transformer be regarded as safe and operational, if not, the transformer has to be properly maintained to increase the Ud-value over the level requested by the norm.
The reading of relevant Ud-values is a very simple, widely accepted solution which purports to have no problems. The oil is sampled from the oil filling of the main tank of the transformer under its normal operational conditions and the Ud-value is subsequently measured in the laboratory.
The test procedure is strictly defined by the norm (IEC 60 156) and the direct connection between the allowed Ud-values and the dielectric status of the transformer is defined by e.g. IEC 60 422.
General claims of any technical measuring:

o the repeatability (Can I get , repeatedly, the same Ud-value by the same transformer and by the same laboratory ?)
o the reproducibility (Can I get the same Ud-value by the same transformer with other laboratories?)

But an every day diagnostic practice based on operative oil test results shows something quite different.
The standard procedure often give us scattered Ud-values for the same transformer even if the oil does not contain any particles or other inhomogenities.
From the physical side is the scattering of Ud-values of the transformer which is comprehensible and explainable. The dielectric strength of arbitrary transformer oil always varies with its relative humidity and heavily depends on two basic variables : the water content in the oil and the temperature of the oil.
The problem is, that the major part of water in any power transformer, usually over 97-98 weight %, is always deponed in its hard cellulose insulants as Kraft paper, boards etc, and in others cellulose-based materials as wood etc. According to the temperature of the oil-cellulose insulation system, the water migrates between cellulose-based materials and the oil filling and subsequently, the water content in the oil and the Ud-value of the oil inevitably fluctuates as well.
The real Ud-value of the oil in the specific transformer therefore generally depends on the water content in its cellulose materials and its temperature. It simply means, the Ud-value of oil has to change with the transformer temperature.
In practice , the impact of these basic facts may be very serious even critical - the oil sampled from the same transformer under different operational temperatures, gives us different Ud-values and therefore a different picture of its dielectric behavior.
The standard diagnostic approach: one sample of the oil from the transformer → one Ud-value evaluation of the oil in a lab → one diagnostics result, cannot give us the clear picture and the answer to the fundamental question : can the given transformer be safely operated or not.
Moreover, the Ud-reading of the specific transformer is not only determined by its temperature but by the dynamic change of the transformer temperature.
We will always obtain two different readings of the Ud-values, at the same temperature level of a transformer, wheather with decreasing or increasing temperatures.
The basic request of any relevant diagnostics is the satisfactory repeatibility and the reproducibility of all readings and evaluations. These generally cannot be met. The practice of oil sampling at any predefined temperature of a transformer is extremely difficult.
The dielectric strength of the oil, the Ud-value mainly depends on:

o Relative Humidity of the oil (RH-value) which , for given transformers permanently varies with its temperature
o particles in the oil

If we presuppose that our oil is without particles, then the Ud-value depends only on the RH-value.
It means, the increase of a temperature of the oil means the decrease of its RH-value and the increase of its Ud-value and vice-versa.
But the increase of the temperature of the whole transformer, means the increase of the temperature of its oil-cellulose system and the migration of the water from cellulose materials into the oil filling. The increase of the water content in the oil, means the growth of the RH-value and the drop of the Ud-value.
And vice-versa - the decrease of the transformer temperature, means the increase of its Ud-values.
Both processes in any transformer proceed simultaneously and partially compensate each other. In the end-effect, the "real" Ud-value of the oil (measured at operational temperature of its oil-cellulose system) then slightly grows with the temperature growth and sinks with its decrease. But the "lab" Ud-value, the oil is sampled from the oil-cellulose system at a operational temperature, but measured then at the 20°C in the lab, which creates generally a big difference in the Ud-value.
The present methods used do not consider these facts in detail. The standard diagnostics is mainly limited to sampling the oil under arbitrary operational conditions of a transformer and subsequent evaluating the Ud-value in the lab at a predetermined temperature of 20°C.
Theoretically this approach means :

