FR2712092A1 - Method for characterising the degree of inflammability and combustibility of vegetation, wood and cellulose fibres, and means for monitoring and warning of fire risks - Google Patents

Abstract

It consists in subjecting a sample of these substances to a radiative flux. The delays, temperature and loss of dry material to combustion are determined and these parameters are combined to give a measure of the fire risk levels. In particular, the loss to combustion is used to define a criticality ratio. The method may be used in particular for establishing a system for warning of forest fires, by virtue of the construction of charts and other mechanical or computerised means.

Description

We know that there are a number of systems for assessing the risks of forest fires. These systems are designed on the basis of various parameters:
- meteorological (wind speed, relative air humidity, precipitation, soil water reserve ...),
- topographic (relief, orientation ...),
- biological (water content, nature and quantity of fuel, etc.).

 These assessment systems make it possible to establish a daily "risk index", a prevention and forecasting tool which makes it possible to respond to the present situation, to implement appropriate alert means.

Among French methods, that of digital risk has been used since 1988. Combining the DROUET and CARREGA methods, digital risk uses three factors:
- humidity: a "false relative humidity" is calculated from the air dew point temperature and the soil temperature, in order to better represent the water content of the litter;
- wind, introduced in different forms in order to modulate its importance;
- the soil water reserve, also introduced with different functions, to modulate its importance.

The Canadian Weather Forest Index (MFI): developed in 1970 by the Canadian Forest Service, this index of weather conditions favorable to forest fires is based on two parameters:
- the water content of three classes of forest fuels (dry light fuel, light surface humus, deep organic matter);
- the effect of wind on fire behavior (temperature and relative humidity of air, rainfall, wind speed).

To this Canadian index, the American National Fire Danger Rating System (NFDRS) also requires knowledge of:
- the state of the weather (cloud cover, precipitation);
- daily conditions of thunderstorm intensity and duration of rain, maximum and minimum temperatures and relative humidity;
- different fuel models;
- risks caused by man;
- relief.

 As can be seen, the preponderance is given, in all of these systems, to meteorological factors, and to their repercussions on the humidity of the litter. However, the North American indices give less formal importance to the properties of the vegetation.

In the Canadian system, an intermediate index, the available fuel index M represents the amount of forest material available for combustion and is used in the calculation of firepower according to BYRAM's law: P = C x M x V
C is the heat of combustion (J / g),
V is the speed of propagation of the fire front (m / s) ,.

 P is the firepower per unit of front length (W / m).

This index is equivalent to the American burning index. The evolution of this NFDRS index is based on a mathematical model of propagation and intensity of the fire which requires the introduction of parameters describing the fuel. Twenty fuel models have been defined for this:
- dead fuels, classified according to the speed with which the moisture content of each fuel particle responds to precipitation, relative air humidity and temperature.

- living fuels, whose water content depends on vital processes, are classified according to the stratum to which they belong:
- the herbaceous layer, comprising annuals with surface roots, therefore affected first by drought;
- perennials, with deep roots;
- the shrub layer (mainly twigs and foliage) affected last.

 The introduction of several fuel models differentiates the American index (one combustibility index for each fuel type) from the Canadian index (only one final index).

 The disadvantage of these systems is that they give only a small place to the intrinsic properties of the living plant (chemical composition, phenology ...).

 Favoring factors external to combustible systems (wind, temperature, etc.) or external (litter), they can only give qualitative indications, undoubtedly sufficient to trigger an alert, but excluding the quantitative relationship to the occurrence of fires (frequency and burnt surfaces) which can only come from the properties of the fuel.

 Unlike the existing systems, the system that we propose, in this patent application, consists of a method of calculating a criticality index, usable for predicting forest fires, based on the properties of the plant and their evolution during of the year.

This criticality index will be evaluated by subjecting the plant to a calorific contribution, in the form of a radiative flux, and by determining the following characteristics:
- ignition delay (tu),
- ignition temperature (tu),
- loss of dry matter on ignition (7ri).

 Quite unexpectedly, these parameters were found to be interrelated and separately correlated to the characteristic values of forest fire statistics (number of fires, area burned).

 The existence of correlations between these parameters, taken two by two, makes it possible to describe the fire behavior of the plant over the whole year and to define the critical values of these parameters which correspond to different levels of fire risk.

 The excellent correlation found between the values of the measured parameters (according to our patent) and the forest fire statistics taken from the "Prometheus" bank justifies the validity of the system for predicting critical conditions presented in this patent application. Of course, the forecasts given here can be refined, in particular for the short term, by the introduction of meteorological data, and in particular of the wind speed.

 The procedure for determining the ignition characteristics, which will be described later, may also be used for the study of the fire behavior of natural materials such as wood, lignocellulosic materials and textile fibers in order to characterize their flammability and their ability to spread a fire.

 In order to carry out the invention, a method is described below for determining the ignition characteristics of a plant, or of a natural material.