ο the sampling of the oil is performed at an operational temperature of a transformer (which is usually much higher as 20°C: the lab temp.)
o the sampled oil is cooled down at 20°C
o the Ud-reading is performed at 20°C

lab temperature

A standard diagnostic procedure then doesn't represents the real behavior of the oil-cellulose insulation but describes its " the worst possible case": the oil-cellulose system is seemingly "jump-like" cooled down from the operational temperature at the lab temperature 20°C, the RH-value of the oil jump-like increases and its Ud-value simultaneously jump-like decreases. Of course provided that, the lab temperature, 20°C, really represents the lowest achievable temperature of the oil-celluloser system.
Only by the jump-like change of the temperature of this system, the migration of the water between oil filling and the cellulose materials can be stopped.
The oil-cellulose system behaves itself here, as if it consists from the oil only : the temperature change is so fast, that the water from the oil filling doesn't have enough time to migrate back to the cellulose, or to diffuse back to the oil..
At first sight, the jump-like cooling-down of a real transformer looks like as a pure theoretical process, but it isn't. The similar effect can be observed in any transformer. In any oil cooler of a transformer, the oil is cooled down without the presence of cellulose materials and therefore the same effect of the "jump-like" decrease of the oil temperature, equals the increase of the RH-value and the jump-like decrease of the Ud-value .
The standard diagnostic Ud-approach then inevitably represents the worst possible dielectric case of a transformer.
Although this approach has its explicit advantages - it remains simple, available and widely used and accepted and if properly performed gives us some , though, mostly unquantifiable, safety reserves by the Ud-diagnostic conclusions.
But the strong dependency on the temperature , the main drawback of this diagnostic approach remains :

o cold transformers, regardless of the real amount of internally deponed water, always show relatively high Ud-values: water is deponed in cellulose materials and subsequently, a low water content in the oil (at 20°C), results in a low RH-level and a high Ud-level even in wet transformers
o hot transformer, on the other hand, always shows relatively low Ud-values: due to higher temperatures the water migrates from the cellulose into the oil, the water content in the oil is therefore high. Though the RH-level at this operational temperature is relatively low and Ud-level relatively high , the cooling-down of the oil to 20°C (at the lab. temp), results in the strong increase of the RH-level and the decrease of the Ud-level, in spite of the fact that the water content of its cellulose materials can be relatively low.

The standard diagnostic results for the given transformer, if it is based on an occasional or periodic reading of a Ud- level can therefore be full of contradictions.
The same transformer, at low temperatures, satisfies norm-requested Ud-levels very well and presents diagnostics indicates it may be operated without a problem, but on higher temperatures the Ud-value drops down and the transformer, according to the same diagnostics should not be operated.
The drawbacks of present off-line diagnostic methods, if based on occasional samplings and point-like readings of Ud-levels, can be summarized in the following way:

o between two periodically performed sampling procedures (and Ud-readings), the users of the transformer simply don't have any relevant information about the real dielectric behavior of his transformer. Any accidental higher loading and subsequently higher temperatures of the transformer usually means the violation of the Ud-limit requested by the norm.
o single reading of the Ud-value is not representative enough, because it often doesn't properly respect the migration of the water between the cellulose and oil filling of a transformer, especially with dynamic changes of its temperature
o the standard method is not able to describe the dielectric behavior of a transformer for the whole range of its operational temperatures
o the method is not able to control the dielectric behavior of a transformer at higher operational temperatures, because there doesn't exist any continuous reading of Ud-values of the oil enabling subsequent changes of operational regimes of a transformer avoiding dangerous decreases of the Ud-levels under the limit requested by the norm.

Disclosure of the Invention

The invention is a method and device for on-line diagnostics and the control of the dielectric behavior of power transformers using the on-line reading of the relative humidity of the oil and its temperature which determines the dilelectric strength of the oil which enables a change of the operational regime of a transformer to effectively prevent any decrease of the dielectric strength of the oil under limits requested by the norm.
The dielectric strength of any transformer oil (without solid particles) is dominatly governed by its relative humidity and can be expressed, for clarity, as relation (A) : $Ud, t = Ud, \max (1 - RH)$

where :

Ud,t: instantaneous theoretical dielectric sgtrength of oil (e.g. in kV/2.5mm)
Ud, max: maximal achievable dielectric strength of oil for RH → 0
RH: relative humidity of oil (1)

Any relevant on-line diagnostic and control of dielectric behavior of a transformer based on the relation (1), has to meet two basic criterias:

o the Ud,t -value relevantly represents the instantaneous dielectrics of a transformer
o the time-related Ud,t-value can be easily verified and calibrated by means of standard and independent readings of the Ud-value in the lab.