The test device includes:
- an epiradiator recommended by AFNOR standard No. 534 RC 2S with a power of 500 Watts, which emits constant infrared radiation energetically;
- a stainless steel wire basket with two compartments; one serving as a reference and the other receiving the sample to be analyzed; this basket is located at a fixed distance of 30 mm from the base of the epirator; this distance corresponds to a lighting power of 3
W / cm2;
- a differential chromel-alumel thermocouple, the two welds of which are located in the middle of each compartment on the same isotherm;
- a data acquisition system consisting of a multivoltmeter controlled by a computer and its peripherals which collects the temperature difference between the two compartments of the basket every 0.3 seconds;
- a potentiometer connected to another thermocouple located in the sample compartment allows direct measurement of the temperature.

The course of the experience is as follows:
- a sample of approximately l gram (weighed to the nearest half centigram) is weighed at the same time as three others which will be used to measure its water content;
- the epiradiator test can be done in two modes:
* under established conditions: the epiradiator is connected and adjusted so as to balance the temperature in the two compartments; as soon as the temperature of the epiradiator is stabilized, the sample is introduced into the basket and the temperature acquisition system is triggered.

 * in transient mode: the sample is put in the basket then the epiradiator and the temperature acquisition system are started simultaneously; this mode of exposure is closer to the process found in a real fire case where the arrival of the fire front produces a gradual heating of the fuel.

Continuous monitoring of the temperature difference between the sample and the reference position (At) as a function of time leads to the curve At = f (T) -
The following particular values are noted on this curve:
- T1: time after which the temperature reaches its lowest value;
- T2: time at the end of which the curve returns through the ordinate point t = 0 during its growth phase; - T3: time corresponding to At = (Atm + AtM) / 2 (with Atm and AtM respectively representing the minimum and maximum temperature differences reached);
- T4: time corresponding to AtM.

 We define the ignition time xi = 13 conventionally. Indeed, it is difficult to know at what precise moment the flame appears physically, because it can be obscured by the plant mass. In addition, certain combustions are preceded only by the single phenomenon of incandescence.

 A typical curve representative of the thermolysis process of the plant fuel studied is presented in Plate 1. It concerns Aleppo Pine (Pin us Halepensis) pyrolyzed according to the established regime (needles of the year, harvested in August).

 The loss of mass on inflammation is obtained in the following manner. A few preliminary tests are carried out in order to determine approximately the time and temperature value ranges. Three inflammations are then carried out successively on samples of approximately 1 gram weighed at half a centigram and of perfectly known water content.

 At the time of ignition, the samples are quickly withdrawn from the radiation, allowed to cool in the desiccator and weighed. The mass lost at the ignition point is used to calculate an ignition loss rate compared to the starting dry matter.

 It has been found quite unpredictably that. the percentage of dry matter lost to ignition (which will be called in the remainder of this patent "loss to ignition") decreases steadily when the water content of the moist plant (or of the natural material) increases. Better still, this decrease is appreciably linear for most of the plants (and natural materials) studied, and we can even find a relationship of proportionality between the relative loss and the relative humidity.

 The curve representative of this relationship between loss and humidity will be called in the remainder of this patent "characteristic curve of the plant".

It was found quite unpredictably that the characteristic curves had an envelope, represented by the right joining the points:
- 100% loss on ignition (i.e. 7tj = 100) and 0% water (i.e. y = O);
- 0% loss on ignition (i.e. zi = O) and 100% water (i.e. y = 100).

 This curve represents the location of the points where the loss on ignition is maximum for a given humidity of the plant (or natural material), or where the water content is maximum for a loss on ignition of the plant (or natural material).

We also define a criticality rate C by the value:
C = [it / (lOO-y) jx 100
It was found quite unpredictably that the ignition delay Tj was related to the absolute temperature Ti = (tj + 273) by the Loglo xi = (A / Ti) + B relationship (with A being essentially positive).

 In the representation (Logio zi, 1 / Ti), the curves representing each plant species during a desiccation or during a natural biological evolution are often represented by straight lines. As a limit case of straight lines or representative points, we find a straight line which is the location of the points where the ignition delay is maximum for a given ignition temperature.

 The approach of this straight line limited by the figurative point of a plant expresses that of the critical conditions.

It has also been found quite unpredictably that the ignition delay is correlated with the water content by a logarithmic law of the form:
Tj = k Logio [H2O] + cste
When the plant approaches the critical conditions mentioned above, the value of k rises suddenly to take values several times greater.

 It is obvious that the set of relationships found between the parameters taken two by two can be used according to our patent for analytical or preparative applications, especially as the conditions of the pyrolysis can be modified to select, in the dry matter of the plant or in natural material such or such fragile material that we want to highlight, measure or isolate.

 Such methods of analysis by combustion may advantageously be practiced because of their speed, in the fields where knowledge of the composition of the plant material is of interest. In particular, in the food and paper industries.