The "representativness" of a Ud-value here means, that this specific value represents the dielectric behavior of the transformer as the whole.
The Ud-value determined by the relation (A) only, cannot unfortunately met the mentioned requirement, because the real Ud-value in a transformer varies very strongly spatially and in time.
In a transformer and for given time-points always exist the whole field of Ud-values: lower ones in its bottom part of a transformer with lower temperature levels and higher ones for its upper part.
The direct measured RH-value of the oil has therefore to be replaced by another value whose level is the same for the whole transformer.
This requirement meets the Qw-value (water content in oil), due to permanent and strong mixing of the oil inventory in every transformer (and relatively weak migration of water between cellulose and oil). The oil filling of a transformer is that way satisfactorily homogenized and the Qw-value therefore satisfactorily "represents" the whole oil inventory of the transformer.
The verification of the calculated Ud,t-value is then acquired by the direct comparison of the calculated Ud-value and the directly measured Ud-value in the lab, provided that the calculation of the theoretical Ud-value is performed at the time-point of the sampling of oil for a lab.
The suitable sensor immersed in the oil leasing the oil-cellulose systém of a transformer then simultaneously reads the RH-value and temperature of the oil and calculates the relevant water content in the oil.
Subsequently, the "conventional" RH_20C -value (valid for the whole transformer) is calculated as : ${RH}_{20 C} = Qw / Qw, Sat$

where:

RH_20C: instantaneous "conventional" relative humidity of oil (at 20 °C)
Qw: measured water content in oil (ppm)
Qw,sat: water content in oil corresponding its full saturation ( at 20 °C)

The theoretical , "conventional" dielectric strength of the oil is then described by the relation: $Ud, t = Ud, \max (1 - {RH}_{20 C})$
The mutual comparison of both Ud-values (calculated versus lab measured values) enables us then a very interesting and quite new diagnostic approach, See following Table:

Ud,t >> Ud,lab: theoretical (calculated) Ud,t-value is substantially higher than Ud,lab value ⇒ the oil is very likely contaminated by particles and this preliminary result should subsequently be confirmed or falsified by a lab reading. Another alternative: reading of any Ud-values is wrong
Ud,t ≈ Ud,lab: theoretical value of dielectric strength is approximately the same as lab value ⇒ both readings verify each other and the time-related Ud,t-diagnostic is verified and correct
Ud,t << Ud,lab: theoretical, maximal attainable Ud-value, is substantially lower than lab value ⇒ lab reading is probably wrong and should be repeated (or, less likely, the on-line reading is wrong)

The quantitative evaluation and subsequent verification of the Ud,t-value enables us to fulfill most of requirements necessary for a relevant diagnostic and/or control of a transformer:

⇒ diagnostic results are explicit and verifiable: the on-line calculated Ud,t-value truly represents, for given time-period, the dielectric behavior of the transformer
⇒ dielectric behavior of a transformer can be controled by the change of its temperature: calculated Ud,t-values are compared with the norm requested Ud-value and the suitable algorithm converts a corresponding difference into the change of the value of the setting point of the temperature governor of transformer
⇒ the temperature governor then changes the temperature of a transformer in such a way so that the actual Ud,t-value is always higher than the value requested by the norm or a level required by the internal limits determined by the provider.

In practice the new diagnostic method of a transformer and the control of its dielectric behavior proceeds in following steps:

o the RH-value and the temperature of the oil is measured on-line by a suitable capacity sensor, not ballasted by any parasitic effects (e.g. the water in organic acids presented in the aged oil are not read etc,)
o the suitable sensor is immersed directly in the hot oil leaving the oil-cellulose system of the transformer and therefore reads both values correctly, immediately and continuously
o measured values are permanently stored in the memory of a suitable process control device and simultaneously the Ud,t-value of the oil is calculated and stored and transferred to a superior PC
o the process control device then compares, at the moment of sampling, both values and evaluates their mutual relevancy and veracity and subsequently confirms or disproves the relevancy of calculated Ud,t-values for the given time-period
o the verified Ud,t-value is compared with the Ud-value requested by the norm and the process control device subsequently calculates /changes the setting point of the temperature governor of a transformer to ensure its proper operational temperature and the Ud-values of the oil remain always higher than the requested limit
o the sampling of the oil in a transformer is perfomed in pre-determined time-periods to acquire Ud, lab-values necessary for the confirmation of the correctness of the Ud,t-processing procedure.

Brief Description of the Drawings

The exemplary embodiment of the proposed solution is described with reference to the drawing where:

Fig. 1 schematically shows the device for on-line diagnostics and control of dielectric behavior of power transformers.

The Example of the Invention

An embodiment of the invention is shown in FIG.1, and primarily includes the transformer 1, consisting of the main tank 10, where the magnetic circuit 100 and the winding 101 is situated. The conservator 12 is located above the main tank 10 and is connected by the connecting tube 13. The upper and the bottom part of the main tank 10 is connected by upper sleeve 111 and bottom sleeve 112 to upper and botom part of oil coolers 11 provided by the fan casings 113 with the inbuilt fans 114. At the right side of the main tank 10, the upper sleeve 111 is provided with the well 31 and the sleeve 21 and the sampling cock 14 is situated at left side of the main tank 10.
The electrical circuits of the first practical aspect of the invention include measuring and control lines which connect the process control device 4 with the temperature sensor 3, moisture sensor 2, temperature governor 6, remote PC (not shown) and the control port 42.
The temperature governor 6 is also connected to the temperature sensor 3 and both fans 114.
The first measuring line 30 connects the temperature sensor 3, inserted in the well 31 in the sleeve 111, with process control device 4 and simultaneously with the temperature governor 6, the second measuring line connects the moisture sensor 2, inserted into the sleeve 21 in the upper sleeve 111 with process control device 4.
The first data line 41 connects the control port 42 with the process control device 4, the second data line 43 connects process control device 4 with the remote PC (not shown), the third data line 46 connects the process control device 4 with the temperature governor 6.
The first control line 61 and the second control line 62 then connects the temperature governor 6 with both fans 114.
The operation of the of the first embodiment of the invention is focussed on the on-line aquirement of two (from the view of its dielectric behavior), fundamental parameters of any transformer: the relative humidity of the oil (RH- value) via the moisture sensor 2 and temperature of the oil, via the temperature sensor 3.
The process control device 4 stores both values in the pre-defined time-periods into its memory, calculates theoretical dielectric strength of oil (Ud,t-value) and desired temperature of the transformer (Tsoll-value). Simultaneously, all calculated values are stored into the process control device 4 memory and these values are also transfered into the PC memory.
The verification of validity of Ud,t-values is then performed by the process control device 4 which compares the Ud,t-value and the lab value (Ud,lab-value) of the oil. The mutual comparison of sampled and calculated Ud-values is then performed at the same time-point (the time of oil sampling).
The data entry of the Ud, lab-value into process control device 4 is performed manually via control ports 42 and the first data line 41, or this value is manualy entered into PC and transferred via the second data line 43 into process control device 4.
If the deviation between the Ud,t- and Ud,lab-value is lower than the pre-defined value, the process control device 4 evalutes the Ud,t-processing as correct and relevant for the elapsed time-period and for another pre-defined time period in future and subsequently calculates and/or changes of value of the setting point (Tsoll-value) and send this value via third data line 46 to the temperature governor 6.
When this pre-defined time-interval elapses, the new lab reading is necessary to verify the corresponding Ud,t-processing for the next time-period. The new sample of oil has to be drawn from the oil inventory of the transformer and subsequenty the new reading of the Ud,lab-value has to be performed.
The Tsoll-value is then calculated by the process control device 4 by a suitable algorithm in order to satisfy the norm-required (or by internal specification) Ud-value of the oil.
If this Tsoll-value is higher than temperature of the oil (To-value) measured by the temperature sensor 3, the theoretical Ud,t-value (and the real Ud-value) of oil is higher than demanded. The temperature governor 6 does not intervene therefore and the cooling conditions of the transformer 1 remain the same as before.
If the Tsoll-value is lower than the To-value, the Ud,t-value is lower than demanded and the oil and the transformer 1 has to be cooled down. The temperature governor 6 turn both fans 114 on and cooling effect induced by the forced blowing of oil coolers 11 substantially increases the heat removal from the transformer 1, its temperature and the temperature of oil gradually decreases and the Ud,t-value increases.
The temperature stabilization of transformer 1 is continuously evaluated by process control device 4, which gradually modifies the setting level of temperature governor 6 to reduce the maximum amplitude of the To-variable under ca 2-4 °C and the variation of the Ud,t-value as well.
The basic requierement is, that the instantaneous Ud,t-value has to be always higher than the Ud-value requested by the norm or an internal regulations of the provider.
The temperature ON / OFF control of fans 114 of the transformer 1 shown here is an example only. The real temperature control loop of the transformer 1 and its function is strongly simplified for clarity and a better understanding.
The practical temperature control loop is substantially more sophisticated because has to handle multiparametric problems. The load of the transformer 1 and the surrounding temperature fluctuate and the proper stabilization of the Ud,t-level needs an adaptive control.
If an absolute deviation betwen the calculated and lab-measured Ud-values exceeds pre-defined limits, the calculated Ud,t-value cannot not more be used for the on-line control of the dielectric behavior of given transformer.
The setting point of the temperature governor control is therefore set back to constant, pre-defined Tsoll-value and the process control device 4 reports this error and subsequently performs the preliminary analysis of the problem.
E.g. when the Ud,t-value (the teoretical, highest attainable value of dielectric strength) is substantially higher than the lab-value then the first, most probable, explanation is the contamination of the oil inventory by particles. The second explanation of this discrepancy is then the wrong reading in the lab.
On the other hand, when calculated Ud,t-value is permanently lower than the lab value, then the most probable explanation is wrong on-line reading and/or lab reading.
The process control device 4 reports all these results via second data line 43 to PC, PC alerts the operating staff and recommends how to identify the problem more precisely and what kind of the counteraction will be necessary.