The criticality values calculated from the characteristics of the plants previously defined in our patent were found in close correlation with the occurrence of forest fires. For this correlation, the forest fire statistics used were taken from data from the IT department of the Boûches du Rhône department (Base Prométhée). The data used relate to all the fires in the 15 departments of the interdepartmental agreement for the defense of the forest against fire for the period from June 1992 to June 1993:
- the burned areas being obtained by totaling by fortnight of all the burned areas for fires of more than 10 hectares;
- the outbreaks being obtained by totalization by fortnight of all the fires whatever their hearing, whatever the cause.

 The characteristic parameters of the inflammation defined above and their combinations may be used to determine the fire risk levels in a given geographical area and the chronology of their appearance. We can also establish a risk forecasting system using control plants taken in this geographic area according to a grid of the territory appropriate to the geographic conditions (relief, orientation, prevailing winds).

 Nonlimiting examples of embodiment of the invention are given below. They concern all the fuels made up of Aleppo Pine needles from the years N-1 (1991), N (1992) and N + 1 (1993).

Example # 1:
This example illustrates the use of the diagrams: loss of dry matter on ignition - water content of the wet plant, for monitoring the plant phenology in the vicinity of critical conditions.

 Plate 2 represents the loss on ignition of Aleppo pine needles outside the growth periods, depending on its water content.

This curve was obtained on the one hand by monitoring the biological evolution (interval b-c), and by dehydradation in an oven (interval a-b) on the other hand. It has been extrapolated to 100% water according to the interval c-d.

 Plate 3 shows the evolution of Aleppo pine needles during the growing period: those of year N in summer period (July to September 1992) and those of young shoots the following year (N + 1 ) during bud break in winter (November 1992 to February 1993).

Example # 2:
This example illustrates the use of diagrams (Logio xi, 1 / Ti) for monitoring the plant phenology in the vicinity of critical conditions.

 Each plant in its different states of hydration and at a determined phenological stage is represented by a point in this plane.

The evolution of the plant towards the critical conditions corresponds to the displacement of this point upwards to reach the right line. This type of representation also makes it possible to follow the chemical evolution imposed on the plant when it is dehydrated in the oven. Plate 4 represents these two cases of evolution of Aleppo Pine needles.

Example # 3:
This example is intended to illustrate the correlation existing between the criticality rates and the areas covered by the lights.

 In Plate 5, Figure 5 represents the evolution of the criticality rates of needles N for the summer period, N + 1 for the winter period (beginning of bud burst in November). Figure 6 superimposes the useful part of these curves on the statistical curves of the gross burnt areas (accounting for flows of more than 10 hectares).

 For a more detailed correlation study, the average area covered by fire was evaluated. Its correlation with the criticality rate N for summer fires is clearly shown in Plate 6 with a criticality threshold of 90%. Rates below 90% seem to favor small fires as we will see in the following example.

Example 4:
This example is intended to illustrate the correlation existing between the criticality rates and the frequencies of the fires.

 For the period June 1992 - November 1992, the number of fires started is represented by fortnight. These are well correlated with the criticality rate N-1; as can be seen on Plate 7. The correlation shown in Plate 8 shows a criticality threshold N-l close to 78%.

 As can be seen in the comparison of the curves of plates 6 and 8, the criticality rate N-1 already reached 80.5% at the end of June, while the rate N is still less than 70% on this date. A high frequency of fires is therefore manifested in July by a relatively high value of the rate N-1 while the rate N is still low. These criticality rates therefore clearly reflect the appearance of numerous small fires at the start of the summer, a significant advance of 1 to 2 months over large fires, which accounts for the same criticality rates.

Example 5:
This example is intended to illustrate how the comparison of the ignition characteristics of the green parts of a plant from the current year and from previous years can serve as a precursor of fire risk.

We followed throughout the year, for Aleppo Pine needles, the differences (see Plate 9): - A #i = #in - # in-1;
- Aiti = 7 (jfl - 7tn-1;
- ATj = Tin Ti "-l.

 At the peak of the critical summer and winter periods, these differences are either maximum (maximum fire frequency) or minimum (maximum burnt area). The fall in property differences is not, however, simultaneous; the fall in ATi and ATi preceding that of A at most by one month in summer.

 Thus, monitoring ATi or ATi can be used to anticipate the maximum risk of several decades (approximately 30 days).

Claims (1)

 CLAIMS 1: Process for determining the flammability and combustibility characteristics of vegetation, wood and cellulosic materials, for estimating and predicting the fire risk levels, characterized in that it consists of:  - subject the sample of these materials to a radiative flux,  - determine the time, temperature and dry matter loss parameters at the time of ignition,  - to combine two or two or separately each of said parameters with the composition of the plant (or materials), in particular with its water content so as to obtain a system for predicting or preventing the fire risk. 2: A method according to claim 1 characterized in that the conditions of exposure to the radiative flux are adjusted so as to control the pyrolysis of each fragile constituent of the dry matter. 3: A method according to claims 1 or 2 characterized in that the loss on ignition is compared to the maximum loss that can be obtained under critical conditions to define a criticality rate, measure of fire risk, closely related, in the plant, frequency of fires and burnt surfaces.