List of Reference Symbols

1: transformer
10: main tank
11: oil cooler
111: upper sleeve
112: bottom sleeve
113: fan casting
114: fan
12: conservator
13: connecting tube
14: sampling cock
100: magnetic circuit
101: winding
102: oil filling
2: humidity sensor
20: second measuring line
21: sleeve
3: temperature sensor
30: first measuring line
31: well
4: process control device
41: first data line
42: control port
43: second data line
46: third data line
6: temperature governor
61: first control line
62: second control line

RH: relative humidity of the oil
Ud,t: the theoretical dielectric strength of the oil
Ud,min: lowest allowed dielectric strength of the oil
Ud,max: maximal achievable dielectric strength of oil
Ud,lab: dielectric strength of the oil measured in a lab
Tsoll: desired temperature of the transformer
Qw: the water content in oil
Qw,sat: water content in oil corresponding its full saturation (at 20 °C)
To: oil temperature

Claims

On-line diagnostic and control of dielectric behavior of power transformers based on on-line reading of the relative humidity (RH) of the oil in oil filling (102) of main tank (10) of transformer (1) by humidity sensor (2)
and
the on-line reading of temperature of this oil by temperature sensor (3), where both sensors are immersed in oil leaving the upper part of the cellulose insulation system of transformer (1) where both measured variables are periodically sampled by proces control device (4) and
the value of the oil temperature is also transfered into temperature governor (6) of transformer (1),
wehereas
proces control device (4), in the predetermined time-intervals, stores values of measured variables into its memory and calculates the Qw-value, which represents the water content in oil, and the Ud,t-value, which represents the theoretical dielectric strength of the oil,
and verifies the veracity and relevancy of calculated Ud,t-value by the comparison of this theoretical value with the Ud,lab-value acquired by the direct reading of the dielectric strength of oil in lab, at the moment of the sampling of oil from transformer (1)
and
the results of this verification process are futher utilized not only for the on-line diagnostics and control of the dielectric behavior of transformer (1), but also for the critical evaluation of mutual veracity of both Ud-reading methods: the theoretical one versus direct measurement,
and
proces control device (4) compares the processed Ud,t-value with the Ud-value requested by the norm or by internal regulations of the provider, calculating the Tsoll-value, the value of the setting point of temperature governor (6)
and
temperature governor (6) then controls the temperature of transformer (1) in such a way, that the real dielectric strength of the oil never decreaeses under pre-determined limits.
A device to implement the on-line diagnostics and the control of the dielectric behavior of power transformers as set out in claim 1 consisting of transformer (1), to which is connected humidity sensor (2) of oil filling (102), temperature sensor (3), process control device (4) and temperature governor (6) characterized in that the main tank (10) includes the active part of transformer (1), consisting of magnetic circuit (100) and winding (101) and both immersed in oil filling (102) of transformer(1),
while the upper part of main tank (10) is connected by connecting tube (13) to conservator (12) and simultaneously right and left side of said main tank (10) is by upper sleeve (111) and bottom sleeve (112) hydraulically connected to oil coolers (11) where fan casting (113) and fan (114) is situated, and the bottom part of said main tank (10) is also provided with sampling cock (14),
while right upper sleeve (111) is provided with well (31) of temperature sensor (3) and with sleeve (21) of humidity sensor (2) whereas process control device (4) is connected by first measuring line (30) to temperature sensor (3), by second measuring line (20) to humidity sensor (2), by first data line (41) to the control port (42), by second data line (43) to external PC, by third data line (46) to temperature governor (6) which is also connected by first measuring line (30) to temperature sensor (3) and said temperature governor (6) is connected by first control line (61) and by second control line (62) to fans (114) of oil coolers (11) of oil filling (102).